**
******************************************************
-There is at least 4 projects in the tree:
+There is at least 5 sub-projects in the tree:
- XBT: eXtended Bundle of Tools (low-level toolbox: logging, datatypes).
- - GRAS: Grid Reality And Simulation (message passing API with two
- implementations allowing to compile programs on top of the
- simulator or for the real life without code modification)
- - SURF: Server for the Use of Resource Fictions (the simulator used
- in GRAS and more)
- - AMOK: Advanced Metacomputing Overlay Kit (high level toolbox; Grid
- application elements such as distributed database, topology
- discovery service, and so on)
-
-They are all in the same tree because GRAS depends on SURF which depends on
-GRAS (that's the only cycle here, we're not *that* vicious).
-
-The tree is not splited on projects, but on file finality:
+ - SURF: a SimUlation aRtiFact. This is the simulation kernel.
+ - MSG: originally MetaSimGrid, MSG is a simple distributed application
+ simulator.
+ - SMPI: Simulated MPI, to run MPI application using emulation technics;
+ - MC: model-checker;
+ - SIMIX: basix interface for simulated processes. This layer defines simcalls
+ (simulation calls) exposed to the simulated processes by the SIMIX "kernel".
+ This interface is used to implement the MSG, SMPI layers.
+ - SIMDAG;
+
+They are all in the same tree because they are complementary tools and
+having all of them in the same package makes the installation easier
+for end-users. Moreover, it enables to share the compilation chain and
+eases the development.
+
+The tree is not split on projects, but on file finality:
include/ -> all *public* headers
include/xbt/*.h -> one file per module
- include/gros.h -> file including all modules headers
- (same for gras, surf and amok instead of gros)
-
- src/Makefile.am -> main makefile. All projects should fit in only one
- library (I mean 2, RL+SG), which is compiled here.
-
- Since all object.o files are placed here, you should
- choose the name of c files carfully to avoid
- conflict.
-
- src/gras/DataDesc -> typical project module
- src/gras/DataDesc/datadesc_interface.h -> visible to any GRAS modules;
- masked to the user and GROS/AMOK/SURF
- src/gras/DataDesc/datadesc_private.h -> visible only from this module
-
- So, the modules have 3 levels of publicity for their interface.
- Private, internal to GRAS, public. Of course, I try to keep as much stuff
- private as possible.
src/include -> another location for protected headers. Used by SURF, and
other should be converted, since this is the Right Thing.
- testsuite/ -> The more test the better.
- Same organization than src/ and include/
- Tests are allowed to load some headers of the module they test.
- All tests should be listed in run_test.in so that they get
- run on 'make check'.
- They are not listed directly in the check_PROGRAMS part of
- the makefile because run_test knows about the gras logging
- features and relaunch with full details the failed tests.
-
examples/ -> Supposed to be copy/pastable by the user, so keep it clear and
avoid any kind of trick. In particular, do only include the
public headers here.
+ teshsuite/ -> The more test the better. Put in there any strange test
+ doing things that the users are not supposed to do,
+ just to see if our framework is robust to incorrect and
+ unusual behaviors. All tests written in this section
+ should leverage our tesh(1) utility.
+
+**
+** Indentation standard
+**
+*****************************************************
+
+Most files use the Kernighan & Ritchie coding style with 2 spaces of
+indentation. The indent program can help you to stick to it:
+
+indent -kr -l120 -nut -i2 -lps -npcs -br -brs -ce -cdw -bbo -npsl <myfile>
+
+The script ./tools/internal/indent runs indent with the appropriate options.
+
+If you use Eclipse, please import the settings in ./tools/internal/eclipse-formating.xml
+
**
** Type naming standard
**
*****************************************************
It may sound strange, but the type naming convention was source of intense
-discution between da GRAS posse members. The convention we came to may not
+discussion between da SimGrid posse members. The convention we came to may not
be the best solution, but it has the merit to exist and leave everyone work.
So please stick to it.
- - ???_t is a valid type (builded with typedef)
+ - ???_t is a valid type (built with typedef)
- s_toto_t is a structure (access to fields with .)
- s_toto is a structure needing 'struct' keyword to be used
- e_toto_t is an enum
- u_toto_t is an union
- u_toto is an union needing 'union' keyword to be used
- toto_t is an 'object' (struct*)
-
+
Please to not call toto_t something else than an 'object' (ie, something you
have to call _new and _free on it).
-Exemple:
+Example:
typedef struct s_toto {} s_toto_t, *toto_t;
typedef enum {} e_toto_t;
-
+
Moreover, only toto_t (and e_toto_t) are public. The rest (mainly s_toto_t)
is private.
**
*****************************************************
-MALLOC:
+MALLOC
Don't use it, or you'll have to check the result (and do some dirty stuff
on AIX). Use xbt_malloc (or even better, xbt_new) instead.
-SIZE_T
+SIZE_T (FIXME: obsolete?)
If possible, avoid size_t and use unsigned long instead. If not,
#include <sys/types.h> in all files manipulating size_t
- do cast it to unsigned long before printing (and use %lu)
-
-PRINTF pointer difference
+ do cast it to unsigned long before printing (and use %lu),
+ or use %zu.
+
+INTEGERS
+ Please avoid to use long ints. This is the source of many compatibility
+ problems between 32 bits and 64 bits archs. Either use plain ints (generally
+ 32 bits wide) or long long ints (64 bits wide, at least). At last resort
+ consider using integer types defined in C99 by <stdint.h>.
+
+PRINTF pointer difference (FIXME: advertise %td instead?)
printf ("diff = %ld\n", (long) (pointer2 - pointer1));
-
-**
+**
** Commenting the source: doxygen
**
****************************************************
-The global structure of the documentation is in doc/modules.doc
+The global structure of the documentation is in doc/modules.doc
-The structure of each module (xbt, gras, etc) is in doc/module-<module>.doc
+The structure of each module (xbt, msg, etc) is in doc/module-<module>.doc
The structure of a module is in its public header. This way, you're sure to
see all the public interface (and only it). The different parts of the
The documentation of each type and macro are also in the public header since
this is were they live.
-The documentation of each function must be in the C file were it lives.
+The documentation of each function must be in the C file were it lives.
Any public element (function, type and macro) must have a @brief part.
-
-**
-** Using the logs
+**
+** XBT virtualization mechanism (FIXME: this section is deprecated)
**
****************************************************
-Dans gras, tu ne te contente pas d'écrire des choses à l'écran, mais tu
-écris sur un sujet particulier (notion de canal) des choses d'une gravité
-particulière. Il y a 7 niveaux de gravité.
- trace: tracer les entrées dans une fonction, retour de fonction
- (famille de macros XBT_IN/XBT_OUT)
- debug: pour t'aider à mettre au point le module, potentiellement tres bavard
- verbose: quelques infos succintes sur les internals du module
- info: niveau normal, ton de la conversation
- warning: problème potentiel, mais auquel on a su faire face
- error: problème qui t'as empêché de faire ton job
- critical: juste avant de mourir
-
-Quand on compile avec -DNDEBUG (par défaut dans le paquet Debian), tout ce
-qui est '>= verbose' est supprimé au moment de la compilation. Retiré du
-binaire, killé.
-
-Ensuite, tu écris dans un canal particulier. Tous les canaux sont rangés en
-arbre. Il faudrait faire un ptit script qui fouille les sources à la
-recherche des macros XBT_LOG_NEW_* utilisées pour créer des canaux. Le
-dernier argument de ces macros est ignoré dans le source. Il est destiné à
-être la documentation de la chose en une ligne. En gros, ca fait:
-root
- +--xbt
- | +--config
- | +--dict
- | | +--dict_cursor
- | | +--dict_elm
- | | ...
- | +--dynar
- | +--set
- | +--log
- | +--module
- +--gras
- +--datadesc
- | +--ddt_cbps
- | +--ddt_convert
- | +--ddt_exchange
- | +--ddt_parse
- | +--lexer
- +--msg
- +--transport
- +--raw_trp (Je devrais tuer ce module, un jour)
- +--trp_buf
- +--trp_sg
- +--trp_file
- +--trp_tcp
-
-Et ensuite les utilisateurs peuvent choisir le niveau de gravité qui les
-interresse sur tel ou tel sujet.
-
-Toute la mécanique de logging repose sur des variables statiques dont le nom
-dépend du nom du canal.
- => attention aux conflits de nom de canal
- => il faut une macro XBT_LOG dans chaque fichier où tu fais des logs.
-
-XBT_LOG_NEW_CATEGORY: nouveau canal sous "root". Rare, donc.
-XBT_LOG_NEW_SUBCATEGORY: nouveau canal dont on précise le père.
-XBT_LOG_DEFAULT_CATEGORY: indique quel est le canal par défaut dans ce fichier
-XBT_LOG_NEW_DEFAULT_CATEGORY: Crèe un canal et l'utilise par défaut
-XBT_LOG_NEW_DEFAULT_SUBCATEGORY: devine
-XBT_LOG_EXTERNAL_CATEGORY: quand tu veux utiliser par défaut un canal créé
- dans un autre fichier.
-
-Une fois que ton canal est créé, tu l'utilise avec les macros LOG, DEBUG,
-VERB, WARN, ERROR et CRITICAL. Il faut que tu donne le nombre d'arguments
-après le nom de macro. Exemple: LOG2("My name is %s %s","Martin","Quinson")
-Si tu veux préciser explicitement le canal où écrire, ajoute un C devant le
-nom de la macro. Exemple: CCRITICAL0(module, "Cannot initialize GRAS")
-
-Toutes ces macros (enfin, ce en quoi elles se réécrivent) vérifient leurs
-arguments comme printf le fait lorsqu'on compile avec gcc.
-LOG1("name: %d","toto"); donne un warning, et donc une erreur en mode
-mainteneur.
-
-Enfin, tu peux tester si un canal est ouvert à une priorité donnée (pour
-préparer plus de débug, par exemple. Dans le parseur, je fais du pretty
-printing sur ce qu'il faut parser dans ce cas).
-XBT_LOG_ISENABLED(catName, priority) Le second argument doit être une valeur
-de e_xbt_log_priority_t (log.h). Par exemple: xbt_log_priority_verbose
-
-Voila sur comment mettre des logs dans ton code. N'hesite pas à faire pleins
-de canaux différents pour des aspects différents de ton code. En
-particulier, dans les dict, j'ai un canal pour l'ajout, le retrait, le
-netoyage du code après suppression et ainsi de suite. De cette façon, je
-peux choisir qui m'interresse.
-
-
-Pour utiliser les logs, tu déjà faire, non ? Tu colle sur la ligne de
-commande un ou plusieurs arguments de la forme
- --gras-log="<réglage> [<reglage>+]" (ou sans " si t'as pas d'espace)
-chaque réglage étant de la forme:
- <canal>.thres=<priorité>
-Les différents réglages sont lus de gauche à droite.
-"root.thres=debug root.thres=critical" ferme tout, normalement.
+There is some functionalities that we want to virtualize in XBT. We
+want xbt_time to give the simulated clock when running on top of the
+simulator, and the host clock when running on a real system. This
+could be placed in GRAS (and was, historically), but there is some
+reason to lower it down to XBT.
+
+Here is the used naming scheme:
+
+ - xbt_<module>_<func>(): functions working both in SG and RL
+ - xbt_os_<module>_<func>(): RL functions usable even in simulator
+
+ That way, in libsimgrid, we still can use native functions if we
+ want to. It may for example be useful to get the real time when
+ implementing the simulator. Think of the SIGINT handler, which
+ wants to see if the user pressed the key twice in a 5 seconds
+ interval. This is of little use to check the simulated time here.
+
+Here is the file layout:
+
+ - xbt_rl_<module>.c: native implementation (xbt_<module>_<func>()).
+ Simply call the corresponding xbt_os_<module>_<func>.
+ Part only of libgras.so
+
+ - xbt_sg_<module>.c: SIMIX implementation xbt_<module>_<func>()).
+ Simply call the corresponding SIMIX implementation.
+ Part only of libsimgrid.so
+
+ - xbt_os_<module>.c: body of the functions implementing natively the
+ stuff (xbt_os_<module>_<func>()).
+ Part of both libgras.so and libsimgrid.so
+
+Since there is almost nothing in xbt_rl_module.c and xbt_sg_module.c,
+it'd be better to use symbol aliasing here (to declare in the object
+code that the same function have two names), but I'm still
+investigating the portability of the thing to windows.
+
+
+*
+* SimGrid Hacker Survival Guide (FIXME: should be betterly placed)
+********************************
+
+* Before pushing any change, don't forget to check if the compilation
+ passes with compiler optimizations and warnings turned on:
+ cmake -Denable_compile_optimizations=ON \
+ -Denable_compile_warnings=ON
+
+* Your commit message should follow the git habits, explained eg here:
+ http://tbaggery.com/2008/04/19/a-note-about-git-commit-messages.html
+
+* When you add/remove files, and/or make changes in the lists of files to build,
+ please check that "make distcheck" still succeeds. This is needed to ensure
+ that the generated archive is consistent.
+
+* If you want to debug memory allocation problems, here are a few hints:
+ - disable compiler optimizations, to have better backtraces;
+ - disable the mallocators, or it will be hard to match malloc's with
+ free's;
+ - disable model checking, unless your problem lies in the model
+ checker part of SimGrid (MC brings its own malloc implementation,
+ which valgrind doesn't understand).
+ All this is configured with:
+ cmake -Denable_model-checking=OFF \
+ -Denable_mallocators=OFF \
+ -Denable_compile_optimizations=OFF
+
+* If you break the logs (for example while hacking in the dynars), you
+ want to define XBT_LOG_MAYDAY at the beginning of log.h. It will
+ deactivate the whole logging mechanism, switching to printfs
+ instead. SimGrid becomes incredibly verbose when doing so, but it
+ you let you fixing the dynars.
