The Tiger Compiler Project Assignment

Table of Contents


Nul n’est censé ignorer la loi.

Everything exposed in this document is expected to be known.

Next: , Up: (dir)   [Contents][Index]

The Tiger Compiler Project

This document, revision EPITA-2017-20-g8bd97f7 of February 6, 2015, details the various tasks EPITA students must complete. It is available under various forms:


Next: , Previous: , Up: Top   [Contents][Index]

1 Introduction

This document presents the Tiger Project as part of the EPITA curriculum. It aims at the implementation of a Tiger compiler (see Modern Compiler Implementation) in C++.


Next: , Up: Introduction   [Contents][Index]

1.1 How to Read this Document

If you are a newcomer, you might be afraid by its sheer size. Don’t worry, but in any case, do not give up: as stated in the very beginning of this document,

Nul n’est censé ignorer la loi.

That is to say everything exposed in this document is considered to be known. If it is written but you didn’t know, you are wrong. If it is not written and was not clearly reported in the news, we are wrong.

Basically this document contains three kinds of information:

Initial and Permanent

What you must read and know since the very beginning of the project. This includes most the following chapters: Introduction (except the History section), Instructions, and Evaluation.

Incremental

You should read these parts as and when needed. This includes mostly Compiler Stages.

Auxiliary

This information is provided to help you: just go there when you feel the need, Tools, and Source Code. If you want to have a better understanding of the project, if you are about to criticize something, be sure to read History beforehand.

There is additional material on the Internet:


Next: , Previous: , Up: Introduction   [Contents][Index]

1.2 Why the Tiger Project

This project is quite different from most other EPITA projects, and has aims at several different goals, in different areas:

Several iterations

This project is about the only one with which you will live for 9 months, with the constant needs to fix errors found in earlier stages.

Complete Project

While the evaluation of most student projects is based on the code, this project restores the deserved emphasis on documentation and testing. Because of the duration of the project, you will value the importance of a good (developer’s) documentation (why did we write this 4 months ago?), and of a good test suite (why does TC-2 fails now that we implemented TC-4? When did we break it?).

This also means that you have to design a test suite, and maintain it through out the project. The test suite is an integral part of the project.

Team Management

The Tiger Compiler is a long project, running from January to September (and optionally further). Each four person team is likely to experience nasty “human problems”. This is explicitly a part of the project: the team management is a task you have to address. That may well include exclusion of lazy members.

C++

C++ is by no means an adequate language to study compilers (C would be even worse). Languages such as Haskell, Ocaml, Stratego are much better suited (actually the latter is even designed to this end). But, as already said, the primary goal is not to learn how to write a compiler: for an EPITA student, learning C++, Design Patterns, and Object Oriented Design is much more important.

Note, however, that implementing an industrial strength compiler in C++ makes a lot of sense1. Bjarne Stroustrup’s list of C++ Applications mentions Metrowerks (CodeWarrior), HP, Sun, Intel, M$ as examples.

Understanding Computers

Too many students still have a very fuzzy mental picture of what a computer is, and how a program runs. Studying compilers helps understanding how it works, and therefore how to perform a good job. Although most students will never be asked to write a single line of assembly during their whole lives, knowing assembly is also of help. See Bjarne Stroustrup, for instance, says:

Q: What is your opinion, is knowing assembly language useful for programmers nowadays?

BS: It is useful to understand how machines work and knowing assembler is almost essential for that.

English

English is the language for this project, starting with this very document, written by a French person, for French students. You cannot be a good computer scientist with absolutely no fluency in English. The following quote is from Bjarne Stroustrup, who is danish (The Design and Evolution of C++, 6.5.3.2 Extended Character Sets):

English has an important role as a common language for programmers, and I suspect that it would be unwise to abandon that without serious consideration.

Any attempt to break the importance of English is wrong. For instance, do not translate this document nor any other. Ask support to the Yakas, or to the English team. By the past, some oral and written examinations were made in English. It may well be back some day. Some books will help you to improve your English, see The Elements of Style.

Compiler

The project aims at the implementation of a compiler, but this is a minor issue. The field of compilers is a wonderful place where most of computer science is concentrated, that’s why this topic is extremely convenient as long term project. But it is not the major goal, the full list of all these items is.

The Tiger project is not unique in these regards, see Cool - The Classroom Object-Oriented Compiler, for instance, with many strikingly similar goals, and some profound differences. See also Making Compiler Design Relevant for Students who will (Most Likely) Never Design a Compiler, for an explanation of why compilation techniques have a broader influence than they seem.


Next: , Previous: , Up: Introduction   [Contents][Index]

1.3 What the Tiger Project is not

This section could have been named “What Akim did not say”, or “Common misinterpretations”.

The first and foremost misinterpretation would be “Akim says C sucks and is useless”. Wrong. C sucks, definitely, but today C is probably the first employer of programmers in the world, so let’s face it: C is mandatory in your education. The fact that C++ is studied afterward does not mean that learning C is a loss of time, it means that since C is basically a subset of C++ it makes sense to learn it first, it also means that (let it be only because it is a superset) C++ provides additional services so it is often a better choice, but even more often you don’t have the choice.

C++ is becoming a common requirement for programmers, so you also have to learn it, although it “features” many defects (but heredity was not in its favor...). It’s an industrial standard, so learn it, and learn it well: know its strengths and weaknesses.

And by the way, of course C++ sucks++.

Another common rumor in EPITA has it that “C/Unix programming does not deserve attention after the first period”. Wrong again. First of all its words are wrong: it is a legacy belief that C and Unix require each other: you can implement advanced system features using other languages than C (starting with C++, of course), and of course C can be used for other tasks than just system programming. For instance Bjarne Stroustrup’s list of C++ Applications includes:

Apple

OS X is written in a mix of language, but a few important parts are C++. The two most interesting are:

  • − Finder
  • − IOKit device drivers. (IOKit is the only place where we use C++ in the kernel, though.)[...]
Ericsson
  • − TelORB - Distributed operating system with object oriented
Microsoft

Literally everything at Microsoft is built using various flavors of Visual C++ - mostly 6.0 and 7.0 but we do have a few holdouts still using 5.0 :-( and some products like Windows XP use more recent builds of the compiler. The list would include major products like:

  • − Windows XP
  • − Windows NT (NT4 and 2000)
  • − Windows 9x (95, 98, Me)
  • − Microsoft Office (Word, Excel, Access, PowerPoint, Outlook)[...]
CDE

The CDE desktop (the standard desktop on many UNIX systems) is written in C++.


Know C. Learn when it is adequate, and why you need it.

Know C++. Learn when it is adequate, and why you need it.

Know other languages. Learn when they are adequate, and why you need them.

And then, if you are asked to choose, make an educated choice. If there is no choice to be made, just deal with Real Life.


Previous: , Up: Introduction   [Contents][Index]

1.4 History

The Tiger Compiler Project evolves every year, so as to improve its infrastructure, to demonstrate more instructional material and so forth. This section tries to keep a list of these changes, together with the most constructive criticisms from students (or ourselves).

If you have information, including criticisms, that should be mentioned here, please send it to us.

The years correspond to the class, e.g., Tiger 2005 refers to EPITA class 2005, i.e., the project ran from October 2002 to July (previously September) 2003.


Next: , Up: History   [Contents][Index]

1.4.1 Fair Criticism

Before diving into the history of the Tiger Compiler Project in EPITA, a whole project in itself for ourselves, with experimental tries and failures, it might be good to review some constraints that can explain why things are the way they are. Understanding these constraints will make it easier to criticize actual flaws, instead of focusing on issues that are mandated by other factors.

Bear in mind that Tiger is an instructional project, the purpose of which is detailed above, see Why the Tiger Project. Because the input is a stream of students with virtually no knowledge whatsoever in C++, and our target is a stream of students with good fluency in many constructs and understanding of complex matters, we have to gradually transform them via intermediate forms with increasing skills. In particular this means that by the end of the project, evolved techniques can and should be used, but at the beginning only introductory knowledge should be needed. As an example of a consequence, we cannot have a nice and high-tech AST.

Because the insight of compilers is not the primary goal, when a choice is to be made between (i) more interesting work on compiler internals with little C++ novelty, and (ii) providing most of this work and focusing on something else, then we are most likely to select the second option. This means that the Tiger Project is doomed to be a low-tech featureless compiler, with no call graph, no default optimization, no debugging support, no bells, no whistles, and even no etc. Hence, most interested students will sometimes feel we “stole” the pleasure to write nice pieces of code from them; understand that we actually provided code to the other students: you are free to rewrite everything if you wish.


Next: , Previous: , Up: History   [Contents][Index]

1.4.2 Tiger 2002

This is not standard C++

We used to run the standard compiler from NetBSD: egcs 1.1.2. This was not standard C++ (e.g., we used to include ‘<iostream.h>’, we could use members of the std name space unqualified etc.). In addition, we were using hash_map which is an SGI extension that is not available in standard C++. It was therefore decided to upgrade the compiler in 2003, and to upgrade the programming style.

Wrapping a tarball is impossible

During the first edition of the Tiger Compiler project, students had to write their own Makefiles — after all, knowing Make is considered mandatory for an Epitean. This had the most dramatic effects, with a wide range of creative and imaginative ways to have your project fail; for instance:

As a result Akim grew tired of fixing the tarballs, and in order to have a robust, efficient (albeit some piece of pain in the neck sometimes) distribution 2 we moved to using Automake, and hence Autoconf.

There are reasons not to be happy with it, agreed. But there are many more reasons to be sad without it. So Autoconf and Automake are here to stay.

Note, however, that you are free to use another system if you wish. Just obey the standard package interface (see Submission).

The SemantVisitor is a nightmare to maintain

The SemantVisitor, which performs both the type checking and the translation to intermediate code, was near to impossible to deliver in pieces to the students: because type checking and translation were so much intertwined, it was not possible to deliver as a first step the type checking machinery template, and then the translation pieces. Students had to fight with non applicable patches. This was fixed in Tiger 2003 by splitting the SemantVisitor into TypeVisitor and TranslationVisitor. The negative impact, of course, is a performance loss.

Akim is tired during the student defenses

Seeing every single group for each compiler stage is a nightmare. Sometimes Akim was not enough aware.


Next: , Previous: , Up: History   [Contents][Index]

1.4.3 Tiger 2003

During this year, Akim was helped by:

Comaintainers

Alexandre Duret-Lutz, Thierry Géraud.

Submission dates were:

StageSubmission
TC-1Monday, December 18th 2000 at noon
TC-2Friday, February 23th 2001 at noon
TC-3Friday, March 30th 2001 at noon
TC-4Tuesday, June 12th 2001 at noon
TC-5Monday, September 17th 2001 at noon

Some groups have reached TC-6.

Criticisms include:

The C++ compiler is broken

Akim had to install an updated version of the C++ compiler since the system team did not want non standard software. Unfortunately, NetBSD turned out to be seriously incompatible with this version of the C++ compiler (its crt1.o dumped core on the standard stream constructors, way before calling main). We had to revert to using the bad native C++ compiler.

It is to be noted that some funny guy once replaced the g++ executable from Akim’s account into ‘rm -rf ~’. Some students and Akim himself have been bitten. The funny thing is that this is when the system administration realized the teacher accounts were not backed up.

Fortunately, since that time, decent compilers have been made available, and the Tiger Compiler is now written in strictly standard C++.

The AST is rigid

Because the members of the AST objects were references, it was impossible to implement any change on it: simplifications, optimization etc. This is fixed in Tiger 2004 where all the members are now pointers, but the interface to these classes still uses references.

Akim is even more tired during the student defenses

Just as the previous year, see Tiger 2002, but with more groups and more stages. But now there are enough competent students to create a group of assistants, the Yakas, to help the students, and to share the load of defenses.

Upgrading is not easy

Only tarballs were submitted, making upgrades delicate, error prone, and time consuming. The systematic use of patches between tarballs since the 2004 edition solves this issue.

Upgraded tarballs don’t compile

Students would like at least to be able to compile a tarball with its holes. To this end, much of the removed code is now inside functions, leaving just what it needed to satisfy the prototype. Unfortunately this is not very easy to do, and conflicts with the next complaint:

Filling holes is not interesting

In order to scale down the amount of code students have to write, in order to have them focus on instructional material, more parts are submitted almost complete except for a few interesting places. Unfortunately, some students decided to answer the question completely mechanically (copy, paste, tweak until it compiles), instead of focusing of completing their own education. There is not much we can do about this. Some parts will therefore grow; typically some files will be left empty instead of having most of the skeleton ready (prototypes and so forth). This means more work, but more interesting I (Akim) guess. But it conflicts with the previous item...


Next: , Previous: , Up: History   [Contents][Index]

1.4.4 Tiger 2004

During this year, Akim was helped by:

Comaintainers

Alexandre Duret-Lutz, Raphaël Poss, Robert Anisko, Yann Régis-Gianas,

Assistants

Arnaud Dumont, Pascal Guedon, Samuel Plessis-Fraissard,

Students

Cédric Bail, Sébastien Broussaud (Darks Bob), Stéphane Molina (Kain), William Fink.

Submission dates were:

StageSubmission
TC-2Tuesday, March 4th 2002 at noon
TC-3Friday, March 15th 2002 at noon
TC-4Friday, April 12th 2002 at noon
TC-5Friday, June 14th 2002, at noon
TC-6Monday, July 15th 2002 at noon

Criticisms include:

The driver is not maintainable

The compiler driver was a nightmare to maintain, extend etc. when delivering additional modules etc. This was fixed in 2005 by the introduction of the Task model.

No sane documentation

This was addressed by the use of Doxygen in 2005.

No UML documentation

The solution is yet to be found.

Too many visitors

It seems that some students think there were too many visitors to implement. I (Akim) do not subscribe to this view (after all, why not complain that “there are too many programs to implement”, or, in a more C++ vocabulary “there are too many classes to implement”), nevertheless in Tiger 2005 this was addressed by making the EscapeVisitor “optional” (actually it became a rush).

Too many memory leaks

The only memory properly reclaimed is that of the AST. No better answer for the rest of the compiler. This is the most severe flaw in this project, and definitely the worst thing to remember of: what we showed is not what student should learn to do.

Though a garbage collector is tempting and well suited for our tasks, its pedagogical content is less interesting: students should be taught how to properly manage the memory.

Upgraded tarballs don’t compile
Filling holes is not interesting

Cannot be solved, see Tiger 2003.

Ending on TC-6 is frustrating

Several students were frustrated by the fact we had to stop at TC-6: the reference compiler did not have any back-end. Continuing onto TC-7 was offered to several groups, and some of them actually finished the compiler. We took their work, adjusted it, and it became the base of the reference compiler of 2005. The most significant effort was made by Daniel Gazard.

Double submission is intractable

Students were allowed to deliver twice their project — with a small penalty — if they failed to meet the so-called “first submission deadline”, or if they wanted to improve their score. But it was impossible to organize, and led to too much sloppiness from some students. These problems were addressed with the introduction of “uploads” in Tiger 2005.


Next: , Previous: , Up: History   [Contents][Index]

1.4.5 Tiger 2005

A lot of the following material is the result of discussion with several people, including, but not limited to3:

Comaintainers

Benoît Perrot, Raphaël Poss,

Assistants

Alexis Brouard, Sébastien Broussaud (Darks Bob), Stéphane Molina (Kain), William Fink,

Students

Claire Calméjane, David Mancel, Fabrice Hesling, Michel Loiseleur.

I (Akim) here thank all the people who participated to this edition of this project. It has been a wonderful vintage, thanks to the students, the assistants, and the members of the LRDE.

Deliveries were:

StageSubmission
TC-0Friday, January 24th 2003 12:00
TC-1Friday, February 14th 2003 12:00
TC-2Friday, March 14th 2003 12:00
TC-4Friday, April 25th 2003 12:00
TC-3Rush from Saturday, May 24th at 18:00 to Sunday 12:00
TC-56Friday, June 20th 2003, 12:00
TC-7Friday, July 4th 2003 12:00
TC-78Friday, July 18th 2003 12:00
TC-9Monday, September 8th 2003 12:00

Criticisms about Tiger 2005 include:

Too many memory leaks

See Tiger 2004. This is the most significant failure of Tiger as an instructional project: we ought to demonstrate the proper memory management in big project, and instead we demonstrate laziness. Please, criticize us, denounce us, but do not reproduce the same errors.

The factors that had pushed to a weak memory management is mainly a lack of coordination between developers: we should have written more things. So don’t do as we did: define the memory management policy for each module, and write it.

The 2006 edition pays strict attention to memory allocation.

Too long to compile

Too much code was in *.hh files. Since then the policy wrt file contents was defined (see File Conventions), and in Tiger 2006 was adjusted to obey these conventions. Unfortunately, although the improvement was significant, it was not measured precisely.

The interfaces between modules have also been cleaned to avoid excessive inter dependencies. Also, when possible, opaque types are used to avoid additional includes. Each module exports forward declarations in a fwd.hh file to promote this. For instance, ast/tasks.hh today includes:

// Forward declarations of ast:: items.
#include "ast/fwd.hh"
// ...
    /// Global root node of abstract syntax tree.
    extern ast::Exp* the_program;
// ...

where it used to include all the AST headers to define exactly the type ast::Exp.

Upgraded tarballs don’t compile
Filling holes is not interesting

Cannot be solved, see Tiger 2003.

No written conventions

Since its inception, the Tiger Compiler Project lacked this very section (see History) and that dedicated to coding style (see Coding Style) until the debriefing of 2005. As a result, some students or even so co-developers of our own tc reproduced errors of the past, changed something for lack of understanding, slightly broke the homogeneity of the coding style etc. Do not make the same mistake: write down your policy.

The AST is too poor

One would like to insert annotations in the AST, say whether a variable is escaping (to know whether it cannot be in a register, see TC-3, and TC-5), or whether the left hand side of an assignment in Void (in which case the translation must not issue an actual assignment), or whether ‘a < b’ is about strings (in which case the translation will issue a hidden call to strcmp), or the type of a variable (needed when implementing object oriented Tiger), etc., etc.

As you can see, the list is virtually infinite. So we would need an extensible system of annotation of the AST. As of September 2003 no solution has been chosen. But we must be cautious not to complicate TC-2 too much (it is already a very steep step).

People don’t learn enough C++

It seems that the goal of learning object oriented programming and C++ is sometimes hidden behind the difficult understanding of the Tiger compiler itself. Sometimes students just fill the holes.

To avoid this:

The computation of the escapes is too hard
The computation of the escapes is too easy

If you understood what it means that a variable escapes, then the implementation is so straightforward that it’s almost boring. If you didn’t understand it, you’re dead. Because the understanding of escapes needs a good understanding of the stack management (explained more in details way afterward, during TC-5), many students are deadly lost.

We are considering splitting TC-5 into two: TC-5- which would be limited to programs without escaping variables, and TC-5+ with escaping variables and the computation of the escapes.

The static-link optimization pass is improperly documented

Todo.

The use of references is confusing

We used to utilize references instead of pointers when the arity of the relation is one; in other words, we used pointers iff 0 was a valid value, and references otherwise. This is nice and clean, but unfortunately it caused great confusion amongst students (who were puzzled before ‘*new’, and, worse yet, ended believing that’s the only way to instantiate objects, even automatic!), and also confused some of the maintainers (for whom a reference does not propagate the responsibility wrt memory allocation/deallocation).

Since Tiger 2006, the coding style enforces a more conventional style.

Not enough freedom

The fact that the modelisation is already settled, together with the extensive skeletons, results in too tight a space for a programmer to experiment alternatives. We try to break these bounds for those who want by providing a generic interface: if you comply with it, you may interchange with your full re-implementation. We also (now explicitly) allow the use of a different tool set. Hints at possible extensions are provided, and finally, alternative implementation are suggested for each stage, for instance see TC-2 Improvements.


Next: , Previous: , Up: History   [Contents][Index]

1.4.6 Tiger 2006

Akim has been helped by:

Assistants

Claire Calméjane, Fabrice Hesling, Marco Tessari, Tristan Lanfrey

Deliveries:

StageKindSubmissionSupervisor
TC-0Wednesday, 2004-02-04 12:00Anne-Lise Brourhant
TC-1Sunday, 2004-02-08 12:00Tristan Lanfrey
TC-2Sunday, 2004-03-07 12:00Anne-Lise Brourhant, Tristan Lanfrey
TC-3RushFr., 2004-03-19 18:30 to Sun., 2004-03-21 19:00Fabrice Hesling
TC-4Sunday, 2004-04-11 19:00Tristan Lanfrey
TC-5Sunday, 2004-06-06 12:00Fabrice Hesling
TC-6Sunday, 2004-06-27 12:00Marco Tessari
TC-7OptSunday, 2004-07-11 12:00Marco Tessari or Fabrice Hesling
TC-89OptThursday, 2004-07-29 12:00Marco Tessari

Criticisms about Tiger 2006 include:

The interface of symbol::Table should be provided

On the one hand side, we meant to have students implement it from scratch so we shouldn’t provide the header, and on the other hand, the rest of the (provided) code expects a well defined interface, so we should publish it! The result was confusion and loss of time.

The problem actually disappeared: Tiger 2007 no longer depends so heavily on scoped symbol tables.

Some examples are incorrectly rejected by the reference compiler

The Tiger reference manual does not exclude sick examples such as:

let
  type rec = {}
in
  rec {}
end

where the type rec escapes its scope since the type checker will assign the type rec to the let construct. Given the suggested implementation, which reclaims memory allocated by the declarations when closing the scope, the compiler dumps core.

The new implementation, tested with 2005b, copes with this gracefully: types are destroyed when the AST is. This does not cure the example, which should be invalid IMHO. The following example, from Arnaud Fabre, amplifies the problem.

let
  var box :=
    let
      type box = {val: string}
      var box := box {val = "42\n"}
    in
       box
    end
in
  print(box.val)
end
TC-5 is too hard a stage

This is a recurrent complaint. We tried to make it easier by moving more material into earlier stages (e.g., scopes are no longer dealt with by the TranslateVisitor: the Binder did it all).

Multiple inheritance is not demonstrated

There are several nice opportunities of factoring the AST using multiple inheritance. Tiger 2007 uses them (e.g., Escapable, Bindable etc.).

The coding style for types is inconsistent

The sources are ambivalent wrt to pointer and reference types. Sometimes ‘type *var’, sometimes ‘type* var’. Obviously the latter is the more “logical”: the space separates the type from the variable name. Unfortunately the declaration semantics in C/C++ introduces pitfalls: ‘int* ip, i’ is equivalent to ‘int* ip; int i;’. That is why I, Akim, was using the ‘type *var’ style, and resisted to expressing the coding style on this regard. The resulting mix of styles was becoming chronic: defining a rule was needed... In favor of ‘type* var’, with the provision that multiple variable declarations are forbidden.

More enthusiasm from the assistants

It has been suggested that assistants should show more motivation for the Tiger Project. It was suggested that they were not enough involved in the process. For Tiger 2007, there are no less than 10 Tiger assistants (as opposed to 4), and two of them are co-maintaining the reference compiler. Assistants will also be kept more informed of code changes than before.

More technical lectures

Some regret when programming techniques (e.g., object functions, ‘#include <functional>’) are not taught. My (Akim’s) personal opinion is that students should learn to learn by themselves. It was decided to more emphasize these goals. Also, oral examinations should be ahead the code submission, and that should ensure that students have understood what is expected from them.

Formal definition of Booleans

The Tiger language enjoys well defined semantics: a given program has a single defined behavior... except if the value of ‘a & b’ or ‘a | b’ is used. To fix this issue, in Tiger 2007 they return either 0 or 1.

Amongst other noteworthy changes, after five years of peaceful existence, the stages of the compiler were renamed from T1, T4 etc. to TC-1, TC-4... EPITA moved from “periods” (P1, P2...) to “trimesters” and they stole T1 and so forth from Tiger.


Next: , Previous: , Up: History   [Contents][Index]

1.4.7 Tiger 2005b

Akim has been helped by:

Comaintainers

Arnaud Fabre, Gilles Walbrou, Roland Levillain

Assistants

Charles Rathouis, Claire Calméjane, Fabrice Hesling, Marco Tessari, Tristan Carel, Tristan Lanfrey,

Deliveries:

StageKindSubmission
TC-1Sun 2004-10-10 12:00
TC-2Sun 2004-10-24 12:00
TC-3Sun 2004-11-7 12:00
TC-4Sun 2004-11-28 12:00

Criticisms about Tiger 2006 include:

Use of misc::ident

Some examples would be most welcome. Well, there is misc/test-indent.cc, and now the PrintVisitor code includes a few examples.

test-ref.cc

This file is used only in TC-5, yet it is submitted at TC-1, so students want to fix it, which is too soon. Tarballs will be adjusted to avoid this.


Next: , Previous: , Up: History   [Contents][Index]

1.4.8 Tiger 2007

Akim has been helped by:

Comaintainers

Arnaud Fabre, Roland Levillain, Gilles Walbrou

Assistants

Arnaud Fabre, Bastien Gueguen, Benoît Monin, Chloé Boivin, Fanny Ricour, Gilles Walbrou, Julien Nesme, Philippe Kajmar, Tristan Carel

Deliveries:

StageKindLaunchSubmissionSupervisor
TC-0Wed 2005-03-09Tue 2005-03-15 23:42Bastien Gueguen
TC-1RushFri 2005-03-18Sun 2005-03-19 9:00Guillaume Bousquet
TC-2Mon 2005-03-21Sun 2005-04-03Nicolas Rateau
TC-3RushFri 2005-04-08 20:00Sun 2005-04-10 12:00Fanny Ricour
TC-4Mon 2005-04-18Sun 2005-05-01Julien Nesme
TC-5Mon 2005-05-09Sun 2005-06-05Benoît Monin
TC-6Mon 2005-06-06Sun 2005-06-12Philippe Kajmar
TC-7Mon 2005-06-13Sun 2005-06-19Gilles Walbrou
TC-8Mon 2005-06-20Mon 2005-06-27Arnaud Fabre
TC-9Mon 2005-06-20Sun 2005-07-03Arnaud Fabre
Final submissionWed 2005-07-06

Criticisms about Tiger 2007 include:

Cheating

Too much cheating during TC-5. Some would like more repression; that’s fair enough. We will also be stricter during the exams.

Debriefing

After a submission, there should be longer debriefings, including details about common errors. Some of the mysterious test cases should be explained (but not given in full). Maybe some bits of C++ code too.

Design of the compiler

More justification of the overall design is demanded. Some selected parts, typically TC-5, should have a UML presentation.

Tarball

Keep the tarball simple to use. We have to improve the case of tcsh. Also: give the tarball before the presentation by the assistants.

Oral examinations

Assistants should be given a map of where to look at. The test suite should be evaluated at each submission. The use of version control too.

Optional parts

They want more of them! We have more: see TC-R, TC-D, and TC-I.

misc:: tools

There should be a presentation of them.

TC-3 is too long

TC-3, a rush, took several groups by surprise.

Some groups would have liked to have the files earlier: in the future we will publish them on the Wednesday, instead of the last minute.

Some groups have found it very difficult to be several working together on the same file (binder.cc of course). This is also a problem in the group management, and use of version control: when tasks are properly assigned, and using a tool such as Subversion, such problems should be minimal. In particular, merges resulting from updates should not be troublesome! Difficult updates result from disordered edition of the files. Dropping the use of a version control manager is not an answer: you will be bitten one day if two people edit concurrently the same file. One option is to split the file, say binder-exp.cc and binder-dec.cc for instance. I (Akim) leave this to students.

The template method template is too hard

Some students would have preferred not to have the declaration of Binder::decs_visit, but the majority prefers: we will stay on this version, but we will emphasize that students are free not to follow our suggestions.

TC-5

Several people would like more time to do it. But let’s face it: the time most student spend on the project is independent of the amount of available time. Rather, early oral exams about TC-5 should suffice to prompt students to start earlier.

People agree it is harder, and mainly because of compiler construction issues, not C++ issues. But many students prefer to keep it this way, rather than completely giving away the answers to compiler construction related problems.


Next: , Previous: , Up: History   [Contents][Index]

1.4.9 Tiger 2008

We have been helped by:

Comaintainers

Christophe Duong, Fabien Ouy

Assistants

Deliveries:

StageKindLaunchSubmissionSupervisor
TC-0Tue 01-03Fri 01-13 23:42Christophe Duong
TC-1RushFri 03-17Sun 03-19 12:12Renaud Lienhart
TC-2Mon 03-20Thu 03-30 23:42David Doukhan
TC-3RushFri 03-31Sun 04-02 12:12Frederick Mousnier-Lompre
TC-4Tue 04-04Mon 04-24 23:42Guillaume Deslandes
TC-5Mon 05-01Sun 05-28 23:42Alexis Sebbane
TC-6Mon 05-29Sun 06-11 23:42Christophe Duong
TC-7Wed 06-14Wed 06-21 12:00
TC-8Wed 06-21Sun 07-2 12:00
TC-9Mon 07-03Sun 07-16 12:00
Final

Some of the noteworthy changes compared to Tiger 2007:

Simplification of the parser

The parser is simplified in a number of ways. First the old syntax for imported files, let <decs> end, is simplified into <decs>. We also use GLR starting at TC-2. &, | and the unary minus operator are desugared using concrete syntax transformations.

See TC-R, Unique identifiers

This new optional part should be done during TC-3. Leave TC-E for later (with TC-5 or maybe TC-4).

Concrete syntax

Transformations can now be written using Tiger concrete syntax rather than explicit AST construction in C++. This applies to the DesugarVisitor, BoundsCheckingVisitor and InlineVisitor.


Next: , Previous: , Up: History   [Contents][Index]

1.4.10 Leopard 2009

We have been helped by:

Comaintainers

Benoît Tailhades, Alain Vongsouvanh, Razik Yousfi, Benoît Perrot, Benoît Sigoure

Assistants

Deliveries:

StageKindLaunchSubmissionSupervisor
LC-0Mon 03-05Fri 03-16 12:00
LC-1RushFri 03-23Sun 03-25 12:00
LC-2Mon 03-26Fri 04-06 12:00
LC-3 & LC-RRushFri 04-06Sun 04-08 12:00
LC-4Mon 04-23Sun 05-06 12:00
LC-5Mon 05-15Sun 06-03 12:00
LC-6Mon 06-04Sun 06-10 12:00
LC-7Mon 06-11Wed 06-20 12:00
LC-8Thu 06-21Sun 07-01 12:00
LC-9Mon 07-02Sun 07-15 12:00

Some of the noteworthy changes compared to Tiger 2008:

Object-Oriented Programming

The language is extended with object-oriented features, as described by Andrew Appel in chapter 14 of Modern Compiler Implementation. The syntax is close to Appel’s, with small modifications, see See Syntactic Specifications in Tiger Compiler Reference Manual.

Leopard

To reflect this major addition, the language (and thus the project) is given a new name, Leopard. These changes was announced at TC-2, (renamed LC-2).

LC-R

LC-R is a mandatory part of the LC-3 assignment.


Next: , Previous: , Up: History   [Contents][Index]

1.4.11 Tiger 2010

We have been helped by:

Comaintainers

Benoît Perrot, Benoît Sigoure, Guillaume Duhamel, Yann Grandmaître, Nicolas Teck

Assistants

Deliveries:

StageKindLaunchSubmissionSupervisor
TC-0Mon Nov 05, 2007Sun Nov 25, 2007 12:00
TC-1Mon Dec 10, 2007Sun Dec 16, 2007 12:00
TC-2Mon Feb 25, 2008Wed Mar 05, 2008 12:00
TC-3 & TC-RRushFri Mar 07, 2008Sun Mar 09, 2008 12:00
TC-4Mon Mar 10, 2008Sun Mar 23, 2008 12:00
TC-5Mon Mar 24, 2008Sun Apr 06, 2008 12:00
TC-6Mon Apr 14, 2008Sun Apr 20, 2008 12:00
TC-7Mon Apr 21, 2008Sun May 04, 2008 12:00
TC-8Mon May 05, 2008Sun May 18, 2008 12:00
TC-9Mon May 19, 2008Sun Jun 01, 2008 12:00

Some of the noteworthy changes compared to Leopard 2009:

The Tiger is back

The project is renamed back to its original name.


Next: , Previous: , Up: History   [Contents][Index]

1.4.12 Tiger 2011

This is the tenth year of the Tiger Project.

We have been helped by:

Assistants

Adrien Biarnes, Medhi Ellaffet, Vincent Nguyen-Huu, Yann Grandmaître, Nicolas Teck

Deliveries:

StageKindLaunchSubmissionSupervisor
.tigRushDec 20, 2008Dec 21, 2008
TC-0Jan 05, 2009Jan 16, 2009 at 12:00
TC-1RushJan 16, 2009Jan 18, 2009 at 12:00
TC-2Feb 16, 2009Feb 25, 2009 at 23:42
TC-3 & TC-RRushFeb 27, 2009Mar 01, 2009 at 11:42
TC-4 & TC-EMar 02, 2009Mar 15, 2009 at 11:42
TC-5Mar 16, 2009Mar 25, 2009 at 23:42
TC-6Apr 23, 2009May 03, 2009 at 12:00
TC-7May 04, 2009May 17, 2009
TC-8May 18, 2009May 31, 2009
TC-9Jun 29, 2009Jul 12, 2009

Some of the noteworthy changes compared to Tiger 2010:

The Bistromatig

A new assignment is given for the .tig project: The Bistromatig. It consists in implementing an arbitrary-radix infinite-precision calculator. The project is an adaptation of the famous Bistromathic project, that used to be one of the first C assignments at EPITA in the Old Days. The name was borrowed from Douglas Adams’s invention from Life, the Universe and Everything.

TC-E

TC-E is a mandatory part of the TC-4 assignment.


Next: , Previous: , Up: History   [Contents][Index]

1.4.13 Tiger 2012

This is the eleventh year of the Tiger Project.

We have been helped by:

Assistants

Adrien Biarnes, Rémi Chaintron, Julien Delhommeau, Thomas Joly, Alexandre Laurent, Vincent Lechemin, Matthieu Martin

Deliveries:

StageKindLaunchSubmissionSupervisor
.tigRushDec 02, 2009Dec 04, 2009
TC-0Dec 11, 2009Dec 20, 2009
TC-1Jan 11, 2010Jan 17, 2010
TC-2Feb 01, 2010Feb 17, 2010
TC-3 & TC-RRushFeb 19, 2010Feb 26, 2010
TC-4 & TC-EFeb 22, 2010Mar 07, 2010
TC-5Mar 11, 2010Mar 22, 2010
TC-6Apr 19, 2010May 02, 2010
TC-7May 12, 2010May 25, 2010
TC-8May 25, 2010Jun 06, 2010
TC-9Jun 07, 2010Jun 12, 2010

Some of the noteworthy changes compared to Tiger 2011:

Shorter mandatory assignment

By decision of the department of studies, the mandatory assignment ends after TC-3.


Next: , Previous: , Up: History   [Contents][Index]

1.4.14 Tiger 2013

This is the twelfth year of the Tiger Project.

We have been helped by:

Assistants

Rémi Chaintron, Julien Grall

Deliveries:

StageKindLaunchSubmissionSupervisor
.tigRush
TC-0
TC-1
TC-2
TC-3 & TC-RRush
TC-4 & TC-E
TC-5
TC-6
TC-7
TC-8
TC-9

Some of the noteworthy changes compared to Tiger 2012:

Build overhaul

Silent rules, fewer Makefiles.

Bison Variant

The parser is storing objects on its stacks, not only pointers. Other recent Bison features are also used.


Next: , Previous: , Up: History   [Contents][Index]

1.4.15 Tiger 2014

This is the thirteenth year of the Tiger Project.

We have been helped by:

Assistants

Jonathan Aigrain, Jules Bovet, Hugo Damme, Michael Denoun, Julien Grall, Christophe Pierre, Paul Similowski

CSI students

Félix Abecassis

Deliveries for Ing1 students:

StageKindLaunchSubmissionSupervisor
.tigLabNov 16, 2011Nov 16, 2011
TC-0Dec 05, 2011Dec 18, 2011 at 23:42
TC-1RushJan 30, 2012 at 19:00Feb 02, 2012 at 18:42
TC-2Feb 02, 2012 at 19:00Feb 10, 2012 at 18:42
TC-3 & TC-RRushFeb 10, 2012 at 19:00Feb 12, 2012 at 11:42
TC-4 & TC-EFeb 20, 2012 at 19:00Mar 04, 2012 at 11:42
TC-5Mar 05, 2012 at 19:00Mar 18, 2012 at 11:42
TC-6Apr 23, 2012 at 19:00May 06, 2012 at 11:42
TC-7May 21, 2012 at 19:00Jun 03, 2012 at 11:42
TC-8Jun 04, 2012 at 19:00Jun 17, 2012 at 11:42
TC-9Jul 02, 2012 at 19:00Jul 15, 2012 at 11:42

Deliveries for AppIng1 students:

StageKindLaunchSubmissionSupervisor
.tigLabNov 19, 2011Nov 19, 2011
TC-0Dec 05, 2011Dec 18, 2011 at 23:42
TC-1Jan 28, 2012 at 10:00Feb 05, 2012 at 11:42
TC-2Feb 08, 2012 at 19:00Feb 17, 2012 at 18:42
TC-3 & TC-RRushFeb 17, 2012 at 19:00Feb 19, 2012 at 11:42

Some of the noteworthy changes compared to Tiger 2013:

The Logomatig

Due to time constraints, the Bistromatig assignment that has been previously used in the past three years for the .tig rush has been replaced by a 4-hour lab assignment: The Logomatig. This assignment is about implementing a small interpreter in Tiger for a subset of the Logo language. The name of this project is a tribute to Logo, Tiger and the Bistromathic (though there are very few calculations in it).

Introduction of C++ 2011 features

Since a new C++ standard has been released this year (September 11, 2011), we are introducing some of its features in the Tiger project, namely range-based for-loops, auto-typed variables, use of the nullptr literal constant, use of explicitly defaulted and deleted functions, template metaprogramming traits provided by the standard library, and use of consecutive right angle brackets in templates. This set of features has been chosen for it is supported both by GCC 4.6 and Clang 3.0.

Git

Git has replaced Subversion as version control system at EPITA. As of this year, we also provide the code with gaps through a public Git repository. This method makes the integration of the code provided at the beginning of each stage easier (with the exception of TC-0, which is still to be done from scratch).


Next: , Previous: , Up: History   [Contents][Index]

1.4.16 Tiger 2015

This is the fourteenth year of the Tiger Project.

We have been helped by:

Assistants

Laurent Gourvénec, Xavier Grand, Frédéric Lefort, Théophile Ranquet, Robin Wils

Deliveries for Ing1 students:

StageKindLaunchSubmissionSupervisor
.tigRushNov 23, 2012 at 18:42Nov 25, 2012 at 11:42
PTHL (TC-0)Dec 10, 2012 at 18:42Dec 23, 2012 at 11:42
TC-1RushFeb 11, 2013 at 18:42Feb 13, 2013 at 23:42
TC-2Feb 14, 2013 at 18:42Feb 24, 2013 at 11:42
TC-3 & TC-RMar 4, 2013 at 18:42Mar 10, 2013 at 11:42
TC-4 & TC-EMar 11, 2013 at 18:42Mar 24, 2013 at 11:42
TC-5Apr 22, 2013 at 18:42May 5, 2013 at 11:42
TC-6May 20, 2013 at 19:00Jun 2, 2013 at 11:42
TC-7Jun 2, 2013 at 19:00Jun 16, 2013 at 11:42
TC-8Jun 28, 2013 at 19:00Jul 11, 2013 at 11:42
TC-9Jul 12, 2013 at 19:00Jul 21, 2013 at 11:42

Deliveries for AppIng1 students:

StageKindLaunchSubmissionSupervisor
.tigRushNov 23, 2012 at 18:42Nov 25, 2012 at 11:42
PTHL (TC-0)Dec 10, 2012 at 18:42Dec 23, 2012 at 11:42
TC-1Feb 11, 2013 at 18:42Feb 17, 2013 at 11:42
TC-2Feb 18, 2013 at 18:42Feb 28, 2013 at 11:42
TC-3 & TC-RMar 11, 2013 at 18:42Mar 20, 2013 at 23:42

Some of the noteworthy changes compared to Tiger 2014:

TC-0 renamed as PTHL

In an effort to emphasize the link between the THL (Formal Languages) lecture and the first stage of the Tiger project, the latter has been renamed as PTHL (“THL Project”).

TC-3 is no longer a rush

TC-3 has not been a successful step among many students for several years now. It has been deemed by many of them as too complex to be understood and implemented in a couple of days. Therefore we decided to extend the time allotted to this stage so as to give students more chance to pass TC-3.

Extension of the mandatory assignment to TC-5

By decision of the department of studies, all Ing1 are required to work on the Tiger project up to TC-5. Subsequent steps remain optional.

Use of more C++ 2011 features

This year, explicit template instantiation declarations (extern template clauses) are introduced in the project to control template instantiations in lieu of *.hcc files. The set of C++ features used in the Tiger compiler is still supported by both GCC 4.6 and Clang 3.0.


Next: , Previous: , Up: History   [Contents][Index]

1.4.17 Tiger 2016

This is the fifteenth year of the Tiger Project.

We have been helped by:

Beta testers

Anthony Seure, Rémi Weng

Assistants

Aurélien Baud, Alexis Chotard, Baptiste Covolato, Arnaud Farbos, Laurent Gourvénec, Frédéric Lefort, Vincent Mirzaian-Dehkordi

Deliveries for Ing1 students:

StageKindLaunchSubmission
.tigRushNov 22, 2013 at 21:00Nov 24, 2013 at 11:42
PTHL (TC-0)Dec 9, 2013 at 18:42Dec 22, 2013 at 11:42
TC-1RushFeb 17, 2014 at 14:00Feb 19, 2014 at 23:42
TC-2Feb 20, 2014 at 09:00Mar 2, 2014 at 11:42
TC-3 & TC-RMar 3, 2014 at 19:00Mar 16, 2014 at 11:42
TC-4 & TC-EMar 14, 2014 at 19:00May 4, 2014 at 11:42
TC-5May 5, 2014 at 19:00May 24, 2014 at 23:42
TC-6May 23, 2014 at 19:00Jun 8, 2014 at 11:42
TC-7Jun 9, 2014 at 19:00Jun 22, 2014 at 11:42
TC-8Jul 7, 2014 at 19:00Jul 13, 2014 at 11:42
TC-9Jul 15, 2014 at 10:00Jul 20, 2014 at 11:42

Deliveries for AppIng1 students:

StageKindLaunchSubmission
.tigRushNov 22, 2013 at 21:00Nov 24, 2013 at 11:42
PTHL (TC-0)Dec 9, 2013 at 18:42Dec 22, 2013 at 11:42

Some of the noteworthy changes compared to Tiger 2015:

Use of even more C++ 2011 features

The compiler introduces the following C++ 2011 features:

The whole set of C++ features used in the Tiger compiler is supported by both GCC 4.8 and Clang 3.3.

C++ scanner

We introduce a C++ scanner this year, still generated by Flex, but implemented as classes. The management of the scanner’s inputs has been improved and responsibilities shared between the scanner and the driver (parse::TigerParser).

More Git Usage

Starting this year, we deliver code with gaps exclusively through the tc-base public Git repository. We no longer provide tarballs nor patches as a means to update students’ code bases.

Changes in the language regarding object-oriented constructs

The nil keyword has been made compatible with objects.

Style

Many stylistics changes have been performed, mainly to match the EPITA Coding Style.


Previous: , Up: History   [Contents][Index]

1.4.18 Tiger 2017

This is the sixteenth year of the Tiger Project.

We have been helped by:

Assistants

Aurélien Baud, Baptiste Covolato, Pierre De Abreu, Léo Ercolanelli, Arnaud Farbos, Axel Manuel, Vincent Mirzaian-Dehkordi, Matthieu Simon, Jérémie Simon

Deliveries for Ing1 students:

StageKindLaunchSubmission
.tigRushNov 21, 2014 at 21:00Nov 23, 2014 at 11:42
PTHL (TC-0)Dec 8, 2014 at 18:42Dec 21, 2014 at 11:42
TC-1RushFeb 4, 2015 at 22:00Feb 8, 2015 at 11:42
TC-2Feb 13, 2015 at 22:00Feb 22, 2015 at 11:42
TC-3 & TC-RFeb 23, 2015 at 22:00Mar 1, 2015 at 11:42
TC-4 & TC-EMar 9, 2015 at 19:00Mar 22, 2015 at 11:42
TC-5Avr 20, 2015 at 19:00May 3, 2015 at 11:42
TC-6May 25, 2015 at 19:00May 31, 2015 at 11:42
TC-7Jun 1, 2015 at 19:00Jun 7, 2015 at 11:42
TC-8Jun 8, 2015 at 19:00Jun 14, 2015 at 11:42
TC-9Jul 6, 2015 at 10:00Jul 19, 2015 at 11:42

Deliveries for AppIng1 students:

StageKindLaunchSubmission
.tigRushNov 21, 2014 at 21:00Nov 23, 2014 at 11:42
PTHL (TC-0)Dec 8, 2014 at 18:42Dec 21, 2014 at 11:42

Some of the noteworthy changes compared to Tiger 2016:

Use of even more C++ 2011 features

The compiler introduces the following C++ 2011 features:

The C++ features used in the Tiger compiler are supported by both GCC 4.8 and Clang 3.3.

Style

Many stylistics changes have been performed.


Next: , Previous: , Up: Top   [Contents][Index]

2 Instructions


Next: , Up: Instructions   [Contents][Index]

2.1 Interactions

Bear in mind that if you are writing, it is to be read, so pay attention to your reader.

The right place

Using mails is almost always wrong: first ask around you, then try to find the assistants in their lab, and finally post into epita.cours.compile. You need to have a very good reason to send a message to the assistants or to Akim or Roland, as it usually annoys us, which is not in your interest.

The news group epita.cours.compile is dedicated to the Compiler Construction lecture, the Tiger project, and related matters (e.g., assignments in Tiger itself). Any other material is off topic.

A meaningful title

Find a meaningful subject.

Don’t do thatDo this
Problem in TC-1Cannot generate location.hh
make checkmake check fails on test-ref
A legal content

Pieces of critical code (e.g., precedence section in the parser, the string handling in the scanner, or whatever you are supposed to find by yourself) are not to be published.

This includes the test cases. While posting a simple test case is tolerated, sending many of them, or simply one that addresses a specific common failure (e.g., some obscure cases for escapes) is strictly forbidden.

A complete content

If you experience a problem that you fail to solve, make a report as complete as possible: include pieces of code (unless the code is critical and shall not be published) and the full error message from the compiler/tool. The following text by Simon Tatham is enlightening; its scope goes way beyond the Tiger Project: How to Report Bugs Effectively. See also How not to go about a programming assignment, item “Be clever when using electronic mail”.

A legible content

Use French or English. Epitean is definitely not a language.

A pertinent content

Trolls are not welcome.


Next: , Previous: , Up: Instructions   [Contents][Index]

2.2 Rules of the Game

As any other assignment, the Tiger Project comes with its rules to follow.

Rule: Thou Shalt Not Copy Code
Rule: Thou Shalt Not Possess Thy Neighbor's Code

It is strictly forbidden to possess code that is not yours. You are encouraged to work with others, but don’t get a copy of their code. See How not to go about a programming assignment, for more hints on what will not be accepted.

Rule: Tests are part of the project
Rule: Do not copy tests or test frame works

Test cases and test engines development are parts of the Tiger Project. As such the same rules apply as for code.

Rule: If something is fishy, say it

If something illegal happened in the course of a stage, let us know, arrangements might be possible. If we find out, the rules will be strictly applied. It already happened that third year students have had to redo the Tiger Project because their code was found in another group: -42/20 is seldom benign.

Rule: Don't hesitate working with other groups

Don’t bother everybody instead of trying first. Conversely, once you did your best, don’t hesitate working with others.


Next: , Previous: , Up: Instructions   [Contents][Index]

2.3 Groups

Starting with TC-1, assignments are to be done by groups of four.

The first cause of failures to the Tiger project is human problems within the groups. We cannot stress too much the importance of constituting a good group of four people. The Tiger project starts way before your first line of code: it begins with the selection of your partners.

Here are a few tips, collected wisdom from the previous failures.

You work for yourself, not for grades

Yes, we know, when you’re a student grades are what matters. But close your eyes, make a step backwards, and look at yourself for a minute, from behind. You see a student, some sort of a larva, which will turn into a grownup. The larva stage lasts 3 to 4 years, while the hard working social insect is there for 40+ years: a 5% ratio without the internships. Three minutes out of an hour. These years are made to prepare you to the rest of your life, to provide you with what it takes to enjoy a lifelong success in jobs. So don’t waste these three minutes by just cheating, paying little attention to what you are given, or by just waiting for this to end. The opportunity to learn is a unique moment in life: treasure it, even if it hurts, if it’s hard, because you may well regret these three minutes for much of your life.

Start recruiting early

Making a team is not easy. Take the time to know the people, talk with them, and prepare your group way before beginning the project. The whole TC-0 is a test bed for you to find good partners.

Don’t recruit good lazy friends

If s/he’s lazy, you’ll have to scold her/him. If s/he’s a friend, that will be hard. Plus it will be even harder to report your problems to us.

Recruit people you can depend on

Trust should be your first criterion.

Members should have similar programming skills
Weak programmers should run away from skilled programmers

The worst “good idea” is “I’m a poor programmer, I should be in a group of skilled programmers: I will learn a lot from them”. Experience shows this is wrong. What actually happens is as follows.

At the first stage, the leader assigns you a task. You try and fail, for weeks. In the meanwhile, the other members teach you lots of facts, but (i) you can’t memorize everything and end up saying “hum hum” without having understood, and (ii) because they don’t understand what you don’t understand, they are often poor teachers. The day before the submission, the leader does your assignments to save the group. You learned nothing, or quite. Second stage: same beginning, you are left with your assignment, but the other members are now bothered by your asking questions: why should they answer, since you don’t understand what they say (remember: they are poor teachers because they don’t understand your problems), and you don’t seem to remember anything! The day before the submission, they do your work. From now on, they won’t even ask you for anything: “fixing” you is much more time consuming than just doing it by themselves. Oral examinations reveal you neither understand nor do anything, hence your grades are bad, and you win another round of first year...

Take our advice: if you have difficulties with programming, be with other people like you. Your chances are better together, and anyway you are allowed to ask for assistance from other groups.

Don’t mix repeaters with first year students

Repeaters have a much better understanding of the project than they think: they know its history, some parts of the code, etc. This will introduce a difference of skills from the beginning, which will remain till the end. It will result in the first year students having not participated enough to learn what was to be learned. Three first year students with one repeater is OK, but a different ratio is asking for troubles.

Don’t pick up old code

This item is especially intended to repeaters: you might be tempted to keep the code from last year, believing this will spare you some work. It may not be so. Indeed, every year the specifications and the provided code change, sometimes with dramatic impact on the whole project. Struggling with an old code base to meet the new standard is a long, error prone, and uninteresting work. You might spend more time trying to preserve your old code than what is actually needed to implement the project from scratch. Not to mention that of course the latter has a much stronger educational impact.

Diagnose and cure drifts

When a dysfunction appears, fix it, don’t let it grow. For instance, if a member never works in spite of the warnings, don’t cover him: he will have the whole group drown. It usually starts with one member making more work on Tiger, less on the rest of the curriculum, and then he gets tired all the time, with bad mood etc. Don’t walk that way: denounce the problems, send ultimatums to this person, and finally, warn the assistants you need to reconfigure your group.

Reconfigure groups when needed

Members can leave a group for many reasons: dropped EPITA, dropped Tiger, joined one of the schools’ laboratories, etc. If your group is seriously unbalanced (three skilled people is OK, otherwise be four), ask for a reconfiguration in the news.

Tiger is a part of your curriculum

Tiger should neither be 0 nor 100% of your curriculum: find the balance. It is not easy to find it, but that’s precisely one thing EPITA teaches: balancing overloads.


Next: , Previous: , Up: Instructions   [Contents][Index]

2.4 Coding Style

This section could have been named “Strong and Weak Requirements”, as it includes not only mandatory features from your compiler (memory management), but also tips and advice. As the captain Barbossa would put it, “actually, it’s more of a guideline than a rule.”


Next: , Up: Coding Style   [Contents][Index]

2.4.1 No Draft Allowed

The code you deliver must be clean. In particular, when some code is provided, and you have to fill in the blanks denoted by ‘FIXME: Some code has been deleted.’. Sometimes you will have to write the code from scratch.

In any case, dead code and dead comments must be removed. You are free to leave comments spotting places where you fixed a ‘FIXME:’, but never leave a fixed ‘FIXME:’ in your code. Nor any irrelevant comment.

The official compiler for this project, is GNU C++ Compiler, 4.8 or higher (see GCC).


Next: , Previous: , Up: Coding Style   [Contents][Index]

2.4.2 Use of Foreign Features

If, and only if, you already have enough fluency in C++ to be willing to try something wilder, then the following exception is made for you. Be warned: along the years the Tiger project was polished to best fit the typical epitean learning curve, trying to escape this curve is also taking a major risk. By the past, some students tried different approaches, and ended with unmaintainable pieces of code.

If you and your group are sure you can afford some additional difficulty (for additional benefits), then you may use the following extra tools. You have to warn the examiners that you use these tools. You also have to take care of harnessing configure.ac to make sure that what you need is available on the testing environment. Be also aware that you are likely to obtain less help from us if you use tools that we don’t master: You are on your own, but, hey!, that’s what you’re looking for, ain’t it?

The Loki Library

See Modern C++ Design, for more information about Loki.

The Boost Library

As provided by the unstable Debian packages libboost-*. See Boost.org.

Any Other Parser or Scanner Generator

If you dislike Flex and/or Bison but you already know how to use them, then you are welcome to use other technologies.

If you think about something not listed here, please send us your proposal; acceptance is required to use them.


Next: , Previous: , Up: Coding Style   [Contents][Index]

2.4.3 File Conventions

There are some strict conventions to obey wrt the files and their contents.

Rule: One class LikeThis per files like-this.*

Each class LikeThis is implemented in a single set of file named like-this.*. Note that the mixed case class names are mapped onto lower case words separated by dashes.

There can be exceptions, for instance auxiliary classes used in a single place do not need a dedicated set of files.

Rule: *.hh: Declarations

The *.hh should contain only declarations, i.e., prototypes, extern for variables etc. Inlined short methods are accepted when there are few of them, otherwise, create an *.hxx file. The documentation should be here too.

There is no good reason for huge objects to be defined here.

As much as possible, avoid including useless headers (GotW007, GotW034):

Rule: *.hxx: Inlined definitions

Some definitions should be loaded in different places: templates, inline functions etc. Declare and document them in the *.hh file, and implement them in the *.hxx file. The *.hh file last includes the *.hxx file, conversely *.hxx first includes *.hh. Read below.

Rule: *.cc: Definitions of functions and variables

Big objects should be defined in the *.cc file corresponding to the declaration/documentation file *.hh.

There are less clear cut cases between *.hxx and *.cc. For instance short but time consuming functions should stay in the *.cc files, since inlining is not expected to speed up significantly. As another example features that require massive header inclusions are better defined in the *.cc file.

As a concrete example, consider the accept methods of the AST classes. They are short enough to be eligible for an *.hxx file:

void
LetExp::accept(Visitor& v)
{
  v(*this);
}

We will leave them in the *.cc file though, since this way only the *.cc file needs to load ast/visitor.hh; the *.hh is kept short, both directly (its contents) and indirectly (its includes).

Rule: Explicit template instantiation

There are several strategies to compile templates. The most common strategy consists in leaving the code in a *.hxx file, and letting every user of the class template instantiate the code. While correct, this approach has several drawbacks:

To circumvent these problems, we may control template instantiations using explicit template instantiation definitions (available since C++ 1998) and declarations (introduced by C++ 2011).

This mechanism is compatible with the way templates are usually handled in the Tiger compiler, i.e., where both template declarations and definitions are accessible from the included header, though often indirectly (see above). We use the following two-fold strategy:

You will notice that both the approach and the syntax used here recall the ones used to declare and define global variables in C and C++.

We can further improve the previous design by factoring explicit instantiation code using the preprocessor.

/**
 ** \file temp/temp.hh
 ** \brief Fresh temps.
 */

#ifndef TEMP_TEMP_HH
# define TEMP_TEMP_HH

# include <misc/endomap.hh>

# ifndef MAYBE_EXTERN
#  define MAYBE_EXTERN extern
# endif

namespace temp
{
  struct Temp { /* ... */ };
}

// ...

namespace misc
{
  // Explicit template instantiation declaration.
  MAYBE_EXTERN template class endo_map<temp::Temp>;
}

#endif // !TEMP_TEMP_HH

File 2.3: temp-factored.hh

/**
 ** \file temp/temp.cc
 ** \brief temp::Temp.
 */

#define MAYBE_EXTERN
#include <temp/temp.hh>
#undef MAYBE_EXTERN

// ...

File 2.4: temp-factored.cc

Explicit template instantiation declarations (not definitions) are only available since C++ 2011. Before that, we used to introduce a fourth type of file, *.hcc: files that had to be compiled once for each concrete template parameter.

Rule: Guard included files (*.hh & *.hxx)

Use the preprocessor to ensure the contents of a file is read only once. This is critical for *.hh and *.hxx files that include one another.

One typically has:

/**
 ** \file sample/sample.hxx
 ** \brief Declaration of sample::Sample.
 **/

#ifndef SAMPLE_SAMPLE_HH
# define SAMPLE_SAMPLE_HH

// ...

#include <sample/sample.hxx>

#endif // !SAMPLE_SAMPLE_HH

File 2.5: sample/sample.hh

/**
 ** \file sample/sample.hxx
 ** \brief Inlined definition of sample::Sample.
 **/

#ifndef SAMPLE_SAMPLE_HXX
# define SAMPLE_SAMPLE_HXX

#include <sample/sample.hh>

// ...

#endif // !SAMPLE_SAMPLE_HXX

File 2.6: sample/sample.hxx

Rule: fwd.hh: forward declarations

Dependencies can be a major problem during big project developments. It is not acceptable to “recompile the world” when a single file changes. To fight this problem, you are encouraged to use fwd.hh files that contain simple forward declarations. Everything that defeat the interest of fwd.hh file must be avoided, e.g., including actual header files. These forward files should be included by the *.hh instead of more complete headers.

The expected benefit is manifold:

Consider for example ast/visitor.hh, which is included directly or indirectly by many other files. Since it needs a declaration of each AST node one could be tempted to use ast/all.hh which includes virtually all the headers of the ast module. Hence all the files including ast/visitor.hh will bring in the whole ast module, where the much shorter and much simpler ast/fwd.hh would suffice.

Of course, usually the *.cc files need actual definitions.

Rule: Module, namespace, and directory likethis

The compiler is composed of several modules that are dedicated to a set of coherent specific tasks (e.g., parsing, AST handling, register allocation etc.). A module name is composed of lower case letters exclusively, likethis, not like_this nor like-this. This module’s files are stored in the directory with the same name, which is also that of the namespace in which all the symbols are defined.

Contrary to file names, we do not use dashes to avoid clashes with Swig and namespace.

Rule: libmodule.*: Pure interface

The interface of the module module contains only pure functions: these functions should not depend upon globals, nor have side effects of global objects. Global variable are forbidden here.

Rule: tasks.*: Impure interface

Tasks are the place for side effects. That’s where globals such as the current AST, the current assembly program, etc., are defined and modified.


Next: , Previous: , Up: Coding Style   [Contents][Index]

2.4.4 Name Conventions

Rule: Stay out of reserved names

The standard reserves a number of identifier classes, most notably ‘_*’ [17.4.3.1.2]:

Each name that begins with an underscore is reserved to the implementation for use as a name in the global namespace.

Using ‘_*’ is commonly used for CPP guards (‘_FOO_HH_’), private members (‘_foo’), and internal functions (‘_foo ()’): don’t.

Rule: Name your classes LikeThis

Class should be named in mixed case; for instance Exp, StringExp, TempMap, InterferenceGraph etc. This applies to class templates. See CStupidClassName.

Rule: Name public members like_this

No upper case letters, and words are separated by an underscore.

Rule: Name private/protected members like_this_

It is extremely convenient to have a special convention for private and protected members: you make it clear to the reader, you avoid gratuitous warnings about conflicts in constructors, you leave the “beautiful” name available for public members etc. We used to write _like_this, but this goes against the standard, see Stay out of reserved names.

For instance, write:

class IntPair
{
public:
  IntPair(int first, int second)
    : first_(first)
    , second_(second)
  {
  }
protected:
  int first_, second_;
}

See CStupidClassName.

Rule: Name your typedef foo_type

When declaring a typedef, name the type foo_type (where foo is obviously the part that changes). For instance:

typedef std::map<const symbol, Entry_T> map_type;
typedef std::list<map_type> symtab_type;

We used to use foo_t, unfortunately this (pseudo) name space is reserved by POSIX.

Rule: Name the parent class super_type

It is often handy to define the type of “the” super class (when there is a single one); use the name super_type in that case. For instance most Visitors of the AST start with:

class TypeChecker: public ast::DefaultVisitor
{
public:
  typedef ast::DefaultVisitor super_type;
  using super_type::operator();
  // ...

(Such using clauses are subject to the current visibility modifier, hence the public beforehand.)

Rule: Hide auxiliary classes

Hide auxiliary/helper classes (i.e., classes private to a single compilation unit, not declared in a header) in functions, or in an anonymous namespace. Instead of:

struct Helper { ... };

void
doit()
{
  Helper h;
  ...
}

write:

namespace { struct Helper { ... }; }

void
doit()
{
  Helper h;
  ...
}

or

void
doit()
{
  struct Helper { ... } h;
  ...
}

The risk otherwise is to declare two classes with the same name: the linker will ignore one of the two silently. The resulting bugs are often difficult to understand.


Next: , Previous: , Up: Coding Style   [Contents][Index]

2.4.5 Use of C++ Features

Rule: Hunt Leaks

Use every possible means to release the resources you consume, especially memory. Valgrind can be a nice assistant to track memory leaks (see Valgrind). To demonstrate different memory management styles, you are invited to use different features in the course of your development: proper use of destructors for the AST, use of a factory for symbol, Temp etc., use of std::auto_ptr starting with the Translate module, and finally use of reference counting via smart pointers for the intermediate representation.

Rule: Hunt code duplication

Code duplication is your enemy: the code is less exercised (if there are two routines instead of one, then the code is run half of the time only), and whenever an update is required, you are likely to forget to update all the other places. Strive to prevent code duplication from sneaking into your code. Every C++ feature is good to prevent code duplication: inheritance, templates etc.

Rule: Prefer dynamic_cast of references

Of the following two snippets, the first is preferred:

const IntExp& ie = dynamic_cast<const IntExp&>(exp);
int val = ie.value_get();
const IntExp* iep = dynamic_cast<const IntExp*>(&exp);
assert(iep);
int val = iep->value_get();

While upon type mismatch the second aborts, the first throws a std::bad_cast: they are equally safe.

Rule: Use virtual methods, not type cases

Do not use type cases: if you want to dispatch by hand to different routines depending upon the actual class of objects, you probably have missed some use of virtual functions. For instance, instead of

bool
compatible_with(const Type& lhs, const Type& rhs)
{
  if (&lhs == &rhs)
    return true;
  if (dynamic_cast<Record*>(&lhs))
    if (dynamic_cast<Nil*>(&rhs))
       return true;
  if (dynamic_cast<Record*>(&rhs))
    if (dynamic_cast<Nil*>(&lhs))
       return true;
  return false;
}

write

bool
Record::compatible_with(const Type& rhs)
{
  return &rhs == this || dynamic_cast<const Nil*>(&rhs);
}

bool
Nil::compatible_with(const Type& rhs)
{
  return &rhs == this || dynamic_cast<const Record*>(&rhs);
}

bool
compatible_with(const Type& lhs, const Type& rhs)
{
  return lhs.compatible_with(rhs);
}
Rule: Use dynamic_cast for type cases

Did you read the previous item, “Use virtual methods, not type cases”? If not, do it now.

If you really need to write type dispatching, carefully chose between typeid and dynamic_cast. In the case of tc, where we sometimes need to down cast an object or to check its membership to a specific subclass, we don’t need typeid, so use dynamic_cast only.

They address different needs:

dynamic_cast for (sub-)membership, typeid for exact type

The semantics of testing a dynamic_cast vs. a comparison of a typeid are not the same. For instance, think of a class A with subclass B with subclass C; then compare the meaning of the following two snippets:

// Is `a' containing an object of exactly the type B?
bool test1 = typeid(a) == typeid(B);
// Is `a' containing an object of type B, or a subclass of B?
bool test2 = dynamic_cast<B*>(&a);
Non polymorphic entities

typeid works on hierarchies without vtable, or even builtin types (int etc.). dynamic_cast requires a dynamic hierarchy. Beware of typeid on static hierarchies; for instance consider the following code, courtesy from Alexandre Duret-Lutz:

#include <iostream>

struct A
{
  // virtual ~A() {};
};

struct B: A
{
};

int
main()
{
  A* a = new B;
  std::cout << typeid(*a).name() << std::endl;
}

it will “answer” that the typeid of ‘*a’ is A(!). Using dynamic_cast here will simply not compile4. If you provide A with a virtual function table (e.g., uncomment the destructor), then the typeid of ‘*a’ is B.

Compromising the future for the sake of speed

Because the job performed by dynamic_cast is more complex, it is also significantly slower that typeid, but hey! better slow and safe than fast and furious.

You might consider that today, a strict equality test of the object’s class is enough and faster, but can you guarantee there will never be new subclasses in the future? If there will be, code based dynamic_cast will probably behave as expected, while code based typeid will probably not.

More material can be found the chapter 8 of see Thinking in C++ Volume 2: Run-time type identification.

Rule: Use const references in arguments to save copies (EC22)

We use const references in arguments (and return value) where otherwise a passing by value would have been adequate, but expensive because of the copy. As a typical example, accessors ought to return members by const reference:

const Exp&
OpExp::lhs_get() const
{
  return lhs_;
}

Small entities can be passed/returned by value.

Rule: Use references for aliasing

When you need to have several names for a single entity (this is the definition of aliasing), use references to create aliases. Note that passing an argument to a function for side effects is a form of aliasing. For instance:

template <typename T>
void
swap(T& a, T& b)
{
  T c = a;
  a = b;
  b = c;
}
Rule: Use pointers when passing an object together with its management

When an object is created, or when an object is given (i.e., when its owner leaves the management of the object’s memory to another entity), use pointers. This is consistent with C++: new creates an object, returns it together with the responsibility to call delete: it uses pointers. For instance, note the three pointers below, one for the return value, and two for the arguments:

OpExp*
opexp_builder(OpExp::Oper oper, Exp* lhs, Exp* rhs)
{
  return new OpExp(oper, lhs, rhs);
}
Rule: Avoid class members (EC47)

More generally, “Ensure that non-local static objects are initialized before they’re used”, as reads the title of EC47.

Non local static objects (such as std::cout etc.) are initialized by the C++ system even before main is called. Unfortunately there is no guarantee on the order of their initialization, so if you happen to have a static object which initialization depends on that of another object, expect the worst. Fortunately this limitation is easy to circumvent: just use a simple Singleton implementation, that relies on a local static variable.

This is covered extensively in EC47.

Rule: Use foo_get, not get_foo

Accessors have standardized names: foo_get and foo_set.

There is an alternative attractive standard, which we don’t follow:

class Class
{
public:
  int foo();
  void foo(int foo);
private:
  int foo_;
}

or even

class Class
{
public:
  int foo();
  Class& foo(int foo);  // Return *this.
private:
  int foo_;
}

which enables idioms such as:

{
  Class obj;
  obj.foo(12)
     .bar(34)
     .baz(56)
     .qux(78)
     .quux(90);
}
Rule: Use dump as a member function returning a stream

You should always have a means to print a class instance, at least to ease debugging. Use the regular operator<< for standalone printing functions, but dump as a member function. Use this kind of prototype:

std::ostream& Class::dump(std::ostream& ostr [, ...]) const

where the ellipsis denote optional additional arguments. dump returns the stream.


Next: , Previous: , Up: Coding Style   [Contents][Index]

2.4.6 Use of STL

Rule: Specify comparison types for associative containers of pointers (ES20)

For instance, instead of declaring

typedef set::set<const Temp*> temp_set_type;

declare

/// Object function to compare two Temp*.
struct temp_compare
{
  bool
  operator()(const Temp* s1, const Temp* s2) const
  {
    return *s1 < *s2;
  }
};

typedef set::set<const Temp* , temp_compare> temp_set_type;

Scott Meyers mentions several good reasons, but leaves implicit a very important one: if you don’t, since the outputs will be based on the order of the pointers in memory, and since (i) this order may change if your allocation pattern changes and (ii) this order depends of the environment you run, then you cannot compare outputs (including traces). Needless to say that, at least during development, this is a serious misfeature.

Rule: Prefer standard algorithms to hand-written loops (ES43)

Using for_each, find, find_if, transform etc. is preferred over explicit loops. This is for (i) efficiency, (ii) correctness, and (iii) maintainability. Knowing these algorithms is mandatory for who claims to be a C++ programmer.

Rule: Prefer member functions to algorithms with the same names (ES44)

For instance, prefer ‘my_set.find(my_item)’ to ‘find (my_item, my_set.begin(), my_set.end())’. This is for efficiency: the former has a logarithmic complexity, versus... linear for the latter! You may find the Item 44 of Effective STL on the Internet.


Next: , Previous: , Up: Coding Style   [Contents][Index]

2.4.7 Matters of Style

The following items are more a matter of style than the others. Nevertheless, you are asked to follow this style.

Rule: 80 columns maximum

Stick to 80 column programming. As a matter of fact, stick to 76 or 78 columns most of the time, as it makes it easier to keep the diffs within the limits. And if you post/mail these diffs, people are likely to reply to the message, hence the suggestion of 76 columns, as for emails.

Rule: Order class members by visibility first

When declaring a class, start with public members, then protected, and last private members. Inside these groups, you are invited to group by category, i.e., methods, types, and members that are related should be grouped together. The motivation is that private members should not even be visible in the class declaration (but of course, it is mandatory that they be there for the compiler), and therefore they should be “hidden” from the reader.

This is an example of what should not be done:

class Foo
{
public:
  Foo(std::string, int);
  virtual ~Foo();

private:
  typedef std::string string_type;
public:
  std::string bar_get() const;
  void bar_set(std::string);
private:
  string_type bar_;

public:
  int baz_get() const;
  void baz_set(int);
private:
  int baz_;
}

rather, write:

class Foo
{
public:
  Foo(std::string, int);
  virtual ~Foo();

  std::string bar_get() const;
  void bar_set(std::string);

  int baz_get() const;
  void baz_set(int);

private:
  typedef std::string string_type;
  string_type bar_;
  int baz_;
}

and add useful Doxygen comments.

Rule: Keep superclasses on the class declaration line

When declaring a derived class, try to keep its list of superclasses on the same line. Leave a space at least on the right hand side of the colon. If there is not enough room to do so, leave the colon on the class declaration line (the opposite applies for constructor, see Put initializations below the constructor declaration).

class Derived: public Base
{
  // ...
};

/// Object function to compare two Temp*.
struct temp_ptr_less
{
  bool operator()(const Temp* s1, const Temp* s2) const;
};
Rule: Don't use inline in declarations

Use inline in implementations (i.e., *.hxx, possibly *.cc)), not during declarations (*.hh files).

Rule: Repeat virtual in subclass declarations

If a method was once declared virtual, it remains virtual. Nevertheless, as an extra bit of documentation to your fellow developers, repeat this virtual:

class Base
{
public:
  // ...
  virtual foo();
};

class Derived: public Base
{
public:
  // ...
  virtual foo();
};
Rule: Pointers and references are part of the type

Pointers and references are part of the type, and should be put near the type, not near the variable.

int* p;        // not `int *p;'
list& l;       // not `list &l;'
void* magic(); // not `void *magic();'
Rule: Do not declare many variables on one line

Use

int* p;
int* q;

instead of

int *p, *q;

The former declarations also allow you to describe each variable.

Rule: Leave no space between template name and effective parameters

Write

std::list<int> l;
std::pair<std::list<int>, int> p;

with a space after the comma. There is no need for a space between two closing ‘>’ (since C++ 2011):

std::list<std::list<int>> ls;

These rules apply for casts:

// Come on baby, light my fire.
int* p = static_cast<int*>(42);
Rule: Leave one space between TEMPLATE and formal parameters

Write

template <class T1, class T2>
struct pair;

with one space separating the keyword template from the list of formal parameters.

Rule: Leave no space between a function name and its argument(s), either formal or actual
int
foo(int n)
{
  return bar(n);
}

The ‘()’ operator is not a list of arguments.

class Foo
{
public:
  Foo();
  virtual ~Foo();
  bool operator()(int n);
};
Rule: Put initializations below the constructor declaration

Don’t put or initializations or constructor invocations on the same line as you declare the constructor. As a matter of fact, don’t even leave the colon on that line. Instead of ‘A::A(): B(), C()’, write either:

A::A()
  : B()
  , C()
{
}

or

A::A()
  : B(), C()
{
}

The rationale is that the initialization belongs more to the body of the constructor than its signature. And when dealing with exceptions leaving the colon above would yield a result even worse than the following.

A::A()
try
  : B()
  , C()
{
}
catch (...)
{
}

Previous: , Up: Coding Style   [Contents][Index]

2.4.8 Documentation Style

Rule: Write correct English

Nowadays most editors provide interactive spell checking, including for sources (strings and comments). For instance, see flyspell-mode in Emacs, and in particular the flyspell-prog-mode. To trigger this automatically, install the following in your ~/.emacs.el:

(add-hook 'c-mode-hook        'flyspell-prog-mode 1)
(add-hook 'c++-mode-hook      'flyspell-prog-mode 1)
(add-hook 'cperl-mode-hook    'flyspell-prog-mode 1)
(add-hook 'makefile-mode-hook 'flyspell-prog-mode 1)
(add-hook 'python-mode-hook   'flyspell-prog-mode 1)
(add-hook 'sh-mode-hook       'flyspell-prog-mode 1)

and so forth.

End comments with a period.

Rule: Be concise

For documentation as for any other kind of writing, the shorter, the better: hunt useless words. See The Elements of Style, for an excellent set of writing guidelines.

Here are a few samples of things to avoid:

Don’t document the definition instead of its object

Don’t write:

/// Declaration of the Foo class.
class Foo
{
  ...
};

Of course you’re documenting the definition of the entities! “Declaration of the” is totally useless, just use ‘/// Foo class’. But read bellow.

Don’t qualify obvious entity kinds

Don’t write:

/// Foo class.
class Foo
{
public:
  /// Construct a Foo object.
  Foo(Bar& bar)
  ...
};

It is so obvious that you’re documenting the class and the constructor that you should not write it down. Instead of documenting the kind of an entity (class, function, namespace, destructor...), document its goal.

/// Wrapper around Bar objects.
class Foo
{
public:
  /// Bind to \a bar.
  Foo(Bar& bar)
  ...
};
Rule: Use the Imperative

Use the imperative when documenting, as if you were giving order to the function or entity you are describing. When describing a function, there is no need to repeat “function” in the documentation; the same applies obviously to any syntactic category. For instance, instead of:

/// \brief Swap the reference with another.
/// The method swaps the two references and returns the first.
ref& swap(ref& other);

write:

/// \brief Swap the reference with another.
/// Swap the two references and return the first.
ref& swap(ref& other);

The same rules apply to ChangeLogs.

Rule: Use rebox.el to mark up paragraphs

Often one wants to leave a clear markup to separate different matters. For declarations, this is typically done using the Doxygen ‘\name ... \{ ... \}’ sequence; for implementation files use rebox.el (see rebox.el).

Rule: Write Documentation in Doxygen

Documentation is a genuine part of programming, just as testing. We use Doxygen (see Doxygen) to maintain the developer documentation of the Tiger Compiler. The quality of this documentation can change the grade.

Beware that Doxygen puts the first letter of documentation in upper case. As a result,

/// \file  ast/arrayexp.hh
/// \brief ast::ArrayExp declaration.

will not work properly, since Doxygen will transform ast::ArrayExp into ‘Ast::ArrayExp’, which will not be recognized as an entity name. As a workaround, write the slightly longer:

/// \file  ast/arrayexp.hh
/// \brief Declaration of ast::ArrayExp.

Of course, Doxygen documentation is not appropriate everywhere.

Rule: Document namespaces in lib*.hh files
Rule: Document classes in their *.hh file

There must be a single location, that’s our standard.

Rule: Use ‘\directive

Prefer backslash (‘\’) to the commercial at (‘@’) to specify directives.

Rule: Prefer C Comments for Long Comments

Prefer C comments (‘/** ... */’) to C++ comments (‘/// ...’). This is to ensure consistency with the style we use.

Rule: Prefer C++ Comments for One Line Comments

Because it is lighter, instead of

/** \brief Name of this program. */
extern const char* program_name;

prefer

/// Name of this program.
extern const char* program_name;

For instance, instead of

/* Construct an InterferenceGraph. */
InterferenceGraph(const std::string& name,
                  const assem::instrs_t& instrs, bool trace = false);

or

/** @brief Construct an InterferenceGraph.
 ** @param name    its name, hopefully based on the function name
 ** @param instrs  the code snippet to study
 ** @param trace   trace flag
 **/
InterferenceGraph(const std::string& name,
                  const assem::instrs_t& instrs, bool trace = false);

or

/// \brief Construct an InterferenceGraph.
/// \param name    its name, hopefully based on the function name
/// \param instrs  the code snippet to study
/// \param trace   trace flag
InterferenceGraph(const std::string& name,
                  const assem::instrs_t& instrs, bool trace = false);

write

/** \brief Construct an InterferenceGraph.
    \param name    its name, hopefully based on the function name
    \param instrs  the code snippet to study
    \param trace   trace flag
  */
InterferenceGraph(const std::string& name,
                  const assem::instrs_t& instrs, bool trace = false);

Next: , Previous: , Up: Instructions   [Contents][Index]

2.5 Tests

As stated in Rules of the Game, writing a test framework and tests is part of the exercise.

As a starting point, we provide a tarball containing a few Tiger files, see Given Test Cases. They are not enough: your test suite should be continually expanding.


Next: , Up: Tests   [Contents][Index]

2.5.1 Writing Tests

In three occasions tests are “easy” to write:

See Testing student-made compilers, for many hints on what tests you need to write.


Previous: , Up: Tests   [Contents][Index]

2.5.2 Generating the Test Driver

Unless your whole test infrastructure is embedded in a single file (which is not a good idea), we advise you to generate any script used to run your tests so that they can be run from a directory other than the source directory where they reside. This is especially useful to maintain several builds (e.g. with different compilers or compiler flags) in parallel (see the section on VPATH Builds in Automake’s manual) and when running ‘make distcheck’ (see the section on Checking the Distribution), as source and build directories are distinct in these circumstances.

The simplest way to generate a script is to rely on configure. For instance, the following line in configure.ac generates a script tests/testsuite from the input tests/testsuite.in, while performing variables substitutions (in particular ‘@srcdir@’ and similar variables):

AC_CONFIG_FILES([tests/testsuite], [chmod a=rx tests/testsuite])

The template file tests/testsuite.in can then leverage this information to find data in the source directory. E.g., if tests are located in the tests/ subdirectory of the top source directory, the beginning of tests/testsuite.in might look like this:

#! /bin/sh
# @configure_input@

# Where the tests can be found.
testdir="@abs_top_srcdir@/tests"

# ...

Another strategy to generate scripts is to use make, as suggested by Autoconf’s manual (see the section on Installation Directory Variables).


Next: , Previous: , Up: Instructions   [Contents][Index]

2.6 Submission

We use two kinds of project submissions in the project.


Previous: , Up: Instructions   [Contents][Index]

2.7 Evaluation

Some stages are evaluated only by a program, and others are evaluated both by humans, and a program.


Next: , Up: Evaluation   [Contents][Index]

2.7.1 Automated Evaluation

Each stage of the compiler will be evaluated by an automatic corrector. Soon after your work is submitted, the logs are available on the assistants’ intranet.

Automated evaluation enforces the requirements: you must stick to what is being asked. For instance, for TC-E it is explicitly asked to display something like:

var /* escaping */ i : int := 2

so if you display any of the following outputs

var i : int /* escaping */ := 2
var i /* escaping */ : int := 2
var /* Escapes */ i : int := 2

be sure to fail all the tests, even if the computation is correct.


Next: , Previous: , Up: Evaluation   [Contents][Index]

2.7.2 During the Examination

When you are defending your projects, here are a few rules to follow:

Don’t talk

Don’t talk unless you are asked to: when a person is asked a question, s/he is the only one to answer. You must not talk to each other either: often, when one cannot answer a question, the question is asked to another member. It is then obvious why the members of the group shall not talk.

Don’t touch the screen

Don’t touch my display! You have nice fingers, but I don’t need their prints on my screen.

Tell the truth

If there is something the examiner must know (someone did not work on the project at all, some files are coming from another group etc.), say it immediately, for, if we discover that by ourselves, you will be severely sanctioned.

Learn

It is explicitly stated that you can not have worked on a stage provided this was an agreement with the group. But it is also explicitly stated that you must have learned what was to be learned from that compiler stage, which includes C++ techniques, Bison and Flex mastering, object oriented concepts, design patterns and so forth.

Complain now!

If you don’t agree with the notation, say it immediately. Private messages about “this is unfair: I worked much more than bardec_f but his grade is better than mine” are thrown away.


Conversely, there is something we wish to make clear: examiners will probably be harsh (maybe even very harsh), but this does not mean they disrespect you, or judge you badly.

You are here to defend your project and knowledge, they are here to stress them, to make sure they are right. Learning to be strong under pressure is part of the exercise. Don’t burst into tears, react! Don’t be shy, that’s not the proper time: you are selling them something, and they will never buy something from someone who cries when they are criticizing his product.

You should also understand that human examination is the moment where we try to evaluate who, or what group, needs help. We are here to diagnose your project and provide solutions to your problems. If you know there is a problem in your project, but you failed to fix it, tell it to the examiner! Work with her/him to fix your project.


Next: , Previous: , Up: Evaluation   [Contents][Index]

2.7.3 Human Evaluation

The point of this evaluation is to measure, among other things:

the quality of the code

How clean it is, amount of code duplication, bad hacks, standards violations (e.g., ‘stderr’ is forbidden in proper C++ code) and so forth. It also aims at detecting cheaters, who will be severely punished (mark = -42).

the knowledge each member acquired

While we do not require that each member worked on a stage, we do require that each member (i) knows how the stage works and (ii) has perfectly understood the (C++, Bison etc.) techniques needed to implement the stage. Each stage comes with a set of goals (see PTHL Goals, for instance) on which you will be interrogated.


Examiners: the human grade.

The examiner should not take (too much) the automated tests into account to decide the mark: the mark is computed later, taking this into account, so don’t do it twice.

Examiners: broken tarballs.

If you fixed the tarball or made whatever modification, runmake distcheck’ again, and update the delivered tarball. Do not keep old tarballs, do not install them in a special place: just replace the first tarball with it, but say so in the ‘eval’ file.

The rationale is simple: only tarballs pass the tests, and every tarball must be able to pass the tests. If you don’t do that, then someone else will have to do it again.


Previous: , Up: Evaluation   [Contents][Index]

2.7.4 Marks Computation

Because the Tiger Compiler is a project with stages, the computation of the marks depends on the stages too. To spell it out explicitly:

A stage is penalized by bad results on tests performed for previous stages.

It means, for instance, that a TC-3 compiler will be exercised on TC-1, TC-2, and TC-3. If there are still errors on TC-1 and TC-2 tests, they will pessimize the result of TC-3 tests. The older the errors are, the more expensive they are.

As an example, here are the formulas to compute the global success rate of TC-3 and TC-5:

global-rate-TC-3 := rate-TC-3 * (+ 2 * rate-TC-1
                             + 1 * rate-TC-2) / 3
global-rate-TC-5 := rate-TC-5 * (+ 4 * rate-TC-1
                             + 3 * rate-TC-2
                             + 2 * rate-TC-3
                             + 1 * rate-TC-4) / 10

Because a project which fail half of the time is not a project that deserves half of 20, the global-rate is elevated to 1.7 before computing the mark:

mark-TC-3 := roundup(power(global-rate-TC-3, 1.7) * 20 - malus-TC-3, 1)

where ‘roundup(x, 1)’ is x rounded up to one decimal (‘roundup(15, 1) = 15’, ‘roundup(15.01, 1) = 15.1’).


When the project is also evaluated by a human, ‘power’ is not used. Rather, the success rate modifies the mark given by the examiner:

mark-TC-2 := roundup(eval-TC-2 * global-rate-TC-2 - malus-TC-2, 1)

Next: , Previous: , Up: Top   [Contents][Index]

3 Source Code


Next: , Up: Source Code   [Contents][Index]

3.1 Given Code

Starting with TC-1, code with gaps is provided through the tc-base public Git repository. We used to provide code through tarballs and patches before, but we only rely on Git now. This approach is the best one, as git merge is arguably simpler than patch and has other advantages (like preserving the execution bit of scripts, identifying the origin of every line of code using git blame, etc.). Each commit containing the contents of a new stage is labeled with a ‘class-tc-base-x.y’ tag.

Here is the recommended strategy to use this repository.

  1. At TC-1, subscribe to the repository, fetch its contents and integrate the given code using git merge with the commit labeled ‘2017-tc-base-1.0’ into your ‘master’ branch:
    $ git remote add tc-base https://gitlab.lrde.epita.fr/tiger/tc-base.git
    $ git fetch tc-base
    $ git merge 2017-tc-base-1.0
    

    Fix the conflicts and record the merge commit:

    $ git add src/tc.cc ...
    $ git commit
    
  2. For any subsequent stage m, all you will need to do is fetch the new commits from the ‘tc-base’ repository and merge the code given at stage m into yours (and of course, fix the conflicts). E.g.:
    $ git fetch tc-base
    $ git merge 2017-tc-base-m.0
    

Next: , Previous: , Up: Source Code   [Contents][Index]

3.2 Project Layout

This section describes the mandatory layout of the package.


Next: , Up: Project Layout   [Contents][Index]

3.2.1 The Top Level

AUTHORS.txt

In the top level of the distribution, there must be a file AUTHORS.txt which contents is as follows:

Fabrice Bardèche     <bardec_f@epita.fr>
Jean-Paul Sartre     <sartre_j@epita.fr>
Jean-Paul Deux       <deux_j@epita.fr>
Jean-Paul Belmondo   <belmon_j@epita.fr>

The group leader is first. Do not include emails other than those of EPITA. We repeat: give the ‘6_1@epita.fr’ address. Starting from TC-1, the file AUTHORS.txt is distributed thanks to the EXTRA_DIST variable in the top-level Makefile.am, but pay attention to the spelling.

ChangeLog

Optional. The list of the changes made in the compiler, with the dates and names of the people who worked on it. See the Emacs key binding ‘C-x 4 a’.

README.txt

Various free information.

NEWS.txt

Optional. Summary of changes introduced by each release.

lib/

This directory contains helping tools, that are not specific to the project.

src/

All the sources are in this directory.

tests/

Your own test suite. You should make it part of the project, and ship it like the rest of the package. Actually, it is abnormal not to have a test suite here.


Next: , Previous: , Up: Project Layout   [Contents][Index]

3.2.2 The build-aux Directory

File: bison++.in (build-aux/bin)

This is a wrapper around Bison, tailored to produce C++ parsers. Compared to bison, bison++ updates the output files only if changed. For a file such as location.hh, virtually included by the whole front-end, this is a big win.

Also, bison outputs ‘\file location.hh’ in Doxygen documentation, which clashes with ast/location.hh. bison++ changes this into ‘\file parse/location.hh’.

File: flex++.in (build-aux/bin)

A wrapper around Flex, to simplify and improve the generation of C++ scanners.

File: monoburg++.in (build-aux/bin)

Likewise for MonoBURG.

File: rebox.el (build-aux/)

This file provides two new Emacs functions, ‘M-x rebox-comment’ and ‘M-x rebox-region’. They build and maintain nice looking boxed comments in most languages. Once installed (read it for instructions), write a simple comment such as:

// Comments end with a period.

then move your cursor into this comment and press ‘C-u 2 2 3 M-q’ to get:

/*-----------------------------.
| Comments end with a period.  |
`-----------------------------*/

2 2 3’ specifies the style of the comment you want to build. Once the comment built, ‘M-q’ suffices to refill it. Run ‘C-u - M-q’ for an interactive interface.

File: tiger.el (build-aux)
File: panther.el (build-aux)

Theses files provide Emacs major modes for Tiger programs (*.tig) and Panther (“object-less” Tiger) programs (*.pan files). Read them to get installation instructions.

File: tiger-ftdetect.vim (build-aux)
File: tiger-syntax.vim (build-aux)

Vim scripts to detect and enable syntax hilighting for Tiger files.


Next: , Previous: , Up: Project Layout   [Contents][Index]

3.2.3 The lib Directory


Next: , Previous: , Up: Project Layout   [Contents][Index]

3.2.4 The lib/misc Directory

Convenient C++ tools.

File: contract.* (lib/misc/)

A useful improvement over cassert.

File: error.* (lib/misc/)

The class misc::error implements an error register. Because libraries are expected to be pure, they cannot issue error messages to the error output, nor exit with failure. One could pass call-backs (as functions or as objects) to set up error handling. Instead, we chose to register the errors in an object, and have the library functions return this register: it is up to the caller to decide what to do with these errors. Note also that direct calls to std::exit bypass stack unwinding. In other words, with std::exit (instead of throw) your application leaks memory.

An instance of misc::error can be used as if it were a stream to output error messages. It also keeps the current exit status until it is “triggered”, i.e., until it is thrown. Each module has its own error handler. For instance, the Binder has an error_ attribute, and uses it to report errors:

void
Binder::error(const ast::Ast& loc, const std::string& msg)
{
  error_ << misc::error::bind
         << loc.location_get() << ": " << msg << std::endl;
}

Then the task system fetches the local error handler, and merges it into the global error handler error (see common.*). Some tasks trigger the error handler: if errors were registered, an exception is raised to exit the program cleanly. The following code demonstrates both aspects.

void
bindings_compute()
{
  // bind::bind returns the local error handler.
  error << ::bind::bind(*ast::tasks::the_program);
  error.exit_on_error();
}
File: escape.* (lib/misc/)

This file implements a means to output string while escaping non printable characters. An example:

  std::cout << "escape(\"\111\") = " << escape("\"\111\"") << std::endl;

Understanding how escape works is required starting from TC-2.

File: flex-lexer.hh (lib/misc/)

The skeleton of the C++ scanner. Adapted from Flex’s FlexLexer.h and used as a replacement, thanks to flex++ (see flex++.in).

File: graph.* (lib/misc/)

This file contains a generic implementation of oriented and undirected graphs.

Understanding how graph works is required starting from TC-8.

File: indent.* (lib/misc/)

Exploiting regular std::ostream to produce indented output.

File: ref.* (lib/misc/)

Smart pointers implementing reference counting.

File: set.* (lib/misc/)

A wrapper around std::set that introduce convenient operators (operator+ and so forth).

File: scoped-map.* (lib/misc/)

The handling of misc::scoped_map<Key, Data>, generic scoped map, serving as a basis for symbol tables used by the Binder. misc::scoped_map maps a Key to a Data (that should ring a bell...). You are encouraged to implement something simple, based on stacks (see std::stack, or better yet, std::list) and maps (see std::map).

It must provide this interface:

void: put (const Key& key, const Data& value)

Associate value to key in the current scope.

Data: get (const Key& key) const

If key was associated to some Data in the open scopes, return the most recent insertion. Otherwise, if Data is a pointer type, then return the empty pointer, else throw a std::range_error. To implement this feature, see misc/traits.

std::ostream&: dump (std::ostream& ostr) const

Send the content of this table on ostr in a human-readable manner, and return the stream.

void: scope_begin ()

Open a new scope.

void: scope_end ()

Close the last scope, forgetting everything since the latest scope_begin().

File: symbol.* (lib/misc/)

In a program, the rule for identifiers is to be used many times: at least once for its definition, and once for each use. Just think about the number of occurrences of size_t in a C program for instance.

To save space one keeps a single copy of each identifier. This provides additional benefits: the address of this single copy can be used as a key: comparisons (equality or order) are much faster.

The class misc::symbol is an implementation of this idea. See the lecture notes, scanner.pdf. misc::symbol is based on misc::unique.

File: timer.* (lib/misc/)

A class that makes it possible to have timings of processes, similarly to gcc’s --time-report, or bison’s --report=time. It is used in the Task machinery, but can be used to provide better timings (e.g., separating the scanner from the parser).

File: traits.* (lib/misc/)

A simple traits to learn whether a type is a pointer type. See Traits, for more about traits.

File: unique.* (lib/misc/)

A generic class implementing the Flyweight design pattern. It maps identical objects to a unique reference.

File: variant.* (lib/misc/)

A wrapper over boost::variant (with two types) supporting conversion operators.


Next: , Previous: , Up: Project Layout   [Contents][Index]

3.2.5 The src Directory

File: common.hh (src/)

Used throughout the project.

File: tc (src/)

Your compiler.

File: tc.cc (src/)

Main entry. Called, the driver.


Next: , Previous: , Up: Project Layout   [Contents][Index]

3.2.6 The src/task Directory

No namespace for the time being, but it should be task. Delivered for TC-1. A generic scheme to handle the components of our compiler, and their dependencies.


Next: , Previous: , Up: Project Layout   [Contents][Index]

3.2.7 The src/parse Directory

Namespace ‘parse’. Delivered during TC-1.

File: scantiger.ll (src/parse/)

The scanner.

File: parsetiger.yy (src/parse/)

The parser.

File: position.hh (src/ast/)

Keeping track of a point (cursor) in a file.

File: location.hh (src/ast/)

Keeping track of a range (two cursors) in a (or two) file.

File: libparse.hh (src/ast/)

which prototypes what tc.cc needs to know about the module ‘parse’.


Next: , Previous: , Up: Project Layout   [Contents][Index]

3.2.8 The src/ast Directory

Namespace ‘ast’, delivered for TC-2. Implementation of the abstract syntax tree. The file ast/README gives an overview of the involved class hierarchy.

File: location.hh (src/ast/)

Imports Bison’s parse::location.

File: visitor.hh (src/ast/)

Abstract base class of the compiler’s visitor hierarchy. Actually, it defines a class template GenVisitor, which expects an argument which can be either misc::constify_traits or misc::id_traits. This allows to define two parallel hierarchies: ConstVisitor and Visitor, similar to iterator and const_iterator.

The understanding of the template programming used is not required at this stage as it is quite delicate, and goes far beyond your (average) current understanding of templates.

File: default-visitor.* (src/ast/)

Implementation of the GenDefaultVisitor class template, which walks the abstract syntax tree, doing nothing. This visitor do not define visit methods for nodes related to object-oriented constructs (classes, methods, etc.); thus it is an abstract class, and is solely used as a basis for deriving other visitors. It is instantiated twice: GenDefaultVisitor<misc::constify_traits> and GenDefaultVisitor<misc::id_traits>.

File: non-object-visitor.* (src/ast/)

Implementation of the GenNonObjectVisitor class template, which walks the abstract syntax tree, doing nothing, but aborting on nodes related to object-oriented constructs (classes, methods, etc.). This visitor is abstract and is solely used as a basis for deriving other visitors (see TC-2 FAQ). It is instantiated twice: GenNonObjectVisitor<misc::constify_traits> and GenNonObjectVisitor<misc::id_traits>.

File: object-visitor.* (src/ast/)

Implementation of the GenObjectVisitor class template, which walks object-related nodes of an abstract syntax tree, doing nothing. This visitor is abstract and is solely used as a basis for deriving other visitors. It is instantiated twice: GenObjectVisitor<misc::constify_traits> and GenObjectVisitor<misc::id_traits>.

File: pretty-printer.* (src/ast/)

The PrettyPrinter class, which pretty-prints an AST back into Tiger concrete syntax.

File: typable.* (src/ast/)

This class is not needed before TC-4 (see TC-4).

Auxiliary class from which typable AST node classes should derive. It has a simple interface made to manage a pointer to the type of the node:

void: type_set (const type::Type*)
const: type::Type* type_get () const

Accessors to the type of this node.

void: accept (ConstVisitor& v) const
void: accept (Visitor& v)

These methods are abstract, as in ast::Ast.

File: type-constructor.* (src/ast/)

This class is not needed before TC-4 (see TC-4).

Auxiliary class from which should derive AST nodes that construct a type (e.g., ast::ArrayTy). Its interface is similar to that of ast::Typable with one big difference: ast::TypeConstructor is responsible for de-allocating that type.

void: created_type_set (const type::Type*)
const: type::Type* created_type_get () const

Accessors to the created type of this node.

void: accept (ConstVisitor& v) const
void: accept (Visitor& v)

It is convenient to be able to visit these, but it is not needed.

File: escapable.* (src/ast/)

This class is needed only for TC-E (see TC-E).

Auxiliary class from which AST node classes that denote the declaration of variables and formal arguments should derive. Its role is to encode a single Boolean value: whether the variable escapes or not. The natural interface includes escape_get and escape_set methods.


Next: , Previous: , Up: Project Layout   [Contents][Index]

3.2.9 The src/bind Directory

Namespace ‘bind’. Binding uses to definitions.

File: binder.* (src/bind/)

The bind::Binder visitor. Binds uses to definitions (works on syntax without object).

File: renamer.* (src/bind/)

The bind::Renamer visitor. Renames every identifier to a unique name (works on syntax without object).


Next: , Previous: , Up: Project Layout   [Contents][Index]

3.2.10 The src/escapes Directory

Namespace ‘escapes’. Compute the escaping variables.

File: escapes-visitor.* (src/escapes/)

The escapes::EscapesVisitor.


Next: , Previous: , Up: Project Layout   [Contents][Index]

3.2.11 The src/type Directory

Namespace ‘type’. Type checking.

File: libtype.* (src/type/)

The interface of the Type module. It exports a single procedure, types_check.

File: types.hh (src/type/)
File: type.* (src/type/)
File: array.* (src/type/)
File: attribute.* (src/type/)
File: builtin-types.* (src/type/)
File: class.* (src/type/)
File: field.* (src/type/)
File: function.* (src/type/)
File: method.* (src/type/)
File: named.* (src/type/)
File: record.* (src/type/)

The definitions of all the types. Built-in types (Int, String, Nil and Void) are defined in src/type/builtin-types.*.

File: type-checker.* (src/type/)

The type::TypeChecker visitor. Computes the types of an AST and adds type labels to the corresponding nodes (works on syntax without object).


Next: , Previous: , Up: Project Layout   [Contents][Index]

3.2.12 The src/object Directory

File: binder.* (src/object/)

The object::Binder visitor. Binds uses to definitions (works on syntax with objects). Inherits from bind::Binder.

File: type-checker.* (src/object/)

The object::TypeChecker visitor. Computes the types of an AST and adds type labels to the corresponding nodes (works on syntax with objects). Inherits from type::TypeChecker.

File: renamer.* (src/object/)

The object::Renamer visitor. Renames every identifier to a unique name (works on syntax with objects), and keeps a record of the names of the renamed classes. Inherits from bind::Renamer.

File: desugar-visitor.* (src/object/)

The object::DesugarVisitor visitor. Transforms an AST with objects into an AST without objects.


Next: , Previous: , Up: Project Layout   [Contents][Index]

3.2.13 The src/overload Directory

Namespace ‘overload’. Overloading function support.


Next: , Previous: , Up: Project Layout   [Contents][Index]

3.2.14 The src/astclone Directory

File: cloner.* (src/astclone)

The astclone::Cloner visitor. Duplicate an AST. This copy is purely structural: the clone is similar to the original tree, but any existing binding or type information is not preserved.


Next: , Previous: , Up: Project Layout   [Contents][Index]

3.2.15 The src/desugar Directory

File: desugar-visitor.* (src/desugar)

The desugar::DesugarVisitor visitor. Remove constructs that can be considered as syntactic sugar using other language constructs. For instance, turn for loops into while loops, string comparisons into function calls. Inherits from astclone::Cloner, so the desugared AST is a modified copy of the initial tree.

File: bounds-checking-visitor.* (src/desugar)

The desugar::BoundsCheckingVisitor visitor. Add dynamic array bounds checks while duplicating an ast. Inherits from astclone::Cloner, so the result is a modified copy of the input AST.


Next: , Previous: , Up: Project Layout   [Contents][Index]

3.2.16 The src/inlining Directory

File: inliner.* (src/inlining)

The desugar::Inliner visitor. Perform inline expansion of functions.

File: pruner.* (src/inlining)

The desugar::Pruner visitor. Prune useless function declarations within an ast.


Next: , Previous: , Up: Project Layout   [Contents][Index]

3.2.17 The src/temp Directory

Namespace temp, delivered for TC-5.

File: identifier.* (src/temp/)

Provides the class template Identifier built upon boost::variant and used to implement temp::Temp and temp::Label. Also contains the generic IdentifierCompareVisitor, used to compare two identifiers.

Identifier handles maps of Identifiers. For instance, the Temp t5 might be allocated the register $t2, in which case, when outputting t5, we should print $t2. Maps stored in the xalloc’d slot Identifier::map of streams implements such a correspondence. In addition, the operator<< of the Identifier class template itself "knows" when such a mapping is active, and uses it.

File: label.* (src/temp/)

We need labels for jumps, for functions, strings etc. Implemented as an instantiation of the temp::Identifier scheme.

File: temp.* (src/temp/)

So called temporaries are pseudo-registers: we may allocate as many temporaries as we want. Eventually the register allocator will map those temporaries to either an actual register, or it will allocate a slot in the activation block (aka frame) of the current function. Implemented as an instantiation of the temp::Identifier scheme.

File: temp-set.* (src/temp/)

A set of temporaries, along with its operator<<.


Next: , Previous: , Up: Project Layout   [Contents][Index]

3.2.18 The src/tree Directory

Namespace tree, delivered for TC-5. The implementation of the intermediate representation. The file tree/README should give enough explanations to understand how it works.

Reading the corresponding explanations in Appel’s book is mandatory.

It is worth noting that contrary to A. Appel, just as we did for ast, we use n-ary structures. For instance, where Appel uses a binary seq, we have an n-ary seq which allows us to put as many statements as we want.

To avoid gratuitous name clashes, what Appel denotes exp is denoted sxp (Statement Expression), implemented in translate::Sxp.

Please, pay extra attention to the fact that there are temp::Temp used to create unique temporaries (similar to misc::symbol), and tree::Temp which is the intermediate representation instruction denoting a temporary (hence a tree::Temp needs a temp::Temp). Similarly, on the one hand, there is temp::Label which is used to create unique labels, and on the other hand there are tree::Label which is the IR statement to define to a label, and tree::Name used to refer to a label (typically, a tree::Jump needs a tree::Name which in turn needs a temp::Label).

File: fragment.* (src/tree/)

It implements tree::Fragment, an abstract class, tree::DataFrag to store the literal strings, and tree::ProcFrag to store the routines.

File: fragments.* (src/tree/)

Lists of tree::Fragment.

File: visitor.* (src/tree/)

Implementation of tree::Visitor and tree::ConstVisitor to implement function objects on tree::Fragments. In other words, these visitors implement polymorphic operations on tree::Fragment.


Next: , Previous: , Up: Project Layout   [Contents][Index]

3.2.19 The src/frame Directory

Namespace ‘frame’, delivered for TC-5.

File: access.* (src/frame/)

An Access is a location of a variable: on the stack, or in a temporary.

File: frame.* (src/frame/)

A Frame knows only what are the “variables” it contains.


Next: , Previous: , Up: Project Layout   [Contents][Index]

3.2.20 The src/translate Directory

Namespace ‘translate’. Translation to intermediate code translation. It includes:

File: libtranslate.* (src/translate/)

The interface.

File: access.* (src/translate/)

Static link aware versions of level::Access.

File: level.* (src/translate/)

translate::Level are wrappers frame::Frame that support the static links, so that we can find an access to the variables of the “parent function”.

File: exp.hh (src/translate/)

Implementation of translate::Ex (expressions), Nx (instructions), Cx (conditions), and Ix (if) shells. They wrap tree::Tree to delay their translation until the actual use is known.

File: translation.hh (src/translate/)

functions used by the translate::Translator to translate the AST into HIR. For instance, it contains ‘Exp* simpleVar(const Access& access, const Level& level)’, ‘Exp* callExp(const temp::Label& label, std::list<Exp*> args)’ etc. which are routines that produce some ‘Tree::Exp’. They handle all the unCx etc. magic.

File: translator.hh (src/translate/)

Implements the class ‘Translator’ which performs the IR generation thanks to translation.hh. It must not be polluted with translation details: it is only coordinating the AST traversal with the invocation of translation routines. For instance, here is the translation of a ‘ast::SimpleVar’:

virtual void operator()(const SimpleVar& e)
{
  exp_ = simpleVar(*var_access_[e.def_get()], *level_);
}

Next: , Previous: , Up: Project Layout   [Contents][Index]

3.2.21 The src/canon Directory

Namespace canon.


Next: , Previous: , Up: Project Layout   [Contents][Index]

3.2.22 The src/assem Directory

Namespace assem, delivered for TC-7.

This directory contains the implementation of the Assem language: yet another intermediate representation that aims at encoding an assembly language, plus a few needed features so that register allocation can be performed afterward. Given in full.

File: instr.* (src/assem/)
File: move.* (src/assem/)
File: oper.* (src/assem/)
File: label.* (src/assem/)
File: comment.* (src/assem/)

Implementation of the basic types of assembly instructions.

File: fragment.* (src/assem/)

Implementation of assem::Fragment, assem::ProcFrag, and assem::DataFrag. They are very similar to tree::Fragment: aggregate some information that must remain together, such as a frame::Frame and the instructions (a list of assem::Instr).

File: visitor.hh (src/assem/)

The root of assembler visitors.

File: layout.hh (src/assem/)

A pretty printing visitor for assem::Fragment.

File: libassem.* (src/assem/)

The interface of the module, and its implementation.


Next: , Previous: , Up: Project Layout   [Contents][Index]

3.2.23 The src/target Directory

Namespace target, delivered for TC-7. Some data on the back end.

File: cpu.* (src/target/)

Description of a CPU: everything about its registers, and its word size.

File: target.* (src/target/)

Description of a target (language): its CPU, its assembly (target::Assembly), and it translator (target::Codegen).

File: assembly.* (src/target/)

The abstract class target::Assembly, the interface for elementary assembly instructions generation.

File: codegen.* (src/target/)

The abstract class target::Codegen, the interface for all our back ends.

Directory: mips (src/target/)
Directory: ia32 (src/target/)

The instruction selection per se split into a generic part, and a target specific (MIPS and IA-32) part. See src/target/mips, and src/target/ia32.

File: libtarget.* (src/target/)

Converting tree::Fragments into assem::Fragments.

File: tiger-runtime.c (src/target/)

This is the Tiger runtime, written in C, based on Andrew Appel’s runtime.c. The actual runtime.s file for MIPS was written by hand, but the IA-32 was a compiled version of this file. It should be noted that:

Strings

Strings are implemented as 4 bytes to encode the length, and then a 0-terminated à la C string. The length part is due to conformance to the Tiger Reference Manual, which specifies that 0 is a regular character that can be part of the strings, but it is nevertheless terminated by 0 to be compliant with SPIM/Nolimips’ print syscall. This might change in the future.

Special Strings

There are some special strings: 0 and 1 character long strings are all implemented via a singleton. That is to say there is only one allocated string ‘""’, a single ‘"1"’ etc. These singletons are allocated by main. It is essential to preserve this invariant/convention in the whole runtime.

strcmp vs. stringEqual

We don’t know how Appel wants to support ‘"bar" < "foo"’ since he doesn’t provide strcmp. We do. His implementation of equality is more efficient than ours though, since he can decide just be looking at the lengths. That could be improved in the future...

main

The runtime has some initializations to make, such as strings singletons, and then calls the compiled program. This is why the runtime provides main, and calls tc_main, which is the “main” that your compiler should provide.


Next: , Previous: , Up: Project Layout   [Contents][Index]

3.2.24 The src/target/mips Directory

Namespace target::mips, delivered for TC-7. Code generation for MIPS R2000.

File: cpu.* (src/target/mips/)

The description of the MIPS (actually, SPIM/Nolimips) CPU.

File: spim-assembly.* (src/target/mips/)

Our assembly language (syntax, opcodes and layout); it abstracts the generation of MIPS R2000 instructions. target::mips::SpimAssembly derives from target::Assembly.

File: spim-layout.* (src/target/mips/)

How MIPS (and SPIM/Nolimips) fragments are to be displayed. In other words, that’s where the (global) syntax of the target assembly file is selected.

File: codegen.* (src/target/mips/)
File: tree.brg (src/target/mips/)
File: exp.brg (src/target/mips/)
File: binop.brg (src/target/mips/)
File: call.brg (src/target/mips/)
File: temp.brg (src/target/mips/)
File: mem.brg (src/target/mips/)
File: stm.brg (src/target/mips/)
File: move.brg (src/target/mips/)
File: move_load.brg (src/target/mips/)
File: move_store.brg (src/target/mips/)
File: cjump.brg (src/target/mips/)
File: prologue.hh (src/target/mips/)
File: epilogue.cc (src/target/mips/)

A translator from LIR to ASSEM using the MIPS R2000 instruction set defined by target::mips::SpimAssembly. It is implemented as a dynamic programming algorithm generated by MonoBURG from a set of brg files. target::mips::Codegen derives from target::Codegen.

File: target.* (src/target/mips/)

The main back end, based on a MIPS CPU and a MIPS code generator.

File: runtime.s (src/target/mips/)
File: runtime.cc (src/target/mips/)

The Tiger runtime in MIPS assembly language: print etc. The C++ file runtime.cc is built from runtime.s: do not edit the former. See src/target, tiger-runtime.c.


Next: , Previous: , Up: Project Layout   [Contents][Index]

3.2.25 The src/target/ia32 Directory

Namespace target::ia32, delivered for TC-7. Code generation for IA-32. This is not part of the student project, but it is left to satisfy their curiosity. In addition its presence is a sane invitation to respect the constraints of a multi-back-end compiler.

File: cpu.* (src/target/ia32)

Description of the i386 CPU.

File: gas-assembly.* (src/target/ia32/)

The IA-32 assembly language (syntax, opcodes and layout); it abstracts the generation of IA-32 instructions using the GNU Assembler (Gas) syntax. target::ia32::GasAssembly derives from target::Assembly.

File: gas-layout.* (src/target/ia32/)

How IA-32 fragments are to be displayed. In other words, that’s where the (global) syntax of the target assembly file is selected.

File: codegen.* (src/target/ia32/)
File: tree.brg (src/target/ia32/)
File: exp.brg (src/target/ia32/)
File: binop.brg (src/target/ia32/)
File: call.brg (src/target/ia32/)
File: temp.brg (src/target/ia32/)
File: mem.brg (src/target/ia32/)
File: stm.brg (src/target/ia32/)
File: move.brg (src/target/ia32/)
File: move_load.brg (src/target/ia32/)
File: move_store.brg (src/target/ia32/)
File: cjump.brg (src/target/ia32/)
File: prologue.hh (src/target/ia32/)
File: epilogue.cc (src/target/ia32/)

A translator from LIR to ASSEM using the IA-32 instruction set defined by target::ia32::GasAssembly. It is implemented as a dynamic programming algorithm generated by MonoBURG from a set of brg files. target::ia32::Codegen derives from target::Codegen.

File: target.* (src/target/ia32/)

The IA-32 back-end, based on an IA-32 CPU and an IA-32 code generator.

File: runtime-gnu-linux.s (src/target/ia32/)
File: runtime-freebsd.s (src/target/ia32/)

The GNU/Linux and FreeBSD Tiger runtimes in IA-32 assembly language: print etc. The C++ files runtime-gnu-linux.cc and runtime-freebsd.cc are built from runtime-gnu-linux.s and runtime-freebsd.s: do not edit the former. See src/target, tiger-runtime.c.


Next: , Previous: , Up: Project Layout   [Contents][Index]

3.2.26 The src/liveness Directory

Namespace liveness, delivered for TC-8.

File: flowgraph.* (src/liveness/)

FlowGraph implementation.

File: test-flowgraph.cc (src/liveness/)

FlowGraph test.

File: liveness.* (src/liveness/)

Computing the live-in and live-out information from the FlowGraph.

File: interference-graph.* (src/liveness/)

Computing the InterferenceGraph from the live-in/live-out information.


Previous: , Up: Project Layout   [Contents][Index]

3.2.27 The src/regalloc Directory

Namespace regalloc, register allocation, delivered for TC-9.

File: color.* (src/regalloc/)

Coloring an interference graph.

File: regallocator.* (src/regalloc/)

Repeating the coloration until it succeeds (no spills).

File: libregalloc.* (src/regalloc/)

Removing useless moves once the register allocation performed, and allocating the register for fragments.

File: test-regalloc.cc (src/regalloc/)

Exercising this.


Previous: , Up: Source Code   [Contents][Index]

3.3 Given Test Cases

We provide a few test cases: you must write your own tests. Writing tests is part of the project. Do not just copy test cases from other groups, as you will not understand why they were written.

The initial test suite is available for download at tests.tgz. It contains the following directories:

good

These programs are correct.

scan

These programs have lexical errors.

parse

These programs have syntax errors.

type

These programs contain type mismatches.


Next: , Previous: , Up: Top   [Contents][Index]

4 Compiler Stages

The compiler will be written in several steps, described below.


Next: , Up: Compiler Stages   [Contents][Index]

4.1 Stage Presentation

The following sections adhere to a standard layout in order to present each stage n:

Introduction

The first few lines specify the last time the section was updated, the class for which it is written, and the submission dates. It also briefly describes the stage.

Tn Goals, What this stage teaches

This section details the goals of the stage as a teaching exercise. Be sure that examiners will make sure you understood these points. They also have instructions to ask questions about previous stages.

Tn Samples, See Tn work

Actual examples generated from the reference compilers are exhibited to present and “specify” the stage.

Tn Given Code, Explanation on the provided code

This subsection points to the on line material we provide, introduces its components, quickly presents their designs and so forth. Check out the developer documentation of the Tiger Compiler for more information, as the code is (hopefully) properly documented.

Tn Code to Write, Explanation on what you have to write

But of course, this code is not complete; this subsection provides hints on what is expected, and where.

Tn Options, Want some more?

During some stages, those who find the main task too easy can implement more features. These sections suggest possible additional features.

Tn FAQ, Questions not to ask

Each stage sees a blossom of new questions, some of which being extremely pertinent. We selected the most important ones, those that you should be aware of, contrary to many more questions that you ought to find and ask yourselves. These sections answer this few questions. And since they are already answered, you should not ask them...

Tn Improvements, Other Designs

The Tiger Compiler is an instructional project the audience of which is learning C++. Therefore, although by the end of the development, in the latter stages, we can expect able C++ programmers, most of the time we have to refrain from using advanced designs, or intricate C++ techniques. These sections provide hints on what could have been done to improve the stage. You can think of these sections as material you ought to read once the project is over and you are a grown-up C++ programmer.


Next: , Previous: , Up: Compiler Stages   [Contents][Index]

4.2 PTHL (TC-0), Naive Scanner and Parser

2017-PTHL submission is Sunday, December 21st 2014 at 11:42.

This section has been updated for EPITA-2017 on 2014-02-18.

TC-0 is a weak form of TC-1: the scanner and the parser are written, but the framework is simplified (see TC-1 Code to Write). The grammar is also simpler: object-related productions are not to be supported at this stage (see PTHL Improvements). No command line option is supported.


Next: , Up: PTHL (TC-0)   [Contents][Index]

4.2.1 PTHL Goals

Things to learn during this stage that you should remember:


Next: , Previous: , Up: PTHL (TC-0)   [Contents][Index]

4.2.2 PTHL Samples

First, please note that all the samples, including in this section, are generated with a TC-1+ compliant compiler: its behavior differs from that of a TC-0 compiler. In particular, for the time being, forget about the options (-X and --parse).

Running TC-0 basically consists in looking at exit values:

print("Hello, World!\n")

File 4.1: simple.tig

$ tc simple.tig

Example 4.1: tc simple.tig

The following example demonstrates the scanner and parser tracing. The glyphs “error→” and “⇒” are typographic conventions to specify respectively the standard error stream and the exit status. They are not part of the output per se.

$ SCAN=1 PARSE=1 tc -X --parse simple.tig
error→Parsing file: simple.tig
error→Starting parse
error→Entering state 0
error→Reading a token: --(end of buffer or a NUL)
error→--accepting rule at line 174("print")
error→Next token is token "identifier" (simple.tig:1.1-5: print)
error→Shifting token "identifier" (simple.tig:1.1-5: print)
error→Entering state 2
error→Reading a token: --accepting rule at line 94("(")
error→Next token is token "(" (simple.tig:1.6: )
error→Reducing stack 0 by rule 104 (line 678):
error→   $1 = token "identifier" (simple.tig:1.1-5: print)
error→-> $$ = nterm funid (simple.tig:1.1-5: print)
error→Entering state 40
error→Next token is token "(" (simple.tig:1.6: )
error→Shifting token "(" (simple.tig:1.6: )
error→Entering state 92
error→Reading a token: --accepting rule at line 175(""")
error→--accepting rule at line 245("Hello, World!")
error→--accepting rule at line 232("\n")
error→--accepting rule at line 207(""")
error→Next token is token "string" (simple.tig:1.7-23: Hello, World!
error→)
error→Shifting token "string" (simple.tig:1.7-23: Hello, World!
error→)
error→Entering state 1
error→Reducing stack 0 by rule 22 (line 367):
error→   $1 = token "string" (simple.tig:1.7-23: Hello, World!
error→)
error→-> $$ = nterm exp (simple.tig:1.7-23: "Hello, World!\n")
error→Entering state 140
error→Reading a token: --accepting rule at line 95(")")
error→Next token is token ")" (simple.tig:1.24: )
error→Reducing stack 0 by rule 53 (line 474):
error→   $1 = nterm exp (simple.tig:1.7-23: "Hello, World!\n")
error→-> $$ = nterm args.1 (simple.tig:1.7-23: "Hello, World!\n")
error→Entering state 142
error→Next token is token ")" (simple.tig:1.24: )
error→Reducing stack 0 by rule 52 (line 469):
error→   $1 = nterm args.1 (simple.tig:1.7-23: "Hello, World!\n")
error→-> $$ = nterm args (simple.tig:1.7-23: "Hello, World!\n")
error→Entering state 141
error→Next token is token ")" (simple.tig:1.24: )
error→Shifting token ")" (simple.tig:1.24: )
error→Entering state 185
error→Reducing stack 0 by rule 24 (line 375):
error→   $1 = nterm funid (simple.tig:1.1-5: print)
error→   $2 = token "(" (simple.tig:1.6: )
error→   $3 = nterm args (simple.tig:1.7-23: "Hello, World!\n")
error→   $4 = token ")" (simple.tig:1.24: )
error→-> $$ = nterm exp (simple.tig:1.1-24: print("Hello, World!\n"))
error→Entering state 27
error→Reading a token: --(end of buffer or a NUL)
error→--accepting rule at line 90("
error→")
error→--(end of buffer or a NUL)
error→--EOF (start condition 0)
error→Now at end of input.
error→Reducing stack 0 by rule 1 (line 300):
error→   $1 = nterm exp (simple.tig:1.1-24: print("Hello, World!\n"))
error→-> $$ = nterm program (simple.tig:1.1-24: )
error→Entering state 26
error→Now at end of input.
error→Shifting token "end of file" (simple.tig:2.1: )
error→Entering state 69
error→Cleanup: popping token "end of file" (simple.tig:2.1: )
error→Cleanup: popping nterm program (simple.tig:1.1-24: )
error→Parsing string: function _main() = (_exp(0); ())
error→Starting parse
error→Entering state 0
error→Reading a token: --(end of buffer or a NUL)
error→--accepting rule at line 120("function")
error→Next token is token "function" (:1.1-8: )
error→Shifting token "function" (:1.1-8: )
error→Entering state 8
error→Reading a token: --accepting rule at line 89(" ")
error→--accepting rule at line 174("_main")
error→Next token is token "identifier" (:1.10-14: _main)
error→Shifting token "identifier" (:1.10-14: _main)
error→Entering state 47
error→Reading a token: --accepting rule at line 94("(")
error→Next token is token "(" (:1.15: )
error→Shifting token "(" (:1.15: )
error→Entering state 100
error→Reading a token: --accepting rule at line 95(")")
error→Next token is token ")" (:1.16: )
error→Reducing stack 0 by rule 100 (line 658):
error→-> $$ = nterm funargs (:1.16: )
error→Entering state 153
error→Next token is token ")" (:1.16: )
error→Shifting token ")" (:1.16: )
error→Entering state 197
error→Reading a token: --accepting rule at line 89(" ")
error→--accepting rule at line 108("=")
error→Next token is token "=" (:1.18: )
error→Reducing stack 0 by rule 88 (line 602):
error→-> $$ = nterm typeid.opt (:1.17: )
error→Entering state 226
error→Next token is token "=" (:1.18: )
error→Shifting token "=" (:1.18: )
error→Entering state 242
error→Reading a token: --accepting rule at line 89(" ")
error→--accepting rule at line 94("(")
error→Next token is token "(" (:1.20: )
error→Shifting token "(" (:1.20: )
error→Entering state 12
error→Reading a token: --accepting rule at line 164("_exp")
error→Next token is token "_exp" (:1.21-24: )
error→Shifting token "_exp" (:1.21-24: )
error→Entering state 21
error→Reading a token: --accepting rule at line 94("(")
error→Next token is token "(" (:1.25: )
error→Shifting token "(" (:1.25: )
error→Entering state 64
error→Reading a token: --accepting rule at line 146("0")
error→Next token is token "integer" (:1.26: 0)
error→Shifting token "integer" (:1.26: 0)
error→Entering state 113
error→Reading a token: --accepting rule at line 95(")")
error→Next token is token ")" (:1.27: )
error→Shifting token ")" (:1.27: )
error→Entering state 173
error→Reducing stack 0 by rule 44 (line 443):
error→   $1 = token "_exp" (:1.21-24: )
error→   $2 = token "(" (:1.25: )
error→   $3 = token "integer" (:1.26: 0)
error→   $4 = token ")" (:1.27: )
error→-> $$ = nterm exp (:1.21-27: print("Hello, World!\n"))
error→Entering state 52
error→Reading a token: --accepting rule at line 104(";")
error→Next token is token ";" (:1.28: )
error→Reducing stack 0 by rule 56 (line 481):
error→   $1 = nterm exp (:1.21-27: print("Hello, World!\n"))
error→-> $$ = nterm exps.1 (:1.21-27: print("Hello, World!\n"))
error→Entering state 53
error→Next token is token ";" (:1.28: )
error→Shifting token ";" (:1.28: )
error→Entering state 106
error→Reading a token: --accepting rule at line 89(" ")
error→--accepting rule at line 94("(")
error→Next token is token "(" (:1.30: )
error→Shifting token "(" (:1.30: )
error→Entering state 12
error→Reading a token: --accepting rule at line 95(")")
error→Next token is token ")" (:1.31: )
error→Reducing stack 0 by rule 60 (line 493):
error→-> $$ = nterm exps.0.2 (:1.31: )
error→Entering state 55
error→Next token is token ")" (:1.31: )
error→Shifting token ")" (:1.31: )
error→Entering state 107
error→Reducing stack 0 by rule 4 (line 309):
error→   $1 = token "(" (:1.30: )
error→   $2 = nterm exps.0.2 (:1.31: )
error→   $3 = token ")" (:1.31: )
error→-> $$ = nterm exp (:1.30-31: ())
error→Entering state 162
error→Reading a token: --(end of buffer or a NUL)
error→--accepting rule at line 95(")")
error→Next token is token ")" (:1.32: )
error→Reducing stack 0 by rule 59 (line 488):
error→   $1 = nterm exps.1 (:1.21-27: print("Hello, World!\n"))
error→   $2 = token ";" (:1.28: )
error→   $3 = nterm exp (:1.30-31: ())
error→-> $$ = nterm exps.2 (:1.21-31: print("Hello, World!\n"), ())
error→Entering state 54
error→Reducing stack 0 by rule 61 (line 494):
error→   $1 = nterm exps.2 (:1.21-31: print("Hello, World!\n"), ())
error→-> $$ = nterm exps.0.2 (:1.21-31: print("Hello, World!\n"), ())
error→Entering state 55
error→Next token is token ")" (:1.32: )
error→Shifting token ")" (:1.32: )
error→Entering state 107
error→Reducing stack 0 by rule 4 (line 309):
error→   $1 = token "(" (:1.20: )
error→   $2 = nterm exps.0.2 (:1.21-31: print("Hello, World!\n"), ())
error→   $3 = token ")" (:1.32: )
error→-> $$ = nterm exp (:1.20-32: (
error→  print("Hello, World!\n");
error→  ()
error→))
error→Entering state 250
error→Reading a token: --(end of buffer or a NUL)
error→--EOF (start condition 0)
error→Now at end of input.
error→Reducing stack 0 by rule 97 (line 650):
error→   $1 = token "function" (:1.1-8: )
error→   $2 = token "identifier" (:1.10-14: _main)
error→   $3 = token "(" (:1.15: )
error→   $4 = nterm funargs (:1.16: )
error→   $5 = token ")" (:1.16: )
error→   $6 = nterm typeid.opt (:1.17: )
error→   $7 = token "=" (:1.18: )
error→   $8 = nterm exp (:1.20-32: (
error→  print("Hello, World!\n");
error→  ()
error→))
error→-> $$ = nterm fundec (:1.1-32: 
error→function _main() =
error→  (
error→    print("Hello, World!\n");
error→    ()
error→  ))
error→Entering state 39
error→Now at end of input.
error→Reducing stack 0 by rule 95 (line 645):
error→   $1 = nterm fundec (:1.1-32: 
error→function _main() =
error→  (
error→    print("Hello, World!\n");
error→    ()
error→  ))
error→-> $$ = nterm fundecs (:1.1-32: 
error→function _main() =
error→  (
error→    print("Hello, World!\n");
error→    ()
error→  ))
error→Entering state 38
error→Now at end of input.
error→Reducing stack 0 by rule 20 (line 360):
error→-> $$ = nterm decs (:1.33: )
error→Entering state 90
error→Reducing stack 0 by rule 16 (line 346):
error→   $1 = nterm fundecs (:1.1-32: 
error→function _main() =
error→  (
error→    print("Hello, World!\n");
error→    ()
error→  ))
error→   $2 = nterm decs (:1.33: )
error→-> $$ = nterm decs (:1.1-32: 
error→function _main() =
error→  (
error→    print("Hello, World!\n");
error→    ()
error→  ))
error→Entering state 28
error→Reducing stack 0 by rule 2 (line 302):
error→   $1 = nterm decs (:1.1-32: 
error→function _main() =
error→  (
error→    print("Hello, World!\n");
error→    ()
error→  ))
error→-> $$ = nterm program (:1.1-32: )
error→Entering state 26
error→Now at end of input.
error→Shifting token "end of file" (:1.33: )
error→Entering state 69
error→Cleanup: popping token "end of file" (:1.33: )
error→Cleanup: popping nterm program (:1.1-32: )

Example 4.2: SCAN=1 PARSE=1 tc -X --parse simple.tig

A lexical error must be properly diagnosed and reported. The following (generated) examples display the location: this is not required for TC-0; nevertheless, an error message on the standard error output is required.

"\z does not exist."

File 4.2: back-zee.tig

$ tc -X --parse back-zee.tig
error→back-zee.tig:1.1-3: unrecognized escape: \z
⇒2

Example 4.3: tc -X --parse back-zee.tig

Similarly for syntactical errors.

a++

File 4.3: postinc.tig

$ tc -X --parse postinc.tig
error→postinc.tig:1.3: syntax error, unexpected +
⇒3

Example 4.4: tc -X --parse postinc.tig


Next: , Previous: , Up: PTHL (TC-0)   [Contents][Index]

4.2.3 PTHL Code to Write

We don’t need several directories, you can program in the top level of the package.

You must write:

src/scantiger.ll

The scanner.

lval supports strings, integers and even symbols. Nevertheless, symbols (i.e., identifiers) are returned as plain C++ strings for the time being: the class misc::symbol is introduced in TC-1.

If the environment variable SCAN is defined (to whatever value) Flex scanner debugging traces are enabled, i.e., set the variable yy_flex_debug to 1.

src/parsetiger.yy

The parser, and maybe main if you wish. Bison advanced features will be used in TC-1.

src/tc.cc

You may write your driver, i.e., main, in this file. Putting it into src/parsetiger.yy is OK in TC-0 as it is reduced to its simplest form with no option support. Of course the exit status must conform to the standard (see Errors in Tiger Compiler Reference Manual).

Makefile

This file is mandatory. Running make must build an executable thl in the root directory. The GNU Build System is not mandatory: TC-1 introduces Autoconf, Automake etc. You may use it, in which case we will run configure before make.


Next: , Previous: , Up: PTHL (TC-0)   [Contents][Index]

4.2.4 PTHL FAQ

Translating escapes in the scanner (or not)

Escapes in string can be translated at the scanning stage, or kept as is. That is, the string "\n" can produce a token STRING with the semantic value \n (translation) or \\n (no translation). You are free to choose your favorite implementation, but keep in mind that if you translate, you’ll have to “untranslate” later (i.e., convert \n back to \\n).

We encourage you to do this translation, but the other solution is also correct, as long as the next steps of your compiler follow the same conventions as your input.

You must check for bad escapes whatever solution you choose.

Must lexical and syntactic extensions be implemented?

No. Language extensions (see Language Extensions in Tiger Compiler Reference Manual) such as metavariables keywords (‘_decs’, ‘_exp’, ‘_lvalue’, ‘_namety’) and casts (‘_cast’) are not required for PTHL.

Handling metavariables constructs becomes mandatory at TC-2 (see TC-2 Code to Write) where they are used within TWEASTs (Text With Embedded AST, see ast.pdf), while casts are only needed for the optional bounds checking assignment (see TC-B).

What values can be represented by an int?

The set of valid integer values is the set of signed 32-bit integers in 2’s complement, that is the integer interval [-2^{31}, 2^{31}-1].

What values can be represented by an integer literal?

Although an integer value can be any number in [-2^{31}, 2^{31}-1], it is however not possible to represent the literal -2^{31} (= -2147483648) for technical reasons. It is however possible to create an integer value representing this number.

To put it in nutshell, the following declaration is not valid:

var i := -2147483648

whereas this one is:

var i := -2147483647 - 1

Previous: , Up: PTHL (TC-0)   [Contents][Index]

4.2.5 PTHL Improvements

Possible improvements include:

Using %destructor

You may use %destructor to reclaim the memory lost during the error recovery. It is mandated in TC-2, see TC-2 FAQ.

Parser driver

You may implement a parser driver to handle the parsing context (flags, open files, etc.). Note that a driver class will be (partially) provided at TC-1.

Handling object-related constructs from PTHL

Your scanner and parser are not required to support OO constructs at PTHL, but you can implement them in your LALR(1) parser if you want. (Fully supporting them will be mandatory at TC-2 though, during the conversion of your LALR(1) parser to a GLR one.)

Object-related productions from the Tiger grammar are:

# Class definition (canonical form).
ty ::= class [ extends type-id ] { classfields }
# Class definition (alternative form).
dec ::= class id [ extends type-id ] { classfields }
classfields ::= { classfield }
# Class fields.
classfield ::=
  # Attribute declaration.
    vardec
  # Method declaration.
  | method id ( tyfields ) [ : type-id ] = exp
# Object creation.
exp ::= new type-id

# Method call.
exp ::= lvalue . id ( [ exp { , exp }] )

Next: , Previous: , Up: Compiler Stages   [Contents][Index]

4.3 TC-1, Scanner and Parser

2017-TC-1 submission for Ing1 students is Sunday, February 8th 2015 at 11:42.

This section has been updated for EPITA-2017 on 2015-02-02.

Scanner and parser are properly running, but the abstract syntax tree is not built yet. Differences with PTHL (TC-0) include:

GNU Build System

Autoconf, Automake are used.

Options, Tasks

The compiler supports basic options via in the Task module. See Invoking tc in Tiger Compiler Reference Manual, for the list of options to support.

Locations

The locations are properly computed and reported in the error messages.

Relevant lecture notes include dev-tools.pdf and scanner.pdf.


Next: , Up: TC-1   [Contents][Index]

4.3.1 TC-1 Goals

Things to learn during this stage that you should remember:

Basic use of the GNU Build System

Autoconf, Automake. The initial set up of the project will best be done via ‘autoreconf -fvim’, but once the project initiated (i.e., configure and the Makefile.ins exist) you should depend on make only. See The GNU Build System.

Integration into an existing framework

Putting your own code into the provided code base.

Basic C++ classes

The classes Location and Position provide a good start to study foreign C++ classes. Your understanding them will be controlled, including the ‘operator’s.

Location Tracking

Issues within the scanner and the parser.

Implementation of a few simple C++ classes

The code for misc::symbol and misc::unique is incomplete.

A first standard container: std::set

The implementation of the misc::unique class relies on std::set.

The Flyweight design pattern

The misc::unique class is an implementation of the Flyweight design pattern.

Version Control System

Using the Git version control system is mandatory. Your understanding of it will be checked.


Next: , Previous: , Up: TC-1   [Contents][Index]

4.3.2 TC-1 Samples

The only information the compiler provides is about lexical and syntax errors. If there are no errors, the compiler shuts up, and exits successfully:

/* An array type and an array variable. */
let
  type  arrtype = array of int
  var arr1 : arrtype := arrtype [10] of 0
in
  arr1[2]
end

File 4.4: test01.tig

$ tc -X --parse test01.tig

Example 4.5: tc -X --parse test01.tig

If there are lexical errors, the exit status is 2, and an error message is output on the standard error output. Its format is standard and mandatory: file, (precise) location, and then the message (see Errors in Tiger Compiler Reference Manual).

1
 /* This comments starts at /* 2.2 */

File 4.5: unterminated-comment.tig

$ tc -X --parse unterminated-comment.tig
error→unterminated-comment.tig:2.2-3.0: unexpected end of file in a comment
⇒2

Example 4.6: tc -X --parse unterminated-comment.tig

If there are syntax errors, the exit status is set to 3:

let var a : nil := ()
in
  1
end

File 4.6: type-nil.tig

$ tc -X --parse type-nil.tig
error→type-nil.tig:1.13-15: syntax error, unexpected nil, expecting identifier or _namety
⇒3

Example 4.7: tc -X --parse type-nil.tig

If there are errors which are non lexical, nor syntactic (Windows will not pass by me):

$ tc C:/TIGER/SAMPLE.TIG
error→tc: cannot open `C:/TIGER/SAMPLE.TIG': No such file or directory
⇒1

Example 4.8: tc C:/TIGER/SAMPLE.TIG

The option --parse-trace, which relies on Bison’s %debug and %printer directives, must work properly5:

a + "a"

File 4.7: a+a.tig

$ tc -X --parse-trace --parse a+a.tig
error→Parsing file: a+a.tig
error→Starting parse
error→Entering state 0
error→Reading a token: Next token is token "identifier" (a+a.tig:1.1: a)
error→Shifting token "identifier" (a+a.tig:1.1: a)
error→Entering state 2
error→Reading a token: Next token is token "+" (a+a.tig:1.3: )
error→Reducing stack 0 by rule 92 (line 620):
error→   $1 = token "identifier" (a+a.tig:1.1: a)
error→-> $$ = nterm varid (a+a.tig:1.1: a)
error→Entering state 35
error→Reducing stack 0 by rule 45 (line 448):
error→   $1 = nterm varid (a+a.tig:1.1: a)
error→-> $$ = nterm lvalue (a+a.tig:1.1: a)
error→Entering state 29
error→Next token is token "+" (a+a.tig:1.3: )
error→Reducing stack 0 by rule 42 (line 441):
error→   $1 = nterm lvalue (a+a.tig:1.1: a)
error→-> $$ = nterm exp (a+a.tig:1.1: a)
error→Entering state 27
error→Next token is token "+" (a+a.tig:1.3: )
error→Shifting token "+" (a+a.tig:1.3: )
error→Entering state 80
error→Reading a token: Next token is token "string" (a+a.tig:1.5-7: a)
error→Shifting token "string" (a+a.tig:1.5-7: a)
error→Entering state 1
error→Reducing stack 0 by rule 22 (line 367):
error→   $1 = token "string" (a+a.tig:1.5-7: a)
error→-> $$ = nterm exp (a+a.tig:1.5-7: "a")
error→Entering state 128
error→Reading a token: Now at end of input.
error→Reducing stack 0 by rule 36 (line 418):
error→   $1 = nterm exp (a+a.tig:1.1: a)
error→   $2 = token "+" (a+a.tig:1.3: )
error→   $3 = nterm exp (a+a.tig:1.5-7: "a")
error→-> $$ = nterm exp (a+a.tig:1.1-7: (a + "a"))
error→Entering state 27
error→Now at end of input.
error→Reducing stack 0 by rule 1 (line 300):
error→   $1 = nterm exp (a+a.tig:1.1-7: (a + "a"))
error→-> $$ = nterm program (a+a.tig:1.1-7: )
error→Entering state 26
error→Now at end of input.
error→Shifting token "end of file" (a+a.tig:2.1: )
error→Entering state 69
error→Cleanup: popping token "end of file" (a+a.tig:2.1: )
error→Cleanup: popping nterm program (a+a.tig:1.1-7: )
error→Parsing string: function _main() = (_exp(0); ())
error→Starting parse
error→Entering state 0
error→Reading a token: Next token is token "function" (:1.1-8: )
error→Shifting token "function" (:1.1-8: )
error→Entering state 8
error→Reading a token: Next token is token "identifier" (:1.10-14: _main)
error→Shifting token "identifier" (:1.10-14: _main)
error→Entering state 47
error→Reading a token: Next token is token "(" (:1.15: )
error→Shifting token "(" (:1.15: )
error→Entering state 100
error→Reading a token: Next token is token ")" (:1.16: )
error→Reducing stack 0 by rule 100 (line 658):
error→-> $$ = nterm funargs (:1.16: )
error→Entering state 153
error→Next token is token ")" (:1.16: )
error→Shifting token ")" (:1.16: )
error→Entering state 197
error→Reading a token: Next token is token "=" (:1.18: )
error→Reducing stack 0 by rule 88 (line 602):
error→-> $$ = nterm typeid.opt (:1.17: )
error→Entering state 226
error→Next token is token "=" (:1.18: )
error→Shifting token "=" (:1.18: )
error→Entering state 242
error→Reading a token: Next token is token "(" (:1.20: )
error→Shifting token "(" (:1.20: )
error→Entering state 12
error→Reading a token: Next token is token "_exp" (:1.21-24: )
error→Shifting token "_exp" (:1.21-24: )
error→Entering state 21
error→Reading a token: Next token is token "(" (:1.25: )
error→Shifting token "(" (:1.25: )
error→Entering state 64
error→Reading a token: Next token is token "integer" (:1.26: 0)
error→Shifting token "integer" (:1.26: 0)
error→Entering state 113
error→Reading a token: Next token is token ")" (:1.27: )
error→Shifting token ")" (:1.27: )
error→Entering state 173
error→Reducing stack 0 by rule 44 (line 443):
error→   $1 = token "_exp" (:1.21-24: )
error→   $2 = token "(" (:1.25: )
error→   $3 = token "integer" (:1.26: 0)
error→   $4 = token ")" (:1.27: )
error→-> $$ = nterm exp (:1.21-27: (a + "a"))
error→Entering state 52
error→Reading a token: Next token is token ";" (:1.28: )
error→Reducing stack 0 by rule 56 (line 481):
error→   $1 = nterm exp (:1.21-27: (a + "a"))
error→-> $$ = nterm exps.1 (:1.21-27: (a + "a"))
error→Entering state 53
error→Next token is token ";" (:1.28: )
error→Shifting token ";" (:1.28: )
error→Entering state 106
error→Reading a token: Next token is token "(" (:1.30: )
error→Shifting token "(" (:1.30: )
error→Entering state 12
error→Reading a token: Next token is token ")" (:1.31: )
error→Reducing stack 0 by rule 60 (line 493):
error→-> $$ = nterm exps.0.2 (:1.31: )
error→Entering state 55
error→Next token is token ")" (:1.31: )
error→Shifting token ")" (:1.31: )
error→Entering state 107
error→Reducing stack 0 by rule 4 (line 309):
error→   $1 = token "(" (:1.30: )
error→   $2 = nterm exps.0.2 (:1.31: )
error→   $3 = token ")" (:1.31: )
error→-> $$ = nterm exp (:1.30-31: ())
error→Entering state 162
error→Reading a token: Next token is token ")" (:1.32: )
error→Reducing stack 0 by rule 59 (line 488):
error→   $1 = nterm exps.1 (:1.21-27: (a + "a"))
error→   $2 = token ";" (:1.28: )
error→   $3 = nterm exp (:1.30-31: ())
error→-> $$ = nterm exps.2 (:1.21-31: (a + "a"), ())
error→Entering state 54
error→Reducing stack 0 by rule 61 (line 494):
error→   $1 = nterm exps.2 (:1.21-31: (a + "a"), ())
error→-> $$ = nterm exps.0.2 (:1.21-31: (a + "a"), ())
error→Entering state 55
error→Next token is token ")" (:1.32: )
error→Shifting token ")" (:1.32: )
error→Entering state 107
error→Reducing stack 0 by rule 4 (line 309):
error→   $1 = token "(" (:1.20: )
error→   $2 = nterm exps.0.2 (:1.21-31: (a + "a"), ())
error→   $3 = token ")" (:1.32: )
error→-> $$ = nterm exp (:1.20-32: (
error→  (a + "a");
error→  ()
error→))
error→Entering state 250
error→Reading a token: Now at end of input.
error→Reducing stack 0 by rule 97 (line 650):
error→   $1 = token "function" (:1.1-8: )
error→   $2 = token "identifier" (:1.10-14: _main)
error→   $3 = token "(" (:1.15: )
error→   $4 = nterm funargs (:1.16: )
error→   $5 = token ")" (:1.16: )
error→   $6 = nterm typeid.opt (:1.17: )
error→   $7 = token "=" (:1.18: )
error→   $8 = nterm exp (:1.20-32: (
error→  (a + "a");
error→  ()
error→))
error→-> $$ = nterm fundec (:1.1-32: 
error→function _main() =
error→  (
error→    (a + "a");
error→    ()
error→  ))
error→Entering state 39
error→Now at end of input.
error→Reducing stack 0 by rule 95 (line 645):
error→   $1 = nterm fundec (:1.1-32: 
error→function _main() =
error→  (
error→    (a + "a");
error→    ()
error→  ))
error→-> $$ = nterm fundecs (:1.1-32: 
error→function _main() =
error→  (
error→    (a + "a");
error→    ()
error→  ))
error→Entering state 38
error→Now at end of input.
error→Reducing stack 0 by rule 20 (line 360):
error→-> $$ = nterm decs (:1.33: )
error→Entering state 90
error→Reducing stack 0 by rule 16 (line 346):
error→   $1 = nterm fundecs (:1.1-32: 
error→function _main() =
error→  (
error→    (a + "a");
error→    ()
error→  ))
error→   $2 = nterm decs (:1.33: )
error→-> $$ = nterm decs (:1.1-32: 
error→function _main() =
error→  (
error→    (a + "a");
error→    ()
error→  ))
error→Entering state 28
error→Reducing stack 0 by rule 2 (line 302):
error→   $1 = nterm decs (:1.1-32: 
error→function _main() =
error→  (
error→    (a + "a");
error→    ()
error→  ))
error→-> $$ = nterm program (:1.1-32: )
error→Entering state 26
error→Now at end of input.
error→Shifting token "end of file" (:1.33: )
error→Entering state 69
error→Cleanup: popping token "end of file" (:1.33: )
error→Cleanup: popping nterm program (:1.1-32: )

Example 4.9: tc -X --parse-trace --parse a+a.tig

Note that (i), --parse is needed, (ii), it cannot see that the variable is not declared nor that there is a type checking error, since type checking... is not implemented, and (iii), the output might be slightly different, depending upon the version of Bison you use. But what matters is that one can see the items: ‘"identifier" a’, ‘"string" a’.


Next: , Previous: , Up: TC-1   [Contents][Index]

4.3.3 TC-1 Given Code

Some code is provided through the ‘tc-base’ repository; use tags ‘2017-tc-base-1.0’ and ‘2017-tc-base-1.1’ to integrate it with your existing code base. See Given Code for more information on using the ‘tc-base’ Git repository.

See The Top Level, src, src/parse, lib/misc.


Next: , Previous: , Up: TC-1   [Contents][Index]

4.3.4 TC-1 Code to Write

Be sure to read Flex and Bison documentations and tutorials, see Flex & Bison.

configure.ac
Makefile.am

Include your own test suite in the tests directory, and hook it to make check.

src/parse/scantiger.ll

The scanner must be completed to read strings, identifiers etc. and track locations.

src/parse/parsetiger.yy
src/parse/tiger-parser.cc

The class TigerParser drives the lexing and parsing of input file. Its implementation in src/parse/tiger-parser.cc is incomplete.

lib/misc/symbol.*
lib/misc/unique.*

The class misc::symbol keeps a single copy of identifiers, see lib/misc. Its implementation in lib/misc/symbol.hxx and lib/misc/symbol.cc is incomplete. Note that running ‘make check’ in lib/misc exercises lib/misc/test-symbol.cc: having this unit test pass should be a goal by itself. As a matter of fact, unit tests were left to help you: once they pass successfully you may proceed to the rest of the compiler. misc::symbol’s implementation is based on misc::unique, a generic class implementing the Flyweight design pattern. The definition of this class, lib/misc/unique.hxx, is also to be completed.

lib/misc/variant.*

The implementation of the class template misc::variant<T0, Ts...> lacks a couple of conversion operators that you have to supply.


Next: , Previous: , Up: TC-1   [Contents][Index]

4.3.5 TC-1 FAQ

Bison reports type clashes

Bison may report type clashes for some actions. For instance, if you have given a type to "string", but none to exp, then it will choke on:

exp: "string";

because, unless you used ‘%define variant’, it actually means

exp: "string" { $$ = $1; };

which is not type consistent. So write this instead:

exp: "string" {};
Where is ast::Exp?

Its real definition will be provided with TC-2, so meanwhile you have to provide a fake. We recommend for a forward declaration of ‘ast::Exp’ in libparse.hh.

Finding prelude.tih

When run, the compiler needs the file prelude.tih that includes the signature of all the primitives. But the executable tc is typically run in two very different contexts:

installed

An installed binary will look for an installed prelude.tih, typically in /usr/local/share/tc/. The cpp macro PKGDATADIR is set to this directory. Its value depends on the use of configure’s option --prefix, defaulting to /usr/local.

compiled, not installed

When compiled, the binary will look for the installed prelude.tih, and of course will fail if it has never been installed. There are two means to address this issue:

The environment variable TC_PKGDATADIR

If set, it overrides the value of PKGDATADIR.

The option --library-prepend/-p

Using this option you may set the library file search path to visit the given directory before the built-in default value. For instance ‘tc -p /tmp foo.tig’ will first look for prelude.tih in /tmp.

Must import be functional?

Yes. Read the previous item.


Previous: , Up: TC-1   [Contents][Index]

4.3.6 TC-1 Improvements

Possible improvements include:


Next: , Previous: , Up: Compiler Stages   [Contents][Index]

4.4 TC-2, Building the Abstract Syntax Tree

2017-TC-2 submission for Ing1 students is Wednesday, February 22th 2015 at 11:42.

This section has been updated for EPITA-2017 on 2015-02-02.

At the end of this stage, the compiler can build abstract syntax trees of Tiger programs and pretty-print them. The parser is now a GLR parser and equipped with error recovery. The memory is properly deallocated on demand.

The code must follow our coding style and be documented, see Coding Style, and Doxygen.

Relevant lecture notes include dev-tools.pdf, ast.pdf.


Next: , Up: TC-2   [Contents][Index]

4.4.1 TC-2 Goals

Things to learn during this stage that you should remember:

Strict Coding Style

Following a strict coding style is an essential part of collaborative work. Understanding the rationales behind rules is even better. See Coding Style.

Memory Leak Trackers

Using tools such as Valgrind (see Valgrind) to track memory leaks.

Understanding the use of a GLR Parser

The parser should now use all the possibilities of a GLR parser.

Error recovery with Bison

Using the error token, and building usable ASTs in spite of lexical/syntax errors.

Using STL containers

The AST uses std::list, misc::symbol uses std::set.

Inheritance

The AST hierarchy is typical example of a proper use of inheritance, together with...

Inclusion polymorphism

An intense use of inclusion polymorphism for accept.

Use of constructors and destructors

In particular using the destructors to reclaim memory bound to components.

virtual

Dynamic and static bindings.

misc::indent

misc::indent extends std::ostream with indentation features. Use it in the PrettyPrinter to pretty-print. Understanding how misc::indent works will be checked later, see TC-3 Goals.

The Composite design pattern

The AST hierarchy is an implementation of the Composite pattern.

The Visitor design pattern

The PrettyPrinter is an implementation of the Visitor pattern.

Writing good developer documentation (using Doxygen)

The AST must be properly documented.


Next: , Previous: , Up: TC-2   [Contents][Index]

4.4.2 TC-2 Samples

Here are a few samples of the expected features.


Next: , Up: TC-2 Samples   [Contents][Index]

4.4.2.1 TC-2 Pretty-Printing Samples

The parser builds abstract syntax trees that can be output by a pretty-printing module:

/* Define a recursive function. */
let
  /* Calculate n!. */
  function fact (n : int) : int =
    if  n = 0
      then 1
      else n * fact (n - 1)
in
  fact (10)
end

File 4.8: simple-fact.tig

$ tc -XA simple-fact.tig
/* == Abstract Syntax Tree. == */

function _main() =
  (
    let
      function fact(n : int) : int =
        (if (n = 0)
          then 1
          else (n * fact((n - 1))))
    in
      fact(10)
    end;
    ()
  )

Example 4.10: tc -XA simple-fact.tig

Passing -D, --ast-delete, reclaims the memory associated to the AST. Valgrind will be used to hunt memory leaks, see Valgrind.

No heroic effort is asked for silly options combinations.

$ tc -D simple-fact.tig

Example 4.11: tc -D simple-fact.tig

$ tc -DA simple-fact.tig
error→../../src/ast/tasks.cc:24: Precondition `the_program' failed.
error→Aborted
⇒134

Example 4.12: tc -DA simple-fact.tig


The pretty-printed output must be valid and equivalent.

Valid means that any Tiger compiler must be able to parse with success your output. Pay attention to the banners such as ‘== Abstract...’: you should use comments: ‘/* == Abstract... */’. Pay attention to special characters too.

print("\"\x45\x50ITA\"\n")

File 4.9: string-escapes.tig

$ tc -XAD string-escapes.tig
/* == Abstract Syntax Tree. == */

function _main() =
  (
    print("\"EPITA\"\n");
    ()
  )

Example 4.13: tc -XAD string-escapes.tig

Equivalent means that, except for syntactic sugar, the output and the input are equal. Syntactic sugar refers to ‘&’, ‘|’, unary ‘-’, etc.

1 = 1 & 2 = 2

File 4.10: 1s-and-2s.tig

$ tc -XAD 1s-and-2s.tig
/* == Abstract Syntax Tree. == */

function _main() =
  (
    (if (1 = 1)
      then ((2 = 2) <> 0)
      else 0);
    ()
  )

Example 4.14: tc -XAD 1s-and-2s.tig

$ tc -XAD 1s-and-2s.tig >output.tig

Example 4.15: tc -XAD 1s-and-2s.tig >output.tig

$ tc -XAD output.tig
/* == Abstract Syntax Tree. == */

function _main() =
  (
    (if (1 = 1)
      then ((2 = 2) <> 0)
      else 0);
    ()
  )

Example 4.16: tc -XAD output.tig

Beware that for loops are encoded using a ast::VarDec: do not display the ‘var’:

for i := 0 to 100 do
   (print_int (i))

File 4.11: for-loop.tig

$ tc -XAD for-loop.tig
/* == Abstract Syntax Tree. == */

function _main() =
  (
    (for i := 0 to 100 do
      print_int(i));
    ()
  )

Example 4.17: tc -XAD for-loop.tig

Parentheses must not stack for free; you must even remove them as the following example demonstrates.

((((((((((0))))))))))

File 4.12: parens.tig

$ tc -XAD parens.tig
/* == Abstract Syntax Tree. == */

function _main() =
  (
    0;
    ()
  )

Example 4.18: tc -XAD parens.tig

This is not a pretty-printer trick: the ASTs of this program and that of ‘0’ are exactly the same: a single ast::IntExp.

As a result, anything output by ‘tc -AD’ is equal to what ‘tc -AD | tc -XAD -’ displays!


Next: , Previous: , Up: TC-2 Samples   [Contents][Index]

4.4.2.2 TC-2 Chunks

The type checking rules of Tiger, or rather its binding rules, justify the contrived parsing of declarations. This is why this section uses -b/--bindings-compute, implemented later (see TC-3).

In Tiger, to support recursive types and functions, continuous declarations of functions and continuous declarations of types are considered “simultaneously”. For instance in the following program, foo and bar are visible in each other’s scope, and therefore the following program is correct wrt type checking.

let function foo() : int = bar()
    function bar() : int = foo()
in
  0
end

File 4.13: foo-bar.tig

$ tc -b foo-bar.tig

Example 4.19: tc -b foo-bar.tig

In the following sample, because bar is not declared in the same bunch of declarations, it is not visible during the declaration of foo. The program is invalid.

let function foo() : int = bar()
    var stop := 0
    function bar() : int = foo()
in
  0
end

File 4.14: foo-stop-bar.tig

$ tc -b foo-stop-bar.tig
error→foo-stop-bar.tig:1.28-32: undeclared function: bar
⇒4

Example 4.20: tc -b foo-stop-bar.tig

The same applies to types.

We shall name chunk a continuous series of type (or function) declaration.


A single name cannot be defined more than once in a chunk.

let function foo() : int = 0
    function bar() : int = 1
    function foo() : int = 2
    var stop := 0
    function bar() : int = 3
in
  0
end

File 4.15: fbfsb.tig

$ tc -b fbfsb.tig
error→fbfsb.tig:3.5-28: redefinition: foo
error→fbfsb.tig:1.5-28: first definition
⇒4

Example 4.21: tc -b fbfsb.tig

It behaves exactly as if chunks were part of embedded let in end, i.e., as if the previous program was syntactic sugar for the following one (in fact, in 2006-tc used to desugar it that way).

let
  function foo() : int = 0
  function bar() : int = 1
in
  let
     function foo() : int = 2
  in
    let
      var stop := 0
    in
      let
        function bar() : int = 3
      in
        0
      end
    end
  end
end

File 4.16: fbfsb-desugared.tig

Given the type checking rules for variables, whose definitions cannot be recursive, chunks of variable declarations are reduced to a single variable.


Previous: , Up: TC-2 Samples   [Contents][Index]

4.4.2.3 TC-2 Error Recovery

Your parser must be robust to (some) syntactic errors. Observe that on the following input several parse errors are reported, not merely the first one:

(
  1;
  (2, 3);
  (4, 5);
  6
)

File 4.17: multiple-parse-errors.tig

$ tc multiple-parse-errors.tig
error→multiple-parse-errors.tig:3.5: syntax error, unexpected ",", expecting ;
error→multiple-parse-errors.tig:4.5: syntax error, unexpected ",", expecting ;
⇒3

Example 4.22: tc multiple-parse-errors.tig

Of course, the exit status still reveals the parse error. Error recovery must not break the rest of the compiler.

$ tc -XAD multiple-parse-errors.tig
error→multiple-parse-errors.tig:3.5: syntax error, unexpected ",", expecting ;
error→multiple-parse-errors.tig:4.5: syntax error, unexpected ",", expecting ;
/* == Abstract Syntax Tree. == */

function _main() =
  (
    (
      1;
      ();
      ();
      6
    );
    ()
  )
⇒3

Example 4.23: tc -XAD multiple-parse-errors.tig


Next: , Previous: , Up: TC-2   [Contents][Index]

4.4.3 TC-2 Given Code

Code is provided through the ‘tc-base’ repository, using tag ‘2017-tc-base-2.0’.

For a description of the new modules, see lib/misc, and src/ast.


Next: , Previous: , Up: TC-2   [Contents][Index]

4.4.4 TC-2 Code to Write

What is to be done:

src/parse/parsetiger.yy
Build the AST

Complete actions to instantiate AST nodes.

Support object-related syntax

Supporting object constructs, an improvement suggested for TC-0 (see PTHL Improvements), is now mandatory.

Support metavariable constructs

Augment your scanner and your parser to support the (reserved) keywords ‘_decs’, ‘_exp’, ‘_lvalue’ and ‘_namety’ and implement the corresponding grammar rules (see Language Extensions in Tiger Compiler Reference Manual). The semantic actions of these productions shall use the ‘metavar’ function template to fetch the right AST subtree from the parse::Tweast object attached to the parsing context (parse::TigerParser instance).

Implement error recovery.

There should be at least three uses of the token error. Read the Bison documentation about it.

Use %printer

Extend the use of %printer to display non-terminals.

Use %destructor

Use %destructor to reclaim the memory bound to semantic values thrown away during error recovery.

GLR

Change your skeleton to glr.cc, use the %glr-parser directive. Thanks to GLR, conflicts (S/R and/or R/R) can be accepted. Use %expect and %expect-rr to specify their number. For information, we have no R/R conflicts, and two S/R: one related to the “big lvalue” issue, and the other to the implementation of the two _cast operators (see Additional Syntactic Specifications in Tiger Compiler Reference Manual).

Chunks

In order to implement easily the type checking of declarations and to simplify following modules, adjust your grammar to parse declarations by chunks. The implementations of these chunks are in ast::FunctionDecs, ast::MethodDecs, ast::VarDecs, and ast::TypeDecs; they are implemented thanks to ast::AnyDecs). Note that an ast::VarDecs node appearing in a declaration list shall contain exactly one ast::VarDec object (see TC-2 Chunks); however, an ast::VarDecs used to implement a function’s formal arguments may of course contain several ast::VarDec (one per formal).

src/ast

Complete the abstract syntax tree module: no ‘FIXME:’ should be left. Several files are missing in full. See src/ast/README for additional information on the missing classes.

src/ast/default-visitor.hxx

Complete the GenDefaultVisitor class template. It is the basis for following visitors in the Tiger compiler.

src/ast/object-visitor.hxx

Likewise, complete GenObjectVisitor. This class template is used to instantiate visitors factoring common code (default traversals of object-related nodes) and serves as a base class of ast::PrettyPrinter (and later bind::Binder).

src/ast/pretty-printer.hh
src/ast/pretty-printer.cc

The PrettyPrinter class must be written entirely. It must use the misc::xalloc features to support indentation.


Next: , Previous: , Up: TC-2   [Contents][Index]

4.4.5 TC-2 FAQ

A NameTy, or a symbol

At some places, you may use one or the other. Just ask yourself which is the most appropriate given the context. Appel is not always right.

Bison

Be sure to read its dedicated section: Flex & Bison.

Why make complains about a missing stack.hh?

When using the C++ LALR(1) skeleton, Bison generates and uses a file named stack.hh, containing an auxiliary class stack used by the parser. This file is no longer generated nor used when the C++ GLR skeleton is used, which shall be the case starting from TC-2 (see TC-2 Code to Write). As you must write and maintain an LALR(1) parser during the TC-0 and TC-1 stages, the code given at TC-1 (see TC-1 Given Code) distributes the file src/parse/stack.hh. To avoid distribution issues at TC-2 with the GLR parser, you have to adjust the list of distributed files in src/parse/local.mk to ignore src/parse/stack.hh (see the variable FROM_PARSETIGER_YY).

Memory leaks in the parser during error recovery

To reclaim the memory during error recovery, use the %destructor directive:

%type <ast::Exp*> exp
%type <ast::Var*> lvalue
%destructor { delete $$; } <ast::Exp*> <ast::Var*> /* ... */;
Memory leaks in the standard containers

See Valgrind, for a pointer to the explanation and solution.

How do I use misc::error

See misc/error, for a description of this component. In the case of the parse module, TigerParser aggregates the local error handler. From scan_open, for instance, your code should look like:

if(!yyin)
  error_ << misc::error::failure
         << program_name << ": cannot open `" << name << "': "
         << strerror(errno) << std::endl
         << &misc::error::exit;
ast::fields_type vs. ast::VarDecs
Record definition vs. Function declaration

The grammar of the Tiger language (see Syntactic Specifications in Tiger Compiler Reference Manual) includes:

# Function, primitive and method declarations.
<dec> ::=
    "function"  <id> "(" <tyfields> ")" [ ":" <type-id> ] "=" <exp>
  | "primitive" <id> "(" <tyfields> ")" [ ":" <type-id> ]
<classfield> ::=
    "method"    <id> "(" <tyfields> ")" [ ":" <type-id> ] "=" <exp>

# Record type declaration.
<ty> ::= "{" <tyfields>  "}"

# List of “id : type”.
<tyfields> ::= [ <id> ":" <type-id> { "," <id> ":" <type-id> } ]

This grammar snippet shows that we used tyfields several times, in two very different contexts: a list of formal arguments of a function, primitive or method; and a list of record fields. The fact that the syntax is similar in both cases is an “accident”: it is by no means required by the language. A. Appel could have chosen to make them different, but what would have been the point then? It does make sense, sometimes, to make two different things look alike, that’s a form of economy — a sane engineering principle.

If the concrete syntaxes were chosen to be identical, should it be the case for abstract too? We would say it depends: the inert data is definitely the same, but the behaviors (i.e., the handling in the various visitors) are very different. So if your language features “inert data”, say C or ML, then keeping the same abstract syntax makes sense; if your language features “active data” — let’s call this... objects — then it is a mistake. Sadly enough, the first edition of Red Tiger book made this mistake, and we also did it for years.

The second edition of the Tiger in Java introduces a dedicated abstract syntax for formal arguments; we made a different choice: there is little difference between formal arguments and local variables, so we use a VarDecs, which fits nicely with the semantics of chunks.

Regarding the abstract syntax of a record type declaration, we use a list of Fields (aka fields_type).

Of course this means that you will have to duplicate your parsing of the tyfields non-terminal in your parser.

ast::DefaultVisitor and ast::NonObjectVisitor

The existence of ast::NonObjectVisitor is the result of a reasonable compromise between (relative) safety and complexity.

The problem is: as object-aware programs are to be desugared into object-free ones, (a part of) our front-end infrastructure must support two kinds of traversals:

The first category has visit methods for all type of nodes of our (object-oriented) AST, so they raise no issue. On the other hand, the second category of visitors knows nothing about objects, and should either be unable to visit AST w/ objects (static solution) or raise an error if they encounter objects (dynamic solution).

Which led us to several solutions:

  1. Consider that we have two kinds of visitors, and thus two hierarchies of visitors. Two hierarchies might confuse the students, and make the maintenance harder. Hooks in the AST nodes (accept methods) must be duplicated, too.
  2. Have a single hierarchy of visitors, but equip all concrete visitors traversing ASTs w/o objects with methods visiting object-related node aborting at run time.
  3. Likewise, but factor the aborting methods in a single place, namely ast::NonObjectVisitor. That is the solution we chose.

Solutions 2 and 3 let us provide a default visitor for ASTs without objects, but it’s harder to have a meaningful default visitor for ASTs with objects: indeed, concrete visitors on ASTs w/ objects inherit from their non-object counterparts, where methods visiting object nodes are already defined! (Though they abort at run time.)

We have found that having two visitors (ast::DefaultVisitor and ast::NonObjectVisitor) to solve this problem was more elegant, rather than merging both of them in ast::DefaultVisitor. The pros are that ast::DefaultVisitor remains a default visitor; the cons are that this visitor is now abstract, since object-related nodes have no visit implementation. Therefore, we also introduced an ast::ObjectVisitor performing default visits of the remaining node types; the combined inheritance of both ast::DefaultVisitor and ast::ObjectVisitor provides a complete default visitor.


Previous: , Up: TC-2   [Contents][Index]

4.4.6 TC-2 Improvements

Possible improvements include:

Desugar Boolean operators and unary minus in concrete syntax

In the original version of the exercise, the | and & operators and the unary minus operator are desugared in abstract syntax (i.e., using explicit instantiations of AST nodes). Using TigerInput, you can desugar using Tiger’s concrete syntax instead. This second solution is advised.

Introduce an Error class

When syntactic errors are caught, a valid AST must be built anyway, hence a critical question is: what value should be given to the missing bits? If your error recovery is not compatible with what the user meant, you are likely to create artificial type errors with your invented value.

While this behavior is compliant with the assignment, you may improve this by introducing an Error class (one?), which will never trigger type checking errors.

Using Generic Visitors

Andrei Alexandrescu has done a very interesting work on generic implementation of Visitors, see Modern C++ Design. It does require advanced C++ skills, since it is based on type lists, which requires heavy use of templates.

Using Visitor Combinators

Going even further that Andrei Alexandrescu, Nicolas Tisserand proposes an implementation of Visitor combinators, see Generic Visitors in C++.


Next: , Previous: , Up: Compiler Stages   [Contents][Index]

4.5 TC-3, Bindings

2017-TC-3 submission for Ing1 students is Sunday, March 1th 2015 at 11:42.
TC-R is part of the mandatory assignment of 2017-TC-3.

This section has been updated for EPITA-2017 on 2015-02-02.

At the end of this stage, the compiler must be able to compute and display the bindings. These features are triggered by the options -b/--bindings-compute, --object-bindings-compute and -B/--bindings-display.

Relevant lecture notes include: names.pdf.


Next: , Up: TC-3   [Contents][Index]

4.5.1 TC-3 Goals

Things to learn during this stage that you should remember:

The Command design pattern

The Task module is based on the Command design pattern.

Writing a Container Class Template

Class template are most useful to implement containers such as misc::scoped_map.

Using methods from parents classes

super_type and qualified method invocation to factor common code.

Traits

Traits are a useful technique that allows to write (compile time) functions ranging over types. See Glossary. The implementation of both hierarchies of visitors (const or not) relies on traits. You are expected to understand the code.

Streams’ internal extensible arrays

C++ streams allows users to dynamically store information within themselves thanks to std::ios::xalloc, std::stream::iword, and std::stream::pword (see ios_base documentation by Cplusplus Ressources). Indented output can use it directly in operator<<, see lib/misc/indent.* and lib/misc/test-indent.cc. More generally, if you have to resort to using print because you need additional arguments than the sole stream, consider using this feature instead.

Use this feature so that the PrettyPrinter can be told from the std::ostream whether escapes and bindings should be displayed.


Next: , Previous: , Up: TC-3   [Contents][Index]

4.5.2 TC-3 Samples

Binding is relating a name use to its definition.

let
  var me := 0
in
  me
end

File 4.18: me.tig

$ tc -XbBA me.tig
/* == Abstract Syntax Tree. == */

function _main /* 0x189eb60 */() =
  (
    let
      var me /* 0x18903c0 */ := 0
    in
      me /* 0x18903c0 */
    end;
    ()
  )

Example 4.24: tc -XbBA me.tig

This is harder when there are several occurrences of the same name. Note that primitive types are accepted, but have no pre-declaration, contrary to primitive functions.

let
  var me := 0
  function id(me : int) : int = me
in
  me
end

File 4.19: meme.tig

$ tc -XbBA meme.tig
/* == Abstract Syntax Tree. == */

function _main /* 0x21eabb0 */() =
  (
    let
      var me /* 0x21dc3c0 */ := 0
      function id /* 0x21ecb00 */(me /* 0x21ec910 */ : int /* 0 */) : int /* 0 */ =
        me /* 0x21ec910 */
    in
      me /* 0x21dc3c0 */
    end;
    ()
  )

Example 4.25: tc -XbBA meme.tig

TC-3 is in charge of incorrect uses of the names, such as undefined names,

me

File 4.20: nome.tig

$ tc -bBA nome.tig
error→nome.tig:1.1-2: undeclared variable: me
⇒4

Example 4.26: tc -bBA nome.tig

or redefined names.

let
  type me = {}
  type me = {}
  function twice(a: int, a: int) : int = a + a
in
  me {} = me {}
end

File 4.21: tome.tig

$ tc -bBA tome.tig
error→tome.tig:3.3-14: redefinition: me
error→tome.tig:2.3-14: first definition
error→tome.tig:4.25-31: redefinition: a
error→tome.tig:4.18-23: first definition
⇒4

Example 4.27: tc -bBA tome.tig


In addition to binding names, --bindings-compute is also in charge of binding the break to their corresponding loop construct.

let var x := 0 in
  while 1 do
  (
    for i := 0 to 10 do
    (
      x := x + i;
      if x >= 42 then
        break
    );
    if x >= 51 then
      break
  )
end

File 4.22: breaks-in-embedded-loops.tig

$ tc -XbBA breaks-in-embedded-loops.tig
/* == Abstract Syntax Tree. == */

function _main /* 0x1b77b80 */() =
  (
    let
      var x /* 0x1b69480 */ := 0
    in
      (while /* 0x1b7a220 */ 1 do
        (
          (for /* 0x1b79ee0 */ i /* 0x1b79e70 */ := 0 to 10 do
            (
              (x /* 0x1b69480 */ := (x /* 0x1b69480 */ + i /* 0x1b79e70 */));
              (if (x /* 0x1b69480 */ >= 42)
                then break /* 0x1b79ee0 */
                else ())
            ));
          (if (x /* 0x1b69480 */ >= 51)
            then break /* 0x1b7a220 */
            else ())
        ))
    end;
    ()
  )

Example 4.28: tc -XbBA breaks-in-embedded-loops.tig

break

File 4.23: break.tig

$ tc -b break.tig
error→break.tig:1.1-5: `break' outside any loop
⇒4

Example 4.29: tc -b break.tig

Embedded loops show that there is scoping for breaks. Beware that there are places, apparently inside loops, where breaks make no sense too.


Although it is a matter of definitions and uses of names, record members are not bound here, because it is easier to implement during type checking. Likewise, duplicate fields are to be reported during type checking.

let
  type     box = { value : int }
  type     dup = { value : int, value : string }
  var      box := box { value = 51 }
in
  box.head
end

File 4.24: box.tig

$ tc -XbBA box.tig
/* == Abstract Syntax Tree. == */

function _main /* 0xfcabe0 */() =
  (
    let
      type box /* 0xfcc910 */ = { value : int /* 0 */ }
      type dup /* 0xfccc30 */ = {
        value : int /* 0 */,
        value : string /* 0 */
      }
      var box /* 0xfccf10 */ := box /* 0xfcc910 */ { value = 51 }
    in
      box /* 0xfccf10 */.head
    end;
    ()
  )

Example 4.30: tc -XbBA box.tig

$ tc -T box.tig
error→box.tig:3.33-46: identifier multiply defined: value
error→box.tig:6.3-10: invalid field: head
⇒5

Example 4.31: tc -T box.tig

But apart from these field-specific checks delayed at TC-4, TC-3 should report other name-related errors. In particular, a field with an invalid type name is a binding error (related to the field’s type, not the field itself), to be reported at TC-3.

let
  type rec = { a : unknown }
in
  rec { a = 42 }
end

File 4.25: unknown-field-type.tig

$ tc -XbBA unknown-field-type.tig
error→unknown-field-type.tig:2.20-26: undeclared type: unknown
⇒4

Example 4.32: tc -XbBA unknown-field-type.tig


Likewise, class members (both attributes and methods) are not to be bound at TC-3, but at the type-checking stage (see TC-4).

let
  type C = class {}
  var c := new C
in
  c.missing_method();
  c.missing_attribute
end

File 4.26: bad-member-bindings.tig

$ tc -X --object-bindings-compute -BA bad-member-bindings.tig
/* == Abstract Syntax Tree. == */

function _main /* 0x12bdc90 */() =
  (
    let
      type C /* 0x12bf9c0 */ = 
      class extends Object /* 0 */
      {
      }
      var c /* 0x12bfb90 */ := new C /* 0x12bf9c0 */
    in
      (
        c /* 0x12bfb90 */.missing_method /* 0 */();
        c /* 0x12bfb90 */.missing_attribute
      )
    end;
    ()
  )

Example 4.33: tc -X --object-bindings-compute -BA bad-member-bindings.tig

$ tc --object-types-compute bad-member-bindings.tig
error→bad-member-bindings.tig:5.3-20: unknown method: missing_method
error→bad-member-bindings.tig:6.3-21: unknown attribute: missing_attribute
⇒5

Example 4.34: tc --object-types-compute bad-member-bindings.tig

Concerning the super class type, the compiler should just check that this type exists in the environment at TC-3. Other checks are left to TC-4 (see TC-4 Samples).

let
  /* Super class doesn't exist.  */
  class Z extends Ghost {}
in
end

File 4.27: missing-super-class.tig

$ tc -X --object-bindings-compute -BA missing-super-class.tig
error→missing-super-class.tig:3.19-23: undeclared type: Ghost
⇒4

Example 4.35: tc -X --object-bindings-compute -BA missing-super-class.tig


Next: , Previous: , Up: TC-3   [Contents][Index]

4.5.3 TC-3 Given Code

Code is provided through the ‘tc-base’ repository, using tag ‘2017-tc-base-3.0’. For a description of the new module, see src/bind.


Next: , Previous: , Up: TC-3   [Contents][Index]

4.5.4 TC-3 Code to Write

misc::scoped_map<Key, Data>

Complete the class template misc::scoped_map in lib/misc/scoped-map.hh and lib/misc/scoped-map.hxx. See lib/misc, See scoped_map, for more details.

Equip ast

Augment constructs “using” an identifier, such as CallExp, with def_, def_get, and def_set to be able to set a reference to their definition, here a FunctionDec.

ast::PrettyPrinter

Implement --bindings-display support in the PrettyPrinter. Be sure to display the addresses exactly as displayed in this document: immediately after the identifier.

Complete the bind::Binder

Most of the assignment is here...

Complete the object::Binder

...and here. object::Binder inherits from bind::Binder so as to factor common parts.

Implement renaming to unique identifiers.

TC-R is a mandatory assignment. Once TC-3 completed, implementing TC-R is straightforward, see TC-R. Note that --rename is helpful to write a test suite for TC-3.

Complete auxiliary code

Write the tasks, libbind.* etc.


Next: , Previous: , Up: TC-3   [Contents][Index]

4.5.5 TC-3 FAQ

Ambiguous resolution of operator<< for ast::VarDec

Starting from TC-3, ast::VarDec inherits both from ast::VarDec and ast::Escapable. Printing an ast::VarDec using operator<< can be troublesome as this operator may be overloaded for both ast::VarDec’s base classes, but not for ast::VarDec itself, resulting in an ambiguous overload resolution. The simplest way to get rid of this ambiguity is to convert the ast::VarDec object to the type of one of its base classes (“upcast”) before printing it, either by creating a alias or (more simply) by using the static_cast operator:

const ast::VarDec& vardec = ...

// Printing VARDEC as an ast::Dec using an intermediate
// variable (alias).
const ast::Dec& dec = vardec;
ostr << dec;

// Printing VARDEC as an ast::Escapable using an
// on-the-fly conversion.
ostr << static_cast<const ast::Escapable&>(vardec);
What is the purpose of the ‘bound’ task?

The computation of name bindings can be carried out in different ways, depending on the input language: Tiger without object constructs (“Panther”), Tiger with object constructs and Tiger with support for function overloading. These different flavors of the binding computation are performed by options --bindings-compute, --object-bindings-compute and --overfun-bindings-compute respectively (see Invoking tc in Tiger Compiler Reference Manual).

However, some subsequent task may later just require that an AST is annotated with bindings (“bound”) regardless of the technique used to compute these bindings. The purpose of the ‘bound’ task is to address this need: ensuring that one of the bindings task has been executed. This task can be considered as a disjunction (logical “or”) of the ‘bindings-compute’, ‘object-bindings-compute’ and ‘overfun-bindings-compute’ tasks, the first one being the default binding strategy.


Previous: , Up: TC-3   [Contents][Index]

4.5.6 TC-3 Improvements

Possible improvements include:

Factoring the binding interface

In the ast module, several classes need to be changed to be “bindable”, i.e., to have new data and function members to set, store, and retrieve their associated definition. Instead of changing several classes in a very similar fashion, introduce a Bindable template class and derive from its instantiation.

Hash tables

How about using true hash tables (aka “unordered associative containers” in Boost parlance) instead of trees? You might also want to try Google’s Sparse Hash Tables.

Escaping Variables Computation

Once TC-3 completed, you might consider the TC-E option now, see TC-E. It takes about 100 lines to make it.


Next: , Previous: , Up: Compiler Stages   [Contents][Index]

4.6 TC-R, Unique Identifiers

2017-TC-3 submission for Ing1 students is Sunday, March 1th 2015 at 11:42.
TC-R is part of the mandatory assignment of 2017-TC-3.

This section has been updated for EPITA-2017 on 2015-02-02.

At the end of this stage, when given the option --rename, the compiler produces an AST such that no identifier is defined twice.

Relevant lecture notes include: names.pdf.


Next: , Up: TC-R   [Contents][Index]

4.6.1 TC-R Samples

Note that the transformation does not apply to field names.

let
  type a = { a: int }
  function a(a: a): a = a{ a = a + a }
  var a : a := a(1, 2)
in
  a.a
end

File 4.28: as.tig

$ tc -X --rename -A as.tig
/* == Abstract Syntax Tree. == */

function _main() =
  (
    let
      type a_0 = { a : int }
      function a_2(a_1 : a_0) : a_0 =
        a_0 { a = (a_1 + a_1) }
      var a_3 : a_0 := a_2(1, 2)
    in
      a_3.a
    end;
    ()
  )

Example 4.36: tc -X --rename -A as.tig


Next: , Previous: , Up: TC-R   [Contents][Index]

4.6.2 TC-R Given Code

No additional code is provided, see TC-3 Given Code.


Next: , Previous: , Up: TC-R   [Contents][Index]

4.6.3 TC-R Code to Write

bind::Renamer

Write it from scratch.

Complete auxiliary code

Write the tasks, libbind.* etc.


Previous: , Up: TC-R   [Contents][Index]

4.6.4 TC-R FAQ

Should I rename primitives (builtins) or _main?

No, you shall not rename them; you have to keep the interface of the Tiger runtime. Likewise for _main.


Next: , Previous: , Up: Compiler Stages   [Contents][Index]

4.7 TC-E, Computing the Escaping Variables

2017-TC-E submission is Sunday, Mar 22th 2015 at 11:42.
TC-E is part of the mandatory assignment of 2017-TC-4.

This section has been updated for EPITA-2017 on 2015-02-02.

At the end of this stage, the compiler must be able to compute and display the escaping variables. These features are triggered by the options --escapes-compute/-e and --escapes-display/-E.

Relevant lecture notes include: names.pdf and intermediate.pdf.


Next: , Up: TC-E   [Contents][Index]

4.7.1 TC-E Goals

Things to learn during this stage that you should remember:

Understanding escaping variables

In TC-E, we consider the case of non-local variables, i.e., variables that are defined in a function, but used (at least once) in another function, nested in the first one. This possibility for an inner function to use variables declared in outer functions is called block structure. Because such variables are used outside of their host function, they are qualified as “escaping”. This information will be necessary during the translation to the intermediate representation (see TC-5) when variables (named temporaries a that stage) are assigned a location (in the stack or in a register). Escaping variables shall indeed be stored in memory, so that non-local uses of such variables can actually have a means to access them.

Writing a Visitor from scratch

The escapes::EscapesVisitor provided is almost empty. A goal of TC-E is to write a complete visitor (though a small one). Do not forget to use ast::DefaultVisitor to factor as much code as possible.


Next: , Previous: , Up: TC-E   [Contents][Index]

4.7.2 TC-E Samples

This example demonstrates the computation and display of escaping variables (and formal arguments). By default, all the variables must be considered as escaping, since it is safe to put a non escaping variable onto the stack, while the converse is unsafe.

let
  var one := 1
  var two := 2
  function incr(x: int) : int = x + one
in
  incr(two)
end

File 4.29: variable-escapes.tig

$ tc -XEAeEA variable-escapes.tig
/* == Abstract Syntax Tree. == */

function _main() =
  (
    let
      var /* escaping */ one := 1
      var /* escaping */ two := 2
      function incr(/* escaping */ x : int) : int =
        (x + one)
    in
      incr(two)
    end;
    ()
  )
/* == Abstract Syntax Tree. == */

function _main() =
  (
    let
      var /* escaping */ one := 1
      var two := 2
      function incr(x : int) : int =
        (x + one)
    in
      incr(two)
    end;
    ()
  )

Example 4.37: tc -XEAeEA variable-escapes.tig

Compute the escapes after binding, so that the AST is known to be sane enough (type checking is irrelevant): the EscapeVisitor should not bother with undeclared entities.

undeclared

File 4.30: undefined-variable.tig

$ tc -e undefined-variable.tig
error→undefined-variable.tig:1.1-10: undeclared variable: undeclared
⇒4

Example 4.38: tc -e undefined-variable.tig

Run your compiler on merge.tig and to study its output. There is a number of silly mistakes that people usually make on TC-E: they are all easy to defeat when you do have a reasonable test suite, and once you understood that torturing your project is a good thing to do.


Next: , Previous: , Up: TC-E   [Contents][Index]

4.7.3 TC-E Given Code

No additional code is provided, see TC-3 Given Code.


Next: , Previous: , Up: TC-E   [Contents][Index]

4.7.4 TC-E Code to Write

See src/ast, and src/escapes.

ast::PrettyPrinter

Implement --escapes-display support in the PrettyPrinter. Follow strictly the output format, since we parse your output to check it. Display the ‘/* escaping */’ flag where needed, and only where needed: each definition of an escaping variable/formal is preceded by the comment ‘/* escaping */’. Do not display meaningless flags due to implementation details. How this pretty-printing is implemented is left to you, but factor common code.

escapes::EscapesVisitor

Write the class escapes::EscapesVisitor in src/escapes/escapes-visitor.hh and src/escapes/escapes-visitor.cc.

Introduce ast::Escapable

Ensure ast::VarDec inherits from ast::Escapable. See Escapable.


Next: , Previous: , Up: TC-E   [Contents][Index]

4.7.5 TC-E FAQ


Previous: , Up: TC-E   [Contents][Index]

4.7.6 TC-E Improvements

Possible improvements include:


Next: , Previous: , Up: Compiler Stages   [Contents][Index]

4.8 TC-4, Type Checking

2017-TC-4 submission is Sunday, Mar 22th 2014 at 11:42.
TC-E is part of the mandatory assignment of 2017-TC-4.

This section has been updated for EPITA-2017 on 2015-02-02.

At the end of this stage, the compiler type checks Tiger programs, and annotates the AST. Clear error messages are required.

Relevant lecture notes include names.pdf, type-checking.pdf.


Next: , Up: TC-4   [Contents][Index]

4.8.1 TC-4 Goals

Things to learn during this stage that you should remember:

Function template and member function templates

Functions template are quite convenient to factor code that looks alike but differs by the nature of its arguments. Member function templates are used to factor error handling the TypeChecker.

Virtual member function templates

You will be asked why there can be no such thing in C++.

Template specialization

Although quite different in nature, types and functions are processed in a similar fashion in a Tiger compiler: first one needs to visit the headers (to introduce the names in the scope, and to check that names are only defined once), and then to visit the bodies (to bind the names to actual values). We use templates and template specialization to factor this. See also the Template Method.

The Template Method design pattern

The Template Method allows to factor a generic algorithm, the steps of which are specific. This is what we use to type check function and type declarations. Do not confuse Template Method with member function template, the order matters. Remember that in English the noun is usually last, preceded by qualifier.

Type-checking

What it is, how to implement it.

Stack unwinding

What it means, and when the C++ standard requires it from the compiler.


Next: , Previous: , Up: TC-4   [Contents][Index]

4.8.2 TC-4 Samples

Type checking is optional, invoked by --types-compute. As for the computation of bindings, this option only handles programs with no object construct. To perform the type-checking of programs with objects, use --object-types-compute.

Implementing overloaded functions in Tiger is an option, which requires the implementation of a different type checker, triggered by --overfun-types-compute (see TC-A). The option --typed/-T makes sure one of them was run.

Currently, the compiler cannot perform the type-checking with both overloading and object support enabled.

1 + "2"

File 4.31: int-plus-string.tig

$ tc int-plus-string.tig

Example 4.39: tc int-plus-string.tig

$ tc -T int-plus-string.tig
error→int-plus-string.tig:1.5-7: type mismatch
error→  right operand type: string
error→  expected type: int
⇒5

Example 4.40: tc -T int-plus-string.tig

The type checker shall ensure loop index variables are read-only.

/* error: index variable erroneously assigned to.  */
for i := 10 to 1 do
  i := i - 1

File 4.32: assign-loop-var.tig

$ tc -T assign-loop-var.tig
error→assign-loop-var.tig:3.3-12: variable is read only
⇒5

Example 4.41: tc -T assign-loop-var.tig

When there are several type errors, it is admitted that some remain hidden by others.

unknown_function(unknown_variable)

File 4.33: unknowns.tig

$ tc -T unknowns.tig
error→unknowns.tig:1.1-34: undeclared function: unknown_function
⇒4

Example 4.42: tc -T unknowns.tig

Be sure to check the type of all the constructs.

if 1 then 2

File 4.34: bad-if.tig

$ tc -T bad-if.tig
error→bad-if.tig:1.1-11: type mismatch
error→  then clause type: int
error→  else clause type: void
⇒5

Example 4.43: tc -T bad-if.tig

Be aware that type and function declarations are recursive by chunks. For instance:

let
  type one = { hd : int, tail : two }
  type two = { hd : int, tail : one }
  function one(hd : int, tail : two) : one
     = one { hd = hd, tail = tail }
  function two(hd : int, tail : one) : two
     = two { hd = hd, tail = tail }
  var one := one(11, two(22, nil))
in
  print_int(one.tail.hd); print("\n")
end

File 4.35: mutuals.tig

$ tc -T mutuals.tig

Example 4.44: tc -T mutuals.tig

In case you are interested, the result is:

$ tc -H mutuals.tig >mutuals.hir

Example 4.45: tc -H mutuals.tig >mutuals.hir

$ havm mutuals.hir
22

Example 4.46: havm mutuals.hir

The type-checker must catch erroneous inheritance relations.

let
  /* Mutually recursive inheritance.  */
  type A = class extends A {}

  /* Mutually recursive inheritance.  */
  type B = class extends C {}
  type C = class extends B {}

  /* Class inherits from a non-class type.  */
  type E = class extends int {}
in
end

File 4.36: bad-super-type.tig

$ tc --object-types-compute bad-super-type.tig
error→bad-super-type.tig:3.12-29: recursive inheritance: A
error→bad-super-type.tig:6.12-29: recursive inheritance: C
error→bad-super-type.tig:10.26-28: class type expected, got: int
⇒5

Example 4.47: tc --object-types-compute bad-super-type.tig

Handle the type-checking of TypeDecs with care in object::TypeChecker: they are processed in three steps, while other declarations use a two-step visit. The object::TypeChecker visitor proceeds as follows when it encounters a TypeDecs:

  1. Visit the headers of all types in the block.
  2. Visit the bodies of all types in the block, but ignore members for each type being a class.
  3. For each type of the block being a class, visit its members.

This three-pass visit allows class members to make forward references to other types defined in the same block of types, for instance, instantiate a class B from a class A (defined in the same block), even if B is defined after A.

let
  /* A block of types.  */
  class A
  {
    /* Valid forward reference to B, defined in the same block
       as the class enclosing this member.  */
    var b := new B
  }
  type t = int
  class B
  {
  }
in
end

File 4.37: forward-reference-to-class.tig

$ tc --object-types-compute forward-reference-to-class.tig

Example 4.48: tc --object-types-compute forward-reference-to-class.tig

(See object::TypeChecker::operator()(ast::TypeDecs&) for more details.


Next: , Previous: , Up: TC-4   [Contents][Index]

4.8.3 TC-4 Given Code

Some code is provided through the ‘tc-base’ repository, using tag ‘2017-tc-base-4.0’. For a description of the new module, see src/type.


Next: , Previous: , Up: TC-4   [Contents][Index]

4.8.4 TC-4 Code to Write

What is to be done.

ast::Typable
ast::TypeConstructor

Because many AST nodes will be annotated with their type, the feature is factored by these two classes. See Typable, and TypeConstructor, for details.

ast::Exp, ast::Dec, ast::Ty

These are typable.

ast::FunctionDec, ast::TypeDec, ast::Ty

These build types.

src/type/type.*,
src/type/array.*,
src/type/builtin-types.*,
src/type/class.*,
src/type/field.*,
src/type/function.*,
src/type/method.*,
src/type/named.*,
src/type/record.*

Implement the Singletons type::String, type::Int, and type::Void. Using templates would be particularly appreciated to factor the code between the four singleton classes, see TC-4 Options.

The remaining classes are incomplete.

Pay extra attention to type::operator==(const Type& a, const Type& b) and type::Type::compatible_with.

type::TypeChecker
object::TypeChecker

Of course this is the most tricky part. We hope there are enough comments in there so that you understand what is to be done. Please, post your questions and help us improve it.

Computing the Escaping Variables

The implementation of TC-E, suggested at TC-3, becomes a mandatory assignment at TC-4.


Next: , Previous: , Up: TC-4   [Contents][Index]

4.8.5 TC-4 Options

These are features that you might want to implement in addition to the core features.

type::Error

One problem is that type error recovery can generate false errors. For instance our compiler usually considers that the type for incorrect constructs is Int, which can create cascades of errors:

"666" = if 000 then 333 else "666"

File 4.38: is_devil.tig

$ tc -T is_devil.tig
error→is_devil.tig:1.9-34: type mismatch
error→  then clause type: int
error→  else clause type: string
⇒5

Example 4.49: tc -T is_devil.tig

One means to avoid this issue consists in introducing a new type, type::Error, that the type checker would never complain about. This can be a nice complement to ast::Error.

Various Desugaring

See TC-D, for more details. This is quite an easy option, and a very interesting one. Note that implementing desugaring makes TC-5 easier.

Bounds Checking

If you felt TC-D was easy, then implementing bounds checking should be easy too. See TC-B.

Overloaded Tiger

See TC-A, for a description of this ambitious option.

Renaming object-oriented constructs

Like TC-R, this task consists in writing a visitor renaming AST nodes holding names (either defined or used), this time with support for object-oriented constructs (option --object-rename). This visitor, object::Renamer, shall also update named types (type::Named) and collect the names of all (renamed) classes. This option is essentially a preliminary step of TC-O (see the next item).

Desugaring Tiger to Panther

If your compiler is complete w.r.t. object constructs (in particular, the type-checking and the renaming of objects is a requirement), then you can implement this very ambitious option, whose goal is to convert a Tiger program with object constructs into a program with none of them (i.e., in the subset of Tiger called Panther). This work consists in completing the object::DesugarVisitor and implementing the --object-desugar option. See TC-O.


Next: , Previous: , Up: TC-4   [Contents][Index]

4.8.6 TC-4 FAQ

Stupid Types

One can legitimately wonder whether the following program is correct:

let type weirdo = array of weirdo
in
  print("I'm a creep.\n")
end

the answer is "yes", as nothing prevents this in the Tiger specifications. This type is not usable though.

Is type::Field useful?

Using std::pair in type::Record is probably enough, and simpler.

Is nil compatible with objects?

For instance, is the following example valid?

var a : Object := nil"

The answer is yes: nil is both compatible with records and objects.

Can one redefine the built-in class Object?

Yes, if the rules of the Tiger Compiler Reference Manual are honored, notably:

For example,

let class Object {} in end

is invalid, since it is similar to

let class Object extends Object {} in end

and recursive inheritance is invalid.

One can try and introduce a Dummy type as a workaround

let
  class Dummy {}
  class Object extends Dummy {}
in
end

but this is just postponing the problem, since the code above is the same as the following:

let
  class Dummy  extends Object {}
  class Object extends Dummy  {}
in
end

where there is still a recursive inheritance.

The one solution is to define our Dummy type beforehand (i.e., in its own block of type declarations), then to redefine Object.

/* Valid.  */
let
  class Dummy {}
in
  let
    class Object extends Dummy {}
  in
  end
end

Take care: this new Object type is different from the built-in one. The code below gives an example of an invalid mix of these two types.

let
  class Dummy {}
  function get_builtin_object() : Object = new Object /* builtin */
in
  let
    class Object extends Dummy {} /* custom */

    /* Invalid assignment, since an instance of the builtin Object
       is *not* an instance of the custom Object.  */
    var o : Object /* custom */ := get_builtin_object() /* builtin */
  in
  end
end

Previous: , Up: TC-4   [Contents][Index]

4.8.7 TC-4 Improvements

Possible improvements include:

A Singleton template

Implementations of the Singleton design pattern are frequently needed; the type module alone requires four instances! Therefore a template to generate such singletons is desirable. There are two ways to address this issue: tailored to type (directly in src/type/builtin-types.*), or in a completely generic way (in lib/misc/singleton.*). See Modern C++ Design, for a topnotch implementation.

A more verbose type display

When reporting a type, one must be careful with recursive definitions that could produce never ending outputs. The suggested simple implementation ensure this by limiting the Named-depth (i.e., the number of Named objects traversed) to one. Another, nicer possibility, would be to limit the expansion to once per Named.

A Graphical User Interface

tcsh is up and running. You might want to use it to implement a GUI using Python’s Tkinter.


Next: , Previous: , Up: Compiler Stages   [Contents][Index]

4.9 TC-D, Removing the syntactic sugar from the Abstract Syntax Tree

TC-D is an optional assignment.

This section has been updated for EPITA-2009 on 2007-04-26.

At the end of this stage, the compiler must be able to remove syntactic sugar from a type-checked AST. These features are triggered by the options --desugar and --overfun-desugar.


Up: TC-D   [Contents][Index]

4.9.1 TC-D Samples

String comparisons can be translated to an equivalent AST using function calls, before the translation to HIR.

"foo" = "bar"

File 4.39: string-equality.tig

$ tc --desugar-string-cmp --desugar -A string-equality.tig
/* == Abstract Syntax Tree. == */

primitive print(string_0 : string)
primitive print_err(string_1 : string)
primitive print_int(int_2 : int)
primitive flush()
primitive getchar() : string
primitive ord(string_3 : string) : int
primitive chr(code_4 : int) : string
primitive size(string_5 : string) : int
primitive streq(s1_6 : string, s2_7 : string) : int
primitive strcmp(s1_8 : string, s2_9 : string) : int
primitive substring(string_10 : string, start_11 : int, length_12 : int) : string
primitive concat(fst_13 : string, snd_14 : string) : string
primitive not(boolean_15 : int) : int
primitive exit(status_16 : int)
function _main() =
  (
    streq("foo", "bar");
    ()
  )

Example 4.50: tc --desugar-string-cmp --desugar -A string-equality.tig

"foo" < "bar"

File 4.40: string-less.tig

$ tc --desugar-string-cmp --desugar -A string-less.tig
/* == Abstract Syntax Tree. == */

primitive print(string_0 : string)
primitive print_err(string_1 : string)
primitive print_int(int_2 : int)
primitive flush()
primitive getchar() : string
primitive ord(string_3 : string) : int
primitive chr(code_4 : int) : string
primitive size(string_5 : string) : int
primitive streq(s1_6 : string, s2_7 : string) : int
primitive strcmp(s1_8 : string, s2_9 : string) : int
primitive substring(string_10 : string, start_11 : int, length_12 : int) : string
primitive concat(fst_13 : string, snd_14 : string) : string
primitive not(boolean_15 : int) : int
primitive exit(status_16 : int)
function _main() =
  (
    (strcmp("foo", "bar") < 0);
    ()
  )

Example 4.51: tc --desugar-string-cmp --desugar -A string-less.tig

for loops can be seen as sugared while loops, and be transformed as such.

for i := 0 to 10 do print_int(i)

File 4.41: simple-for-loop.tig

$ tc --desugar-for --desugar -A simple-for-loop.tig
/* == Abstract Syntax Tree. == */

primitive print(string_0 : string)
primitive print_err(string_1 : string)
primitive print_int(int_2 : int)
primitive flush()
primitive getchar() : string
primitive ord(string_3 : string) : int
primitive chr(code_4 : int) : string
primitive size(string_5 : string) : int
primitive streq(s1_6 : string, s2_7 : string) : int
primitive strcmp(s1_8 : string, s2_9 : string) : int
primitive substring(string_10 : string, start_11 : int, length_12 : int) : string
primitive concat(fst_13 : string, snd_14 : string) : string
primitive not(boolean_15 : int) : int
primitive exit(status_16 : int)
function _main() =
  (
    let
      var _lo := 0
      var _hi := 10
      var i_17 := _lo
    in
      (if (_lo <= _hi)
        then (while 1 do
            (
              print_int(i_17);
              (if (i_17 = _hi)
                then break
                else ());
              (i_17 := (i_17 + 1))
            ))
        else ())
    end;
    ()
  )

Example 4.52: tc --desugar-for --desugar -A simple-for-loop.tig


Next: , Previous: , Up: Compiler Stages   [Contents][Index]

4.10 TC-I, Function inlining

TC-I is an optional assignment.

This section has been updated for EPITA-2009 on 2007-04-26.

At the end of this stage, the compiler inlines function bodies where functions are called. In a later pass, useless functions can be pruned from the AST. These features are triggered by the options --inline and --prune. If you also implemented function overloading (see TC-A), use the options --overfun-inline and --overfun-prune.


Up: TC-I   [Contents][Index]

4.10.1 TC-I Samples

let
  function sub(i: int, j: int) :int = i + j
in
  sub(1, 2)
end

File 4.42: sub.tig

$ tc -X --inline -A sub.tig
/* == Abstract Syntax Tree. == */

function _main() =
  (
    let
      function sub_2(i_0 : int, j_1 : int) : int =
        (i_0 + j_1)
    in
      let
        var a_3 : int := 1
        var a_4 : int := 2
      in
        (a_3 + a_4)
      end
    end;
    ()
  )

Example 4.53: tc -X --inline -A sub.tig

Recursive functions cannot be inlined.


Next: , Previous: , Up: Compiler Stages   [Contents][Index]

4.11 TC-B, Array bounds checking

TC-B is an optional assignment.

This section has been updated for EPITA-2017 on 2015-01-31.

At the end of this stage, the compiler adds dynamic checks of the bounds of arrays to the AST. Every access (either on read or write) is checked, and the program should stops with the runtime exit code (120) on out-of-bounds access. This feature is triggered by the options --bounds-checks-add and --overfun-bounds-checks-add.


Next: , Up: TC-B   [Contents][Index]

4.11.1 TC-B Samples

Here is an example with an out-of-bounds array subscript, run with HAVM.

let
  type int_array = array of int
  var foo := int_array [10] of 3
in
  /* Out-of-bounds access.  */
  foo[20]
end

File 4.43: subscript-read.tig

$ tc --bounds-checks-add -A subscript-read.tig
error→tc: unrecognized option `--bounds-checks-add'
error→Try `tc --help' or `tc --usage' for more information.
⇒64

Example 4.54: tc --bounds-checks-add -A subscript-read.tig

$ tc --bounds-checks-add -L subscript-read.tig >subscript-read.lir
error→tc: unrecognized option `--bounds-checks-add'
error→Try `tc --help' or `tc --usage' for more information.
⇒64

Example 4.55: tc --bounds-checks-add -L subscript-read.tig >subscript-read.lir

$ havm subscript-read.lir
error→havm: Error: no label main
⇒1

Example 4.56: havm subscript-read.lir

And here is an example with an out-of-bounds assignment to an array cell, tested with Nolimips.

let
  type int_array = array of int
  var foo := int_array [10] of 3
in
  /* Out-of-bounds assignment.  */
  foo[42] := 51
end

File 4.44: subscript-write.tig

$ tc --bounds-checks-add -A subscript-write.tig
error→tc: unrecognized option `--bounds-checks-add'
error→Try `tc --help' or `tc --usage' for more information.
⇒64

Example 4.57: tc --bounds-checks-add -A subscript-write.tig

$ tc --bounds-checks-add -S subscript-write.tig >subscript-write.s
error→tc: unrecognized option `--bounds-checks-add'
error→Try `tc --help' or `tc --usage' for more information.
⇒64

Example 4.58: tc --bounds-checks-add -S subscript-write.tig >subscript-write.s

$ nolimips -l nolimips -Nue subscript-write.s
error→No `main' label in assembly file.
⇒5

Example 4.59: nolimips -l nolimips -Nue subscript-write.s


Previous: , Up: TC-B   [Contents][Index]

4.11.2 TC-B FAQ

The bounds checking extension relies on the use of casts (see TC-B Samples), see See Language Extensions in Tiger Compiler Reference Manual. However, a simplistic implementation of casts introduces ambiguities in the grammar that even a GLR parser cannot resolve dynamically.

Consider the following example, where foo is an l-value :

_cast(foo, string)

This piece of code can be parsed in two different ways:

  1. exp -> cast-exp -> exp -> lvalue (foo)
  2. exp -> lvalue -> cast-lvalue -> lvalue (foo)

As the cast must preserve the l-value nature of foo, it must itself produce an l-value. Hence we want the latter interpretation. This is a true ambiguity, not a local ambiguity that GLR can resolve simply by “waiting for enough look-ahead”.

To help it take the right decision, you can favor the right path by assigning dynamic priorities to relevant rules, using Bison’s %dprec keyword. See Bison’s manual (see Flex & Bison) for more information on this feature.


Next: , Previous: , Up: Compiler Stages   [Contents][Index]

4.12 TC-A, Ad Hoc Polymorphism (Function Overloading)

TC-A is an optional assignment.

This section has been updated for EPITA-2009 on 2007-04-26.

At the end of this stage, the compiler must be able to resolve overloaded function calls. These features are triggered by the options --overfun-bindings-compute and --overfun-types-compute/-O.

Relevant lecture notes include: names.pdf.


Next: , Up: TC-A   [Contents][Index]

4.12.1 TC-A Samples

Overloaded functions are not supported in regular Tiger.

let
  function null(i: int) : int    = i = 0
  function null(s: string) : int = s = ""
in
  null("123") = null(123)
end

File 4.45: sizes.tig

$ tc -Xb sizes.tig
error→sizes.tig:3.3-41: redefinition: null
error→sizes.tig:2.3-40: first definition
⇒4

Example 4.60: tc -Xb sizes.tig

Instead of regular binding, overloaded binding binds each function call to the set of active function definitions. Unfortunately displaying this set is not implemented, so we cannot see them in the following example:

$ tc -X --overfun-bindings-compute -BA sizes.tig
/* == Abstract Syntax Tree. == */

function _main /* 0x24c2bb0 */() =
  (
    let
      function null /* 0x24c4a90 */(i /* 0x24b4630 */ : int /* 0 */) : int /* 0 */ =
        (i /* 0x24b4630 */ = 0)
      function null /* 0x24c4e90 */(s /* 0x24c4c10 */ : string /* 0 */) : int /* 0 */ =
        (s /* 0x24c4c10 */ = "")
    in
      (null /* 0 */("123") = null /* 0 */(123))
    end;
    ()
  )

Example 4.61: tc -X --overfun-bindings-compute -BA sizes.tig

The selection of the right binding cannot be done before type-checking, since precisely overloading relies on types to distinguish the actual function called. Therefore it is the type checker that “finishes” the binding.

$ tc -XOBA sizes.tig
/* == Abstract Syntax Tree. == */

function _main /* 0x1461c10 */() =
  (
    let
      function null /* 0x1463a90 */(i /* 0x1453630 */ : int /* 0 */) : int /* 0 */ =
        (i /* 0x1453630 */ = 0)
      function null /* 0x1463eb0 */(s /* 0x1463c10 */ : string /* 0 */) : int /* 0 */ =
        (s /* 0x1463c10 */ = "")
    in
      (null /* 0x1463eb0 */("123") = null /* 0x1463a90 */(123))
    end;
    ()
  )

Example 4.62: tc -XOBA sizes.tig


There can be ambiguous (overloaded) calls.

let
  type foo = {}
  function empty(f: foo) : int = f = nil
  type bar = {}
  function empty(b: bar) : int = b = nil
in
  empty(foo {});
  empty(bar {});
  empty(nil)
end

File 4.46: over-amb.tig

$ tc -XO over-amb.tig
error→over-amb.tig:9.3-12: nil ambiguity calling `empty'
error→matching declarations: 
error→    empty @ 
error→    {
error→      f : foo = 
error→      {
error→      }
error→    }
error→    empty @ 
error→    {
error→      b : bar = 
error→      {
error→      }
error→    }
⇒5

Example 4.63: tc -XO over-amb.tig

The spirit of plain Tiger is kept: a “chunk” is not allowed to redefine a function with the same signature:

let
  function foo(i: int) = ()
  function foo(i: int) = ()
in
  foo(42)
end

File 4.47: over-duplicate.tig

$ tc -XO over-duplicate.tig
error→over-duplicate.tig:3.3-27: function complete redefinition: foo
error→over-duplicate.tig:2.3-27: first definition
⇒5

Example 4.64: tc -XO over-duplicate.tig

but a signature can be defined twice in different blocks of function definitions.

let
  function foo(i: int) = ()
in
  let
    function foo(i: int) = ()
  in
    foo(51)
  end
end

File 4.48: over-scoped.tig

$ tc -XOBA over-scoped.tig
/* == Abstract Syntax Tree. == */

function _main /* 0x23cbc10 */() =
  (
    let
      function foo /* 0x23cd970 */(i /* 0x23bd630 */ : int /* 0 */) =
        ()
    in
      let
        function foo /* 0x23cdcb0 */(i /* 0x23cdb50 */ : int /* 0 */) =
          ()
      in
        foo /* 0x23cdcb0 */(51)
      end
    end;
    ()
  )

Example 4.65: tc -XOBA over-scoped.tig


Next: , Previous: , Up: TC-A   [Contents][Index]

4.12.2 TC-A Given Code

No additional code is provided.


Previous: , Up: TC-A   [Contents][Index]

4.12.3 TC-A Code to Write

See src/ast, and src/overload.


Next: , Previous: , Up: Compiler Stages   [Contents][Index]

4.13 TC-O, Desugaring object constructs

TC-O is an optional assignment.

This section has been updated for EPITA-2012 on 2015-01-21.

At the end of this stage, the compiler must be able to desugar object constructs into plain Tiger without objects, a.k.a. Panther. This feature is triggered by the option --object-desugar. Do not forget that you need to complete and write all missing parts of the object support (parser, ast, binder, type-checker, etc...). Make sure that all of these are correctly working before starting this bonus.

This a very hard assignment. If you plan to work on it, start with very simple programs, and progressively add new desugaring patterns. Be sure to keep a complete test suite to cover all cases and avoid regressions.

Achieving a faithful and complete translation from Tiger to Panther requires a lot of work. Even the reference implementation of the object-desugar pass (about 1,000 lines of code) is not perfect, as some inputs may generate invalid Tiger code after desugaring objects (in particular when playing with scopes).


Up: TC-O   [Contents][Index]

4.13.1 TC-O Samples

Be warned: even Small object-oriented Tiger programs may generate complicated desugared outputs.

let
  class A {}
in
end

File 4.49: empty-class.tig

$ tc -X --object-desugar -A empty-class.tig
/* == Abstract Syntax Tree. == */

function _main() =
  let
    type _variant_Object = { exact_type : int }
    type _variant_A_0 = { exact_type : int }
    var _id_Object := 0
    var _id_A_0 := 1
    function _new_Object() : _variant_Object =
      _variant_Object { exact_type = _id_Object }
  in
    (
      let
        function _new_A_0() : _variant_A_0 =
          let
          in
            _variant_A_0 { exact_type = _id_A_0 }
          end
        function _upcast_A_0_to_Object(source : _variant_A_0) : _variant_Object =
          _variant_Object { exact_type = _id_A_0 }
      in
        ()
      end;
      ()
    )
  end

Example 4.66: tc -X --object-desugar -A empty-class.tig

let
  class B
  {
    var a := 42
    method m() : int = self.a
  }
  var b := new B
in
  b.a := 51
end

File 4.50: simple-class.tig

$ tc -X --object-desugar -A simple-class.tig
/* == Abstract Syntax Tree. == */

function _main() =
  let
    type _variant_Object = {
      exact_type : int,
      field_B_1 : _contents_B_1
    }
    type _contents_B_1 = { a : int }
    type _variant_B_1 = {
      exact_type : int,
      field_B_1 : _contents_B_1
    }
    var _id_Object := 0
    var _id_B_1 := 1
    function _new_Object() : _variant_Object =
      _variant_Object {
        exact_type = _id_Object,
        field_B_1 = nil
      }
  in
    (
      let
        function _new_B_1() : _variant_B_1 =
          let
            var contents_B_1 := _contents_B_1 { a = 42 }
          in
            _variant_B_1 {
              exact_type = _id_B_1,
              field_B_1 = contents_B_1
            }
          end
        function _upcast_B_1_to_Object(source : _variant_B_1) : _variant_Object =
          _variant_Object {
            exact_type = _id_B_1,
            field_B_1 = source.field_B_1
          }
        function _method_B_1_m(self : _variant_B_1) : int =
          self.field_B_1.a
        function _dispatch_B_1_m(self : _variant_B_1) : int =
          _method_B_1_m(self)
        var b_2 := _new_B_1()
      in
        (b_2.field_B_1.a := 51)
      end;
      ()
    )
  end

Example 4.67: tc -X --object-desugar -A simple-class.tig

let
  class C
  {
    var a := 0
    method m() : int = self.a
  }
  class D extends C
  {
    var b := 9
    /* Override C.m().  */
    method m() : int = self.a + self.b
  }
  var d : D := new D
  /* Valid upcast due to inclusion polymorphism.  */
  var c : C := d
in
  c.a := 42;
  /* Note that accessing `c.b' is not allowed, since `c' is
     statically known as a `C', even though it is actually a `D'
     at run time.  */
  let
    /* Polymorphic call.  */
    var res := c.m()
  in
    print_int(res);
    print("\n")
  end
end

File 4.51: override.tig

$ tc --object-desugar -A override.tig
/* == Abstract Syntax Tree. == */

primitive print(string_0 : string)
primitive print_err(string_1 : string)
primitive print_int(int_2 : int)
primitive flush()
primitive getchar() : string
primitive ord(string_3 : string) : int
primitive chr(code_4 : int) : string
primitive size(string_5 : string) : int
primitive streq(s1_6 : string, s2_7 : string) : int
primitive strcmp(s1_8 : string, s2_9 : string) : int
primitive substring(string_10 : string, start_11 : int, length_12 : int) : string
primitive concat(fst_13 : string, snd_14 : string) : string
primitive not(boolean_15 : int) : int
primitive exit(status_16 : int)
function _main() =
  let
    type _variant_Object = {
      exact_type : int,
      field_C_18 : _contents_C_18,
      field_D_20 : _contents_D_20
    }
    type _contents_C_18 = { a : int }
    type _variant_C_18 = {
      exact_type : int,
      field_C_18 : _contents_C_18,
      field_D_20 : _contents_D_20
    }
    type _contents_D_20 = { b : int }
    type _variant_D_20 = {
      exact_type : int,
      field_D_20 : _contents_D_20,
      field_C_18 : _contents_C_18
    }
    var _id_Object := 0
    var _id_C_18 := 1
    var _id_D_20 := 2
    function _new_Object() : _variant_Object =
      _variant_Object {
        exact_type = _id_Object,
        field_C_18 = nil,
        field_D_20 = nil
      }
  in
    (
      let
        function _new_C_18() : _variant_C_18 =
          let
            var contents_C_18 := _contents_C_18 { a = 0 }
          in
            _variant_C_18 {
              exact_type = _id_C_18,
              field_C_18 = contents_C_18,
              field_D_20 = nil
            }
          end
        function _upcast_C_18_to_Object(source : _variant_C_18) : _variant_Object =
          _variant_Object {
            exact_type = _id_C_18,
            field_C_18 = source.field_C_18,
            field_D_20 = source.field_D_20
          }
        function _downcast_C_18_to_D_20(source : _variant_C_18) : _variant_D_20 =
          _variant_D_20 {
            exact_type = _id_D_20,
            field_D_20 = source.field_D_20,
            field_C_18 = source.field_C_18
          }
        function _method_C_18_m(self : _variant_C_18) : int =
          self.field_C_18.a
        function _dispatch_C_18_m(self : _variant_C_18) : int =
          (if (self.exact_type = _id_C_18)
            then _method_C_18_m(self)
            else _method_D_20_m(_downcast_C_18_to_D_20(self)))
        function _new_D_20() : _variant_D_20 =
          let
            var contents_D_20 := _contents_D_20 { b = 9 }
            var contents_C_18 := _contents_C_18 { a = 0 }
          in
            _variant_D_20 {
              exact_type = _id_D_20,
              field_D_20 = contents_D_20,
              field_C_18 = contents_C_18
            }
          end
        function _upcast_D_20_to_C_18(source : _variant_D_20) : _variant_C_18 =
          _variant_C_18 {
            exact_type = _id_D_20,
            field_C_18 = source.field_C_18,
            field_D_20 = source.field_D_20
          }
        function _upcast_D_20_to_Object(source : _variant_D_20) : _variant_Object =
          _variant_Object {
            exact_type = _id_D_20,
            field_C_18 = source.field_C_18,
            field_D_20 = source.field_D_20
          }
        function _method_D_20_m(self : _variant_D_20) : int =
          (self.field_C_18.a + self.field_D_20.b)
        function _dispatch_D_20_m(self : _variant_D_20) : int =
          _method_D_20_m(self)
        var d_21 : _variant_D_20 := _new_D_20()
        var c_22 : _variant_C_18 := _upcast_D_20_to_C_18(d_21)
      in
        (
          (c_22.field_C_18.a := 42);
          let
            var res_23 := _dispatch_C_18_m(c_22)
          in
            (
              print_int(res_23);
              print("\n")
            )
          end
        )
      end;
      ()
    )
  end

Example 4.68: tc --object-desugar -A override.tig

$ tc --object-desugar -L override.tig >override.lir

Example 4.69: tc --object-desugar -L override.tig >override.lir

$ havm override.lir
51

Example 4.70: havm override.lir


Next: , Previous: , Up: Compiler Stages   [Contents][Index]

4.14 TC-5, Translating to the High Level Intermediate Representation

2017-TC-5 submission is Saturday, May 3th 2015 at 23:42

This section has been updated for EPITA-2017 on 2015-02-02.

At the end of this stage the compiler translates the AST into the high level intermediate representation, HIR for short.

Relevant lecture notes include intermediate.pdf.


Next: , Up: TC-5   [Contents][Index]

4.14.1 TC-5 Goals

Things to learn during this stage that you should remember:

Smart pointers

The techniques used to implement reference counting via the redefinition of operator-> and operator*. std::auto_ptr are also smart pointers.

std::auto_ptr

The intermediate translation is stored in an auto_ptr to guarantee it is released (delete) at the end of the run. The translate module interface also uses auto_ptr to specify sources and sinks.

Reference counting

The class template misc::ref provides reference counting smart pointers to ease the memory management. It is used to handle nodes of the intermediate representation, especially because during TC-6 some rewriting might transform this tree into an DAG, in which case memory deallocation is complex.

Variants

C++ features the union keyword, inherited from C. Not only is union not type safe, it also forbids class members. Some people have worked hard to implement union à la C++, i.e., with type safety, polymorphism etc. These union are called “discriminated unions” or “variants” to follow the vocabulary introduced by Caml. See the papers from Andrei Alexandrescu: Discriminated Unions (i), Discriminated Unions (ii), Generic: Discriminated Unions (iii) for an introduction to the techniques. We use boost::variant (see Boost.org) in temp.

I (Akim) strongly encourage you to read these enlightening articles.

Default copy constructor, default assignment operator

The C++ standard specifies that unless specified, default implementations of the copy constructor and assignment operator must be provided by the compiler. There are some pitfalls though, clearly exhibited in the implementation of misc::ref. You must be able to explain these pitfalls.

Template template parameters

C++ allows several kinds of entities to be used as template parameters. The most well known kind is “type”: you frequently parameterize class templates with types via ‘template <typename T>’ or ‘template <class T>’. But you may also parameterize with a class template. The temp module heavily uses this feature: understand it, and be ready to write similar code.

Explicit template instantiations

You must be able to explain how templates are “compiled”. In addition, you know how to explicitly instantiate templates, and explain what it can be used for. The implementation of temp::Identifier (and temp::Temp and temp::Label) is based on these ideas. See the corresponding rule in File Conventions for some explanations on this topic.

Covariant return

C++ supports covariance of the method return type. This feature is crucial to implement methods such as clone, as in frame::Access::clone(). Understand return type covariance.

Lazy/delayed computation

The ‘Ix’, ‘Cx’, ‘Nx’, and ‘Ex’ classes delay computation to address context-depend issues in a context independent way.

Intermediate Representations
A different approach of hierarchies

In this project, the AST is composed of different classes related by inheritance (as if the kinds of the nodes were class members). Here, the nodes are members of a single class, but their nature is specified by the object itself (as if the kinds of the nodes were object members).

Stack Frame, Activation Record

The implementation of recursion and automatic variables.

Inner functions and their impact on memory management at runtime

Reaching non local variables.


Next: , Previous: , Up: TC-5   [Contents][Index]

4.14.2 TC-5 Samples

TC-5 can be started (and should be started if you don’t want to finish it in a hurry) by first making sure your compiler can handle code that uses no variables. Then, you can complete your compiler to support more and more Tiger features.


Next: , Up: TC-5 Samples   [Contents][Index]

4.14.2.1 TC-5 Primitive Samples

This example is probably the simplest Tiger program.

0

File 4.52: 0.tig

$ tc --hir-display 0.tig
/* == High Level Intermediate representation. == */
# Routine: _main
label main
# Prologue
# Body
seq
  sxp
    const 0
  sxp
    const 0
seq end
# Epilogue
label end

Example 4.71: tc --hir-display 0.tig

You should then probably try to make more difficult programs with literals only. Arithmetics is one of the easiest tasks.

1 + 2 * 3

File 4.53: arith.tig

$ tc -H arith.tig
/* == High Level Intermediate representation. == */
# Routine: _main
label main
# Prologue
# Body
seq
  sxp
    binop add
      const 1
      binop mul
        const 2
        const 3
  sxp
    const 0
seq end
# Epilogue
label end

Example 4.72: tc -H arith.tig

Use havm to exercise your output.

$ tc -H arith.tig >arith.hir

Example 4.73: tc -H arith.tig >arith.hir

$ havm arith.hir

Example 4.74: havm arith.hir

Unfortunately, without actually printing something, you won’t see the final result, which means you need to implement function calls. Fortunately, you can ask havm for a verbose execution:

$ havm --trace arith.hir
error→checkingLow
error→plaining
error→unparsing
error→checking
error→evaling
error→ call ( name main ) []
error→9.6-9.13:  const 1
error→11.8-11.15:  const 2
error→12.8-12.15:  const 3
error→10.6-12.15:  binop mul 2 3
error→8.4-12.15:  binop add 1 6
error→7.2-12.15:  sxp 7
error→14.4-14.11:  const 0
error→13.2-14.11:  sxp 0
error→ end call ( name main ) [] = 0

Example 4.75: havm --trace arith.hir

If you look carefully, you will find an ‘sxp 7’ in there...


Then you are encouraged to implement control structures.

if 101 then 102 else 103

File 4.54: if-101.tig

$ tc -H if-101.tig
/* == High Level Intermediate representation. == */
# Routine: _main
label main
# Prologue
# Body
seq
  seq
    cjump ne
      const 101
      const 0
      name l0
      name l1
    label l0
    sxp
      const 102
    jump
      name l2
    label l1
    sxp
      const 103
    label l2
  seq end
  sxp
    const 0
seq end
# Epilogue
label end

Example 4.76: tc -H if-101.tig

And even more difficult control structure uses:

while 101
  do (if 102 then break)

File 4.55: while-101.tig

$ tc -H while-101.tig
/* == High Level Intermediate representation. == */
# Routine: _main
label main
# Prologue
# Body
seq
  seq
    label l1
    cjump ne
      const 101
      const 0
      name l2
      name l0
    label l2
    seq
      cjump ne
        const 102
        const 0
        name l3
        name l4
      label l3
      jump
        name l0
      jump
        name l5
      label l4
      sxp
        const 0
      label l5
    seq end
    jump
      name l1
    label l0
  seq end
  sxp
    const 0
seq end
# Epilogue
label end

Example 4.77: tc -H while-101.tig

Beware that HAVM has some known bugs with its handling of break, see HAVM Bugs.


Next: , Previous: , Up: TC-5 Samples   [Contents][Index]

4.14.2.2 TC-5 Optimizing Cascading If

Optimize the number of jumps needed to compute nested if, using ‘translate::Ix’. A plain use of ‘translate::Cx’ is possible, but less efficient.

Consider the following sample:

if if 11 < 22 then 33 < 44 else 55 < 66 then print("OK\n")

File 4.56: boolean.tig

a naive implementation will probably produce too many cjump instructions6:

$ tc --hir-naive -H boolean.tig
/* == High Level Intermediate representation. == */
label l7
        "OK\n"
# Routine: _main
label main
# Prologue
# Body
seq
  seq
    cjump ne
      eseq
      seq
        cjump lt
          const 11
          const 22
          name l0
          name l1
        label l0
        move
          temp t0
          eseq
          seq
            move
              temp t1
              const 1
            cjump lt
              const 33
              const 44
              name l3
              name l4
            label l4
            move
              temp t1
              const 0
            label l3
          seq end
            temp t1
        jump
          name l2
        label l1
        move
          temp t0
          eseq
          seq
            move
              temp t2
              const 1
            cjump lt
              const 55
              const 66
              name l5
              name l6
            label l6
            move
              temp t2
              const 0
            label l5
          seq end
            temp t2
        jump
          name l2
        label l2
      seq end
        temp t0
      const 0
      name l8
      name l9
    label l8
    sxp
      call
        name print
        name l7
      call end
    jump
      name l10
    label l9
    sxp
      const 0
    jump
      name l10
    label l10
  seq end
  sxp
    const 0
seq end
# Epilogue
label end

Example 4.78: tc --hir-naive -H boolean.tig

$ tc --hir-naive -H boolean.tig >boolean-1.hir

Example 4.79: tc --hir-naive -H boolean.tig >boolean-1.hir

$ havm --profile boolean-1.hir
error→/* Profiling.  */
error→fetches from temporary : 2
error→fetches from memory    : 0
error→binary operations      : 0
error→function calls         : 1
error→stores to temporary    : 2
error→stores to memory       : 0
error→jumps                  : 2
error→conditional jumps      : 3
error→/* Execution time.  */
error→number of cycles : 19
OK

Example 4.80: havm --profile boolean-1.hir

An analysis of this pessimization reveals that it is related to the computation of an intermediate expression (the value of ‘if 11 < 22 then 33 < 44 else 55 < 66’) later decoded as a condition. A better implementation will produce:

$ tc -H boolean.tig
/* == High Level Intermediate representation. == */
label l0
        "OK\n"
# Routine: _main
label main
# Prologue
# Body
seq
  seq
    seq
      cjump lt
        const 11
        const 22
        name l4
        name l5
      label l4
      cjump lt
        const 33
        const 44
        name l1
        name l2
      label l5
      cjump lt
        const 55
        const 66
        name l1
        name l2
    seq end
    label l1
    sxp
      call
        name print
        name l0
      call end
    jump
      name l3
    label l2
    sxp
      const 0
    label l3
  seq end
  sxp
    const 0
seq end
# Epilogue
label end

Example 4.81: tc -H boolean.tig

$ tc -H boolean.tig >boolean-2.hir

Example 4.82: tc -H boolean.tig >boolean-2.hir

$ havm --profile boolean-2.hir
error→/* Profiling.  */
error→fetches from temporary : 0
error→fetches from memory    : 0
error→binary operations      : 0
error→function calls         : 1
error→stores to temporary    : 0
error→stores to memory       : 0
error→jumps                  : 1
error→conditional jumps      : 2
error→/* Execution time.  */
error→number of cycles : 13
OK

Example 4.83: havm --profile boolean-2.hir


Next: , Previous: , Up: TC-5 Samples   [Contents][Index]

4.14.2.3 TC-5 Builtin Calls Samples

The game becomes more interesting with primitive calls (which are easier to compile than function definitions and function calls).

(print_int(101); print("\n"))

File 4.57: print-101.tig

$ tc -H print-101.tig >print-101.hir

Example 4.84: tc -H print-101.tig >print-101.hir

$ havm print-101.hir
101

Example 4.85: havm print-101.hir

Complex values, arrays and records, also need calls to the runtime system:

let
  type ints = array of int
  var  ints := ints [51] of 42
in
  print_int(ints[ints[0]]); print("\n")
end

File 4.58: print-array.tig

$ tc -H print-array.tig
/* == High Level Intermediate representation. == */
label l0
        "\n"
# Routine: _main
label main
# Prologue
move
  temp t1
  temp fp
move
  temp fp
  temp sp
move
  temp sp
  binop sub
    temp sp
    const 4
# Body
seq
  seq
    move
      mem
        temp fp
      eseq
        move
          temp t0
          call
            name init_array
            const 51
            const 42
          call end
        temp t0
    seq
      sxp
        call
          name print_int
          mem
            binop add
              mem
                temp fp
              binop mul
                mem
                  binop add
                    mem
                      temp fp
                    binop mul
                      const 0
                      const 4
                const 4
        call end
      sxp
        call
          name print
          name l0
        call end
    seq end
  seq end
  sxp
    const 0
seq end
# Epilogue
move
  temp sp
  temp fp
move
  temp fp
  temp t1
label end

Example 4.86: tc -H print-array.tig

$ tc -H print-array.tig >print-array.hir

Example 4.87: tc -H print-array.tig >print-array.hir

$ havm print-array.hir
42

Example 4.88: havm print-array.hir

The case of record is more subtle. Think carefully about the following example

let
  type list = { h: int, t: list }
  var list := list { h = 1,
                     t = list { h = 2,
                               t = nil } }
in
  print_int(list.t.h); print("\n")
end

File 4.59: print-record.tig


Previous: , Up: TC-5 Samples   [Contents][Index]

4.14.2.4 TC-5 Samples with Variables

The following example demonstrates the usefulness of information about escapes: when it is not computed, all the variables are stored on the stack.

let
  var a := 1
  var b := 2
  var c := 3
in
  a := 2;
  c := a + b + c;
  print_int(c);
  print("\n")
end

File 4.60: vars.tig

$ tc -H vars.tig
/* == High Level Intermediate representation. == */
label l0
        "\n"
# Routine: _main
label main
# Prologue
move
  temp t0
  temp fp
move
  temp fp
  temp sp
move
  temp sp
  binop sub
    temp sp
    const 12
# Body
seq
  seq
    move
      mem
        temp fp
      const 1
    move
      mem
        binop add
          temp fp
          const -4
      const 2
    move
      mem
        binop add
          temp fp
          const -8
      const 3
    seq
      move
        mem
          temp fp
        const 2
      move
        mem
          binop add
            temp fp
            const -8
        binop add
          binop add
            mem
              temp fp
            mem
              binop add
                temp fp
                const -4
          mem
            binop add
              temp fp
              const -8
      sxp
        call
          name print_int
          mem
            binop add
              temp fp
              const -8
        call end
      sxp
        call
          name print
          name l0
        call end
    seq end
  seq end
  sxp
    const 0
seq end
# Epilogue
move
  temp sp
  temp fp
move
  temp fp
  temp t0
label end

Example 4.89: tc -H vars.tig

Once escaping variable computation implemented, we know none escape in this example, hence they can be stored in temporaries:

$ tc -eH vars.tig
/* == High Level Intermediate representation. == */
label l0
        "\n"
# Routine: _main
label main
# Prologue
# Body
seq
  seq
    move
      temp t0
      const 1
    move
      temp t1
      const 2
    move
      temp t2
      const 3
    seq
      move
        temp t0
        const 2
      move
        temp t2
        binop add
          binop add
            temp t0
            temp t1
          temp t2
      sxp
        call
          name print_int
          temp t2
        call end
      sxp
        call
          name print
          name l0
        call end
    seq end
  seq end
  sxp
    const 0
seq end
# Epilogue
label end

Example 4.90: tc -eH vars.tig

$ tc -eH vars.tig >vars.hir

Example 4.91: tc -eH vars.tig >vars.hir

$ havm vars.hir
7

Example 4.92: havm vars.hir

Then, you should implement the declaration of functions:

let
  function fact(i: int) : int =
    if i = 0 then 1
             else i * fact(i - 1)
in
  print_int(fact(15));
  print("\n")
end

File 4.61: fact15.tig

$ tc -H fact15.tig
/* == High Level Intermediate representation. == */
# Routine: fact
label l0
# Prologue
move
  temp t1
  temp fp
move
  temp fp
  temp sp
move
  temp sp
  binop sub
    temp sp
    const 8
move
  mem
    temp fp
  temp i0
move
  mem
    binop add
      temp fp
      const -4
  temp i1
# Body
move
  temp rv
  eseq
  seq
    cjump eq
      mem
        binop add
          temp fp
          const -4
      const 0
      name l1
      name l2
    label l1
    move
      temp t0
      const 1
    jump
      name l3
    label l2
    move
      temp t0
      binop mul
        mem
          binop add
            temp fp
            const -4
        call
          name l0
          mem
            temp fp
          binop sub
            mem
              binop add
                temp fp
                const -4
            const 1
        call end
    label l3
  seq end
    temp t0
# Epilogue
move
  temp sp
  temp fp
move
  temp fp
  temp t1
label end

label l4
        "\n"
# Routine: _main
label main
# Prologue
# Body
seq
  seq
    sxp
      call
        name print_int
        call
          name l0
          temp fp
          const 15
        call end
      call end
    sxp
      call
        name print
        name l4
      call end
  seq end
  sxp
    const 0
seq end
# Epilogue
label end

Example 4.93: tc -H fact15.tig

$ tc -H fact15.tig >fact15.hir

Example 4.94: tc -H fact15.tig >fact15.hir

$ havm fact15.hir
2004310016

Example 4.95: havm fact15.hir

Note that the result of 15! (1307674368000) does not fit on a signed 32-bit integer, and is therefore wrapped (to 2004310016).

And finally, you should support escaping variables (see File 4.29).

$ tc -eH variable-escapes.tig
/* == High Level Intermediate representation. == */
# Routine: incr
label l0
# Prologue
move
  temp t2
  temp fp
move
  temp fp
  temp sp
move
  temp sp
  binop sub
    temp sp
    const 4
move
  mem
    temp fp
  temp i0
move
  temp t1
  temp i1
# Body
move
  temp rv
  binop add
    temp t1
    mem
      mem
        temp fp
# Epilogue
move
  temp sp
  temp fp
move
  temp fp
  temp t2
label end

# Routine: _main
label main
# Prologue
move
  temp t3
  temp fp
move
  temp fp
  temp sp
move
  temp sp
  binop sub
    temp sp
    const 4
# Body
seq
  sxp
    eseq
    seq
      move
        mem
          temp fp
        const 1
      move
        temp t0
        const 2
    seq end
      call
        name l0
        temp fp
        temp t0
      call end
  sxp
    const 0
seq end
# Epilogue
move
  temp sp
  temp fp
move
  temp fp
  temp t3
label end

Example 4.96: tc -eH variable-escapes.tig


Next: , Previous: , Up: TC-5   [Contents][Index]

4.14.3 TC-5 Given Code

Some code is provided through the ‘tc-base’ repository, using tags ‘2017-tc-base-5.0’ and ‘2017-tc-base-5.1’. For a description of the new modules, see src/temp, src/tree, src/frame, src/translate.


Next: , Previous: , Up: TC-5   [Contents][Index]

4.14.4 TC-5 Code to Write

You are encouraged to first try very simple examples: ‘nil’, ‘1 + 2’, ‘"foo" < "bar"’ etc. Then consider supporting variables, and finally handle the case of the functions.

temp::Identifier

Their implementations are to be finished. This task is independent of others. Passing test-temp.cc is probably the sign you completed correctly the implementation.

You are invited to follow the best practices for variants, in particular, avoid “type switching” by hand, rather use variant visitors. For instance the IdentifierEqualVisitor can be used this way:

template <template <typename Tag_> class Traits_>
bool
Identifier<Traits_>::operator==(const Identifier<Traits_>& rhs) const
{
  return
    rank_get() == rhs.rank_get() &&
    boost::apply_visitor(IdentifierEqualToVisitor(), value_, rhs.value_);
}
tree::Fragment

There remains to implement tree::ProcFrag::dump that outputs the routine themselves plus the glue code (allocating the frame etc.).

translate/translation.*
translate::Translator

There are holes to fill.


Next: , Previous: , Up: TC-5   [Contents][Index]

4.14.5 TC-5 Options

This section documents possible extensions you could implement in TC-5.


Next: , Up: TC-5 Options   [Contents][Index]

4.14.5.1 TC-5 Bounds Checking

The implementation of the bounds checking can be done when generating the IR. Requirements are the same than for the see TC-B option. You can use HAVM to test the success of your bounds checking.


Previous: , Up: TC-5 Options   [Contents][Index]

4.14.5.2 TC-5 Optimizing Static Links

Warning: this optimization is difficult to do it perfectly, and therefore, expect a big bonus.

In a first and conservative extension, the compiler considers that all the functions (but the builtins!) need a static link. This is correct, but inefficient: for instance, the traditional fact function will spend almost as much time handling the static link, than its real argument.

Some functions need a static link, but don’t need to save it on the stack. For instance, in the following example:

let
  var foo := 1
  function foo() : int = foo
in
  foo()
end

the function foo does need a static link to access the variable foo, but does not need to store its static link on the stack.

It is suggested to address these problems in the following order:

  1. Implement the detection of functions that do not need a static link (see exercise 6.5 in Modern Compiler Implementation), but still consider any static link escapes.
  2. Adjust the output of --escapes-display to display ‘/* escaping sl */before the first formal argument of the functions (declarations) that need the static link:
    $ cat fact.tig
    let
       function fact(n : int) : int =
          if (n = 0)
             then 1
             else n * fact((n - 1))
    in
       fact(10)
    end
    $ tc -XEA fact.tig
    /* == Abstract Syntax Tree. == */
    
    function _main() =
      (
        let
          function fact(/* escaping sl *//* escaping */ n : int) : int =
            (if (n = 0)
              then 1
              else (n * fact((n - 1))))
        in
          fact(10)
        end;
        ()
      )
    $ tc -XeEA fact.tig
    /* == Abstract Syntax Tree. == */
    
    function _main() =
      (
        let
          function fact(n : int) : int =
            (if (n = 0)
              then 1
              else (n * fact((n - 1))))
        in
          fact(10)
        end;
        ()
      )
    
  3. Adjust your call and progFrag prologues.
  4. Improve your computation so that non escaping static links are detected:
    $ cat escaping-sl.tig
    let
       var toto := 1
       function outer() : int =
         let function inner() : int = toto
         in inner() end
    in
      outer()
    end
    $ tc -XeEA escaping-sl.tig
    /* == Abstract Syntax Tree. == */
    
    function _main() =
      (
        let
          var /* escaping */ toto := 1
          function outer(/* escaping sl */ ) : int =
            let
              function inner(/* sl */ ) : int =
                toto
            in
              inner()
            end
        in
          outer()
        end;
        ()
      )
    

    Here, both outer and inner need their static link (so that inner can access toto. However, outer’s static link escapes, while inner’s does not.

    Watch out, it is not trivial to find the minimum. What do you think about the static link of the function sister below?

    let
      var v := 1
      function outer() : int =
        let
          function inner() : int = v
        in
          inner()
        end
      function sister() : int = outer()
    in
      sister()
    end
    

Next: , Previous: , Up: TC-5   [Contents][Index]

4.14.6 TC-5 FAQ

$fp’ or ‘fp’?

Andrew Appel clearly has his HIR/LIR depend on the target in three different ways: the names of the frame pointer and result registers7, and the machine word size.

That would mean that the target module (see src/target) would be given during TC-5, which seemed too difficult and anti-pedagogical, so we used fp and rv where he uses $fp and $v0. While this does make TC-5 more target independent and TC-5 code base lighter, it slightly complicates the rest of the compiler.

There remains one target dependent information wired in hard: the word size is set to 4.

$x13’ or ‘t13’?

Anonymous temporaries should be output as ‘t13’ for HAVM at stages 5 and 6, and as ‘$x13’ for Nolimips, stage 7. The code provided does not support (yet) this double standard, so it always outputs ‘t13’, although the samples provided here use ‘$x13’. Fortunately HAVM supports both standards8, so this does not matter for TC-5 and TC-6. We recommend ‘t13’ though, contrary to our samples, generated with a tc that needs more work.

How to perform the allocation of the static link in a level?

The constructor of translate::Level reads:

  // Install a slot for the static link if needed.
  Level::Level(const misc::symbol& name,
               const Level* parent,
               frame::bool_list_type formal_escapes)
    : parent_(parent)
    , frame_(new frame::Frame(name))
  {
// FIXME: Some code was deleted here (Allocate a formal for the static link).

    // Install translate::Accesses for all the formals.
    for (const bool b : formal_escapes)
      formal_alloc(b);
  }

To allocate a formal for the static link, look at how other formals are allocated, and take these into account:

Obviously, this won’t hold if you plan to optimize the static links (see TC-5 Optimizing Static Links); you’ll have to tweak translate::Level’s constructor.

Why var i := 0 won’t compile?

If you try to compute the intermediate representation for a single variable declaration, you’ll probably run into a SIGSEGV or a failed assertion. For instance, the following command probably won’t work: echo 'var i := 0' | tc --hir-compute -.

Variables must be allocated in a level (see translate::Translator::operator()(const ast::VarDec&)). However, there is no level for global variable declarations (outside _main). The current language specification does not address this case, so you are free to handle it as you wish, though an assertion on the presence of an enclosing level is probably the easiest solution.


Previous: , Up: TC-5   [Contents][Index]

4.14.7 TC-5 Improvements

Possible improvements include:

Maximal node sharing

The proposed implementation of Tree creates new nodes for equal expressions; for instance two uses of the variable foo lead to two equal instantiations of tree::Temp. The same applies to more complex constructs such as the same translation if foo is actually a frame resident variable etc. Because memory consumption may have a negative impact on performances, it is desirable to implement maximal sharing: whenever a Tree is needed, we first check whether it already exists and then reuse it. This must be done recursively: the translation of ‘(x + x) * (x + x)’ should have a single instantiation of ‘x + x’ instead of two, but also a single instantiation of ‘x’ instead of four.

Node sharing makes some algorithms, such as rewriting, more complex, especially wrt memory management. Garbage collection is almost required, but fortunately the node of Tree are reference counted! Therefore, almost everything is ready to implement maximal node sharing. See spot, for an explanation on how this approach was successfully implemented. See The ATerm library for a general implementation of maximally shared trees.


Next: , Previous: , Up: Compiler Stages   [Contents][Index]

4.15 TC-6, Translating to the Low Level Intermediate Representation

2017-TC-6 is a part of the TC Back End option.
2017-TC-6 submission is Sunday, May 31st 2014 at 11:42.

This section has been updated for EPITA-2017 on 2013-12-30.

At the end of this stage, the compiler produces low level intermediate representation: LIR. LIR is a subset of the HIR: some patterns are forbidden. This is why it is also named canonicalization.

Relevant lecture notes include intermediate.pdf.


Next: , Up: TC-6   [Contents][Index]

4.15.1 TC-6 Goals

Things to learn during this stage that you should remember:

Term Rewriting System

Term rewriting system are a whole topic of research in itself. If you need to be convinced, just look for “term rewriting system” on Google.

“Functional” Programming in C++

A lot of TC-6 is devoted to looking for specific nodes in lists of nodes, and splitting, and splicing lists at these places. This could be done by hand, with many hand-written iterations, or using functors and STL algorithms. You are expected to do the latter, and to discover things such as std::splice, std::find_if, std::unary_function, std::not1 etc.


Next: , Previous: , Up: TC-6   [Contents][Index]

4.15.2 TC-6 Samples

There are several stages in TC-6.


Next: , Up: TC-6 Samples   [Contents][Index]

4.15.2.1 TC-6 Canonicalization Samples

The first task in TC-6 is getting rid of all the eseq. To do this, you have to move the statement part of an eseq at the end of the current sequence point, and keeping the expression part in place.

Compare for instance the HIR to the LIR in the following case:

let function print_ints(a: int, b: int) =
    (print_int(a); print(", "); print_int(b); print("\n"))
    var a := 0
in
  print_ints(1, (a := a + 1; a))
end

File 4.62: preincr-1.tig

One possible HIR translation is:

$ tc -eH preincr-1.tig
/* == High Level Intermediate representation. == */
label l1
        ", "
label l2
        "\n"
# Routine: print_ints
label l0
# Prologue
move
  temp t2
  temp fp
move
  temp fp
  temp sp
move
  temp sp
  binop sub
    temp sp
    const 4
move
  mem
    temp fp
  temp i0
move
  temp t0
  temp i1
move
  temp t1
  temp i2
# Body
seq
  sxp
    call
      name print_int
      temp t0
    call end
  sxp
    call
      name print
      name l1
    call end
  sxp
    call
      name print_int
      temp t1
    call end
  sxp
    call
      name print
      name l2
    call end
seq end
# Epilogue
move
  temp sp
  temp fp
move
  temp fp
  temp t2
label end

# Routine: _main
label main
# Prologue
# Body
seq
  seq
    move
      temp t3
      const 0
    sxp
      call
        name l0
        temp fp
        const 1
        eseq
          move
            temp t3
            binop add
              temp t3
              const 1
          temp t3
      call end
  seq end
  sxp
    const 0
seq end
# Epilogue
label end

Example 4.97: tc -eH preincr-1.tig

A possible canonicalization is then:

$ tc -eL preincr-1.tig
/* == Low Level Intermediate representation. == */
label l1
        ", "
label l2
        "\n"
# Routine: print_ints
label l0
# Prologue
move
  temp t2
  temp fp
move
  temp fp
  temp sp
move
  temp sp
  binop sub
    temp sp
    const 4
move
  mem
    temp fp
  temp i0
move
  temp t0
  temp i1
move
  temp t1
  temp i2
# Body
seq
  label l3
  sxp
    call
      name print_int
      temp t0
    call end
  sxp
    call
      name print
      name l1
    call end
  sxp
    call
      name print_int
      temp t1
    call end
  sxp
    call
      name print
      name l2
    call end
  label l4
seq end
# Epilogue
move
  temp sp
  temp fp
move
  temp fp
  temp t2
label end

# Routine: _main
label main
# Prologue
# Body
seq
  label l5
  move
    temp t3
    const 0
  move
    temp t5
    temp fp
  move
    temp t3
    binop add
      temp t3
      const 1
  sxp
    call
      name l0
      temp t5
      const 1
      temp t3
    call end
  label l6
seq end
# Epilogue
label end

Example 4.98: tc -eL preincr-1.tig

The example above is simple because ‘1commutes with ‘(a := a + 1; a)’: the order does not matter. But if you change the ‘1’ into ‘a’, then you cannot exchange ‘a’ and ‘(a := a + 1; a)’, so the translation is different. Compare the previous LIR with the following, and pay attention to

let function print_ints(a: int, b: int) =
    (print_int(a); print(", "); print_int(b); print("\n"))
    var a := 0
in
  print_ints(a, (a := a + 1; a))
end

File 4.63: preincr-2.tig

$ tc -eL preincr-2.tig
/* == Low Level Intermediate representation. == */
label l1
        ", "
label l2
        "\n"
# Routine: print_ints
label l0
# Prologue
move
  temp t2
  temp fp
move
  temp fp
  temp sp
move
  temp sp
  binop sub
    temp sp
    const 4
move
  mem
    temp fp
  temp i0
move
  temp t0
  temp i1
move
  temp t1
  temp i2
# Body
seq
  label l3
  sxp
    call
      name print_int
      temp t0
    call end
  sxp
    call
      name print
      name l1
    call end
  sxp
    call
      name print_int
      temp t1
    call end
  sxp
    call
      name print
      name l2
    call end
  label l4
seq end
# Epilogue
move
  temp sp
  temp fp
move
  temp fp
  temp t2
label end

# Routine: _main
label main
# Prologue
# Body
seq
  label l5
  move
    temp t3
    const 0
  move
    temp t5
    temp fp
  move
    temp t6
    temp t3
  move
    temp t3
    binop add
      temp t3
      const 1
  sxp
    call
      name l0
      temp t5
      temp t6
      temp t3
    call end
  label l6
seq end
# Epilogue
label end

Example 4.99: tc -eL preincr-2.tig

As you can see, the output is the same for the HIR and the LIR:

$ tc -eH preincr-2.tig >preincr-2.hir

Example 4.100: tc -eH preincr-2.tig >preincr-2.hir

$ havm preincr-2.hir
0, 1

Example 4.101: havm preincr-2.hir

$ tc -eL preincr-2.tig >preincr-2.lir

Example 4.102: tc -eL preincr-2.tig >preincr-2.lir

$ havm preincr-2.lir
0, 1

Example 4.103: havm preincr-2.lir


Be very careful when dealing with mem. For instance, rewriting something like:

call(foo, eseq(move(temp t, const 51), temp t))

into

move temp t1, temp t
move temp t, const 51
call(foo, temp t)

is wrong: ‘temp t’ is not a subexpression, rather it is being defined here. You should produce:

move temp t, const 51
call(foo, temp t)

Another danger is the handling of ‘move(mem, )’. For instance:

move(mem foo, x)

must be rewritten into:

move(temp t, foo)
move(mem(temp t), x)

not as:

move(temp t, mem(foo))
move(temp t, x)

In other words, the first subexpression of ‘move(mem(foo), )’ is ‘foo’, not ‘mem(foo)’. The following example is a good crash test against this problem:

let type int_array = array of int
    var tab := int_array [2] of 51
in
  tab[0] := 100;
  tab[1] := 200;
  print_int(tab[0]); print("\n");
  print_int(tab[1]); print("\n")
end

File 4.64: move-mem.tig

$ tc -eL move-mem.tig >move-mem.lir

Example 4.104: tc -eL move-mem.tig >move-mem.lir

$ havm move-mem.lir
100
200

Example 4.105: havm move-mem.lir


You also ought to get rid of nested calls:

print(chr(ord("\n")))

File 4.65: nested-calls.tig

$ tc -L nested-calls.tig
/* == Low Level Intermediate representation. == */
label l0
        "\n"
# Routine: _main
label main
# Prologue
# Body
seq
  label l1
  move
    temp t1
    call
      name ord
      name l0
    call end
  move
    temp t2
    call
      name chr
      temp t1
    call end
  sxp
    call
      name print
      temp t2
    call end
  label l2
seq end
# Epilogue
label end

Example 4.106: tc -L nested-calls.tig

There are only two valid call forms: ‘sxp(call(...))’, and ‘move(temp(...), call(...))’.


Contrary to C, the HIR and LIR always denote the same value. For instance the following Tiger code:

let
  var a := 1
  function a(t: int) : int =
     (a := a + 1;
      print_int(t); print(" -> "); print_int(a); print("\n");
      a)
  var b := a(1) + a(2) * a(3)
in
  print_int(b); print("\n")
end

File 4.66: seq-point.tig

should always produce:

$ tc -L seq-point.tig >seq-point.lir

Example 4.107: tc -L seq-point.tig >seq-point.lir

$ havm seq-point.lir
1 -> 2
2 -> 3
3 -> 4
14

Example 4.108: havm seq-point.lir

independently of the what IR you ran. It has nothing to do with operator precedence!

In C, you have no such guarantee: the following program can give different results with different compilers and/or on different architectures.

#include <stdio.h>

int a_ = 1;
int
a(int t)
{
  ++a_;
  printf("%d -> %d\n", t, a_);
  return a_;
}

int
main(void)
{
  int b = a(1) + a(2) * a(3);
  printf("%d\n", b);
  return 0;
}

Previous: , Up: TC-6 Samples   [Contents][Index]

4.15.2.2 TC-6 Scheduling Samples

Once your eseq and call canonicalized, normalize cjumps: they must be followed by their “false” label. This goes in two steps:

  1. Split in basic blocks.

    A basic block is a sequence of code starting with a label, ending with a jump (conditional or not), and with no jumps, no labels inside.

  2. Build the traces.

    Now put all the basic blocks into a single sequence.


The following example highlights the need for new labels: at least one for the entry point, and one for the exit point:

1 & 2

File 4.67: 1-and-2.tig

$ tc -L 1-and-2.tig
/* == Low Level Intermediate representation. == */
# Routine: _main
label main
# Prologue
# Body
seq
  label l3
  cjump ne
    const 1
    const 0
    name l0
    name l1
  label l1
  label l2
  jump
    name l4
  label l0
  jump
    name l2
  label l4
seq end
# Epilogue
label end

Example 4.109: tc -L 1-and-2.tig


The following example contains many jumps. Compare the HIR to the LIR:

while 10 | 20 do if 30 | 40 then break else break

File 4.68: broken-while.tig

$ tc -H broken-while.tig
/* == High Level Intermediate representation. == */
# Routine: _main
label main
# Prologue
# Body
seq
  seq
    label l1
    seq
      cjump ne
        const 10
        const 0
        name l3
        name l4
      label l3
      cjump ne
        const 1
        const 0
        name l2
        name l0
      label l4
      cjump ne
        const 20
        const 0
        name l2
        name l0
    seq end
    label l2
    seq
      seq
        cjump ne
          const 30
          const 0
          name l8
          name l9
        label l8
        cjump ne
          const 1
          const 0
          name l5
          name l6
        label l9
        cjump ne
          const 40
          const 0
          name l5
          name l6
      seq end
      label l5
      jump
        name l0
      jump
        name l7
      label l6
      jump
        name l0
      label l7
    seq end
    jump
      name l1
    label l0
  seq end
  sxp
    const 0
seq end
# Epilogue
label end

Example 4.110: tc -H broken-while.tig

$ tc -L broken-while.tig
/* == Low Level Intermediate representation. == */
# Routine: _main
label main
# Prologue
# Body
seq
  label l10
  label l1
  cjump ne
    const 10
    const 0
    name l3
    name l4
  label l4
  cjump ne
    const 20
    const 0
    name l2
    name l0
  label l0
  jump
    name l11
  label l2
  cjump ne
    const 30
    const 0
    name l8
    name l9
  label l9
  cjump ne
    const 40
    const 0
    name l5
    name l6
  label l6
  jump
    name l0
  label l5
  jump
    name l0
  label l8
  cjump ne
    const 1
    const 0
    name l5
    name l13
  label l13
  jump
    name l6
  label l3
  cjump ne
    const 1
    const 0
    name l2
    name l14
  label l14
  jump
    name l0
  label l11
seq end
# Epilogue
label end

Example 4.111: tc -L broken-while.tig


Next: , Previous: , Up: TC-6   [Contents][Index]

4.15.3 TC-6 Given Code

Some code is provided through the ‘tc-base’ repository, using tag ‘2017-tc-base-6.0’. For a description of the new module, see src/canon.

It includes most of the canonicalization.


Next: , Previous: , Up: TC-6   [Contents][Index]

4.15.4 TC-6 Code to Write

Everything you need.


Previous: , Up: TC-6   [Contents][Index]

4.15.5 TC-6 Improvements

Possible improvements include:


Next: , Previous: , Up: Compiler Stages   [Contents][Index]

4.16 TC-7, Instruction Selection

2017-TC-7 is a part of the TC Back End option.
2017-TC-7 submission is Sunday, June 7th 2014 at 11:42.

This section has been updated for EPITA-2017 on 2013-12-30.

At the end of this stage, the compiler produces the very low level intermediate representation: ASSEM. This language is basically the target assembly, enhanced with arbitrarily many registers ($x666). This output is obviously target dependent: we aim at MIPS, as we use Nolimips to run it.

Relevant lecture notes include instr-selection.pdf.


Next: , Up: TC-7   [Contents][Index]

4.16.1 TC-7 Goals

Things to learn during this stage that you should remember:

RISC vs. CISC etc.

Different kinds of microprocessors, different spirits in assembly.

Assembly

Understanding how computer actually run.

Memory hierarchy/management at runtime

Recursive languages need memory management to implement automatic variables.

Tree matching, rewriting

Writing/debugging a code generator with MonoBURG.

Use of ios::xalloc

Instr are contained in Instrs, itself in Fragment, itself in Fragments. Suppose you mean to add a debugging flag to print an Instr, what shall you do? Add another argument to all the dump methods in these four hierarchies? The problem with Temp is even worse: they are scattered everywhere, yet we would like to specify how to output them thanks to a std::map. Should we pass this map in each and every single call?

Using ios::xalloc, ostream::pword, and ostream::iword saves the day.


Next: , Previous: , Up: TC-7   [Contents][Index]

4.16.2 TC-7 Samples

The goal of TC-7 is straightforward: starting from LIR, generate the MIPS instructions, except that you don’t have actual registers: we still heavily use Temps. Register allocation will be done in a later stage, TC-9.

let
  var answer := 42
in
  answer := 51
end

File 4.69: the-answer.tig

$ tc --inst-display the-answer.tig
# == Final assembler ouput. == #
# Routine: _main
tc_main:
# Allocate frame
        move    $x11, $ra
        move    $x3, $s0
        move    $x4, $s1
        move    $x5, $s2
        move    $x6, $s3
        move    $x7, $s4
        move    $x8, $s5
        move    $x9, $s6
        move    $x10, $s7
l0:
        li      $x1, 42
        sw      $x1, ($fp)
        li      $x2, 51
        sw      $x2, ($fp)
l1:
        move    $s0, $x3
        move    $s1, $x4
        move    $s2, $x5
        move    $s3, $x6
        move    $s4, $x7
        move    $s5, $x8
        move    $s6, $x9
        move    $s7, $x10
        move    $ra, $x11
# Deallocate frame
        jr      $ra

Example 4.112: tc --inst-display the-answer.tig

At this stage the compiler cannot know what registers are used; that’s why in the previous output it saves "uselessly" all the callee-save registers on main entry. For the same reason, the frame is not allocated.

While Nolimips accepts the lack of register allocation, it does require the frame to be allocated. That is the purpose of --nolimips-display:

$ tc --nolimips-display the-answer.tig
# == Final assembler ouput. == #
# Routine: _main
tc_main:
        sw      $fp, -4 ($sp)
        move    $fp, $sp
        sub     $sp, $sp, 8
        move    $x11, $ra
        move    $x3, $s0
        move    $x4, $s1
        move    $x5, $s2
        move    $x6, $s3
        move    $x7, $s4
        move    $x8, $s5
        move    $x9, $s6
        move    $x10, $s7
l0:
        li      $x1, 42
        sw      $x1, ($fp)
        li      $x2, 51
        sw      $x2, ($fp)
l1:
        move    $s0, $x3
        move    $s1, $x4
        move    $s2, $x5
        move    $s3, $x6
        move    $s4, $x7
        move    $s5, $x8
        move    $s6, $x9
        move    $s7, $x10
        move    $ra, $x11
        move    $sp, $fp
        lw      $fp, -4 ($fp)
        jr      $ra

Example 4.113: tc --nolimips-display the-answer.tig

The final stage, register allocation, addresses both issues. For your information, it results in:

$ tc -sI the-answer.tig
# == Final assembler ouput. == #
# Routine: _main
tc_main:
        sw      $fp, -4 ($sp)
        move    $fp, $sp
        sub     $sp, $sp, 8
l0:
        li      $t0, 42
        sw      $t0, ($fp)
        li      $t0, 51
        sw      $t0, ($fp)
l1:
        move    $sp, $fp
        lw      $fp, -4 ($fp)
        jr      $ra

Example 4.114: tc -sI the-answer.tig


A delicate part of this exercise is handling the function calls:

let function add(x: int, y: int) : int = x + y
in
  print_int(add(1,(add(2, 3)))); print("\n")
end

File 4.70: add.tig

$ tc -e --inst-display add.tig
# == Final assembler ouput. == #
# Routine: add
tc_l0:
# Allocate frame
        move    $x15, $ra
        sw      $a0, ($fp)
        move    $x0, $a1
        move    $x1, $a2
        move    $x7, $s0
        move    $x8, $s1
        move    $x9, $s2
        move    $x10, $s3
        move    $x11, $s4
        move    $x12, $s5
        move    $x13, $s6
        move    $x14, $s7
l2:
        add     $x6, $x0, $x1
        move    $v0, $x6
l3:
        move    $s0, $x7
        move    $s1, $x8
        move    $s2, $x9
        move    $s3, $x10
        move    $s4, $x11
        move    $s5, $x12
        move    $s6, $x13
        move    $s7, $x14
        move    $ra, $x15
# Deallocate frame
        jr      $ra

.data
l1:
        .word 1
        .asciiz "\n"
.text

# Routine: _main
tc_main:
# Allocate frame
        move    $x28, $ra
        move    $x20, $s0
        move    $x21, $s1
        move    $x22, $s2
        move    $x23, $s3
        move    $x24, $s4
        move    $x25, $s5
        move    $x26, $s6
        move    $x27, $s7
l4:
        move    $a0, $fp
        li      $x16, 2
        move    $a1, $x16
        li      $x17, 3
        move    $a2, $x17
        jal     tc_l0
        move    $x4, $v0
        move    $a0, $fp
        li      $x18, 1
        move    $a1, $x18
        move    $a2, $x4
        jal     tc_l0
        move    $x5, $v0
        move    $a0, $x5
        jal     tc_print_int
        la      $x19, l1
        move    $a0, $x19
        jal     tc_print
l5:
        move    $s0, $x20
        move    $s1, $x21
        move    $s2, $x22
        move    $s3, $x23
        move    $s4, $x24
        move    $s5, $x25
        move    $s6, $x26
        move    $s7, $x27
        move    $ra, $x28
# Deallocate frame
        jr      $ra

Example 4.115: tc -e --inst-display add.tig


Once your function calls work properly, you can start using Nolimips (using options --nop-after-branch --unlimited-registers --execute) to check the behavior of your compiler.

$ tc -eR --nolimips-display add.tig >add.nolimips

Example 4.116: tc -eR --nolimips-display add.tig >add.nolimips

$ nolimips -l nolimips -Nue add.nolimips
6

Example 4.117: nolimips -l nolimips -Nue add.nolimips


You must also complete the runtime. No difference must be observable between a run with HAVM and another with Nolimips:

substring("", 1, 1)

File 4.71: substring-0-1-1.tig

$ tc -e --nolimips-display substring-0-1-1.tig
# == Final assembler ouput. == #
.data
l0:
        .word 0
        .asciiz ""
.text

# Routine: _main
tc_main:
# Allocate frame
        move    $x12, $ra
        move    $x4, $s0
        move    $x5, $s1
        move    $x6, $s2
        move    $x7, $s3
        move    $x8, $s4
        move    $x9, $s5
        move    $x10, $s6
        move    $x11, $s7
l1:
        la      $x1, l0
        move    $a0, $x1
        li      $x2, 1
        move    $a1, $x2
        li      $x3, 1
        move    $a2, $x3
        jal     tc_substring
l2:
        move    $s0, $x4
        move    $s1, $x5
        move    $s2, $x6
        move    $s3, $x7
        move    $s4, $x8
        move    $s5, $x9
        move    $s6, $x10
        move    $s7, $x11
        move    $ra, $x12
# Deallocate frame
        jr      $ra

Example 4.118: tc -e --nolimips-display substring-0-1-1.tig

$ tc -eR --nolimips-display substring-0-1-1.tig >substring-0-1-1.nolimips

Example 4.119: tc -eR --nolimips-display substring-0-1-1.tig >substring-0-1-1.nolimips

$ nolimips -l nolimips -Nue substring-0-1-1.nolimips
error→substring: arguments out of bounds
⇒120

Example 4.120: nolimips -l nolimips -Nue substring-0-1-1.nolimips


Next: , Previous: , Up: TC-7   [Contents][Index]

4.16.3 TC-7 Given Code

Some code is provided through the ‘tc-base’ repository, using tag ‘2017-tc-base-7.0’. For more information about the TC-7 code delivered see src/target, src/assem.


Next: , Previous: , Up: TC-7   [Contents][Index]

4.16.4 TC-7 Code to Write

There is not much code to write:

Information on MIPS R2000 assembly instructions may be found in SPIM manual.


Completing the following routines will be needed during register allocation only (see TC-9):


Next: , Previous: , Up: TC-7   [Contents][Index]

4.16.5 TC-7 FAQ

Nolimips ‘Precondition `has_unlimited(reg.get_index())' failed

This lovely error message is the sign you’re using an obsolete version of Nolimips. Update.


Previous: , Up: TC-7   [Contents][Index]

4.16.6 TC-7 Improvements

Possible improvements include:


Next: , Previous: , Up: Compiler Stages   [Contents][Index]

4.17 TC-8, Liveness Analysis

2017-TC-8 is a part of the TC Back End option.
2017-TC-8 submission is Sunday, June 14th 2014 at 11:42.

This section has been updated for EPITA-2017 on 2014-01-18.

At the end of this stage, the compiler computes the input of TC-9: the interference graph (or conflict graph). The options -N and --interference-dump allow the user to see these graphs, one per function. To compute the interference graph, the compiler first computes the liveness of each temporary, i.e., a graph whose nodes are the instructions, and labeled with live temporaries. The options -V, --liveness-dump dumps these graphs. Finally, the structure of the liveness graph is the flow graph: its nodes are the instructions, and edges correspond to control flow. Use options -F, --flowgraph-dump to dump them.

All dumped graphs use the DOT format. You can display them using dotty or convert them to other formats (such as PDF or PNG) using dot, both part of the GraphViz package.

Relevant lecture notes include liveness.pdf.


Next: , Up: TC-8   [Contents][Index]

4.17.1 TC-8 Goals

Things to learn during this stage that you should remember:

Graph handling, using the Boost Graph Library

We use the Boost Graph Library to implement graphs in the Tiger Compiler. You must be able to manipulate Boost Graphs, and understand some aspects of their design.

Flow graph
Liveness
Interference graph/conflict graph

Next: , Previous: , Up: TC-8   [Contents][Index]

4.17.2 TC-8 Samples

First consider simple examples, without any branching:

10 + 20 * 30

File 4.72: tens.tig

$ tc -I tens.tig
# == Final assembler ouput. == #
# Routine: _main
tc_main:
# Allocate frame
        move    $x13, $ra
        move    $x5, $s0
        move    $x6, $s1
        move    $x7, $s2
        move    $x8, $s3
        move    $x9, $s4
        move    $x10, $s5
        move    $x11, $s6
        move    $x12, $s7
l0:
        li      $x1, 10
        li      $x2, 20
        mul     $x3, $x2, 30
        add     $x4, $x1, $x3
l1:
        move    $s0, $x5
        move    $s1, $x6
        move    $s2, $x7
        move    $s3, $x8
        move    $s4, $x9
        move    $s5, $x10
        move    $s6, $x11
        move    $s7, $x12
        move    $ra, $x13
# Deallocate frame
        jr      $ra

Example 4.121: tc -I tens.tig

$ tc -FVN tens.tig

Example 4.122: tc -FVN tens.tig

assignments.ext/tens.main._main.flow

File 4.73: tens.main._main.flow.gv

assignments.ext/tens.main._main.liveness

File 4.74: tens.main._main.liveness.gv

assignments.ext/tens.main._main.interference

File 4.75: tens.main._main.interference.gv

But as you can see, the result is quite hairy, and unreadable, especially for interference graphs:

To circumvent this problem, use --callee-save to limit the number of such registers:

100 + 200 * 300

File 4.76: hundreds.tig

$ tc --callee-save=0 -VN hundreds.tig

Example 4.123: tc --callee-save=0 -VN hundreds.tig

assignments.ext/hundreds.main._main.liveness

File 4.77: hundreds.main._main.liveness.gv

assignments.ext/hundreds.main._main.interference

File 4.78: hundreds.main._main.interference.gv


Branching is of course a most interesting feature to exercise:

1 | 2 | 3

File 4.79: ors.tig

$ tc --callee-save=0 -I ors.tig
# == Final assembler ouput. == #
# Routine: _main
tc_main:
# Allocate frame
        move    $x4, $ra
l5:
        li      $x1, 1
        bne     $x1, 0, l3
l4:
        li      $x2, 2
        bne     $x2, 0, l0
l1:
l2:
        j       l6
l0:
        j       l2
l3:
        li      $x3, 1
        bne     $x3, 0, l0
l7:
        j       l1
l6:
        move    $ra, $x4
# Deallocate frame
        jr      $ra

Example 4.124: tc --callee-save=0 -I ors.tig

$ tc -FVN ors.tig

Example 4.125: tc -FVN ors.tig

assignments.ext/ors.main._main.flow

File 4.80: ors.main._main.flow.gv

assignments.ext/ors.main._main.liveness

File 4.81: ors.main._main.liveness.gv

assignments.ext/ors.main._main.interference

File 4.82: ors.main._main.interference.gv


Next: , Previous: , Up: TC-8   [Contents][Index]

4.17.3 TC-8 Given Code

Some code is provided through the ‘tc-base’ repository, using tag ‘2017-tc-base-8.0’. To read the description of the new modules, see lib/misc, src/liveness.


Next: , Previous: , Up: TC-8   [Contents][Index]

4.17.4 TC-8 Code to Write

lib/misc/graph.*

Implement the topological sort.

src/liveness/flowgraph.*

Write the constructor, which is where the FlowGraph is actually constructed from the assembly fragments.

src/liveness/liveness.*

Write the constructor, which is where the Liveness (a decorated FlowGraph) is built from assembly instructions.

src/liveness/interference-graph.*

In InterferenceGraph::compute_liveness, build the graph.


Next: , Previous: , Up: TC-8   [Contents][Index]

4.17.5 TC-8 FAQ

Why do we have a TempMap, and not Appel?

See $fp or fp, for all the details. Pay special attention to converting the temporaries where needed:

There is another reason to use a TempMap here: to build the liveness graph after register allocation, to check the compiler.

1 & 2

File 4.83: and.tig

$ tc -sV and.tig

Example 4.126: tc -sV and.tig

assignments.ext/and.main._main.liveness

File 4.84: and.main._main.liveness.gv


Previous: , Up: TC-8   [Contents][Index]

4.17.6 TC-8 Improvements

Possible improvements include:


Next: , Previous: , Up: Compiler Stages   [Contents][Index]

4.18 TC-9, Register Allocation

2017-TC-9 is a part of the TC Back End option.
2017-TC-9 submission is Sunday, July 19th 2014 at 11:42.

This section has been updated for EPITA-2017 on 2013-12-30.

At the end of this stage, the compiler produces code that is runnable using Nolimips.

Relevant lecture notes include regalloc.pdf.


Next: , Up: TC-9   [Contents][Index]

4.18.1 TC-9 Goals

Things to learn during this stage that you should remember:


Next: , Previous: , Up: TC-9   [Contents][Index]

4.18.2 TC-9 Samples

This section will not demonstrate the output of the option -S, --asm-display, since it outputs the long Tiger runtime. Once the registers allocated (i.e., once -s, --asm-compute executed) the option -I, --instr-display produces the code without the runtime. In short: we use -sI instead of -S to save place.

Allocating registers in the main function, when there is no register pressure is easy, as, in particular, there are no spills. A direct consequence is that many move are now useless, and have disappeared. For instance (File 4.85, see Example 4.127):

1 + 2 * 3

File 4.85: seven.tig

$ tc -sI seven.tig
# == Final assembler ouput. == #
# Routine: _main
tc_main:
# Allocate frame
l0:
        li      $t1, 1
        li      $t0, 2
        mul     $t0, $t0, 3
        add     $t0, $t1, $t0
l1:
# Deallocate frame
        jr      $ra

Example 4.127: tc -sI seven.tig

$ tc -S seven.tig >seven.s

Example 4.128: tc -S seven.tig >seven.s

$ nolimips -l nolimips -Ne seven.s

Example 4.129: nolimips -l nolimips -Ne seven.s

Another means to display the result of register allocation consists in reporting the mapping from temps to actual registers:

$ tc -s --tempmap-display seven.tig
/* Temporary map. */
fp -> $fp
rv -> $v0
t1 -> $t1
t2 -> $t0
t3 -> $t0
t4 -> $t0
t5 -> $s0
t6 -> $s1
t7 -> $s2
t8 -> $s3
t9 -> $s4
t10 -> $s5
t11 -> $s6
t12 -> $s7
t13 -> $ra

Example 4.130: tc -s --tempmap-display seven.tig

Of course it is much better to see what is going on:

(print_int(1 + 2 * 3); print("\n"))

File 4.86: print-seven.tig

$ tc -sI print-seven.tig
# == Final assembler ouput. == #
.data
l0:
        .word 1
        .asciiz "\n"
.text

# Routine: _main
tc_main:
        sw      $fp, -4 ($sp)
        move    $fp, $sp
        sub     $sp, $sp, 8
        sw      $ra, ($fp)
l1:
        li      $t0, 1
        li      $ra, 2
        mul     $ra, $ra, 3
        add     $a0, $t0, $ra
        jal     tc_print_int
        la      $a0, l0
        jal     tc_print
l2:
        lw      $ra, ($fp)
        move    $sp, $fp
        lw      $fp, -4 ($fp)
        jr      $ra

Example 4.131: tc -sI print-seven.tig

$ tc -S print-seven.tig >print-seven.s

Example 4.132: tc -S print-seven.tig >print-seven.s

$ nolimips -l nolimips -Ne print-seven.s
7

Example 4.133: nolimips -l nolimips -Ne print-seven.s


To torture your compiler, you ought to use many temporaries. To be honest, ours is quite slow, it spends way too much time in register allocation.

let
  var a00 := 00      var a55 := 55
  var a11 := 11      var a66 := 66
  var a22 := 22      var a77 := 77
  var a33 := 33      var a88 := 88
  var a44 := 44      var a99 := 99
in
  print_int(0
            +  a00 + a00 + a55 + a55
            +  a11 + a11 + a66 + a66
            +  a22 + a22 + a77 + a77
            +  a33 + a33 + a88 + a88
            +  a44 + a44 + a99 + a99);
  print("\n")
end

File 4.87: print-many.tig

$ tc -eIs --tempmap-display -I --time-report print-many.tig
error→Execution times (seconds)
error→ 8: liveness edges       : 0      (    0%)   0      (    0%)   0.01   (   50%) 
error→ 9: asm-compute          : 0.02   (  100%)   0      (    0%)   0.02   (  100%) 
error→ 9: coalesce             : 0.01   (   50%)   0      (    0%)   0      (    0%) 
error→ rest                    : 0.02   (  100%)   0      (    0%)   0.02   (  100%) 
error→Cumulated times (seconds)
error→ 7: inst-display         : 0.02   (  100%)   0      (    0%)   0.02   (  100%) 
error→ 8: liveness edges       : 0      (    0%)   0      (    0%)   0.01   (   50%) 
error→ 9: asm-compute          : 0.02   (  100%)   0      (    0%)   0.02   (  100%) 
error→ rest                    : 0.02   (  100%)   0      (    0%)   0.02   (  100%) 
error→ TOTAL (seconds)         : 0.02   user,      0      system,    0.02   wall
# == Final assembler ouput. == #
.data
l0:
        .word 1
        .asciiz "\n"
.text

# Routine: _main
tc_main:
# Allocate frame
        move    $x41, $ra
        move    $x33, $s0
        move    $x34, $s1
        move    $x35, $s2
        move    $x36, $s3
        move    $x37, $s4
        move    $x38, $s5
        move    $x39, $s6
        move    $x40, $s7
l1:
        li      $x0, 0
        li      $x1, 55
        li      $x2, 11
        li      $x3, 66
        li      $x4, 22
        li      $x5, 77
        li      $x6, 33
        li      $x7, 88
        li      $x8, 44
        li      $x9, 99
        li      $x11, 0
        add     $x12, $x11, $x0
        add     $x13, $x12, $x0
        add     $x14, $x13, $x1
        add     $x15, $x14, $x1
        add     $x16, $x15, $x2
        add     $x17, $x16, $x2
        add     $x18, $x17, $x3
        add     $x19, $x18, $x3
        add     $x20, $x19, $x4
        add     $x21, $x20, $x4
        add     $x22, $x21, $x5
        add     $x23, $x22, $x5
        add     $x24, $x23, $x6
        add     $x25, $x24, $x6
        add     $x26, $x25, $x7
        add     $x27, $x26, $x7
        add     $x28, $x27, $x8
        add     $x29, $x28, $x8
        add     $x30, $x29, $x9
        add     $x31, $x30, $x9
        move    $a0, $x31
        jal     tc_print_int
        la      $x32, l0
        move    $a0, $x32
        jal     tc_print
l2:
        move    $s0, $x33
        move    $s1, $x34
        move    $s2, $x35
        move    $s3, $x36
        move    $s4, $x37
        move    $s5, $x38
        move    $s6, $x39
        move    $s7, $x40
        move    $ra, $x41
# Deallocate frame
        jr      $ra
/* Temporary map. */
fp -> $fp
rv -> $v0
t0 -> $t9
t1 -> $t8
t2 -> $t7
t3 -> $t6
t4 -> $t5
t5 -> $t4
t6 -> $t3
t7 -> $t2
t8 -> $t1
t9 -> $t0
t11 -> $ra
t12 -> $ra
t13 -> $ra
t14 -> $ra
t15 -> $ra
t16 -> $ra
t17 -> $ra
t18 -> $ra
t19 -> $ra
t20 -> $ra
t21 -> $ra
t22 -> $ra
t23 -> $ra
t24 -> $ra
t25 -> $ra
t26 -> $ra
t27 -> $ra
t28 -> $ra
t29 -> $ra
t30 -> $ra
t31 -> $a0
t32 -> $a0
t33 -> $s0
t34 -> $s1
t35 -> $s2
t36 -> $s3
t37 -> $s4
t38 -> $s5
t39 -> $s6
t40 -> $s7
t110 -> $ra
t111 -> $ra

# == Final assembler ouput. == #
.data
l0:
        .word 1
        .asciiz "\n"
.text

# Routine: _main
tc_main:
        sw      $fp, -4 ($sp)
        move    $fp, $sp
        sub     $sp, $sp, 8
        sw      $ra, ($fp)
l1:
        li      $t9, 0
        li      $t8, 55
        li      $t7, 11
        li      $t6, 66
        li      $t5, 22
        li      $t4, 77
        li      $t3, 33
        li      $t2, 88
        li      $t1, 44
        li      $t0, 99
        li      $ra, 0
        add     $ra, $ra, $t9
        add     $ra, $ra, $t9
        add     $ra, $ra, $t8
        add     $ra, $ra, $t8
        add     $ra, $ra, $t7
        add     $ra, $ra, $t7
        add     $ra, $ra, $t6
        add     $ra, $ra, $t6
        add     $ra, $ra, $t5
        add     $ra, $ra, $t5
        add     $ra, $ra, $t4
        add     $ra, $ra, $t4
        add     $ra, $ra, $t3
        add     $ra, $ra, $t3
        add     $ra, $ra, $t2
        add     $ra, $ra, $t2
        add     $ra, $ra, $t1
        add     $ra, $ra, $t1
        add     $ra, $ra, $t0
        add     $a0, $ra, $t0
        jal     tc_print_int
        la      $a0, l0
        jal     tc_print
l2:
        lw      $ra, ($fp)
        move    $sp, $fp
        lw      $fp, -4 ($fp)
        jr      $ra

Example 4.134: tc -eIs --tempmap-display -I --time-report print-many.tig


Next: , Previous: , Up: TC-9   [Contents][Index]

4.18.3 TC-9 Given Code

Some code is provided through the ‘tc-base’ repository, using tag ‘2017-tc-base-9.0’. To read the description of the new module, see src/regalloc.


Next: , Previous: , Up: TC-9   [Contents][Index]

4.18.4 TC-9 Code to Write

src/regalloc/color.hh

Implement the graph coloring. The skeleton we provided is an exact copy of the implementation of the code suggest by Andrew Appel in the section 11.4 “Graph Coloring Implementation” of his book. A lot of comments that are verbatim copies of his comments are left in the code. Unfortunately, the books have several nasty mistakes on the algorithm, they reported on his web page (see Modern Compiler Implementation); be sure to fix your books.

Pay attention to misc::set: there is a lot of syntactic sugar provided to implement set operations. The code of Color can range from ugly and obfuscated to readable and very close to its specification.

src/regalloc/regallocator.cc

Run the register allocation on each code fragment. Remove the useless moves.

src/target/mips/epilogue.cc

If your compiler supports spills, implement Codegen::rewrite_program.


Next: , Previous: , Up: TC-9   [Contents][Index]

4.18.5 TC-9 FAQ


Previous: , Up: TC-9   [Contents][Index]

4.18.6 TC-9 Improvements

Possible improvements include:


Previous: , Up: Compiler Stages   [Contents][Index]

4.19 TC-X, IA-32 Back End

TC-X is an optional assignment.

This section has been updated for EPITA-2015 on 2013-07-19.

At the end of this stage, the compiler produces IA-32 code (possibly with infinite registers). Basically, this stage is TC-7 with the IA-32 assembly language instead of MIPS.

The IA-32 architecture is the 32-bit Intel Architecture defined for the Intel 80306 (i386) processors, an extension of the original 16-bit 8086 (x86) architecture. IA-32 may also be referenced as x86, i386 and sometimes x86-32 or even x32, to distinguish it from the original 16-bit (“x86-16”) or the 64-bit (x86-64 or x64) variants of the x86 family.

Relevant lecture notes include instr-selection.pdf.


Next: , Up: TC-X   [Contents][Index]

4.19.1 TC-X Goals

Things to learn during this stage that you should remember:

CISC vs. RISC (again)

MIPS (see TC-7) has shown you an example of RISC architecture. Targeting IA-32 shows you an example of the CISC family of processors.

Compiler toolchain

At the end of the compiler (when register allocation is functional), the IA-32 back end generates code in IA-32 assembly language, which can be assembled and linked to produce a genuine executable program.


Next: , Previous: , Up: TC-X   [Contents][Index]

4.19.2 TC-X Samples

The goal of TC-X is straightforward: starting from LIR, generate the IA-32 instructions, except that you don’t have actual registers: we still heavily use Temps. Register allocation has been (or will be) done in a another stage, TC-9.

let
  var answer := 42
in
  answer := 51
end

File 4.88: the-answer-ia32.tig

$ tc --target-ia32 --inst-display the-answer-ia32.tig
/** Tiger final assembler ouput. */

/** Routine: _main */
        .text
.globl tc_main
        .type   tc_main,@function
tc_main:
# Allocate frame
l0:
        movl    $42, %t1
        movl    %t1, (%ebp)
        movl    $51, %t2
        movl    %t2, (%ebp)
l1:
# Deallocate frame
        ret     $0
l2:
        .size   tc_main,l2-tc_main
        .ident  "LRDE Tiger Compiler"

Example 4.135: tc --target-ia32 --inst-display the-answer-ia32.tig

At this stage the compiler cannot know what registers are used; the frame is not allocated. The final stage, register allocation, addresses this issue. For your information, it results in:

$ tc -sI the-answer-ia32.tig
# == Final assembler ouput. == #
# Routine: _main
tc_main:
        sw      $fp, -4 ($sp)
        move    $fp, $sp
        sub     $sp, $sp, 8
l0:
        li      $t0, 42
        sw      $t0, ($fp)
        li      $t0, 51
        sw      $t0, ($fp)
l1:
        move    $sp, $fp
        lw      $fp, -4 ($fp)
        jr      $ra

Example 4.136: tc -sI the-answer-ia32.tig


A delicate part of this exercise is handling the function calls:

let function add(x: int, y: int) : int = x + y
in
  print_int(add(1,(add(2, 3)))); print("\n")
end

File 4.89: add-ia32.tig

$ tc -e --target-ia32 --inst-display add-ia32.tig
/** Tiger final assembler ouput. */

/** Routine: add */
        .text
.globl tc_l0
        .type   tc_l0,@function
tc_l0:
# Allocate frame
        movl    12(%ebp), %t7
        movl    %t7, (%ebp)
        movl    16(%ebp), %t0
        movl    20(%ebp), %t1
l2:
        movl    %t0, %t6
        addl    %t1, %t6
        movl    %t6, %eax
l3:
# Deallocate frame
        ret     $12
l6:
        .size   tc_l0,l6-tc_l0

        .section        .rodata
l1:
        .long 1
        .asciz "\n"

/** Routine: _main */
        .text
.globl tc_main
        .type   tc_main,@function
tc_main:
# Allocate frame
l4:
        movl    $3, %t8
        pushl   %t8
        movl    $2, %t9
        pushl   %t9
        pushl   %ebp
        call    tc_l0
        movl    %eax, %t4
        pushl   %t4
        movl    $1, %t10
        pushl   %t10
        pushl   %ebp
        call    tc_l0
        movl    %eax, %t5
        pushl   %t5
        call    tc_print_int
        lea     l1, %t11
        pushl   %t11
        call    tc_print
l5:
# Deallocate frame
        ret     $0
l7:
        .size   tc_main,l7-tc_main
        .ident  "LRDE Tiger Compiler"

Example 4.137: tc -e --target-ia32 --inst-display add-ia32.tig


Once your compiler is complete, you can produce an actual IA-32 output, assemble it and link it with gcc to produce a real executable program:

$ tc -e --target-ia32 --asm-compute --inst-display add-ia32.tig
/** Tiger final assembler ouput. */

/** Routine: add */
        .text
.globl tc_l0
        .type   tc_l0,@function
tc_l0:
        pushl   %ebp
        subl    $4, %esp
        movl    %esp, %ebp
        subl    $4, %esp
        movl    12(%ebp), %ebx
        movl    %ebx, (%ebp)
        movl    16(%ebp), %eax
        movl    20(%ebp), %ebx
l2:
        addl    %ebx, %eax
l3:
        addl    $4, %ebp
        leave
        ret     $12
l6:
        .size   tc_l0,l6-tc_l0

        .section        .rodata
l1:
        .long 1
        .asciz "\n"

/** Routine: _main */
        .text
.globl tc_main
        .type   tc_main,@function
tc_main:
        pushl   %ebp
        subl    $4, %esp
        movl    %esp, %ebp
        subl    $0, %esp
l4:
        movl    $3, %ebx
        pushl   %ebx
        movl    $2, %ebx
        pushl   %ebx
        pushl   %ebp
        call    tc_l0
        pushl   %eax
        movl    $1, %ebx
        pushl   %ebx
        pushl   %ebp
        call    tc_l0
        pushl   %eax
        call    tc_print_int
        lea     l1, %ebx
        pushl   %ebx
        call    tc_print
l5:
        addl    $4, %ebp
        leave
        ret     $0
l7:
        .size   tc_main,l7-tc_main
        .ident  "LRDE Tiger Compiler"

Example 4.138: tc -e --target-ia32 --asm-compute --inst-display add-ia32.tig

$ tc -e --target-ia32 --asm-display add-ia32.tig >add-ia32.s

Example 4.139: tc -e --target-ia32 --asm-display add-ia32.tig >add-ia32.s

$ gcc -m32 -oadd-ia32 add-ia32.s

Example 4.140: gcc -m32 -oadd-ia32 add-ia32.s

$ ./add-ia32
6

Example 4.141: ./add-ia32


The runtime must be functional. No difference must be observable in comparison with a run with HAVM:

substring("", 1, 1)

File 4.90: substring-0-1-1-ia32.tig

$ tc -e --target-ia32 --inst-display substring-0-1-1-ia32.tig
/** Tiger final assembler ouput. */

        .section        .rodata
l0:
        .long 0
        .asciz ""

/** Routine: _main */
        .text
.globl tc_main
        .type   tc_main,@function
tc_main:
# Allocate frame
l1:
        movl    $1, %t1
        pushl   %t1
        movl    $1, %t2
        pushl   %t2
        lea     l0, %t3
        pushl   %t3
        call    tc_substring
l2:
# Deallocate frame
        ret     $0
l3:
        .size   tc_main,l3-tc_main
        .ident  "LRDE Tiger Compiler"

Example 4.142: tc -e --target-ia32 --inst-display substring-0-1-1-ia32.tig

$ tc -e --target-ia32 --asm-compute --inst-display substring-0-1-1-ia32.tig
/** Tiger final assembler ouput. */

        .section        .rodata
l0:
        .long 0
        .asciz ""

/** Routine: _main */
        .text
.globl tc_main
        .type   tc_main,@function
tc_main:
        pushl   %ebp
        subl    $4, %esp
        movl    %esp, %ebp
        subl    $0, %esp
l1:
        movl    $1, %ebx
        pushl   %ebx
        movl    $1, %ebx
        pushl   %ebx
        lea     l0, %ebx
        pushl   %ebx
        call    tc_substring
l2:
        addl    $4, %ebp
        leave
        ret     $0
l3:
        .size   tc_main,l3-tc_main
        .ident  "LRDE Tiger Compiler"

Example 4.143: tc -e --target-ia32 --asm-compute --inst-display substring-0-1-1-ia32.tig

$ tc -e --target-ia32 --asm-display substring-0-1-1-ia32.tig >substring-0-1-1-ia32.s

Example 4.144: tc -e --target-ia32 --asm-display substring-0-1-1-ia32.tig >substring-0-1-1-ia32.s

$ gcc -m32 -osubstring-0-1-1-ia32 substring-0-1-1-ia32.s

Example 4.145: gcc -m32 -osubstring-0-1-1-ia32 substring-0-1-1-ia32.s

$ ./substring-0-1-1-ia32
error→substring: arguments out of bounds
⇒120

Example 4.146: ./substring-0-1-1-ia32


The following example illustrates conditional jumps.

if 42 > 51 then "forty-two" else "fifty-one"

File 4.91: condjump-ia32.tig

$ tc -e --target-ia32 --inst-display condjump-ia32.tig
/** Tiger final assembler ouput. */

        .section        .rodata
l0:
        .long 9
        .asciz "forty-two"

        .section        .rodata
l1:
        .long 9
        .asciz "fifty-one"

/** Routine: _main */
        .text
.globl tc_main
        .type   tc_main,@function
tc_main:
# Allocate frame
l5:
        movl    $42, %t1
        cmp     $51, %t1
        jg      l2
l3:
        lea     l1, %t2
l4:
        jmp     l6
l2:
        lea     l0, %t3
        jmp     l4
l6:
# Deallocate frame
        ret     $0
l7:
        .size   tc_main,l7-tc_main
        .ident  "LRDE Tiger Compiler"

Example 4.147: tc -e --target-ia32 --inst-display condjump-ia32.tig

$ tc -e --target-ia32 --asm-compute --inst-display condjump-ia32.tig
/** Tiger final assembler ouput. */

        .section        .rodata
l0:
        .long 9
        .asciz "forty-two"

        .section        .rodata
l1:
        .long 9
        .asciz "fifty-one"

/** Routine: _main */
        .text
.globl tc_main
        .type   tc_main,@function
tc_main:
        pushl   %ebp
        subl    $4, %esp
        movl    %esp, %ebp
        subl    $0, %esp
l5:
        movl    $42, %ebx
        cmp     $51, %ebx
        jg      l2
l3:
        lea     l1, %ebx
l4:
        jmp     l6
l2:
        lea     l0, %ebx
        jmp     l4
l6:
        addl    $4, %ebp
        leave
        ret     $0
l7:
        .size   tc_main,l7-tc_main
        .ident  "LRDE Tiger Compiler"

Example 4.148: tc -e --target-ia32 --asm-compute --inst-display condjump-ia32.tig


Next: , Previous: , Up: TC-X   [Contents][Index]

4.19.3 TC-X Given Code

Some code is provided along with the code given at TC-7 (see TC-7 Given Code). See src/target/ia32.


Next: , Previous: , Up: TC-X   [Contents][Index]

4.19.4 TC-X Code to Write

There is not much code to write:

Information on IA-32 assembly instructions may be found in the Intel® 64 and IA-32 Architectures Software Developer Manuals or in this much shorter IA32 Instruction List form. The documentation of the GNU Assembler (GAS) is also a recommended reading.


Completing the following routines is needed for register allocation only (see TC-9):


Next: , Previous: , Up: TC-X   [Contents][Index]

4.19.5 TC-X FAQ


Previous: , Up: TC-X   [Contents][Index]

4.19.6 TC-X Improvements

Possible improvements include:


Next: , Previous: , Up: Top   [Contents][Index]

5 Tools

This chapter aims at providing some helpful information about the various tools that you are likely to use to implement tc. It does not replace the reading of the genuine documentation, nevertheless, helpful tips are given. Feel free to contribute additional information.


Next: , Up: Tools   [Contents][Index]

5.1 Programming Environment

This section lists the tools you need to work in good conditions.

ToolVersionComment
GCC4.8See GCC.
Clang3.3Optional: See Clang.
Autoconf2.64See The GNU Build System.
Automake1.14.1See The GNU Build System.
Libtool2.2.6See The GNU Build System.
GNU Make3.81
Boost1.53TC >= 5, See Boost.org.
Doxygen1.5.1See Doxygen.
Python2.5See Python.
SWIG2.0Optional: See SWIG.
Flex2.5.35See Flex & Bison.
Bison3.0.4.19-fbafSee Flex & Bison.
HAVM0.26TC >= 5, See HAVM.
MonoBURG1.0.6aTC >= 7, See MonoBURG.
Nolimips0.9aTC >= 7, See Nolimips.
GDB6.6See GDB.
Valgrind3.6See Valgrind.
Git1.7
xvcg3.17Optional: display LALR(1) automata.
GraphViz2.26.3Optional: display DOT graphs.

Next: , Previous: , Up: Tools   [Contents][Index]

5.2 Modern Compiler Implementation

The Tiger Bible exists in two profoundly different versions.


Next: , Up: Modern Compiler Implementation   [Contents][Index]

5.2.1 First Editions

The single most important tool for implementing the Tiger Project is the original book, Modern Compiler Implementation in C/Java/ML, by Andrew W. Appel, published by Cambridge University Press (New York, Cambridge). ISBN 0-521-58388-8/.

It is not possible to finish this project without having at least one copy per group. We provide a convenient mini Tiger Compiler Reference Manual that contains some information about the language but it does not cover all the details, and sometimes digging into the original book is required. This is on purpose, by virtue of due respect to the author of this valuable book.

Several copies are available at the EPITA library.

There are three flavors of this book:

C
assignments.img/appel-c

The code samples are written in C. Avoid this edition, as C is not appropriate to describe the elaborate algorithms involved: most of the time, the simple ideas are destroyed with longuish unpleasant lines of code.

Java, First edition
assignments.img/appel-java-1

The samples are written in Java. This book is the closest to the EPITA Tiger Project, since it is written in an object oriented language. Nevertheless, the modelisation is very poor, and therefore, don’t be surprised if the EPITA project is significantly different. For a start, there is no Visitors at all. Of course the main purpose of the book is compilers, but it is not a reason for such a poor modelisation.

ML
assignments.img/appel-ml

This book, which is the “original”, provides code samples in ML, which is a very adequate language to write compilers. Therefore it is very readable, even if you are not fluent in ML. We recommend this edition, unless you have severe problems with functional programming.

This book addresses many more issues than the sole Tiger Project as we implement it. In other words, it is an extremely interesting book whose provides insights on garbage collection, object oriented and functional languages etc.

There is a dozen copies at the EPITA library, but buying it is a good idea.


Pay extra attention: there are several errors in the books, some of which are reported on Andrew Appel’s pages (C Java, and ML), and others are not.

Because these pages no longer seem to be maintained, additional errors are reported below. “p. C.245” means page 245 in the C book. Please send us additions.

11.3. Example with Precolored Nodes (p. C.245)

The first interference graph presented for this example lacks the interference between r1 and c.

11.4 Graph Coloring Implementation (p. C.248)

In the first sentence, s/inteference/interference/.


Previous: , Up: Modern Compiler Implementation   [Contents][Index]

5.2.2 In Java - Second Edition

assignments.img/appel-java-2

The Second Edition of Modern Compiler Implementation in Java, by Andrew W. Appel and Jens Palsberg, published by Cambridge University Press (New York, Cambridge), ISBN 052182060X, is a very different book from the rest of the series.

While, finally, the design is much better, starting with the introduction of the Visitors, there are many shortcoming for us:

Nevertheless, because we don’t encourage book copying, we now provide a complete definition of the Tiger language in Tiger Language Reference Manual in Tiger Compiler Reference Manual.


Next: , Previous: , Up: Tools   [Contents][Index]

5.3 Bibliography

Below is presented a selection of books, papers and web sites that are pertinent to the Tiger project. Of course, you are not requested to read them all, except Modern Compiler Implementation. A suggested ordered small selection of books is:

  1. Modern Compiler Implementation
  2. C++ Primer
  3. Design Patterns - Elements of Reusable Object-Oriented Software
  4. Effective C++
  5. Effective STL

The books are available at the EPITA Library: you are encouraged to borrow them there. If some of these books are missing, please suggest them to the library’s manager. To buy these books, we recommend Le Monde en “tique”, a bookshop that has demonstrated several times its dedication to its job, and its kindness to EPITA students/members.

Web Site: Autotools Tutorial -- Alexandre Duret-Lutz

The Autotools Tutorial is the best introduction to Autoconf, Automake, and Libtool, that we know. It covers also other components of the GNU Build System. You should read this before diving into the documentation.

Other resources include:

Web Site: Bjarne Stroustrup
assignments.img/bjarne

Bjarne Stroustrup is the author of C++, which he describes as (The C++ Programming Language):

C++ is a general purpose programming language with a bias towards systems programming that

  • − is a better C
  • − supports data abstraction
  • − supports object-oriented programming
  • − supports generic programming.

His web page contains interesting material on C++, including many interviews. The interview by Aleksey V. Dolya for the Linux Journal contains thoughts about C and C++. For instance:

I think that the current mess of C/C++ incompatibilities is a most unfortunate accident of history, without a fundamental technical or philosophical basis. Ideally the languages should be merged, and I think that a merger is barely technically possible by making convergent changes to both languages. It seems, however, that because there is an unwillingness to make changes it is likely that the languages will continue to drift apart–to the detriment of almost every C and C++ programmer. [...] However, there are entrenched interests keeping convergence from happening, and I’m not seeing much interest in actually doing anything from the majority that, in my opinion, would benefit most from compatibility.

His list of C++ Applications is worth the browsing.

Web Site: Boost.org

The Boost.org web site reads:

The Boost web site provides free peer-reviewed portable C++ source libraries. The emphasis is on libraries that work well with the C++ Standard Library. One goal is to establish "existing practice" and provide reference implementations so that the Boost libraries are suitable for eventual standardization. Some of the libraries have already been proposed for inclusion in the C++ Standards Committee’s upcoming C++ Standard Library Technical Report.

In addition to actual code, a lot of good documentation is available. Amongst libraries, you ought to have a look at the Spirit object-oriented recursive-descent parser generator framework, the Boost Graph Library, the Boost Variant Library etc.

Paper: BURG: Fast Optimal Instruction Selection and Tree Parsing -- Christopher W. Fraser, Robert R. Henry, Todd A. Proebsting

SIGPLAN Notices 24(4), 68-76. 1992.

This paper is a description of BURG and an introduction to the concept of code generator generators.

Book: Compilers and Compiler Generators, an introduction with C++ -- P.D. Terry

Its site reads:

This site provides an on-line edition of the text and other material from the book "Compilers and Compiler Generators - an introduction with C++", published in 1997 by International Thomson Computer Press. The original edition is now out of print, and the copyright has reverted to the author.

This book is not very interesting for us: it depends upon tools we don’t use, its C++ is antique, and its approach to compilation is significantly different from Appel’s.

Book: C++ Primer -- Stanley B. Lippman, Josée Lajoie
assignments.img/c++-primer

Published by Addison-Wesley; ISBN 0-201-82470-1.

This book teaches C++ for programmers. It is quite extensive and easy to read. Unfortunately it is not 100% standard compliant, in particular many std:: are missing. Weirdly enough, the authors seems to promote using declarations instead of explicit qualifiers; the page 441 reads:

In this book, to keep the code examples Short, and because many of the examples were compiled with implementations not supporting namespace, we have not explicitly listed the using declarations needed to properly compile the examples. It is assumed that using declarations are provided for the members of namespace std used in the code examples.

It should not be too much of a problem though. This is the book we recommend to learn C++. See the Addison-Wesley C++ Primer Page.

Warning: The French translation is L’Essentiel du C++, which is extremely stupid since Essential C++ is another book from Stanley B. Lippman (but not with Josée Lajoie).

Book: Compilers: Principles, Techniques and Tools -- Alfred V. Aho, Ravi Sethi, and Jeffrey D. Ullman
Book: The Dragon Book
assignments.img/compilers-principles-techniques-and-tools

Published by Addison-Wesley 1986; ISBN 0-201-10088-6.

This book is the bible in compiler design. It has extensive insight on the whole architecture of compilers, provides a rigorous treatment for theoretical material etc. Nevertheless I (Akim) would not recommend this book to EPITA students, because

it is getting old

It doesn’t mention RISC, object orientation, functional, modern optimization techniques such as SSA, register allocation by graph coloring 9 etc.

it is fairly technical

The book can be hard to read for the beginner, contrary to Modern Compiler Implementation.

Nevertheless, curious readers will find valuable information about historically important compilers, people, papers etc. Reading the last section of each chapter (Bibliographical Notes) is a real pleasure for whom is interested.

It should be noted that the French edition, “Compilateurs: Principes, techniques et outils”, was brilliantly translated by Pierre Boullier, Philippe Deschamp, Martin Jourdan, Bernard Lorho and Monique Lazaud: the pleasure is as good in French as it is in English.

Web Site: Cool: The Classroom Object-Oriented Compiler

The Classroom Object-Oriented Compiler, from the University of California, Berkeley, is very similar in its goals to the Tiger project as described here. Unfortunately it seems dead: there are no updates since 1996. Nevertheless, if you enjoy the Tiger project, you might want to see its older siblings.

Paper: CStupidClassName -- Dejan Jelović

This short paper, CStupidClassName, explains why naming classes CLikeThis is stupid, but why lexical conventions are nevertheless very useful. It turns out we follow the same scheme that is emphasized there.

Book: Design Patterns: Elements of Reusable Object-Oriented Software -- Erich Gamma, Richard Helm, Ralph Johnson, and John Vlissides
assignments.img/design-patterns

Published by Addison-Wesley; ISBN: 0-201-63361-2.

A book you must have read, or at least, you must know it. In a few words, let’s say it details nice programming idioms, some of them you should know: the VISITOR, the FLYWEIGHT, the SINGLETON etc. See the Design Patterns Addison-Wesley Page. A pre-version of this book is available on the Internet as a paper: Design Patterns: Abstraction and Reuse of Object-Oriented Design. Surprisingly, The full version of Design Pattern CD is available on the net.

You may find additional information about Design Patterns on the Portland Pattern Repository.

Book: Effective C++ -- Scott Meyers
assignments.img/effective-c++

288 pages; Publisher: Addison-Wesley Pub Co; 2nd edition (September 1997); ISBN: 0-201-92488-9

An excellent book that might serve as a C++ lecture for programmers. Every C++ programmer should have read it at least once, as it treasures C++ recommended practices as a list of simple commandments. Be sure to buy the second edition, as the first predates the C++ standard. See the Effective C++ Addison-Wesley Page.

In this document, ECN refers to item n in Effective C++.

Book: Effective STL -- Scott Meyers
assignments.img/effective-stl

Published by Addison-Wesley; ISBN: 0-201-74962-9

A remarkable book that provides deep insight on the best practice with STL. Not only does it teach what’s to be done, but it clearly shows why. A book that any C++ programmer should have read. See the Effective STL Addison-Wesley Page.

In this document, ESN refers to item n in Effective STL.

Paper: Engineering a simple, efficient code generator generator -- Christopher W. Fraser, David R. Hanson, Todd A. Proebsting

ACM Letters on Programming Languages and Systems 1, 3 (Sep. 1992), 213-226.

This paper describes iburg, a BURG clone that delay dynamic programming at compile time (BURG-like programs use dynamic programming to select the optimum tree tiling during a bottom-up walk).

Technical Report: Generic Visitors in C++ -- Nicolas Tisserand

This report is available on line from Visitors Page: Generic Visitors in C++. Its abstract reads:

The Visitor design pattern is a well-known software engineering technique that solves the double dispatch problem and allows decoupling of two inter-dependent hierarchies. Unfortunately, when used on hierarchies of Composites, such as abstract syntax trees, it presents two major drawbacks: target hierarchy dependence and mixing of traversal and behavioral code.

CWI’s visitor combinators are a seducing solution to these problems. However, their use is limited to specific “combinators aware” hierarchies.

We present here Visitors, our attempt to build a generic, efficient C++ visitor combinators library that can be used on any standard “visitable” target hierarchies, without being intrusive on their codes.

This report is in the spirit of Modern C++ Design, and should probably be read afterward.

News: Guru of the Week

Written by various authors, compiled by Herb Sutter

Guru of the Week (GotW) is a regular series of C++ programming problems created and written by Herb Sutter. Since 1997, it has been a regular feature of the Internet newsgroup comp.lang.c++.moderated, where you can find each issue’s questions and answers (and a lot of interesting discussion).

The Guru of the Week Archive (the famous GotW) is freely available. In this document, GotWn refers to the item number n.

Article: How not to go about a programming assignment -- Agustín Cernuda del Río

This paper provides excellent advice on how to succeed an assignment by showing the converse: how not to go about a programming assignment:

Book: Lex & Yacc -- John R. Levine, Tony Mason, Doug Brown

Published by O’Reilly & Associates; 2nd edition (October 1992); ISBN: 1-565-92000-7.

assignments.img/lex-and-yacc

Because the books aims at a complete treatment of Lex and Yacc on a wide range of platforms, it provides too many details on material with little interest for us (e.g., we don’t care about portability to other Lexes and Yacces), and too few details on material with big interest for us (more about exclusive start condition (Flex only), more about Bison only stuff, interaction with C++ etc.).

Article: Making Compiler Design Relevant for Students who will (Most Likely) Never Design a Compiler -- Saumya K. Debray

This paper about teaching compilers justifies this lecture. This paper is addressing compiler construction lectures, not compiler construction projects, and therefore it misses quite a few motivations we have for the Tiger project.

Book: Modern C++ Design -- Generic Programming and Design Patterns Applied -- Andrei Alexandrescu
assignments.img/modern-c++-design

Published by Addison-Wesley in 2001; ISBN: 0-52201-70431-5

A wonderful book on very advanced C++ programming with a heavy use of templates to achieve beautiful and useful designs (including the classical design patterns, see Design Patterns - Elements of Reusable Object-Oriented Software). The code is available in the form of the Loki Library. The Modern C++ Design Web Site includes pointers to excerpts such as the Smart Pointers chapter.

Read this book only once you have gained good understanding of the C++ core language, and after having read the “Effective C++/STL” books.

Book: Modern Compiler Implementation in C, Java, ML -- Andrew W. Appel

Published by Cambridge University Press; ISBN: 0-521-58390-X

assignments.img/appel-c assignments.img/appel-java-1 assignments.img/appel-ml assignments.img/appel-java-2

See Modern Compiler Implementation. In our humble opinion, most books give way too much emphasis to scanning and parsing, leaving little material to the rest of the compiler, or even nothing for advanced material. This book does not suffer these flaws.

Web Site: Object Management Group

OMG’s Home Page, with a lot of ressources for object-oriented software engineering, particularly on the Unified Modeling Language (UML).

Book: Parsing Techniques -- A Practical Guide -- Dick Grune and Ceriel J. Jacob

Published by the authors; ISBN: 0-13-651431-6

A remarkable review of all the parsing techniques. Because the book is out of print, its authors made it freely available: Parsing Techniques – A Practical Guide.

Book: Programming: Principles and Practice Using C++ -- Bjarne Stroustrup
assignments.img/programming-principles-and-practice-using-c++

This book targets the “advanced beginner” in C++ and covers a wide range of topics including non-core C++ subjects such as GUI programming. A recommended lecture for modern C++ learning.

Published by Addison-Wesley Professional, 2008; ISBN-13: 978-0321543721.

Report: SPOT : une bibliothèque de vérification de propriétés de logique temporelle à temps linéaire -- Alexandre Duret-Lutz & Rachid Rebiha

This report presents SPOT, a model checking library written in C++ and Python. Parts were inspired by the Tiger project, and reciprocally, parts inspired modifications in the Tiger project. For instance, you are encouraged to read the sections about the visitor hierarchy and its implementation. Another useful source of inspiration is the use of Python and Swig to write the command line interface.

Paper: Testing student-made compilers -- José de Oliveira Guimarães

ACM SIGCSE Bulletin archive Volume 26, Issue 3 (September 1994).

This paper gives a classified list of test cases for a small Pascal compiler. It is a good source of inspiration for any other language.

Book: The Design and Evolution of C++ -- Bjarne Stroustrup
assignments.img/the-design-and-evolution-of-c++

Published by Addison-Wesley, ISBN 0-201-54330-3.

This book is definitely worth reading for curious C++ programmers. I (Roland) find it an excellent companion to reference C++ books, or even to the C++ standard. Many aspects of the language that are often criticized find a justification in this book. Moreover, the book not only tells the history of C++ (up to 1994), but it also explains the design choices and reflexions of its authors (and Bjarne Stroustrup’s in the first place), which go far beyond the scope of C++.

However, the book only describes the first 15 years of C++ or so. Recent work on C++ (and especially on the C++0x effort that eventually led to C++ 2011) can be found in Stroustrup’s papers, available online.

Book: The Elements of Style -- William Strunk Jr., E.B. White
assignments.img/the-elements-of-style

Published by Pearson Allyn & Bacon; 4th edition (January 15, 2000); ISBN: 020530902X.

This little book (105 pages) is perfect for people who want to improve their English prose. It is quite famous, and, in addition to providing useful writing thumb rules, it features rules that are interesting as pieces of writing themselves! For instance “The writer must, however, be certain that the emphasis is warranted, lest a clipped sentence seem merely a blunder in syntax or in punctuation”.

You may find the much shorter (43 pages) First Edition of The Elements of Style on line.

Book: Thinking in C++ Volume 1 -- Bruce Eckel

Published by Prentice Hall; ISBN: 0-13-979809-9

Available on the Internet on many Book Download Sites. For instance, Thinking in C++ Volume 1 Zipped.

Book: Thinking in C++ Volume 2 -- Bruce Eckel and Chuck Allison

Available on the Internet on many Book Download Sites. For instance, Thinking in C++ Volume 2 Zipped.

Article: Traits: a new and useful template technique -- Nathan C. Myers

The first presentation of the traits technique is from this paper, Traits: a new and useful template technique. It is now a common C++ programming idiom, which is even used in the C++ standard.

Book: Writing Compilers and Interpreters -- An Applied Approach Using C++ -- Ronald Mak

Published by Wiley; Second Edition, ISBN: 0-471-11353-0

This book is not very interesting for us: the compiler material is not very advanced (no real AST, not a single line on optimization, register allocation is naive as the translation is stack based etc.), and the C++ material is not convincing (for a start, it is not standard C++ as it still uses ‘#include <iostream.h>’ and the like, there is no use of STL etc.).

Web site: STL Home

SGI’s STL Home Page, which includes the complete documentation on line.


Next: , Previous: , Up: Tools   [Contents][Index]

5.4 The GNU Build System

Automake is used to facilitate the writing of power Makefile. Libtool eases the creation of libraries, especially dynamic ones. Autoconf is required by Automake: we do not address portability issues for this project. See Autotools Tutorial, for documentation.

Using info is pleasant, for instance ‘info autoconf’ on any properly set up system.


Next: , Up: The GNU Build System   [Contents][Index]

5.4.1 Package Name and Version

To set the name and version of your package, change the AC_INIT invocation. For instance, TC-4 for the bardec_f group gives:

AC_INIT([Bardeche Group Tiger Compiler], 4, [bardec_f@epita.fr],
        [bardec_f-tc])

Next: , Previous: , Up: The GNU Build System   [Contents][Index]

5.4.2 Bootstrapping the Package

If something goes wrong, or if it is simply the first time you create configure.ac or a Makefile.am, you need to set up the GNU Build System. That’s the goal of the simple script bootstrap, which most important action is invoking:

$ autoreconf -fvi

The various files (configure, Makefile.in, etc.) are created. There is no need to run ‘make distclean’, or aclocal or whatever, before running autoreconf: it knows what to do.

Then invoke configure and make (see GCC):

$ mkdir _build
$ cd _build
$ ../configure CXX=g++-4.8
$ make

Alternatively you may set CC and CXX in your environment:

$ export CXX=g++-4.8
$ mkdir _build
$ cd _build
$ ../configure && make

This solution is preferred since the value of CC etc. will be used by the configure invocation from ‘make distcheck’ (see Making a Tarball).


Next: , Previous: , Up: The GNU Build System   [Contents][Index]

5.4.3 Making a Tarball

Once the package correctly autotool’ed and configured (see Bootstrapping the Package), run ‘make distcheck’ to build the tarball. Contrary to a simple ‘dist’, ‘distcheck’ makes sure everything will work properly. In particular it:

  1. performs some simple checks. For instance, it checks that the NEWS file is about the current version, i.e., it checks that the second argument given to AC_INIT is in the top of NEWS, otherwise it fails with ‘NEWS not updated; not releasing’.
  2. creates the tarball (via ‘make dist’)
  3. untars the tarball
  4. configures the tarball in a separate directory _build (to avoid cluttering the source files with the built files).

    Arguments passed to the top level configure (e.g., ‘CXX=g++-4.8’) will not be taken into account here. Running ‘export CXX=g++-4.8’ is a better way to require these compilers. Alternatively use DISTCHECK_CONFIGURE_FLAGS to specify the arguments of the embedded configure:

    $ make distcheck DISTCHECK_CONFIGURE_FLAGS='--without-swig CXX=g++-4.0'
    
  5. runs ‘make’ (and following targets) in paranoid mode. This mode consists in forbidding any change in the source tree, because if, when you run ‘make’ something must be changed in the sources, then it means something is broken in the tarball. If, for instance, for some reason it wants to run autoconf to recreate configure, or if it complains that autom4te.cache cannot be created, then it means the tarball is broken! So track down the reason of the failure.
  6. runs ‘make check
  7. runs ‘make dist’ again.

If you just run ‘make dist’ instead of ‘make distcheck’, then you might not notice some files are missing in the distribution. If you don’t even run ‘make dist’, the tarball might not compile elsewhere (not to mention that we don’t care about object files etc.).

Running ‘make distcheck’ is the only means for you to check that the project will properly compile on our side. Not running distcheck is like turning off the type checking of your compiler: you hide instead of solving.


At this stage, if running ‘make distcheck’ does not create bardec_f-tc-4.tar.bz2, something is wrong in your package. Do not rename it, do not create the tarball by hand: something is rotten and be sure it will break on the examiner’s machine.


Previous: , Up: The GNU Build System   [Contents][Index]

5.4.4 Setting site defaults using CONFIG_SITE

Another way to pass options to configure is to use a site configuration file. This file will be “sourced” by configure to set some values and options, and will save you some bytes on your command line when you’ll invoke configure.

First, write a config.site file:

# -*- shell-script -*-

echo "Loading config.site for $PACKAGE_TARNAME"
echo "(srcdir: $srcdir)"
echo

package=$PACKAGE_TARNAME

echo "config.site: $package"
echo

# Configuration specific to EPITA KB machines (GNU/Linux on x86-64).
case $package in
  tc)
    # Turn off optimization when building with debugging information
    # (the build dir must have ``debug'' in its name).
    case `pwd` in
      *debug*) :
        : ${CFLAGS="-ggdb -O0"}
        : ${CXXFLAGS="-ggdb -O0 -D_GLIBCXX_DEBUG"}
      ;;
    esac
    # Help configure to find the Boost libraries on NetBSD.
    if test -f /usr/pkg/include/boost/config.hpp; then
      with_boost=/usr/pkg/include
    fi

    # Set CC, CXX, BISON, MONOBURG, and other programs as well.
    : ${CC=/u/prof/acu/pub/NetBSD/bin/gcc}
    : ${CXX=/u/prof/acu/pub/NetBSD/bin/g++}
    : ${BISON=/u/prof/yaka/bin/bison}
    : ${MONOBURG=/u/prof/yaka/bin/monoburg}
    # ...
  ;;
esac

set +vx

Then, set the environment variable CONFIG_SITE to the path to this file, and run configure:

$ export CONFIG_SITE="$HOME/src/config.site"
$ ../configure

or if you use a C-shell:

$ setenv CONFIG_SITE "$HOME/src/config.site"
$ ../configure

This is useful when invoking make distcheck: you don’t need to pollute your environment, nor use Automake’s DISTCHECK_CONFIGURE_FLAGS (see Making a Tarball).

Of course, you can have several config.site files, one for each architecture you work on for example, and set the CONFIG_SITE variable according to the host/system.


Next: , Previous: , Up: Tools   [Contents][Index]

5.5 GCC, The GNU Compiler Collection

We use GCC 4.8, which includes both gcc-4.8 and g++-4.8: the C and C++ compilers. Do not use older versions as they have poor compliance with the C++ standard. You are welcome to use more recent versions of GCC if you can use one, but the tests will be done with 4.8. Using a more recent version is often a good means to get better error messages if you can’t understand what GCC 4.8 is trying to say.

There are good patches floating around to improve GCC. The GCC Bounds Checking Page is an interesting example in this respect. It is however no longer maintained and we advise you to have a look at mudflap instead, which is officially part of GCC.


Next: , Previous: , Up: Tools   [Contents][Index]

5.6 Clang, A C language family front end for LLVM

Clang is a front end for the LLVM compiler infrastructure supporting the C, C++, Objective C and Objective C++ languages. LLVM provides a modern framework written in C++ for creating compiler-related projects.

We advise you to check your code with the clang (C) and clang++ (C++) front ends (version 3.3 or more) in addition to gcc and g++. Clang may indeed report other errors and warnings. Moreover, Clang’s messages are often easier to read than GCC’s.

You can find more information on Clang, LLVM and other related projects on the LLVM Home Page.


Next: , Previous: , Up: Tools   [Contents][Index]

5.7 GDB, The GNU Project Debugger

Every serious project development makes use of a debugger. Such a tool allows the programmer to examine her program, running it step by step, display/change values etc.

GDB is a debugger for programs written in C, C++, Objective-C, Pascal (and other languages). It will help you to track and fix bugs in your project. Don’t forget to pass the option -g (or -ggdb, depending on your linker’s abilities to handle GDB extensions) to your compiler to include useful information into the debugged program.

Pay attention when debugging a libtoolized program, as it may be a shell script wrapper around the real binary. Thus don’t use

$ gdb tc

or expect errors from GDB when running the program. Use libtool’s --mode=execute option to run gdb instead:

$ libtool --mode=execute gdb tc

or the following shortcut:

$ libtool exe gdb tc

Detailed explanations can be found in the Libtool manual.


Next: , Previous: , Up: Tools   [Contents][Index]

5.8 Valgrind, The Ultimate Memory Debugger

Valgrind is an open-source memory debugger for GNU/Linux on x86/x86-64 (and other environments) written by Julian Seward, already known for having committed Bzip2. It is the best news for programmers for years. Valgrind is so powerful, so beautifully designed that you definitely should wander on the Valgrind Home Page.

In the case of the Tiger Compiler Project correct memory management is a primary goal. To this end, Valgrind is a precious tool, as is dmalloc, but because STL implementations are often keeping some memory for efficiency, you might see “leaks” from your C++ library. See its documentation on how to reclaim this memory. For instance, reading the GCC’s C++ Library FAQ, especially the item “memory leaks” in containers is enlightening.

I (Akim) personally use the following shell script to track memory leaks:

#! /bin/sh

exec 3>&1
export GLIBCPP_FORCE_NEW=1
export GLIBCXX_FORCE_NEW=1
exec valgrind --num-callers=20 \
              --leak-check=yes \
              --leak-resolution=high \
              --show-reachable=yes \
              "$@" 2>&1 1>&3 3>&- |
  sed 's/^==[0-9]*==/==/' >&2 1>&2 3>&-

File 5.1: v

For instance on File 4.52,

$ v tc -XA 0.tig
error→== Memcheck, a memory error detector
error→== Copyright (C) 2002-2013, and GNU GPL'd, by Julian Seward et al.
error→== Using Valgrind-3.10.0 and LibVEX; rerun with -h for copyright info
error→== Command: tc -XA 0.tig
error→== 
error→== 
error→== HEAP SUMMARY:
error→==     in use at exit: 208 bytes in 1 blocks
error→==   total heap usage: 1,700 allocs, 1,699 frees, 191,551 bytes allocated
error→== 
error→== 208 bytes in 1 blocks are still reachable in loss record 1 of 1
error→==    at 0x4C298A0: operator new[](unsigned long) (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
error→==    by 0x40B9511: std::ios_base::_M_grow_words(int, bool) (in /lrde/dev/x86_64/stow/gcc-4.9.0/lib64/libstdc++.so.6.0.20)
error→==    by 0x46CB89: iword (ios_base.h:744)
error→==    by 0x46CB89: indent (indent.cc:22)
error→==    by 0x46CB89: misc::iendl(std::ostream&) (indent.cc:50)
error→==    by 0x4686BB: operator<< (ostream:113)
error→==    by 0x4686BB: ast::PrettyPrinter::operator()(ast::FunctionDec const&, std::string const&) (pretty-printer.cc:305)
error→==    by 0x46891D: ast::PrettyPrinter::operator()(ast::FunctionDec const&) (pretty-printer.cc:332)
error→==    by 0x435AC7: decs_visit<ast::AnyDecs<ast::FunctionDec> > (default-visitor.hxx:234)
error→==    by 0x435AC7: ast::GenDefaultVisitor<misc::constify_traits>::operator()(ast::AnyDecs<ast::FunctionDec> const&) (default-visitor.hxx:258)
error→==    by 0x4350EF: operator()<ast::Decs> (visitor.hxx:31)
error→==    by 0x4350EF: for_each<const std::list<ast::Decs*>, ast::GenDefaultVisitor<misc::constify_traits> > (algorithm.hxx:51)
error→==    by 0x4350EF: ast::GenDefaultVisitor<misc::constify_traits>::operator()(ast::DecsList const&) (default-visitor.hxx:203)
error→==    by 0x466773: ast::operator<<(std::ostream&, ast::Ast const&) (libast.cc:36)
error→==    by 0x40BB2E: ast::tasks::ast_display() (tasks.cc:26)
error→==    by 0x412303: task::TaskRegister::execute() (task-register.cc:295)
error→==    by 0x409695: main (tc.cc:31)
error→== 
error→== LEAK SUMMARY:
error→==    definitely lost: 0 bytes in 0 blocks
error→==    indirectly lost: 0 bytes in 0 blocks
error→==      possibly lost: 0 bytes in 0 blocks
error→==    still reachable: 208 bytes in 1 blocks
error→==         suppressed: 0 bytes in 0 blocks
error→== 
error→== For counts of detected and suppressed errors, rerun with: -v
error→== ERROR SUMMARY: 0 errors from 0 contexts (suppressed: 0 from 0)
/* == Abstract Syntax Tree. == */

function _main() =
  (
    0;
    ()
  )

Example 5.1: v tc -XA 0.tig

$ v tc -XAD 0.tig
error→== Memcheck, a memory error detector
error→== Copyright (C) 2002-2013, and GNU GPL'd, by Julian Seward et al.
error→== Using Valgrind-3.10.0 and LibVEX; rerun with -h for copyright info
error→== Command: tc -XAD 0.tig
error→== 
error→== 
error→== HEAP SUMMARY:
error→==     in use at exit: 208 bytes in 1 blocks
error→==   total heap usage: 1,714 allocs, 1,713 frees, 192,172 bytes allocated
error→== 
error→== 208 bytes in 1 blocks are still reachable in loss record 1 of 1
error→==    at 0x4C298A0: operator new[](unsigned long) (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
error→==    by 0x40B9511: std::ios_base::_M_grow_words(int, bool) (in /lrde/dev/x86_64/stow/gcc-4.9.0/lib64/libstdc++.so.6.0.20)
error→==    by 0x46CB89: iword (ios_base.h:744)
error→==    by 0x46CB89: indent (indent.cc:22)
error→==    by 0x46CB89: misc::iendl(std::ostream&) (indent.cc:50)
error→==    by 0x4686BB: operator<< (ostream:113)
error→==    by 0x4686BB: ast::PrettyPrinter::operator()(ast::FunctionDec const&, std::string const&) (pretty-printer.cc:305)
error→==    by 0x46891D: ast::PrettyPrinter::operator()(ast::FunctionDec const&) (pretty-printer.cc:332)
error→==    by 0x435AC7: decs_visit<ast::AnyDecs<ast::FunctionDec> > (default-visitor.hxx:234)
error→==    by 0x435AC7: ast::GenDefaultVisitor<misc::constify_traits>::operator()(ast::AnyDecs<ast::FunctionDec> const&) (default-visitor.hxx:258)
error→==    by 0x4350EF: operator()<ast::Decs> (visitor.hxx:31)
error→==    by 0x4350EF: for_each<const std::list<ast::Decs*>, ast::GenDefaultVisitor<misc::constify_traits> > (algorithm.hxx:51)
error→==    by 0x4350EF: ast::GenDefaultVisitor<misc::constify_traits>::operator()(ast::DecsList const&) (default-visitor.hxx:203)
error→==    by 0x466773: ast::operator<<(std::ostream&, ast::Ast const&) (libast.cc:36)
error→==    by 0x40BB2E: ast::tasks::ast_display() (tasks.cc:26)
error→==    by 0x412303: task::TaskRegister::execute() (task-register.cc:295)
error→==    by 0x409695: main (tc.cc:31)
error→== 
error→== LEAK SUMMARY:
error→==    definitely lost: 0 bytes in 0 blocks
error→==    indirectly lost: 0 bytes in 0 blocks
error→==      possibly lost: 0 bytes in 0 blocks
error→==    still reachable: 208 bytes in 1 blocks
error→==         suppressed: 0 bytes in 0 blocks
error→== 
error→== For counts of detected and suppressed errors, rerun with: -v
error→== ERROR SUMMARY: 0 errors from 0 contexts (suppressed: 0 from 0)
/* == Abstract Syntax Tree. == */

function _main() =
  (
    0;
    ()
  )

Example 5.2: v tc -XAD 0.tig

Starting with GCC 3.4, GLIBCPP_FORCE_NEW is spelled GLIBCXX_FORCE_NEW.

As in the case of GDB, you should be careful when running a libtoolized program in Valgrind. Use the following command to make sure that this is your tc binary (and not the shell) that is checked by Valgrind:

$ libtool exe valgrind tc

You can ask Valgrind to run a debugger when it catches an error, using the --db-attach option. This is useful to inspect a process interactively.

$ valgrind --db-attach=yes ./tc

The default debugger used by Valgrind is GDB. Use the --db-command option to change this.

Another technique to make Valgrind and GDB interact is to use Valgrind’s gdbserver and the vgdb command (see Valgrind’s documentation for detailed explanations).


Next: , Previous: , Up: Tools   [Contents][Index]

5.9 Flex & Bison

We use Bison 3.0.4.19-fbaf, that is able to produce a C++ parser combined with modern features such as GLR, variants and complete symbols. If you don’t use this Bison, you will be in trouble.

The original papers on Lex and Yacc are:

Johnson, Stephen C. [1975].

Yacc: Yet Another Compiler Compiler. Computing Science Technical Report No. 32, Bell Laboratories, Murray hill, New Jersey.

Lesk, M. E. and E. Schmidt [1975].

Lex: A Lexical Analyzer Generator. Computing Science Technical Report No. 39, Bell Laboratories, Murray Hill, New Jersey.

These introductory guides can help beginners:

Thomas Niemann.

A Compact Guide to Lex & Yacc.

An introduction to Lex and Yacc.

Collective Work

Programming with GNU Software.

Contains information about Autoconf, Automake, Gperf, Flex, Bison, and GCC.

The Bison documentation, and the Flex documentation are available for browsing.


Next: , Previous: , Up: Tools   [Contents][Index]

5.10 HAVM

HAVM is a Tree (HIR or LIR) programs interpreter. It was written by Robert Anisko so that EPITA students could exercise their compiler projects before the final jump to assembly code. It is implemented in Haskell, a pure non strict functional language very well suited for this kind of symbolic processing. HAVM was coined on both Haskell, and VM standing for Virtual Machine.

Resources:


Next: , Previous: , Up: Tools   [Contents][Index]

5.11 MonoBURG

MonoBURG is a code generator generator, a tool that produces a function from a tree-pattern description of an instruction set. If you think of Bison being a program generating an AST generator from concrete syntax, you can see MonoBURG as a program generating an Assem generator from LIR trees.

MonoBURG is named after BURG, a program that generates a fast tree parser using BURS (Bottom-Up Rewrite System). MonoBURG is part of the Mono Project and has been extended by Michaël Cadilhac for the needs of the Tiger Project.

Resources:

Some papers on code generator generators are available in the bibliography. See BURG - Fast Optimal Instruction Selection and Tree Parsing, and Engineering a simple efficient code generator generator.


Next: , Previous: , Up: Tools   [Contents][Index]

5.12 Nolimips

Nolimips (formerly Mipsy) is a MIPS simulator designed to execute simple register based MIPS assembly code. It is a minimalist MIPS virtual machine that, contrary to other simulators (see SPIM), supports unlimited registers. The lack of a simulator featuring this prompted the development of Nolimips.

Its features are:

It was written by Benoît Perrot as an LRDE member, so that EPITA students could exercise their compiler projects after instruction selection but before register allocation. It is implemented in C++ and Python.

Resources:


Next: , Previous: , Up: Tools   [Contents][Index]

5.13 SPIM

The SPIM documentation reads:

SPIM S20 is a simulator that runs programs for the MIPS R2000/R3000 RISC computers. SPIM can read and immediately execute files containing assembly language. SPIM is a self-contained system for running these programs and contains a debugger and interface to a few operating system services.

The architecture of the MIPS computers is simple and regular, which makes it easy to learn and understand. The processor contains 32 general-purpose 32-bit registers and a well-designed instruction set that make it a propitious target for generating code in a compiler.

However, the obvious question is: why use a simulator when many people have workstations that contain a hardware, and hence significantly faster, implementation of this computer? One reason is that these workstations are not generally available. Another reason is that these machine will not persist for many years because of the rapid progress leading to new and faster computers. Unfortunately, the trend is to make computers faster by executing several instructions concurrently, which makes their architecture more difficult to understand and program. The MIPS architecture may be the epitome of a simple, clean RISC machine.

In addition, simulators can provide a better environment for low-level programming than an actual machine because they can detect more errors and provide more features than an actual computer. For example, SPIM has a X-window interface that is better than most debuggers for the actual machines.

Finally, simulators are an useful tool for studying computers and the programs that run on them. Because they are implemented in software, not silicon, they can be easily modified to add new instructions, build new systems such as multiprocessors, or simply to collect data.

SPIM is written and maintained by James R. Larus on SourceForge..


Next: , Previous: , Up: Tools   [Contents][Index]

5.14 SWIG

Our compiler provides two different user interfaces: one is a command line interface fully written in C++, using the “Task” system, and the other is a binding of the primary functions into the Python script language (see Python. This binding is automatically extracted from our modules using SWIG.

The SWIG home page reads:

SWIG is a software development tool that connects programs written in C and C++ with a variety of high-level programming languages. SWIG is primarily used with common scripting languages such as Perl, Python, Tcl/Tk, and Ruby, however the list of supported languages also includes non-scripting languages such as Java, OCAML and C#. Also several interpreted and compiled Scheme implementations (Guile, MzScheme, Chicken) are supported. SWIG is most commonly used to create high-level interpreted or compiled programming environments, user interfaces, and as a tool for testing and prototyping C/C++ software. SWIG can also export its parse tree in the form of XML and Lisp s-expressions. SWIG may be freely used, distributed, and modified for commercial and non-commercial use.


Next: , Previous: , Up: Tools   [Contents][Index]

5.15 Python

We promote, but do not require, Python as a scripting language over Perl because in our opinion it is a cleaner language. A nice alternative to Python is Ruby.

The Python Home Page reads:

Python is an interpreted, interactive, object-oriented programming language. It is often compared to Tcl, Perl, Scheme or Java.

Python combines remarkable power with very clear syntax. It has modules, classes, exceptions, very high level dynamic data types, and dynamic typing. There are interfaces to many system calls and libraries, as well as to various windowing systems (X11, Motif, Tk, Mac, MFC). New built-in modules are easily written in C or C++. Python is also usable as an extension language for applications that need a programmable interface.

The Python implementation is portable: it runs on many brands of UNIX, on Windows, OS/2, Mac, Amiga, and many other platforms. If your favorite system isn’t listed here, it may still be supported, if there’s a C compiler for it. Ask around on news:comp.lang.python – or just try compiling Python yourself.

The Python implementation is copyrighted but freely usable and distributable, even for commercial use.


Previous: , Up: Tools   [Contents][Index]

5.16 Doxygen

We use Doxygen as the standard tool for producing the developer’s documentation of the project. Its features must be used to produce good documentation, with an explanation of the role of the arguments etc. The quality of the documentation will be part of the notation. Details on how to use proper comments are given in the Doxygen Manual.

The documentation produced by Doxygen must not be included, but the target html must produce the HTML documentation in the doc/html directory.


Previous: , Up: Top   [Contents][Index]

Appendix A Appendices


Next: , Up: Appendices   [Contents][Index]

A.1 Glossary

Contributions to this section (as for the rest of this documentation) will be greatly appreciated.

activation block

Portion of dynamically allocated memory holding all the information a (recursive) function needs at runtime. It typically contains arguments, automatic local variables etc. Implemented by the class frame::Frame (see TC-5).

build

The machine/architecture on which the program is built. For instance, EPITA students typically build their compiler on GNU/Linux. Contrast with “target” and “host”.

curriculum

From WordNet: n : a course of academic studies; “he was admitted to a new program at the university” (syn: “course of study”, “program”, “syllabus”).

Guru of the Week
GotW

See Bibliography.

HAVM

HAVM is a Tree (HIR or LIR) programs interpreter. See HAVM.

host

The machine/architecture on which the program is run. For instance, EPITA students typically run their Tiger Compiler on GNU/Linux. Contrast with “build and “target”.

IA-32

The official new name for the i386 architecture.

scholarship

It is related to “scholar”, not “school”! It does not mean “scolarité”.

From WordNet:

See “schooling” and “curriculum”.

schooling

From WordNet:

snippet

A piece of something, e.g., “code snippet”.

stack frame

Synonym for “activation block”.

static hierarchy

A hierarchy of classes without virtual methods. In that case there is no (inclusion) polymorphism. For instance:

struct A     { };
struct B: A  { };
SPIM

SPIM S20 is a simulator that runs programs for the MIPS R2R3000 RISC computers. See SPIM.

target

The machine (or language) aimed at by a compiling tool. For instance, our target is principally MIPS. Compare with “build” and “host”.

traits

Traits are a useful technique that allows to write (compile time) functions ranging over types. See Traits, for the original presentation of traits. See Modern C++ Design, for an extensive use of traits.

vtable

For a given class, its table of pointers to virtual methods.


Next: , Previous: , Up: Appendices   [Contents][Index]

A.2 GNU Free Documentation License

Version 1.1, March 2000
Copyright © 2000 Free Software Foundation, Inc.
51 Franklin St, Fifth Floor, Boston, MA  02110-1301, USA

Everyone is permitted to copy and distribute verbatim copies
of this license document, but changing it is not allowed.
  1. PREAMBLE

    The purpose of this License is to make a manual, textbook, or other written document free in the sense of freedom: to assure everyone the effective freedom to copy and redistribute it, with or without modifying it, either commercially or noncommercially. Secondarily, this License preserves for the author and publisher a way to get credit for their work, while not being considered responsible for modifications made by others.

    This License is a kind of “copyleft”, which means that derivative works of the document must themselves be free in the same sense. It complements the GNU General Public License, which is a copyleft license designed for free software.

    We have designed this License in order to use it for manuals for free software, because free software needs free documentation: a free program should come with manuals providing the same freedoms that the software does. But this License is not limited to software manuals; it can be used for any textual work, regardless of subject matter or whether it is published as a printed book. We recommend this License principally for works whose purpose is instruction or reference.

  2. APPLICABILITY AND DEFINITIONS

    This License applies to any manual or other work that contains a notice placed by the copyright holder saying it can be distributed under the terms of this License. The “Document”, below, refers to any such manual or work. Any member of the public is a licensee, and is addressed as “you”.

    A “Modified Version” of the Document means any work containing the Document or a portion of it, either copied verbatim, or with modifications and/or translated into another language.

    A “Secondary Section” is a named appendix or a front-matter section of the Document that deals exclusively with the relationship of the publishers or authors of the Document to the Document’s overall subject (or to related matters) and contains nothing that could fall directly within that overall subject. (For example, if the Document is in part a textbook of mathematics, a Secondary Section may not explain any mathematics.) The relationship could be a matter of historical connection with the subject or with related matters, or of legal, commercial, philosophical, ethical or political position regarding them.

    The “Invariant Sections” are certain Secondary Sections whose titles are designated, as being those of Invariant Sections, in the notice that says that the Document is released under this License.

    The “Cover Texts” are certain short passages of text that are listed, as Front-Cover Texts or Back-Cover Texts, in the notice that says that the Document is released under this License.

    A “Transparent” copy of the Document means a machine-readable copy, represented in a format whose specification is available to the general public, whose contents can be viewed and edited directly and straightforwardly with generic text editors or (for images composed of pixels) generic paint programs or (for drawings) some widely available drawing editor, and that is suitable for input to text formatters or for automatic translation to a variety of formats suitable for input to text formatters. A copy made in an otherwise Transparent file format whose markup has been designed to thwart or discourage subsequent modification by readers is not Transparent. A copy that is not “Transparent” is called “Opaque”.

    Examples of suitable formats for Transparent copies include plain ASCII without markup, Texinfo input format, LaTeX input format, SGML or XML using a publicly available DTD, and standard-conforming simple HTML designed for human modification. Opaque formats include PostScript, PDF, proprietary formats that can be read and edited only by proprietary word processors, SGML or XML for which the DTD and/or processing tools are not generally available, and the machine-generated HTML produced by some word processors for output purposes only.

    The “Title Page” means, for a printed book, the title page itself, plus such following pages as are needed to hold, legibly, the material this License requires to appear in the title page. For works in formats which do not have any title page as such, “Title Page” means the text near the most prominent appearance of the work’s title, preceding the beginning of the body of the text.

  3. VERBATIM COPYING

    You may copy and distribute the Document in any medium, either commercially or noncommercially, provided that this License, the copyright notices, and the license notice saying this License applies to the Document are reproduced in all copies, and that you add no other conditions whatsoever to those of this License. You may not use technical measures to obstruct or control the reading or further copying of the copies you make or distribute. However, you may accept compensation in exchange for copies. If you distribute a large enough number of copies you must also follow the conditions in section 3.

    You may also lend copies, under the same conditions stated above, and you may publicly display copies.

  4. COPYING IN QUANTITY

    If you publish printed copies of the Document numbering more than 100, and the Document’s license notice requires Cover Texts, you must enclose the copies in covers that carry, clearly and legibly, all these Cover Texts: Front-Cover Texts on the front cover, and Back-Cover Texts on the back cover. Both covers must also clearly and legibly identify you as the publisher of these copies. The front cover must present the full title with all words of the title equally prominent and visible. You may add other material on the covers in addition. Copying with changes limited to the covers, as long as they preserve the title of the Document and satisfy these conditions, can be treated as verbatim copying in other respects.

    If the required texts for either cover are too voluminous to fit legibly, you should put the first ones listed (as many as fit reasonably) on the actual cover, and continue the rest onto adjacent pages.

    If you publish or distribute Opaque copies of the Document numbering more than 100, you must either include a machine-readable Transparent copy along with each Opaque copy, or state in or with each Opaque copy a publicly-accessible computer-network location containing a complete Transparent copy of the Document, free of added material, which the general network-using public has access to download anonymously at no charge using public-standard network protocols. If you use the latter option, you must take reasonably prudent steps, when you begin distribution of Opaque copies in quantity, to ensure that this Transparent copy will remain thus accessible at the stated location until at least one year after the last time you distribute an Opaque copy (directly or through your agents or retailers) of that edition to the public.

    It is requested, but not required, that you contact the authors of the Document well before redistributing any large number of copies, to give them a chance to provide you with an updated version of the Document.

  5. MODIFICATIONS

    You may copy and distribute a Modified Version of the Document under the conditions of sections 2 and 3 above, provided that you release the Modified Version under precisely this License, with the Modified Version filling the role of the Document, thus licensing distribution and modification of the Modified Version to whoever possesses a copy of it. In addition, you must do these things in the Modified Version:

    1. Use in the Title Page (and on the covers, if any) a title distinct from that of the Document, and from those of previous versions (which should, if there were any, be listed in the History section of the Document). You may use the same title as a previous version if the original publisher of that version gives permission.
    2. List on the Title Page, as authors, one or more persons or entities responsible for authorship of the modifications in the Modified Version, together with at least five of the principal authors of the Document (all of its principal authors, if it has less than five).
    3. State on the Title page the name of the publisher of the Modified Version, as the publisher.
    4. Preserve all the copyright notices of the Document.
    5. Add an appropriate copyright notice for your modifications adjacent to the other copyright notices.
    6. Include, immediately after the copyright notices, a license notice giving the public permission to use the Modified Version under the terms of this License, in the form shown in the Addendum below.
    7. Preserve in that license notice the full lists of Invariant Sections and required Cover Texts given in the Document’s license notice.
    8. Include an unaltered copy of this License.
    9. Preserve the section entitled “History”, and its title, and add to it an item stating at least the title, year, new authors, and publisher of the Modified Version as given on the Title Page. If there is no section entitled “History” in the Document, create one stating the title, year, authors, and publisher of the Document as given on its Title Page, then add an item describing the Modified Version as stated in the previous sentence.
    10. Preserve the network location, if any, given in the Document for public access to a Transparent copy of the Document, and likewise the network locations given in the Document for previous versions it was based on. These may be placed in the “History” section. You may omit a network location for a work that was published at least four years before the Document itself, or if the original publisher of the version it refers to gives permission.
    11. In any section entitled “Acknowledgments” or “Dedications”, preserve the section’s title, and preserve in the section all the substance and tone of each of the contributor acknowledgments and/or dedications given therein.
    12. Preserve all the Invariant Sections of the Document, unaltered in their text and in their titles. Section numbers or the equivalent are not considered part of the section titles.
    13. Delete any section entitled “Endorsements”. Such a section may not be included in the Modified Version.
    14. Do not retitle any existing section as “Endorsements” or to conflict in title with any Invariant Section.

    If the Modified Version includes new front-matter sections or appendices that qualify as Secondary Sections and contain no material copied from the Document, you may at your option designate some or all of these sections as invariant. To do this, add their titles to the list of Invariant Sections in the Modified Version’s license notice. These titles must be distinct from any other section titles.

    You may add a section entitled “Endorsements”, provided it contains nothing but endorsements of your Modified Version by various parties—for example, statements of peer review or that the text has been approved by an organization as the authoritative definition of a standard.

    You may add a passage of up to five words as a Front-Cover Text, and a passage of up to 25 words as a Back-Cover Text, to the end of the list of Cover Texts in the Modified Version. Only one passage of Front-Cover Text and one of Back-Cover Text may be added by (or through arrangements made by) any one entity. If the Document already includes a cover text for the same cover, previously added by you or by arrangement made by the same entity you are acting on behalf of, you may not add another; but you may replace the old one, on explicit permission from the previous publisher that added the old one.

    The author(s) and publisher(s) of the Document do not by this License give permission to use their names for publicity for or to assert or imply endorsement of any Modified Version.

  6. COMBINING DOCUMENTS

    You may combine the Document with other documents released under this License, under the terms defined in section 4 above for modified versions, provided that you include in the combination all of the Invariant Sections of all of the original documents, unmodified, and list them all as Invariant Sections of your combined work in its license notice.

    The combined work need only contain one copy of this License, and multiple identical Invariant Sections may be replaced with a single copy. If there are multiple Invariant Sections with the same name but different contents, make the title of each such section unique by adding at the end of it, in parentheses, the name of the original author or publisher of that section if known, or else a unique number. Make the same adjustment to the section titles in the list of Invariant Sections in the license notice of the combined work.

    In the combination, you must combine any sections entitled “History” in the various original documents, forming one section entitled “History”; likewise combine any sections entitled “Acknowledgments”, and any sections entitled “Dedications”. You must delete all sections entitled “Endorsements.”

  7. COLLECTIONS OF DOCUMENTS

    You may make a collection consisting of the Document and other documents released under this License, and replace the individual copies of this License in the various documents with a single copy that is included in the collection, provided that you follow the rules of this License for verbatim copying of each of the documents in all other respects.

    You may extract a single document from such a collection, and distribute it individually under this License, provided you insert a copy of this License into the extracted document, and follow this License in all other respects regarding verbatim copying of that document.

  8. AGGREGATION WITH INDEPENDENT WORKS

    A compilation of the Document or its derivatives with other separate and independent documents or works, in or on a volume of a storage or distribution medium, does not as a whole count as a Modified Version of the Document, provided no compilation copyright is claimed for the compilation. Such a compilation is called an “aggregate”, and this License does not apply to the other self-contained works thus compiled with the Document, on account of their being thus compiled, if they are not themselves derivative works of the Document.

    If the Cover Text requirement of section 3 is applicable to these copies of the Document, then if the Document is less than one quarter of the entire aggregate, the Document’s Cover Texts may be placed on covers that surround only the Document within the aggregate. Otherwise they must appear on covers around the whole aggregate.

  9. TRANSLATION

    Translation is considered a kind of modification, so you may distribute translations of the Document under the terms of section 4. Replacing Invariant Sections with translations requires special permission from their copyright holders, but you may include translations of some or all Invariant Sections in addition to the original versions of these Invariant Sections. You may include a translation of this License provided that you also include the original English version of this License. In case of a disagreement between the translation and the original English version of this License, the original English version will prevail.

  10. TERMINATION

    You may not copy, modify, sublicense, or distribute the Document except as expressly provided for under this License. Any other attempt to copy, modify, sublicense or distribute the Document is void, and will automatically terminate your rights under this License. However, parties who have received copies, or rights, from you under this License will not have their licenses terminated so long as such parties remain in full compliance.

  11. FUTURE REVISIONS OF THIS LICENSE

    The Free Software Foundation may publish new, revised versions of the GNU Free Documentation License from time to time. Such new versions will be similar in spirit to the present version, but may differ in detail to address new problems or concerns. See http://www.gnu.org/copyleft/.

    Each version of the License is given a distinguishing version number. If the Document specifies that a particular numbered version of this License “or any later version” applies to it, you have the option of following the terms and conditions either of that specified version or of any later version that has been published (not as a draft) by the Free Software Foundation. If the Document does not specify a version number of this License, you may choose any version ever published (not as a draft) by the Free Software Foundation.

A.2.1 ADDENDUM: How to use this License for your documents

To use this License in a document you have written, include a copy of the License in the document and put the following copyright and license notices just after the title page:

  Copyright (C)  year  your name.
  Permission is granted to copy, distribute and/or modify this document
  under the terms of the GNU Free Documentation License, Version 1.1
  or any later version published by the Free Software Foundation;
  with the Invariant Sections being list their titles, with the
  Front-Cover Texts being list, and with the Back-Cover Texts being list.
  A copy of the license is included in the section entitled ``GNU
  Free Documentation License''.

If you have no Invariant Sections, write “with no Invariant Sections” instead of saying which ones are invariant. If you have no Front-Cover Texts, write “no Front-Cover Texts” instead of “Front-Cover Texts being list”; likewise for Back-Cover Texts.

If your document contains nontrivial examples of program code, we recommend releasing these examples in parallel under your choice of free software license, such as the GNU General Public License, to permit their use in free software.


Next: , Previous: , Up: Appendices   [Contents][Index]

A.3 Colophon

This is version EPITA-2017-20-g8bd97f7 of assignments.texi, last edited on February 6, 2015, and compiled February 6, 2015, using:

$ tc --version
tc (LRDE Tiger Compiler 1.34a)
$Id: d2d741d14b9324c37698db091296e9b0e4b78c9f $

Akim Demaille            Alain Vongsouvanh        Alexandre Duret-Lutz
Alexis Brouard           Arnaud Fabre             Benoît Perrot
Benoît Sigoure           Benoît Tailhades         Cédric Bail
Christophe Duong         Clément Vasseur          Daniel Gazard
Fabien Ouy               Francis Maes             Gilles Walbrou
Guillaume Duhamel        Julien Roussel           Julien Grall
Laurent Gourvénec        Michaël Cadilhac         Nicolas Burrus
Nicolas Pouillard        Nicolas Teck             Pablo Oliveira
Pierre-Yves Strub        Quôc Peyrot              Raphaël Poss
Razik Yousfi             Roland Levillain         Robert Anisko
Sébastien Broussaud      Stéphane Molina          Théophile Ranquet
Thierry Géraud           Valentin David           Yann Grandmaître
Yann Popo                Yann Régis-Gianas

Example A.1: tc --version

$ havm --version
HAVM 0.26
Written by Robert Anisko.

Copyright (C) 2002-2003  Robert Anisko
Copyright (C) 2003-2007, 2009, 2011-2014  EPITA Research and Development Laboratory (LRDE).
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

Example A.2: havm --version

$ nolimips --version
nolimips (Nolimips) 0.9a
Written by Benoit Perrot.

Copyright (C) 2003, 2004, 2005, 2006, 2008, 2009, 2010, 2012 Benoit Perrot.
nolimips comes with ABSOLUTELY NO WARRANTY.
This is free software, and you are welcome to redistribute and modify it 
under certain conditions; see source for details.

Example A.3: nolimips --version


Next: , Previous: , Up: Appendices   [Contents][Index]

A.4 List of Files

File 2.1

temp.cc

File 2.2

temp.hh

File 2.3

temp-factored.hh

File 2.4

temp-factored.cc

File 2.5

sample/sample.hh

File 2.6

sample/sample.hxx

File 4.1

simple.tig

File 4.2

back-zee.tig

File 4.3

postinc.tig

File 4.4

test01.tig

File 4.5

unterminated-comment.tig

File 4.6

type-nil.tig

File 4.7

a+a.tig

File 4.8

simple-fact.tig

File 4.9

string-escapes.tig

File 4.10

1s-and-2s.tig

File 4.11

for-loop.tig

File 4.12

parens.tig

File 4.13

foo-bar.tig

File 4.14

foo-stop-bar.tig

File 4.15

fbfsb.tig

File 4.16

fbfsb-desugared.tig

File 4.17

multiple-parse-errors.tig

File 4.18

me.tig

File 4.19

meme.tig

File 4.20

nome.tig

File 4.21

tome.tig

File 4.22

breaks-in-embedded-loops.tig

File 4.23

break.tig

File 4.24

box.tig

File 4.25

unknown-field-type.tig

File 4.26

bad-member-bindings.tig

File 4.27

missing-super-class.tig

File 4.28

as.tig

File 4.29

variable-escapes.tig

File 4.30

undefined-variable.tig

File 4.31

int-plus-string.tig

File 4.32

assign-loop-var.tig

File 4.33

unknowns.tig

File 4.34

bad-if.tig

File 4.35

mutuals.tig

File 4.36

bad-super-type.tig

File 4.37

forward-reference-to-class.tig

File 4.38

is_devil.tig

File 4.39

string-equality.tig

File 4.40

string-less.tig

File 4.41

simple-for-loop.tig

File 4.42

sub.tig

File 4.43

subscript-read.tig

File 4.44

subscript-write.tig

File 4.45

sizes.tig

File 4.46

over-amb.tig

File 4.47

over-duplicate.tig

File 4.48

over-scoped.tig

File 4.49

empty-class.tig

File 4.50

simple-class.tig

File 4.51

override.tig

File 4.52

0.tig

File 4.53

arith.tig

File 4.54

if-101.tig

File 4.55

while-101.tig

File 4.56

boolean.tig

File 4.57

print-101.tig

File 4.58

print-array.tig

File 4.59

print-record.tig

File 4.60

vars.tig

File 4.61

fact15.tig

File 4.62

preincr-1.tig

File 4.63

preincr-2.tig

File 4.64

move-mem.tig

File 4.65

nested-calls.tig

File 4.66

seq-point.tig

File 4.67

1-and-2.tig

File 4.68

broken-while.tig

File 4.69

the-answer.tig

File 4.70

add.tig

File 4.71

substring-0-1-1.tig

File 4.72

tens.tig

File 4.73

tens.main._main.flow.gv

File 4.74

tens.main._main.liveness.gv

File 4.75

tens.main._main.interference.gv

File 4.76

hundreds.tig

File 4.77

hundreds.main._main.liveness.gv

File 4.78

hundreds.main._main.interference.gv

File 4.79

ors.tig

File 4.80

ors.main._main.flow.gv

File 4.81

ors.main._main.liveness.gv

File 4.82

ors.main._main.interference.gv

File 4.83

and.tig

File 4.84

and.main._main.liveness.gv

File 4.85

seven.tig

File 4.86

print-seven.tig

File 4.87

print-many.tig

File 4.88

the-answer-ia32.tig

File 4.89

add-ia32.tig

File 4.90

substring-0-1-1-ia32.tig

File 4.91

condjump-ia32.tig

File 5.1

v

File 5.2

ineffective-break.tig

File 5.3

ineffective-if.tig


Next: , Previous: , Up: Appendices   [Contents][Index]

A.5 List of Examples

Example 4.1

tc simple.tig

Example 4.2

SCAN=1 PARSE=1 tc -X --parse simple.tig

Example 4.3

tc -X --parse back-zee.tig

Example 4.4

tc -X --parse postinc.tig

Example 4.5

tc -X --parse test01.tig

Example 4.6

tc -X --parse unterminated-comment.tig

Example 4.7

tc -X --parse type-nil.tig

Example 4.8

tc C:/TIGER/SAMPLE.TIG

Example 4.9

tc -X --parse-trace --parse a+a.tig

Example 4.10

tc -XA simple-fact.tig

Example 4.11

tc -D simple-fact.tig

Example 4.12

tc -DA simple-fact.tig

Example 4.13

tc -XAD string-escapes.tig

Example 4.14

tc -XAD 1s-and-2s.tig

Example 4.15

tc -XAD 1s-and-2s.tig >output.tig

Example 4.16

tc -XAD output.tig

Example 4.17

tc -XAD for-loop.tig

Example 4.18

tc -XAD parens.tig

Example 4.19

tc -b foo-bar.tig

Example 4.20

tc -b foo-stop-bar.tig

Example 4.21

tc -b fbfsb.tig

Example 4.22

tc multiple-parse-errors.tig

Example 4.23

tc -XAD multiple-parse-errors.tig

Example 4.24

tc -XbBA me.tig

Example 4.25

tc -XbBA meme.tig

Example 4.26

tc -bBA nome.tig

Example 4.27

tc -bBA tome.tig

Example 4.28

tc -XbBA breaks-in-embedded-loops.tig

Example 4.29

tc -b break.tig

Example 4.30

tc -XbBA box.tig

Example 4.31

tc -T box.tig

Example 4.32

tc -XbBA unknown-field-type.tig

Example 4.33

tc -X --object-bindings-compute -BA bad-member-bindings.tig

Example 4.34

tc --object-types-compute bad-member-bindings.tig

Example 4.35

tc -X --object-bindings-compute -BA missing-super-class.tig

Example 4.36

tc -X --rename -A as.tig

Example 4.37

tc -XEAeEA variable-escapes.tig

Example 4.38

tc -e undefined-variable.tig

Example 4.39

tc int-plus-string.tig

Example 4.40

tc -T int-plus-string.tig

Example 4.41

tc -T assign-loop-var.tig

Example 4.42

tc -T unknowns.tig

Example 4.43

tc -T bad-if.tig

Example 4.44

tc -T mutuals.tig

Example 4.45

tc -H mutuals.tig >mutuals.hir

Example 4.46

havm mutuals.hir

Example 4.47

tc --object-types-compute bad-super-type.tig

Example 4.48

tc --object-types-compute forward-reference-to-class.tig

Example 4.49

tc -T is_devil.tig

Example 4.50

tc --desugar-string-cmp --desugar -A string-equality.tig

Example 4.51

tc --desugar-string-cmp --desugar -A string-less.tig

Example 4.52

tc --desugar-for --desugar -A simple-for-loop.tig

Example 4.53

tc -X --inline -A sub.tig

Example 4.54

tc --bounds-checks-add -A subscript-read.tig

Example 4.55

tc --bounds-checks-add -L subscript-read.tig >subscript-read.lir

Example 4.56

havm subscript-read.lir

Example 4.57

tc --bounds-checks-add -A subscript-write.tig

Example 4.58

tc --bounds-checks-add -S subscript-write.tig >subscript-write.s

Example 4.59

nolimips -l nolimips -Nue subscript-write.s

Example 4.60

tc -Xb sizes.tig

Example 4.61

tc -X --overfun-bindings-compute -BA sizes.tig

Example 4.62

tc -XOBA sizes.tig

Example 4.63

tc -XO over-amb.tig

Example 4.64

tc -XO over-duplicate.tig

Example 4.65

tc -XOBA over-scoped.tig

Example 4.66

tc -X --object-desugar -A empty-class.tig

Example 4.67

tc -X --object-desugar -A simple-class.tig

Example 4.68

tc --object-desugar -A override.tig

Example 4.69

tc --object-desugar -L override.tig >override.lir

Example 4.70

havm override.lir

Example 4.71

tc --hir-display 0.tig

Example 4.72

tc -H arith.tig

Example 4.73

tc -H arith.tig >arith.hir

Example 4.74

havm arith.hir

Example 4.75

havm --trace arith.hir

Example 4.76

tc -H if-101.tig

Example 4.77

tc -H while-101.tig

Example 4.78

tc --hir-naive -H boolean.tig

Example 4.79

tc --hir-naive -H boolean.tig >boolean-1.hir

Example 4.80

havm --profile boolean-1.hir

Example 4.81

tc -H boolean.tig

Example 4.82

tc -H boolean.tig >boolean-2.hir

Example 4.83

havm --profile boolean-2.hir

Example 4.84

tc -H print-101.tig >print-101.hir

Example 4.85

havm print-101.hir

Example 4.86

tc -H print-array.tig

Example 4.87

tc -H print-array.tig >print-array.hir

Example 4.88

havm print-array.hir

Example 4.89

tc -H vars.tig

Example 4.90

tc -eH vars.tig

Example 4.91

tc -eH vars.tig >vars.hir

Example 4.92

havm vars.hir

Example 4.93

tc -H fact15.tig

Example 4.94

tc -H fact15.tig >fact15.hir

Example 4.95

havm fact15.hir

Example 4.96

tc -eH variable-escapes.tig

Example 4.97

tc -eH preincr-1.tig

Example 4.98

tc -eL preincr-1.tig

Example 4.99

tc -eL preincr-2.tig

Example 4.100

tc -eH preincr-2.tig >preincr-2.hir

Example 4.101

havm preincr-2.hir

Example 4.102

tc -eL preincr-2.tig >preincr-2.lir

Example 4.103

havm preincr-2.lir

Example 4.104

tc -eL move-mem.tig >move-mem.lir

Example 4.105

havm move-mem.lir

Example 4.106

tc -L nested-calls.tig

Example 4.107

tc -L seq-point.tig >seq-point.lir

Example 4.108

havm seq-point.lir

Example 4.109

tc -L 1-and-2.tig

Example 4.110

tc -H broken-while.tig

Example 4.111

tc -L broken-while.tig

Example 4.112

tc --inst-display the-answer.tig

Example 4.113

tc --nolimips-display the-answer.tig

Example 4.114

tc -sI the-answer.tig

Example 4.115

tc -e --inst-display add.tig

Example 4.116

tc -eR --nolimips-display add.tig >add.nolimips

Example 4.117

nolimips -l nolimips -Nue add.nolimips

Example 4.118

tc -e --nolimips-display substring-0-1-1.tig

Example 4.119

tc -eR --nolimips-display substring-0-1-1.tig >substring-0-1-1.nolimips

Example 4.120

nolimips -l nolimips -Nue substring-0-1-1.nolimips

Example 4.121

tc -I tens.tig

Example 4.122

tc -FVN tens.tig

Example 4.123

tc --callee-save=0 -VN hundreds.tig

Example 4.124

tc --callee-save=0 -I ors.tig

Example 4.125

tc -FVN ors.tig

Example 4.126

tc -sV and.tig

Example 4.127

tc -sI seven.tig

Example 4.128

tc -S seven.tig >seven.s

Example 4.129

nolimips -l nolimips -Ne seven.s

Example 4.130

tc -s --tempmap-display seven.tig

Example 4.131

tc -sI print-seven.tig

Example 4.132

tc -S print-seven.tig >print-seven.s

Example 4.133

nolimips -l nolimips -Ne print-seven.s

Example 4.134

tc -eIs --tempmap-display -I --time-report print-many.tig

Example 4.135

tc --target-ia32 --inst-display the-answer-ia32.tig

Example 4.136

tc -sI the-answer-ia32.tig

Example 4.137

tc -e --target-ia32 --inst-display add-ia32.tig

Example 4.138

tc -e --target-ia32 --asm-compute --inst-display add-ia32.tig

Example 4.139

tc -e --target-ia32 --asm-display add-ia32.tig >add-ia32.s

Example 4.140

gcc -m32 -oadd-ia32 add-ia32.s

Example 4.141

./add-ia32

Example 4.142

tc -e --target-ia32 --inst-display substring-0-1-1-ia32.tig

Example 4.143

tc -e --target-ia32 --asm-compute --inst-display substring-0-1-1-ia32.tig

Example 4.144

tc -e --target-ia32 --asm-display substring-0-1-1-ia32.tig >substring-0-1-1-ia32.s

Example 4.145

gcc -m32 -osubstring-0-1-1-ia32 substring-0-1-1-ia32.s

Example 4.146

./substring-0-1-1-ia32

Example 4.147

tc -e --target-ia32 --inst-display condjump-ia32.tig

Example 4.148

tc -e --target-ia32 --asm-compute --inst-display condjump-ia32.tig

Example 5.1

v tc -XA 0.tig

Example 5.2

v tc -XAD 0.tig

Example A.1

tc --version

Example A.2

havm --version

Example A.3

nolimips --version


Previous: , Up: Appendices   [Contents][Index]

A.6 Index

Jump to:   %   *   -   8    
A   B   C   D   E   F   G   H   I   K   L   M   N   O   P   R   S   T   U   V   W   X   Y  
Index Entry  Section

%
%require: PTHL Code to Write

*
*.cc: File Conventions
*.cc: Definitions of functions and variables: File Conventions
*.hh: File Conventions
*.hh: Declarations: File Conventions
*.hxx: File Conventions
*.hxx: Inlined definitions: File Conventions

-
--bindings-compute: TC-3
--bindings-display: TC-3
--bounds-checks-add: TC-B
--desugar: TC-D
--desugar-for: TC-D
--desugar-string-cmp: TC-D
--escapes-compute: TC-E
--escapes-display: TC-E
--inline: TC-I
--object-bindings-compute: TC-3
--object-desugar: TC-O
--object-rename: TC-4 Options
--object-types-compute: TC-4 Samples
--overfun-bindings-compute: TC-A
--overfun-bounds-checks-add: TC-B
--overfun-desugar: TC-D
--overfun-inline: TC-I
--overfun-prune: TC-I
--overfun-types-compute: TC-4 Samples
--overfun-types-compute: TC-A
--prune: TC-I
--rename: TC-R
--target-ia32: TC-X
--typed: TC-4 Samples
--types-compute: TC-4 Samples
-T: TC-4 Samples

8
80 columns maximum: Matters of Style

: PTHL Samples

A
accept: src/ast
accept: src/ast
accept: src/ast
accept: src/ast
access.*: src/frame
access.*: src/translate
Accessors: Use of C++ Features
activation block: Glossary
aliasing: Use of C++ Features
array.*: src/type
ASM: TC-9
Assem: src/assem
assembly.*: src/target
attribute.*: src/type
AUTHORS.txt: The Top Level
Autoconf: The GNU Build System
Automake: The GNU Build System
Autotools Tutorial: Bibliography
auxiliary class: Name Conventions
Avoid class members (EC47): Use of C++ Features

B
basic block: TC-6 Scheduling Samples
Be concise: Documentation Style
binder.*: src/bind
binder.*: src/object
binding: TC-3 Samples
binop.brg: src/target/mips
binop.brg: src/target/ia32
Bison: Flex & Bison
bison++.in: build-aux
Bjarne Stroustrup: Bibliography
block structure: TC-E Goals
Bookshop: Bibliography
Boost.org: Bibliography
bounds-checking-visitor.*: src/desugar
build: Glossary
builtin-types.*: src/type
BURG: Fast Optimal Instruction Selection and Tree Parsing: Bibliography

C
C++ Primer: Bibliography
call.brg: src/target/mips
call.brg: src/target/ia32
canonicalization: TC-6
chunk: TC-2 Chunks
chunk: TC-2 Code to Write
cjump.brg: src/target/mips
cjump.brg: src/target/ia32
Clang: Clang
class.*: src/type
cloner.*: src/astclone
Code duplication: Use of C++ Features
codegen.*: src/target
codegen.*: src/target/mips
codegen.*: src/target/ia32
color.*: src/regalloc
comment.*: src/assem
common.hh: src
commute: TC-6 Canonicalization Samples
Compilers and Compiler Generators, an introduction with C++: Bibliography
Compilers: Principles, Techniques and Tools: Bibliography
conflict graph: TC-8
contract.*: lib/misc
Cool: The Classroom Object-Oriented Compiler: Bibliography
cpu.*: src/target
cpu.*: src/target/mips
cpu.*: src/target/ia32
created_type_set: src/ast
CStupidClassName: Bibliography
curriculum: Glossary

D
default-visitor.*: src/ast
Design Patterns: Elements of Reusable Object-Oriented Software: Bibliography
Design Patterns: Elements of Reusable Object-Oriented Software: Bibliography
desugar-visitor.*: src/object
desugar-visitor.*: src/desugar
distcheck: Making a Tarball
DISTCHECK_CONFIGURE_FLAGS: Making a Tarball
dmalloc: Valgrind
Do not copy tests or test frame works: Rules of the Game
Do not declare many variables on one line: Matters of Style
Document classes in their *.hh file: Documentation Style
Document namespaces in lib*.hh files: Documentation Style
Don't hesitate working with other groups: Rules of the Game
Don't use inline in declarations: Matters of Style
Dragon Book: Bibliography
driver: src
dump: Use of C++ Features
dump: lib/misc
dynamic_cast: Use of C++ Features

E
ECN: Bibliography
Effective C++: Bibliography
Effective STL: Bibliography
Engineering a simple, efficient code generator generator: Bibliography
epilogue.cc: src/target/mips
epilogue.cc: src/target/ia32
EPITA Library: Bibliography
error.*: lib/misc
error→: PTHL Samples
escapable.*: src/ast
escape: lib/misc
escape.*: lib/misc
escapes-visitor.*: src/escapes
escapes::EscapesVisitor: TC-E Code to Write
ESN: Bibliography
exp.brg: src/target/mips
exp.brg: src/target/ia32
exp.hh: src/translate
Explicit template instantiation: File Conventions

F
FDL, GNU Free Documentation License: GNU Free Documentation License
field.*: src/type
Flex: Flex & Bison
flex++.in: build-aux
flex-lexer.hh: lib/misc
flow graph: TC-8
flowgraph.*: src/liveness
foo_get: Use of C++ Features
foo_set: Use of C++ Features
fragment.*: src/tree
fragment.*: src/assem
fragments.*: src/tree
frame.*: src/frame
function.*: src/type
fwd.hh: forward declarations: File Conventions

G
gas-assembly.*: src/target/ia32
gas-layout.*: src/target/ia32
GCC: GCC
GDB: GDB
Generic Visitors in C++: Bibliography
get: lib/misc
GLIBCPP_FORCE_NEW: Valgrind
GLIBCXX_FORCE_NEW: Valgrind
GNU Build System: The GNU Build System
GotW: Bibliography
GOTWN: Bibliography
graph: lib/misc
graph.*: lib/misc
Guard included files (*.hh & *.hxx): File Conventions
Guru of the Week: Bibliography

H
havm: TC-5 Primitive Samples
HAVM: HAVM
HAVM: Glossary
helper class: Name Conventions
Hide auxiliary classes: Name Conventions
HIR: TC-5
host: Glossary
How not to go about a programming assignment: Bibliography
Hunt code duplication: Use of C++ Features
Hunt Leaks: Use of C++ Features

I
I386: TC-X
IA-32: TC-X
IA-32: Glossary
ia32: src/target
identifier.*: src/temp
If something is fishy, say it: Rules of the Game
indent.*: lib/misc
inliner.*: src/inlining
INSTR: TC-7
instr.*: src/assem
instruction selection: TC-7
interference graph: TC-8
interference-graph.*: src/liveness

K
Keep superclasses on the class declaration line: Matters of Style

L
label.*: src/temp
label.*: src/assem
layout.hh: src/assem
Le Monde en Tique: Bibliography
Leave no space between a function name and its argument(s), either formal or actual: Matters of Style
Leave no space between template name and effective parameters: Matters of Style
Leave one space between TEMPLATE and formal parameters: Matters of Style
level.*: src/translate
Lex: Flex & Bison
Lex & Yacc: Bibliography
libassem.*: src/assem
libmodule.*: Pure interface: File Conventions
libparse.hh: src/parse
libregalloc.*: src/regalloc
libtarget.*: src/target
Libtool: The GNU Build System
libtranslate.*: src/translate
libtype.*: src/type
LIR: TC-6
liveness analysis: TC-8
liveness.*: src/liveness
LLVM: Clang
location.hh: src/parse
location.hh: src/ast

M
Making Compiler Design Relevant for Students who will (Most Likely) Never Design a Compiler: Bibliography
malloc: TC-5 Builtin Calls Samples
mem.brg: src/target/mips
mem.brg: src/target/ia32
method.*: src/type
mips: src/target
misc::error: lib/misc
misc::scoped_map<Key, Data>: lib/misc
misc::variant<T0, T1>: lib/misc
misc::variant<T0, Ts...>: TC-1 Code to Write
Modern C++ Design -- Generic Programming and Design Patterns Applied: Bibliography
Modern Compiler Implementation in C, Java, ML: Bibliography
Module, namespace, and directory likethis: File Conventions
MonoBURG: MonoBURG
monoburg++.in: build-aux
move.*: src/assem
move.brg: src/target/mips
move.brg: src/target/ia32
move_load.brg: src/target/mips
move_load.brg: src/target/ia32
move_store.brg: src/target/mips
move_store.brg: src/target/ia32

N
Name private/protected members like_this_: Name Conventions
Name public members like_this: Name Conventions
Name the parent class super_type: Name Conventions
Name your classes LikeThis: Name Conventions
Name your typedef foo_type: Name Conventions
named.*: src/type
nested function: TC-E Goals
NEWS: Making a Tarball
Nolimips: Nolimips
non-local variable: TC-E Goals
non-object-visitor.*: src/ast

O
Object Management Group: Bibliography
object-visitor.*: src/ast
object::Renamer: TC-4 Options
One class LikeThis per files like-this.*: File Conventions
oper.*: src/assem
Order class members by visibility first: Matters of Style

P
panther.el: build-aux
parsetiger.yy: src/parse
Parsing Techniques -- A Practical Guide: Bibliography
Pointers and references are part of the type: Matters of Style
Portland Pattern Repository: Bibliography
position.hh: src/parse
Prefer C Comments for Long Comments: Documentation Style
Prefer C++ Comments for One Line Comments: Documentation Style
Prefer dynamic_cast of references: Use of C++ Features
Prefer member functions to algorithms with the same names (ES44): Use of STL
Prefer standard algorithms to hand-written loops (ES43): Use of STL
pretty-printer.*: src/ast
Programming: Principles and Practice Using C++: Bibliography
prologue.hh: src/target/mips
prologue.hh: src/target/ia32
pruner.*: src/inlining
put: lib/misc
Put initializations below the constructor declaration: Matters of Style
Python: Python

R
rebox: Documentation Style
rebox: build-aux
rebox.el: build-aux
record.*: src/type
ref.*: lib/misc
regallocator.*: src/regalloc
register allocation: TC-9
renamer.*: src/bind
renamer.*: src/object
Repeat virtual in subclass declarations: Matters of Style
runtime, Tiger: src/target
runtime-freebsd.s: src/target/ia32
runtime-gnu-linux.s: src/target/ia32
runtime.cc: src/target/mips
runtime.s: src/target/mips

S
scantiger.ll: src/parse
scholarship: Glossary
schooling: Glossary
scoped-map.*: lib/misc
scope_begin: lib/misc
scope_end: lib/misc
sequence point: TC-6 Canonicalization Samples
set.*: lib/misc
snippet: Glossary
Specify comparison types for associative containers of pointers (ES20): Use of STL
SPIM: SPIM
spim-assembly.*: src/target/mips
spim-layout.*: src/target/mips
SPOT : une bibliothèque de vérification de propriétés de logique temporelle à temps linéaire: Bibliography
stack frame: Glossary
static hierarchy: Glossary
Stay out of reserved names: Name Conventions
STL Home: Bibliography
stm.brg: src/target/mips
stm.brg: src/target/ia32
SWIG: SWIG
symbol.*: lib/misc

T
tarball name: Package Name and Version
target: Glossary
target.*: src/target
target.*: src/target/mips
target.*: src/target/ia32
tasks.*: Impure interface: File Conventions
tc: src
tc.cc: src
temp-set.*: src/temp
temp.*: src/temp
temp.brg: src/target/mips
temp.brg: src/target/ia32
test, unit: TC-1 Code to Write
test-flowgraph.cc: src/liveness
test-regalloc.cc: src/regalloc
Testing student-made compilers: Bibliography
Tests are part of the project: Rules of the Game
The Design and Evolution of C++: Bibliography
The Dragon Book: Bibliography
The Elements of Style: Bibliography
Thinking in C++ Volume 1: Bibliography
Thinking in C++ Volume 2: Bibliography
Thou Shalt Not Copy Code: Rules of the Game
Thou Shalt Not Possess Thy Neighbor's Code: Rules of the Game
tiger-ftdetect.vim: build-aux
tiger-runtime.c: src/target
tiger-syntax.vim: build-aux
tiger.el: build-aux
timer.*: lib/misc
traces: TC-6 Scheduling Samples
traits: Bibliography
traits: Glossary
traits.*: lib/misc
Traits: a new and useful template technique: Bibliography
translation.hh: src/translate
translator.hh: src/translate
tree.brg: src/target/mips
tree.brg: src/target/ia32
typable.*: src/ast
type checking: TC-4
type-checker.*: src/type
type-checker.*: src/object
type-constructor.*: src/ast
type.*: src/type
type::Error: TC-4 Options
type::Type*: src/ast
type::Type*: src/ast
typeid: Use of C++ Features
types.hh: src/type
type_set: src/ast

U
unique.*: lib/misc
unit test: TC-1 Code to Write
Use const references in arguments to save copies (EC22): Use of C++ Features
Use dump as a member function returning a stream: Use of C++ Features
Use dynamic_cast for type cases: Use of C++ Features
Use foo_get, not get_foo: Use of C++ Features
Use pointers when passing an object together with its management: Use of C++ Features
Use rebox.el to mark up paragraphs: Documentation Style
Use references for aliasing: Use of C++ Features
Use the Imperative: Documentation Style
Use virtual methods, not type cases: Use of C++ Features
Use ‘\directive: Documentation Style

V
Valgrind: Valgrind
variant: PTHL Code to Write
variant.*: lib/misc
visitor.*: src/tree
visitor.hh: src/ast
visitor.hh: src/assem
vtable: Glossary

W
Write correct English: Documentation Style
Write Documentation in Doxygen: Documentation Style
Writing Compilers and Interpreters -- An Applied Approach Using C++: Bibliography

X
X86: TC-X

Y
Yacc: Flex & Bison

Jump to:   %   *   -   8    
A   B   C   D   E   F   G   H   I   K   L   M   N   O   P   R   S   T   U   V   W   X   Y  

Footnotes

(1)

The fact that the compiler compiles C++ is virtually irrelevant.

(2)

See the shift of language? From tarball to distribution.

(3)

Please, let us know whom we forgot!

(4)

For instance, g++ reports an ‘error: cannot dynamic_cast `a' (of type `struct A*') to type `struct B*' (source type is not polymorphic)’.

(5)

For the time being, forget about -X.

(6)

The option --hir-naive is not to be implemented.

(7)

The case of the stack pointer register is different because it is not used in the actual function body: it is referred to by the “fake” prologue/epilogue output by the ProcFrag.

(8)

Actually temporaries in HAVM may have any name, you might use ‘He110W0r1d13’ as well.

(9)

To be fair, the Dragon Book leaves a single page (not sheet) to graph coloring.