1 SimGrid (2.95) unstable; urgency=low
4 * Implement MSG_get_msgload [MQ]
7 * Bugfix: Allow absolute paths for platform description files [MQ]
10 * <project>.Makefile.local (generated from gras_stub_generator) |MQ]:
13 * bugfix in automatic type parsing: differentiate the types "char[22]"
14 and "unsigned char[22]". "short" and "long" modifiers were also ignored;
15 other modifier (such as reference level) are still ignored. [MQ]
16 * Real (even if simplistic) implementation of gras_os_myname() on RL [MQ]
17 * bugfix: embeed the buffer size within the buffer itself on SG. [MQ]
18 That way, send() are atomic and cannot get intermixed anymore (at least
19 the ones which are less than 100k; bigger messages still have the issue)
23 SimGrid (2.94) unstable; urgency=low
25 The first release candidate for SimGrid 3 !
29 * Update the main page and the FAQ. Adding references to gforge.
32 * Add a gras_os_getpid function.
35 * Add MSG_task_get_compute_duration() and MSG_task_get_data_size()
36 * Extend the logs so that they also print PID, hostname, date, ... if
38 * Convert the MSG example to the use of xbt_logs instead of PRINT_MESSAGE,
39 and kill the old version which were in testsuite/
40 * Rewrite tools/MSG_visualization/colorize.pl for using with logs instead
44 * Add xbt_os_time(). As the rest of xbt/portability, this is not public
45 for users. Instead, each programming environment (GRAS, MSG,...) use it
46 when needed to provide such a feature to users.
47 Don't shortcut the mecanism or you will also shortcut the virtualization
48 you need on the simulator.
52 * Cleanups in configury with regard to compile optimization/warning flags.
53 Also add -fno-loop-optimize to any powerpc since it's the optimization
54 killing gcc (< 3.4.0).
55 * Doxygen cleanups: move MSG examples, kill the second Doxygen phase
56 needed by MSG examples complications
57 * Borrow configury beautifications from PHP
60 * Bugfix: XBT_LOG_NEW_DEFAULT_CATEGORY now compiles without compiler
61 warning (thanks loris for stumbling into this one).
62 * Bugfix: stop loading private headers (gras_config.h) from the public
66 * Change SIMGRID_INSTALL_PATH to GRAS_ROOT in Makefiles generated for user.
67 * Rename gras_get_my_fqdn to gras_os_myname and implement it in the simulator
68 RL would imply a DNS resolver, which is *hard* to do in a portable way
69 (and therefore delayed).
70 * Implement a real timer mecanism and use it in timing macros. This allows
71 to avoid rounding errors and get a 0.000005 sec precision in timing
72 macros. While I was at it, various cleanups:
73 - allow to declare more than one timed section per file (fix a stupid bug)
74 - move some private declaration to the right place
75 - merge conditional execution and timing macros into emulation module
77 - make sure the module cleanups its mess on gras_exit
78 * Documentation improvements:
79 - (new) how to compile applications using GRAS
80 - (new) emulation support (timing macros)
82 -- Da SimGrid team <simgrid-devel@lists.gforge.inria.fr> Fri, 13 May 2005 10:49:31 +0200
84 SimGrid (2.93) unstable; urgency=low
86 Alpha 4 on the path to SimGrid 3 (aka the "neuf-trois" version)
89 - Use Paje properly where used. Still to be sanitized properly.
90 - Portability fix: Add an implementation of the contexts using pthread
94 - Add xbt_procname(): returns the name of the current process.
95 Use it to show the current process's name in all logging.
97 - fix detection of older flex version and the reaction, since we do
98 depend on modern ones (we use lex_destroy)
99 - Better separation of SG and RL in the libs: remove all simulation code
100 from libgras. As a result, this lib is now only 200k when stripped.
101 Some of the xbt modules may also be duplicated (two sets and such) and
102 should be cleaned/killed before SG3.
103 - Insist on using xlC on AIX because of weird problems involving gcc there.
104 - Cleanup the make remote stuff. This is now done by scripts
105 tools/graspe-{master,slave} (GRAS Platform Expender). This is still
106 mainly for our private use, but we're working on changing them to user
109 - Bugfix: flush the socket on close only if there is some *output*.
110 - Bugfix: flush idempotent when there's nothing to send (don't send size=0)
112 - Add MSG_task_get_name. The task names are mainly for debugging purpose,
115 -- SimGrid team <simgrid2-users@listes.ens-lyon.fr> Fri, 4 Mar 2005 14:32:37 -0800
117 SimGrid (2.92) unstable; urgency=low
119 Alpha 3 on the path to SimGrid 3
123 - New! First try of benchmarking macros.
124 - New! First try so that gras_stub_generator generate deployment and
125 remote compilation helpers.
127 - Bugfix: Initialization fix in msg_test.
131 - Bugfix: applied patch to lexer so that it doesn't need a huge heap.
133 - Bugfix: let dicts work with NULL content (_foreach didn't) and cleanups
135 - API Change: gras_os_sleep to take the amount of seconds as a double.
136 Accepting an int was error prone since it was the only location where
137 seconds were coded as such. It leaded to damn rounding errors.
138 - Bugfix: Hard to belive that timers ever worked before this.
140 -- SimGrid team <simgrid2-users@listes.ens-lyon.fr> Wed, 23 Feb 2005 22:09:21 +0100
142 SimGrid (2.91) unstable; urgency=low
144 Alpha 2 on the path to SimGrid 3
148 - Bug fix in the lmm_solver.
150 - New! Interface to Paje (see http://www-id.imag.fr/Logiciels/paje/)
151 through the function MSG_paje_output.
152 - New! Introducing two new functions MSG_process_kill() and MSG_process_killall().
153 - It is possible to bound the rate of a communication in MSG with
154 MSG_task_put_bounded() (was already in the previous version but I had forgotten
155 to write it in the changelog).
156 - Bug fix to let GRAS run on top of MSG until we move it directly on top
161 - Various cleanups to the autotools stuff
162 - Begin to move Gras examples to examples/gras/
163 - Let make distcheck work again (yeah!)
165 - documentation overhauled using doxygen.
166 gtk-doc-tools is dead in SimGrid now.
167 - Automatically extract all existing logging categories, and add the list
168 to the documentation (long standing one, to say the less)
170 - Cleanup the known architecture table. Reorder the entries to group what
171 should be, and use a more consistent naming scheme.
172 (some of the test dataset are still to be regenerated)
173 - New! Allow library to register globals on each process just as userdata
175 This is implemented using a xbt_dict and not a xbt_set, so we loose the
176 lookup time (for now).
177 Use it in msg and trp.
178 This cleans a lot the internals and helps enforcing privacy of the
179 headers between the gras components.
180 - New! Add a timer mechanism, not unlike cron(8) and at(1).
181 - Bugfix: gras_os_time was delirious in RL.
182 - Bugfix: gras_trp_select/RL don't run into the wall when asked to select
184 - Reenable GRAS now that it works.
186 -- Arnaud Legrand <Arnaud.Legrand@imag.fr> Mon, 14 Feb 2005 14:02:13 -0800
188 SimGrid (2.90) unstable; urgency=low
190 Alpha 1 on the path to SimGrid 3
192 * It is a long time since the last release of SimGrid. I'm sorry about
193 that but as I had told you, I was rewriting a lot of things. I apologize
194 to those who had been reporting bugs to me and that I had not answered.
195 If your bug is still in the new version, please tell me. Here is a
196 summary of the main changes.
198 * REVOLUTION 1: The SimGrid project has merged with the GRAS project
199 lead by Martin Quinson. As a consequence SimGrid gains a lot in
200 portability, speed, and a lot more but you'll figure it out later.
201 SimGrid now comprises 3 different projects : MSG, GRAS and SMPI.
202 I wanted to release the new MSG as soon as possible and I have
203 broken GRAS, which is the reason why, for now, only MSG is fully
204 functional. A laconic description of these projects is available
205 in the documentation.
207 * REVOLUTION 2: I have removed SG and I am now using a new simulation
208 kernel optimized for our needs (called SURF but only the developers
209 should use it). Hence, MSG is now roughly 30 times faster and I think
210 that by rewriting a little bit MSG, I could event speed it up a little
211 bit more. Beside the gain in speed, it is also much easier to encode a
212 new platform model with SURF than it was with SG. More to come...
214 * REVOLUTION 3: I have tried to change a little as possible the API of
215 MSG but a few things really had to disappear. The main differences
216 with the previous version are :
217 1) no more m_links_t and the corresponding functions. Platforms are
218 directly read from a XML description and cannot be hard-coded
219 anymore. The same format is used for application deployment
220 description. The new format is described in the documentation.
221 Have a look in tools/platform_generation. There is a tiny script
222 that converts from the old platform format to the new one. Concerning
223 the application deployment format, parsing the old one is tricky.
224 I think most of you should however be able to convert your files. If
225 it is really an issue, I can write a C code that does the conversion.
227 2) the toolbox tbx does not exist anymore. We now have a library
228 with much more data-structures but without the hash-tables (we have
229 dictionaries that are much faster).
231 -- Arnaud Legrand <Arnaud.Legrand@imag.fr> Mon, 31 Jan 2005 10:45:53 -0800
233 *****************************************************************************
234 * Follows the old GRAS changelog. It does not follow the same syntax, but I *
235 * don't feel like converting the oldies. (Mt) *
236 *****************************************************************************
239 Version 2.90: "the long awaited one"
240 - Finished rewriting and debugging MSG. Rewrote the documentation.
241 - disable GRAS for now since it needs to be ported to the newest SG
244 - Finish the port to windows (using mingw32 for cross-compile)
247 - Main loop and datastructures of SURF. A cpu resource object is
248 functional. Surf can thus be used to create cpu's with variable
249 performance on which you can execute some actions.
251 2004-11-15 Martin Quinson
252 - Port to ARM. Simply added the alignment and size descriptions. Should
253 work, but the ARM machines are so slow that I didn't had the opportunity
254 to 'make check' over there yet.
256 2004-11-15 Arnaud Legrand
257 - Trace manager now written. It uses a heap structure and is therefore
258 expected to be efficient. It may however be speeded up (particularly
259 when many events occur at the same date) by using red and black
260 trees. One day maybe...
261 - Max-min linear system solver written. It uses a sparse matrix
262 structure taking advantage of its expected use. Most operations are
263 O(1) and free/calloc are called as few as possible. The computation of
264 the minimum could however be improved by using a red and black tree
267 2004-11-03 Arnaud Legrand
268 - Rename every gras_* function that was in xbt/ to its xbt_
270 - Add a heap and a doubly-linked list to xbt
271 - Added a dichotomy to the dictionaries. make check works as well before
272 so I assume that the patch is correct. I do not know however if things
273 run effectively faster than before now. :)
275 Inclusion of the SimGrid tree in the GRAS one. The archive is renamed to
276 SimGrid, and the version number is bumped to 2.x
278 2004-10-29 Martin Quinson
279 - Introduction of the remote errors.
280 They are the result of a RMI/RPC on the remote machine.
281 ErrCodes being scalar values, you can't get the host on which those
282 errors did happen. Extending the error mechanism as in Gnome is possible.
283 No idea yet whether it is a good idea.
285 2004-10-28 Martin Quinson
286 - Interface revolution: the Starred Structure Eradication.
287 I used to do typedef struct {} toto_t; and then handle *toto_t.
288 Arnaud (and Oli) didn't like it, and I surrendered. Now, you have:
289 - ???_t is a valid type (builded with typedef)
290 - s_toto_t is a structure (access to fields with .)
291 - s_toto is a structure needing 'struct' keyword to be used
292 - e_toto_t is an enum
293 - toto_t is an 'object' (struct*)
295 typedef struct s_toto {} s_toto_t, *toto_t;
296 typedef enum {} e_toto_t;
297 Moreover, only toto_t (and e_toto_t) are public. The rest (mainly
298 s_toto_t) is private.
300 - While I was at it, all gras_<obj>_free() functions want a gras_<obj>_t*
301 so that it can set the variable to NULL. It was so for dicts and sets,
302 it changed for dynars.
304 - Fix a bunch of memleaks in dict_remove
305 - Fix a bug in sg/server_socket opening: it failed all the time.
307 2004-10-07 Martin Quinson
308 - Speed up dynar lookup operation a bit.
310 gras_dynar_get is dead.
312 Now, you can choose between gras_dynar_get_cpy (the old gras_dynar_get
313 but should be avoided for efficiency reasons) and gras_dynar_get_ptr
314 (which gives you the address of the stored data).
316 gras_dynar_get_as is an helpful macro which allows you to retrieve a
317 copy of the data using an affectation to do the job and not a memcpy.
319 int toto = gras_dynar_get_as(dyn,0,int); rewrites itself to
320 int toto = *(int*)gras_dynar_get_ptr(dyn,0);
322 It does not really speedup the dynar test because they are
323 setting elements all the time (and look them seldom). But the dict does
324 far more lookup than setting.
326 So, this brings the dict_crash test from ~33s to ~25s (200000 elms).
328 2004-10-05 Martin Quinson
329 - Allow to (en/dis)able the cycle detection at run time.
331 Whether we should check for cycle or not is now a property of each
332 datatype. When you think there may be some cycle, use datadesc_cycle_set.
333 datadesc_cycle_unset allow to remove this property when previously set.
335 Note that the cycle detection is off by default since it impacts the
336 performance. Watch the data you feed GRAS with ;)
338 This property is hereditary. Any element embedded in a structure having it
339 set have it set for the time of this data exchange.
341 You should set it both on sender and receiver side. If you don't set it on
342 sender side, it will enter an endless loop. If you forget on receiver
343 side, the cycles won't be recreated after communication.
345 - Header reorganization.
346 Kill gras_private.h, each submodule must load the headers it needs.
348 2004-10-04 Martin Quinson
349 - Interface revolution: do not try to survive to malloc failure.
351 Now, gras_malloc and friends call gras_abort() on failure.
352 As a conclusion, malloc_error is not a valid error anymore, and all
353 functions for which it was the only gras_error_t return value are
354 changed. They now return void, or there result directly.
355 This simplify the API a lot.
357 2004-09-29 Martin Quinson
358 - Re-enable raw sockets.
359 Created by gras_socket_{client,server}_ext;
360 Used with gras_raw_{send,recv}
363 It should allow to kill the last bits of gras first version soon.
365 This is not completely satisfactory yet (duplicate code with
366 chunk_{send,recv}; a bit out of the plugin mechanism), but it should
369 - Simplify transport plugin (internal) interface by not passing any
370 argument to _server and _client, but embedding them in the socket
373 2004-09-28 Martin Quinson
374 - Finish the port to AIX.
375 autoconf was my problem (segfault within the malloc replacement
376 function. No idea why)
378 2004-09-16 Martin Quinson
379 - Fix some size_t madness on 64bit architectures.
381 2004-09-08 Martin Quinson
382 - Reduce the number of system headers loaded, overload some more system
383 calls (such as malloc to cast the result of the system one, and work
385 - Fix and reintroduce the config support
387 2004-09-07 Martin Quinson
388 - Source code reorganization to allow Arnaud to surf all over there.
389 - Allow to document the logging categories.
390 - Remove all uppercase from logging categories and useless cleanup in names.
392 2004-08-18 Martin Quinson
393 Version 0.6.2 (protocol not changed; API changed)
394 - Interface cleanup: gras_msgtype_by_name returns the type (instead of a
395 gras_error_t), and NULL when not found. Functions expecting a msgtype
396 as argument (msg_wait; msg_send) deal with NULL argument by providing a
397 hopefully usefull message.
398 - Portability to prehistoric sparcs again
400 2004-08-17 Martin Quinson
401 Version 0.6.1 (protocol not changed; ABI not changed)
402 - prealloc some buffers to speed things up
404 2004-08-11 Martin Quinson
405 Version 0.6 (protocol not changed; ABI expended)
406 - The parsing macro can deal with the references, provided that you add
407 the relevant annotations (using GRAS_ANNOTE(size,field_name))
409 2004-08-09 Martin Quinson
410 Version 0.5 (protocol not changed; ABI changed)
411 - Allow to off turn the cycle detection code in data exchange at
412 compilation time. It should be at run time, but I'm short of time (and
413 the config stuff is still broken). That way, we keep dict out of the
414 critical path, which is good because the performance is poor:
415 - search not dichotomial yet
416 - dynar give no way to access their content and memcpy everytime
417 - In composed data description (struct, ref and so on), stop foolness of
418 keeping the subtype's ID, but store the type itself. This keeps sets out
419 of the critical path, which is good since they rely on dynar and
420 dictionnaries. The only loose of that is that we cannot detect the
421 redeclaration of a structure/union with another content (but I'm not sure
422 the code detected well this error before anyway). We still can detect
423 the redefinition discrepancy for the other types.
424 - Use a whole bunch of optimisation flags (plus -fno-strict-aliasing since
425 it breaks the code because of type-punning used all over the place).
426 This breaks on all non-gcc architectures (for now).
428 All those changes (plus the buffer of last time) allow me to gain 2 order
429 of magnitude on cruel tests consisting of 800000 array of integers on two
430 level of a hierarchical structure (200 secondes -> 4 secondes)
433 - the selector of reference must now return the type it points to, not
436 2004-08-06 Martin Quinson
437 Version 0.4 (protocol changed; ABI not changed)
438 - Allow to pass --gras-log argument to processes in simulation mode. Really.
439 - New debugging level: trace (under debug) to see effect of GRAS_IN/OUT
440 - Implement a buffer transport, and use it by default (it relies on tcp in
441 real life and on sg in simulation).
442 That's a bit hackish since I had a new field to the structure to store
443 its data without interfering with the subtype ones. Inheritance
444 is tricky in C. And that's a kind of reverse inheritance with one class
445 derivating two classes. Or maybe a game with java interfaces. Anyway,
446 that's damn hard in C (at least).
447 Moreover, I got tired while trying to ensure plugin separation and
448 genericity in SG mode. MSG wants me to do weird things, so let's go for
449 cruel hacks (temporarily of course ;).
450 See comment in transport_private.h:71
451 - do not include all the _interface headers in private but in the files
452 which really need them (to cut the compilation time when they are
455 2004-07-26 Martin Quinson
456 Version 0.3 (protocol not changed; ABI changed)
457 - Major overhault of the datadesc interface to simplify it:
458 - shorted the function names:
459 s/gras_datadesc_declare_struct/gras_datadesc_struct/ and so on
460 - add a trivial way to push/pop integers into the cbps without malloc.
461 This allows to make really generic sub_type description, which simply
462 pop their size of the stack.
463 - add a function gras_datadesc_ref_pop_arr() which does what users want
464 most of the time: Declare a dynamic array (which pops its size of the
465 stack) and declare a reference to it. Poor name, but anyway.
466 - kill the post-send callback, add a post-receive one
468 2004-07-23 Martin Quinson
469 Version 0.2 (protocol changed; ABI changed)
470 - add some testing for cpbs in the test cases, and fix some more bugs.
471 This invalidate again the little64 data file, since I cannot regenerate
473 - remove an awfull optimization in the logging stuff, allowing me to:
474 - understand it again
475 - learn gcc how to check that the argument match the provided format
476 - fix all errors revealed by gcc after that
477 - internal keys of dict are not \0 terminated. Deal with it properly in
478 loggings instead of segfaulting when the user want to see the logs :-/
480 2004-07-22 Martin Quinson
481 - Fix some stupid bug preventing cbps (callback postit) from working
483 2004-07-21 Martin Quinson
484 - Some documentation cleanups
485 - remove the useless last argument of msgtype_declare
486 - rename the Virtu functions to fit into the 'os' namespace
487 - move headers src/include -> src/include/gras/ and stop fooling with
488 gras -> . symbolic link
489 - make distcheck is now successful
491 2004-07-19 Martin Quinson
493 - Build shared library also
494 - Install html doc to the right location
495 - stop removing maintainer files in make clean
496 - build tests only on make check
498 2004-07-13 Martin Quinson
500 - No major issue in previous version => change versionning schema
501 - Re-enable little64 convertion test now that Abdou kindly regenerated the
502 corresponding dataset.
504 2004-07-11 Martin Quinson
506 - Get it working with any kind of structure (we can compute the padding
507 bytes remotely for all the architectures I have access to)
508 - Implement the structure parsing macro (still not quite robust/complete)
509 - Improvement to the remote testing toysuite
511 2004-07-10 Martin Quinson
513 - get ride of a bunch of deprecated macros
514 - actually run the test for two-compliment, not only compile it :-/
515 - test whether the structures get packed (and bail out if yes. Damn.
516 Alignment is a serious matter)
517 - test whether the structures get compacted (but respecting the alignment
518 constraints of each types)
519 - test whether the array fields of structures can straddle alignment boundaries
521 - Damnit, double are bigger than float (typo in creation of 'double' datadesc)
522 (took me 2 hours to find that bug, looking at the wrong place)
523 - Add gras_datadesc_declare_{union,struct}_close(). They must be used
524 before sending/receiving and are used to compute the offsets of fields
525 - Given that padding size depend even on compiler options, keep track of
526 alignment and aligned_size only for the current architecture. Not a big
527 deal since we send structure fields one after the other (seems
529 - Add the datastructure used for IEEE paper by the PBIO guys to the test
530 program, let it work on linux/gcc/little32. portability todo.
532 2004-07-08 Martin Quinson
533 - import and improve remote compilation support from FAST
534 - make sure make check works on half a dozen of machines out there
536 2004-07-07 Martin Quinson
537 Let's say it's version 0.0.3 ;)
538 - Implement conversions (yuhu!)
539 - Let it work on solaris (beside conversion, of course)
540 - Stupid me, using rand() to generate the conversion datatests in not wise.
542 2004-07-06 Martin Quinson
543 - Let make dist work, since I'm gonna need it to compile on remote hosts
544 - Let Tests/datadesc_usage write the architecture on which the file was
545 generated as first byte.
546 - Add PowerPC (being also IRIX64), SPARC (also power4) and ALPHA
547 architecture descriptions.
548 - Add datadesc_usage.{i386,ppc,sparc} files being the result of execution
549 on those architectures.
550 - Optimization: send/recv array of scalar in one shoot
552 2004-07-05 Martin Quinson
553 - YEAH! GRAS/SG and GRAS/RL are both able to run the ping example !
555 - Plug a whole bunch of memleaks
556 - each process now have to call gras_{init,exit}. One day, their log
557 settings will be separated
558 - Continue the code factorisation between SG, RL and common in Transport.
560 2004-07-04 Martin Quinson
562 - Redistribution between SG and RL.
563 We wanna have to accept in SG, so move accepted related parts of RL in
564 the common part. (more precisely, the dynar of all known sockets is no
565 more a static in transport.c, but part of the process_data)
567 [gras_stub_generator]
568 - Bug fix: Do call gras_process_init from gras_init (wasnt called in RL).
570 2004-07-03 Martin Quinson
571 - Create a new log channel tbx containing dict, set, log, dynar (to shut
572 them all up in one shot)
574 - Fix the ugly case of reference to dynamic array.
575 - New (semi-public) function gras_datadesc_size to allow the messaging
576 layer to malloc the needed space for the buffer.
578 - gras_socket_close now expect the socket to close (and not its address to
579 put NULL in it after it). This is because the socket passed to handlers
580 is one of their argument (=> not writable).
582 - propagate the interface cleanup from last week in datadesc, ie remove a
583 superfluous level of indirection. User pass adress of variable
584 containing data (both when sending and receiving), and not of a variable
585 being a pointer to the data. Let's say that I like it better ;)
586 The price for that is constructs like "int msg=*(int*)payload" in
587 handlers, but it's a fine price, IMHO.
589 - Let it work in RL (yuhu)
591 2004-06-21 Martin Quinson
593 - porting SG plugin and SG select to new standards (works almost).
594 - plug memleaks and fix bugs around.
597 - cleanup the prototype of data recv and force users to specify when they
598 want to handle references to objects. Test case working even for cycles.
599 - plug memleaks. Valgrind is perfectly ok with this.
601 2004-06-12 Martin Quinson
603 - cleanup the separation between plugin and main code in plugin creation
605 2004-06-11 Martin Quinson
607 - Reput hook for raw sockets, needed for BW experiments
608 - kill a few lines of dead code
609 [Data description] Interface cleanup
610 - gras_datadesc_by_name returns the searched type or NULL.
611 That way, no variable is needed to use a type desc once, which makes
613 - gras_datadesc_declare_[struct|union]_append_name is removed. The last
614 two parameters were strings (field name, type name), leading to
616 [Dicos] Interface cleanup
617 - gras_dico_retrieve -> gras_dico_get ; gras_dico_insert -> gras_dico_set
618 This is consistant with the dynar API.
620 2004-04-21 Martin Quinson
622 - Porting to new standards.
625 There is no bag anymore, no need to take extra provision to mask the
626 pointers behind "ID".
627 Better splitup of functions between files create/exchange/convert.
628 This is still a bit artificial since convert and receive are so
629 interleaved, but anyway.
631 - add a queued message list to procdata (the ones not matching criteria
633 - factorize some more code between SG and RL wrt procdata
635 - use gras_exit in example to track memleaks
636 - get rid of gs_example now that GS is properly integrated into gras
637 - update run_test to integrate the lastest tests (datadesc)
639 - rename WARNINGn macros to WARNn since it prooved error-prone
641 2004-04-19 Martin Quinson
643 - register init/exit functions within gras module mechanism
644 - send/receive function.
645 Convertion is not implemented, but short-cutted if not needed.
646 struct/array elements are sent one by one (instead of block-wise), but
647 nobody really cares (yet). Get a prototype before optimizing.
648 - tests (using a file socket) for DD send/receive on:
649 - base types: int, float
650 - array: fixed size, string (ie ref to dynamic string)
651 - structure: homogeneous, heterogeneous
652 - chained list, graph with cycle
653 Believe it or not, valgrind is not too unhappy with the results. The
654 cycle happily segfaults, but the others are ok. And I'm sick of pointers
658 - Bugfix when using a filename explicitely (instead of '-')
660 2004-04-09 Martin Quinson
662 - factorize more code between RL and SG in socket creation
663 - Complete the implementation and tests of:
665 o file (only in RL, and mainly for debugging)
667 I lost 3 days to design a portable address resolver, and then decided
668 that the prototype mainly have to run on my box.
669 Addressing portability too early may be like optimizing too early :-/
671 - use gras_init in the Tests instead of the crappy parse_log_opt
672 (the latter function is removed)
673 [Conditional execution]
674 - New functions: gras_if_RL/gras_if_SG (basic support for this)
675 [Code reorganisation]
676 - Get rid of libgrasutils.a since it makes more trouble than it solves.
677 Build examples against the RL library, since there is no way to disable
678 its creation for now.