1 SimGrid (2.95) unstable; urgency=low
4 * Implement MSG_get_msgload [MQ, AL]
7 * Bugfix: Allow absolute paths for platform description files [MQ]
8 * Bugfix: do free the variables after use. Leads to drastic performance
12 * gras_datadesc_dynar() builds a dynar type descriptor, allowing to send
13 dynar over the network (yeah) [MQ]
14 * Real (even if simplistic) implementation of gras_os_myname() on RL [MQ]
15 * simple/static token-ring example. [Alexandre Colucci and MQ]
18 * Differentiate the types "char[22]" and "unsigned char[22]" in automatic
19 type parsing. "short" and "long" modifiers were also ignored; other
20 modifier (such as reference level) are still ignored. [MQ]
21 * Embeed the buffer size within the buffer itself on SG. [MQ]
22 That way, send() are atomic and cannot get intermixed anymore (at least
23 the ones which are less than 100k; bigger messages still have the issue)
24 * Array size pushed by the field, not by the field type (or each
25 and every long int will push stuff to the cbps) [MQ]
28 * <project>.Makefile.local (generated from gras_stub_generator) |MQ]:
31 * Callbacks now receive the gras_datadesc_type_t they work on as argument.
32 * type category 'ignored' killed as it was never used and were difficult
34 * whether a type can cycle or not is now a flag, leaving room for more
35 flags (as "ignored", if we feel the need one day ;)
36 * Rename raw sockets to measurement sockets since "raw" has another
37 meaning in networking community.
41 SimGrid (2.94) unstable; urgency=low
43 The first release candidate for SimGrid 3 !
47 * Update the main page and the FAQ. Adding references to gforge.
50 * Add a gras_os_getpid function.
53 * Add MSG_task_get_compute_duration() and MSG_task_get_data_size()
54 * Extend the logs so that they also print PID, hostname, date, ... if
56 * Convert the MSG example to the use of xbt_logs instead of PRINT_MESSAGE,
57 and kill the old version which were in testsuite/
58 * Rewrite tools/MSG_visualization/colorize.pl for using with logs instead
62 * Add xbt_os_time(). As the rest of xbt/portability, this is not public
63 for users. Instead, each programming environment (GRAS, MSG,...) use it
64 when needed to provide such a feature to users.
65 Don't shortcut the mecanism or you will also shortcut the virtualization
66 you need on the simulator.
70 * Cleanups in configury with regard to compile optimization/warning flags.
71 Also add -fno-loop-optimize to any powerpc since it's the optimization
72 killing gcc (< 3.4.0).
73 * Doxygen cleanups: move MSG examples, kill the second Doxygen phase
74 needed by MSG examples complications
75 * Borrow configury beautifications from PHP
78 * Bugfix: XBT_LOG_NEW_DEFAULT_CATEGORY now compiles without compiler
79 warning (thanks loris for stumbling into this one).
80 * Bugfix: stop loading private headers (gras_config.h) from the public
84 * Change SIMGRID_INSTALL_PATH to GRAS_ROOT in Makefiles generated for user.
85 * Rename gras_get_my_fqdn to gras_os_myname and implement it in the simulator
86 RL would imply a DNS resolver, which is *hard* to do in a portable way
87 (and therefore delayed).
88 * Implement a real timer mecanism and use it in timing macros. This allows
89 to avoid rounding errors and get a 0.000005 sec precision in timing
90 macros. While I was at it, various cleanups:
91 - allow to declare more than one timed section per file (fix a stupid bug)
92 - move some private declaration to the right place
93 - merge conditional execution and timing macros into emulation module
95 - make sure the module cleanups its mess on gras_exit
96 * Documentation improvements:
97 - (new) how to compile applications using GRAS
98 - (new) emulation support (timing macros)
100 -- Da SimGrid team <simgrid-devel@lists.gforge.inria.fr> Fri, 13 May 2005 10:49:31 +0200
102 SimGrid (2.93) unstable; urgency=low
104 Alpha 4 on the path to SimGrid 3 (aka the "neuf-trois" version)
107 - Use Paje properly where used. Still to be sanitized properly.
108 - Portability fix: Add an implementation of the contexts using pthread
112 - Add xbt_procname(): returns the name of the current process.
113 Use it to show the current process's name in all logging.
115 - fix detection of older flex version and the reaction, since we do
116 depend on modern ones (we use lex_destroy)
117 - Better separation of SG and RL in the libs: remove all simulation code
118 from libgras. As a result, this lib is now only 200k when stripped.
119 Some of the xbt modules may also be duplicated (two sets and such) and
120 should be cleaned/killed before SG3.
121 - Insist on using xlC on AIX because of weird problems involving gcc there.
122 - Cleanup the make remote stuff. This is now done by scripts
123 tools/graspe-{master,slave} (GRAS Platform Expender). This is still
124 mainly for our private use, but we're working on changing them to user
127 - Bugfix: flush the socket on close only if there is some *output*.
128 - Bugfix: flush idempotent when there's nothing to send (don't send size=0)
130 - Add MSG_task_get_name. The task names are mainly for debugging purpose,
133 -- SimGrid team <simgrid2-users@listes.ens-lyon.fr> Fri, 4 Mar 2005 14:32:37 -0800
135 SimGrid (2.92) unstable; urgency=low
137 Alpha 3 on the path to SimGrid 3
141 - New! First try of benchmarking macros.
142 - New! First try so that gras_stub_generator generate deployment and
143 remote compilation helpers.
145 - Bugfix: Initialization fix in msg_test.
149 - Bugfix: applied patch to lexer so that it doesn't need a huge heap.
151 - Bugfix: let dicts work with NULL content (_foreach didn't) and cleanups
153 - API Change: gras_os_sleep to take the amount of seconds as a double.
154 Accepting an int was error prone since it was the only location where
155 seconds were coded as such. It leaded to damn rounding errors.
156 - Bugfix: Hard to belive that timers ever worked before this.
158 -- SimGrid team <simgrid2-users@listes.ens-lyon.fr> Wed, 23 Feb 2005 22:09:21 +0100
160 SimGrid (2.91) unstable; urgency=low
162 Alpha 2 on the path to SimGrid 3
166 - Bug fix in the lmm_solver.
168 - New! Interface to Paje (see http://www-id.imag.fr/Logiciels/paje/)
169 through the function MSG_paje_output.
170 - New! Introducing two new functions MSG_process_kill() and MSG_process_killall().
171 - It is possible to bound the rate of a communication in MSG with
172 MSG_task_put_bounded() (was already in the previous version but I had forgotten
173 to write it in the changelog).
174 - Bug fix to let GRAS run on top of MSG until we move it directly on top
179 - Various cleanups to the autotools stuff
180 - Begin to move Gras examples to examples/gras/
181 - Let make distcheck work again (yeah!)
183 - documentation overhauled using doxygen.
184 gtk-doc-tools is dead in SimGrid now.
185 - Automatically extract all existing logging categories, and add the list
186 to the documentation (long standing one, to say the less)
188 - Cleanup the known architecture table. Reorder the entries to group what
189 should be, and use a more consistent naming scheme.
190 (some of the test dataset are still to be regenerated)
191 - New! Allow library to register globals on each process just as userdata
193 This is implemented using a xbt_dict and not a xbt_set, so we loose the
194 lookup time (for now).
195 Use it in msg and trp.
196 This cleans a lot the internals and helps enforcing privacy of the
197 headers between the gras components.
198 - New! Add a timer mechanism, not unlike cron(8) and at(1).
199 - Bugfix: gras_os_time was delirious in RL.
200 - Bugfix: gras_trp_select/RL don't run into the wall when asked to select
202 - Reenable GRAS now that it works.
204 -- Arnaud Legrand <Arnaud.Legrand@imag.fr> Mon, 14 Feb 2005 14:02:13 -0800
206 SimGrid (2.90) unstable; urgency=low
208 Alpha 1 on the path to SimGrid 3
210 * It is a long time since the last release of SimGrid. I'm sorry about
211 that but as I had told you, I was rewriting a lot of things. I apologize
212 to those who had been reporting bugs to me and that I had not answered.
213 If your bug is still in the new version, please tell me. Here is a
214 summary of the main changes.
216 * REVOLUTION 1: The SimGrid project has merged with the GRAS project
217 lead by Martin Quinson. As a consequence SimGrid gains a lot in
218 portability, speed, and a lot more but you'll figure it out later.
219 SimGrid now comprises 3 different projects : MSG, GRAS and SMPI.
220 I wanted to release the new MSG as soon as possible and I have
221 broken GRAS, which is the reason why, for now, only MSG is fully
222 functional. A laconic description of these projects is available
223 in the documentation.
225 * REVOLUTION 2: I have removed SG and I am now using a new simulation
226 kernel optimized for our needs (called SURF but only the developers
227 should use it). Hence, MSG is now roughly 30 times faster and I think
228 that by rewriting a little bit MSG, I could event speed it up a little
229 bit more. Beside the gain in speed, it is also much easier to encode a
230 new platform model with SURF than it was with SG. More to come...
232 * REVOLUTION 3: I have tried to change a little as possible the API of
233 MSG but a few things really had to disappear. The main differences
234 with the previous version are :
235 1) no more m_links_t and the corresponding functions. Platforms are
236 directly read from a XML description and cannot be hard-coded
237 anymore. The same format is used for application deployment
238 description. The new format is described in the documentation.
239 Have a look in tools/platform_generation. There is a tiny script
240 that converts from the old platform format to the new one. Concerning
241 the application deployment format, parsing the old one is tricky.
242 I think most of you should however be able to convert your files. If
243 it is really an issue, I can write a C code that does the conversion.
245 2) the toolbox tbx does not exist anymore. We now have a library
246 with much more data-structures but without the hash-tables (we have
247 dictionaries that are much faster).
249 -- Arnaud Legrand <Arnaud.Legrand@imag.fr> Mon, 31 Jan 2005 10:45:53 -0800
251 *****************************************************************************
252 * Follows the old GRAS changelog. It does not follow the same syntax, but I *
253 * don't feel like converting the oldies. (Mt) *
254 *****************************************************************************
257 Version 2.90: "the long awaited one"
258 - Finished rewriting and debugging MSG. Rewrote the documentation.
259 - disable GRAS for now since it needs to be ported to the newest SG
262 - Finish the port to windows (using mingw32 for cross-compile)
265 - Main loop and datastructures of SURF. A cpu resource object is
266 functional. Surf can thus be used to create cpu's with variable
267 performance on which you can execute some actions.
269 2004-11-15 Martin Quinson
270 - Port to ARM. Simply added the alignment and size descriptions. Should
271 work, but the ARM machines are so slow that I didn't had the opportunity
272 to 'make check' over there yet.
274 2004-11-15 Arnaud Legrand
275 - Trace manager now written. It uses a heap structure and is therefore
276 expected to be efficient. It may however be speeded up (particularly
277 when many events occur at the same date) by using red and black
278 trees. One day maybe...
279 - Max-min linear system solver written. It uses a sparse matrix
280 structure taking advantage of its expected use. Most operations are
281 O(1) and free/calloc are called as few as possible. The computation of
282 the minimum could however be improved by using a red and black tree
285 2004-11-03 Arnaud Legrand
286 - Rename every gras_* function that was in xbt/ to its xbt_
288 - Add a heap and a doubly-linked list to xbt
289 - Added a dichotomy to the dictionaries. make check works as well before
290 so I assume that the patch is correct. I do not know however if things
291 run effectively faster than before now. :)
293 Inclusion of the SimGrid tree in the GRAS one. The archive is renamed to
294 SimGrid, and the version number is bumped to 2.x
296 2004-10-29 Martin Quinson
297 - Introduction of the remote errors.
298 They are the result of a RMI/RPC on the remote machine.
299 ErrCodes being scalar values, you can't get the host on which those
300 errors did happen. Extending the error mechanism as in Gnome is possible.
301 No idea yet whether it is a good idea.
303 2004-10-28 Martin Quinson
304 - Interface revolution: the Starred Structure Eradication.
305 I used to do typedef struct {} toto_t; and then handle *toto_t.
306 Arnaud (and Oli) didn't like it, and I surrendered. Now, you have:
307 - ???_t is a valid type (builded with typedef)
308 - s_toto_t is a structure (access to fields with .)
309 - s_toto is a structure needing 'struct' keyword to be used
310 - e_toto_t is an enum
311 - toto_t is an 'object' (struct*)
313 typedef struct s_toto {} s_toto_t, *toto_t;
314 typedef enum {} e_toto_t;
315 Moreover, only toto_t (and e_toto_t) are public. The rest (mainly
316 s_toto_t) is private.
318 - While I was at it, all gras_<obj>_free() functions want a gras_<obj>_t*
319 so that it can set the variable to NULL. It was so for dicts and sets,
320 it changed for dynars.
322 - Fix a bunch of memleaks in dict_remove
323 - Fix a bug in sg/server_socket opening: it failed all the time.
325 2004-10-07 Martin Quinson
326 - Speed up dynar lookup operation a bit.
328 gras_dynar_get is dead.
330 Now, you can choose between gras_dynar_get_cpy (the old gras_dynar_get
331 but should be avoided for efficiency reasons) and gras_dynar_get_ptr
332 (which gives you the address of the stored data).
334 gras_dynar_get_as is an helpful macro which allows you to retrieve a
335 copy of the data using an affectation to do the job and not a memcpy.
337 int toto = gras_dynar_get_as(dyn,0,int); rewrites itself to
338 int toto = *(int*)gras_dynar_get_ptr(dyn,0);
340 It does not really speedup the dynar test because they are
341 setting elements all the time (and look them seldom). But the dict does
342 far more lookup than setting.
344 So, this brings the dict_crash test from ~33s to ~25s (200000 elms).
346 2004-10-05 Martin Quinson
347 - Allow to (en/dis)able the cycle detection at run time.
349 Whether we should check for cycle or not is now a property of each
350 datatype. When you think there may be some cycle, use datadesc_cycle_set.
351 datadesc_cycle_unset allow to remove this property when previously set.
353 Note that the cycle detection is off by default since it impacts the
354 performance. Watch the data you feed GRAS with ;)
356 This property is hereditary. Any element embedded in a structure having it
357 set have it set for the time of this data exchange.
359 You should set it both on sender and receiver side. If you don't set it on
360 sender side, it will enter an endless loop. If you forget on receiver
361 side, the cycles won't be recreated after communication.
363 - Header reorganization.
364 Kill gras_private.h, each submodule must load the headers it needs.
366 2004-10-04 Martin Quinson
367 - Interface revolution: do not try to survive to malloc failure.
369 Now, gras_malloc and friends call gras_abort() on failure.
370 As a conclusion, malloc_error is not a valid error anymore, and all
371 functions for which it was the only gras_error_t return value are
372 changed. They now return void, or there result directly.
373 This simplify the API a lot.
375 2004-09-29 Martin Quinson
376 - Re-enable raw sockets.
377 Created by gras_socket_{client,server}_ext;
378 Used with gras_raw_{send,recv}
381 It should allow to kill the last bits of gras first version soon.
383 This is not completely satisfactory yet (duplicate code with
384 chunk_{send,recv}; a bit out of the plugin mechanism), but it should
387 - Simplify transport plugin (internal) interface by not passing any
388 argument to _server and _client, but embedding them in the socket
391 2004-09-28 Martin Quinson
392 - Finish the port to AIX.
393 autoconf was my problem (segfault within the malloc replacement
394 function. No idea why)
396 2004-09-16 Martin Quinson
397 - Fix some size_t madness on 64bit architectures.
399 2004-09-08 Martin Quinson
400 - Reduce the number of system headers loaded, overload some more system
401 calls (such as malloc to cast the result of the system one, and work
403 - Fix and reintroduce the config support
405 2004-09-07 Martin Quinson
406 - Source code reorganization to allow Arnaud to surf all over there.
407 - Allow to document the logging categories.
408 - Remove all uppercase from logging categories and useless cleanup in names.
410 2004-08-18 Martin Quinson
411 Version 0.6.2 (protocol not changed; API changed)
412 - Interface cleanup: gras_msgtype_by_name returns the type (instead of a
413 gras_error_t), and NULL when not found. Functions expecting a msgtype
414 as argument (msg_wait; msg_send) deal with NULL argument by providing a
415 hopefully usefull message.
416 - Portability to prehistoric sparcs again
418 2004-08-17 Martin Quinson
419 Version 0.6.1 (protocol not changed; ABI not changed)
420 - prealloc some buffers to speed things up
422 2004-08-11 Martin Quinson
423 Version 0.6 (protocol not changed; ABI expended)
424 - The parsing macro can deal with the references, provided that you add
425 the relevant annotations (using GRAS_ANNOTE(size,field_name))
427 2004-08-09 Martin Quinson
428 Version 0.5 (protocol not changed; ABI changed)
429 - Allow to off turn the cycle detection code in data exchange at
430 compilation time. It should be at run time, but I'm short of time (and
431 the config stuff is still broken). That way, we keep dict out of the
432 critical path, which is good because the performance is poor:
433 - search not dichotomial yet
434 - dynar give no way to access their content and memcpy everytime
435 - In composed data description (struct, ref and so on), stop foolness of
436 keeping the subtype's ID, but store the type itself. This keeps sets out
437 of the critical path, which is good since they rely on dynar and
438 dictionnaries. The only loose of that is that we cannot detect the
439 redeclaration of a structure/union with another content (but I'm not sure
440 the code detected well this error before anyway). We still can detect
441 the redefinition discrepancy for the other types.
442 - Use a whole bunch of optimisation flags (plus -fno-strict-aliasing since
443 it breaks the code because of type-punning used all over the place).
444 This breaks on all non-gcc architectures (for now).
446 All those changes (plus the buffer of last time) allow me to gain 2 order
447 of magnitude on cruel tests consisting of 800000 array of integers on two
448 level of a hierarchical structure (200 secondes -> 4 secondes)
451 - the selector of reference must now return the type it points to, not
454 2004-08-06 Martin Quinson
455 Version 0.4 (protocol changed; ABI not changed)
456 - Allow to pass --gras-log argument to processes in simulation mode. Really.
457 - New debugging level: trace (under debug) to see effect of GRAS_IN/OUT
458 - Implement a buffer transport, and use it by default (it relies on tcp in
459 real life and on sg in simulation).
460 That's a bit hackish since I had a new field to the structure to store
461 its data without interfering with the subtype ones. Inheritance
462 is tricky in C. And that's a kind of reverse inheritance with one class
463 derivating two classes. Or maybe a game with java interfaces. Anyway,
464 that's damn hard in C (at least).
465 Moreover, I got tired while trying to ensure plugin separation and
466 genericity in SG mode. MSG wants me to do weird things, so let's go for
467 cruel hacks (temporarily of course ;).
468 See comment in transport_private.h:71
469 - do not include all the _interface headers in private but in the files
470 which really need them (to cut the compilation time when they are
473 2004-07-26 Martin Quinson
474 Version 0.3 (protocol not changed; ABI changed)
475 - Major overhault of the datadesc interface to simplify it:
476 - shorted the function names:
477 s/gras_datadesc_declare_struct/gras_datadesc_struct/ and so on
478 - add a trivial way to push/pop integers into the cbps without malloc.
479 This allows to make really generic sub_type description, which simply
480 pop their size of the stack.
481 - add a function gras_datadesc_ref_pop_arr() which does what users want
482 most of the time: Declare a dynamic array (which pops its size of the
483 stack) and declare a reference to it. Poor name, but anyway.
484 - kill the post-send callback, add a post-receive one
486 2004-07-23 Martin Quinson
487 Version 0.2 (protocol changed; ABI changed)
488 - add some testing for cpbs in the test cases, and fix some more bugs.
489 This invalidate again the little64 data file, since I cannot regenerate
491 - remove an awfull optimization in the logging stuff, allowing me to:
492 - understand it again
493 - learn gcc how to check that the argument match the provided format
494 - fix all errors revealed by gcc after that
495 - internal keys of dict are not \0 terminated. Deal with it properly in
496 loggings instead of segfaulting when the user want to see the logs :-/
498 2004-07-22 Martin Quinson
499 - Fix some stupid bug preventing cbps (callback postit) from working
501 2004-07-21 Martin Quinson
502 - Some documentation cleanups
503 - remove the useless last argument of msgtype_declare
504 - rename the Virtu functions to fit into the 'os' namespace
505 - move headers src/include -> src/include/gras/ and stop fooling with
506 gras -> . symbolic link
507 - make distcheck is now successful
509 2004-07-19 Martin Quinson
511 - Build shared library also
512 - Install html doc to the right location
513 - stop removing maintainer files in make clean
514 - build tests only on make check
516 2004-07-13 Martin Quinson
518 - No major issue in previous version => change versionning schema
519 - Re-enable little64 convertion test now that Abdou kindly regenerated the
520 corresponding dataset.
522 2004-07-11 Martin Quinson
524 - Get it working with any kind of structure (we can compute the padding
525 bytes remotely for all the architectures I have access to)
526 - Implement the structure parsing macro (still not quite robust/complete)
527 - Improvement to the remote testing toysuite
529 2004-07-10 Martin Quinson
531 - get ride of a bunch of deprecated macros
532 - actually run the test for two-compliment, not only compile it :-/
533 - test whether the structures get packed (and bail out if yes. Damn.
534 Alignment is a serious matter)
535 - test whether the structures get compacted (but respecting the alignment
536 constraints of each types)
537 - test whether the array fields of structures can straddle alignment boundaries
539 - Damnit, double are bigger than float (typo in creation of 'double' datadesc)
540 (took me 2 hours to find that bug, looking at the wrong place)
541 - Add gras_datadesc_declare_{union,struct}_close(). They must be used
542 before sending/receiving and are used to compute the offsets of fields
543 - Given that padding size depend even on compiler options, keep track of
544 alignment and aligned_size only for the current architecture. Not a big
545 deal since we send structure fields one after the other (seems
547 - Add the datastructure used for IEEE paper by the PBIO guys to the test
548 program, let it work on linux/gcc/little32. portability todo.
550 2004-07-08 Martin Quinson
551 - import and improve remote compilation support from FAST
552 - make sure make check works on half a dozen of machines out there
554 2004-07-07 Martin Quinson
555 Let's say it's version 0.0.3 ;)
556 - Implement conversions (yuhu!)
557 - Let it work on solaris (beside conversion, of course)
558 - Stupid me, using rand() to generate the conversion datatests in not wise.
560 2004-07-06 Martin Quinson
561 - Let make dist work, since I'm gonna need it to compile on remote hosts
562 - Let Tests/datadesc_usage write the architecture on which the file was
563 generated as first byte.
564 - Add PowerPC (being also IRIX64), SPARC (also power4) and ALPHA
565 architecture descriptions.
566 - Add datadesc_usage.{i386,ppc,sparc} files being the result of execution
567 on those architectures.
568 - Optimization: send/recv array of scalar in one shoot
570 2004-07-05 Martin Quinson
571 - YEAH! GRAS/SG and GRAS/RL are both able to run the ping example !
573 - Plug a whole bunch of memleaks
574 - each process now have to call gras_{init,exit}. One day, their log
575 settings will be separated
576 - Continue the code factorisation between SG, RL and common in Transport.
578 2004-07-04 Martin Quinson
580 - Redistribution between SG and RL.
581 We wanna have to accept in SG, so move accepted related parts of RL in
582 the common part. (more precisely, the dynar of all known sockets is no
583 more a static in transport.c, but part of the process_data)
585 [gras_stub_generator]
586 - Bug fix: Do call gras_process_init from gras_init (wasnt called in RL).
588 2004-07-03 Martin Quinson
589 - Create a new log channel tbx containing dict, set, log, dynar (to shut
590 them all up in one shot)
592 - Fix the ugly case of reference to dynamic array.
593 - New (semi-public) function gras_datadesc_size to allow the messaging
594 layer to malloc the needed space for the buffer.
596 - gras_socket_close now expect the socket to close (and not its address to
597 put NULL in it after it). This is because the socket passed to handlers
598 is one of their argument (=> not writable).
600 - propagate the interface cleanup from last week in datadesc, ie remove a
601 superfluous level of indirection. User pass adress of variable
602 containing data (both when sending and receiving), and not of a variable
603 being a pointer to the data. Let's say that I like it better ;)
604 The price for that is constructs like "int msg=*(int*)payload" in
605 handlers, but it's a fine price, IMHO.
607 - Let it work in RL (yuhu)
609 2004-06-21 Martin Quinson
611 - porting SG plugin and SG select to new standards (works almost).
612 - plug memleaks and fix bugs around.
615 - cleanup the prototype of data recv and force users to specify when they
616 want to handle references to objects. Test case working even for cycles.
617 - plug memleaks. Valgrind is perfectly ok with this.
619 2004-06-12 Martin Quinson
621 - cleanup the separation between plugin and main code in plugin creation
623 2004-06-11 Martin Quinson
625 - Reput hook for raw sockets, needed for BW experiments
626 - kill a few lines of dead code
627 [Data description] Interface cleanup
628 - gras_datadesc_by_name returns the searched type or NULL.
629 That way, no variable is needed to use a type desc once, which makes
631 - gras_datadesc_declare_[struct|union]_append_name is removed. The last
632 two parameters were strings (field name, type name), leading to
634 [Dicos] Interface cleanup
635 - gras_dico_retrieve -> gras_dico_get ; gras_dico_insert -> gras_dico_set
636 This is consistant with the dynar API.
638 2004-04-21 Martin Quinson
640 - Porting to new standards.
643 There is no bag anymore, no need to take extra provision to mask the
644 pointers behind "ID".
645 Better splitup of functions between files create/exchange/convert.
646 This is still a bit artificial since convert and receive are so
647 interleaved, but anyway.
649 - add a queued message list to procdata (the ones not matching criteria
651 - factorize some more code between SG and RL wrt procdata
653 - use gras_exit in example to track memleaks
654 - get rid of gs_example now that GS is properly integrated into gras
655 - update run_test to integrate the lastest tests (datadesc)
657 - rename WARNINGn macros to WARNn since it prooved error-prone
659 2004-04-19 Martin Quinson
661 - register init/exit functions within gras module mechanism
662 - send/receive function.
663 Convertion is not implemented, but short-cutted if not needed.
664 struct/array elements are sent one by one (instead of block-wise), but
665 nobody really cares (yet). Get a prototype before optimizing.
666 - tests (using a file socket) for DD send/receive on:
667 - base types: int, float
668 - array: fixed size, string (ie ref to dynamic string)
669 - structure: homogeneous, heterogeneous
670 - chained list, graph with cycle
671 Believe it or not, valgrind is not too unhappy with the results. The
672 cycle happily segfaults, but the others are ok. And I'm sick of pointers
676 - Bugfix when using a filename explicitely (instead of '-')
678 2004-04-09 Martin Quinson
680 - factorize more code between RL and SG in socket creation
681 - Complete the implementation and tests of:
683 o file (only in RL, and mainly for debugging)
685 I lost 3 days to design a portable address resolver, and then decided
686 that the prototype mainly have to run on my box.
687 Addressing portability too early may be like optimizing too early :-/
689 - use gras_init in the Tests instead of the crappy parse_log_opt
690 (the latter function is removed)
691 [Conditional execution]
692 - New functions: gras_if_RL/gras_if_SG (basic support for this)
693 [Code reorganisation]
694 - Get rid of libgrasutils.a since it makes more trouble than it solves.
695 Build examples against the RL library, since there is no way to disable
696 its creation for now.