1 SimGrid (2.95) unstable; urgency=low
4 * Implement MSG_get_msgload [MQ]
7 * Bugfix: Allow absolute paths for platform description files [MQ]
10 * gras_datadesc_dynar() builds a dynar type descriptor, allowing to send
11 dynar over the network (yeah) [MQ]
12 * Real (even if simplistic) implementation of gras_os_myname() on RL [MQ]
15 * Differentiate the types "char[22]" and "unsigned char[22]" in automatic
16 type parsing. "short" and "long" modifiers were also ignored; other
17 modifier (such as reference level) are still ignored. [MQ]
18 * Embeed the buffer size within the buffer itself on SG. [MQ]
19 That way, send() are atomic and cannot get intermixed anymore (at least
20 the ones which are less than 100k; bigger messages still have the issue)
21 * Array size pushed by the field, not by the field type (or each
22 and every long int will push stuff to the cbps) [MQ]
25 * <project>.Makefile.local (generated from gras_stub_generator) |MQ]:
28 * Callbacks now receive the gras_datadesc_type_t they work on as argument.
29 * type category 'ignored' killed as it was never used and were difficult
31 * whether a type can cycle or not is now a flag, leaving room for more
32 flags (as "ignored", if we feel the need one day ;)
33 * Rename raw sockets to measurement sockets since "raw" has another
34 meaning in networking community.
38 SimGrid (2.94) unstable; urgency=low
40 The first release candidate for SimGrid 3 !
44 * Update the main page and the FAQ. Adding references to gforge.
47 * Add a gras_os_getpid function.
50 * Add MSG_task_get_compute_duration() and MSG_task_get_data_size()
51 * Extend the logs so that they also print PID, hostname, date, ... if
53 * Convert the MSG example to the use of xbt_logs instead of PRINT_MESSAGE,
54 and kill the old version which were in testsuite/
55 * Rewrite tools/MSG_visualization/colorize.pl for using with logs instead
59 * Add xbt_os_time(). As the rest of xbt/portability, this is not public
60 for users. Instead, each programming environment (GRAS, MSG,...) use it
61 when needed to provide such a feature to users.
62 Don't shortcut the mecanism or you will also shortcut the virtualization
63 you need on the simulator.
67 * Cleanups in configury with regard to compile optimization/warning flags.
68 Also add -fno-loop-optimize to any powerpc since it's the optimization
69 killing gcc (< 3.4.0).
70 * Doxygen cleanups: move MSG examples, kill the second Doxygen phase
71 needed by MSG examples complications
72 * Borrow configury beautifications from PHP
75 * Bugfix: XBT_LOG_NEW_DEFAULT_CATEGORY now compiles without compiler
76 warning (thanks loris for stumbling into this one).
77 * Bugfix: stop loading private headers (gras_config.h) from the public
81 * Change SIMGRID_INSTALL_PATH to GRAS_ROOT in Makefiles generated for user.
82 * Rename gras_get_my_fqdn to gras_os_myname and implement it in the simulator
83 RL would imply a DNS resolver, which is *hard* to do in a portable way
84 (and therefore delayed).
85 * Implement a real timer mecanism and use it in timing macros. This allows
86 to avoid rounding errors and get a 0.000005 sec precision in timing
87 macros. While I was at it, various cleanups:
88 - allow to declare more than one timed section per file (fix a stupid bug)
89 - move some private declaration to the right place
90 - merge conditional execution and timing macros into emulation module
92 - make sure the module cleanups its mess on gras_exit
93 * Documentation improvements:
94 - (new) how to compile applications using GRAS
95 - (new) emulation support (timing macros)
97 -- Da SimGrid team <simgrid-devel@lists.gforge.inria.fr> Fri, 13 May 2005 10:49:31 +0200
99 SimGrid (2.93) unstable; urgency=low
101 Alpha 4 on the path to SimGrid 3 (aka the "neuf-trois" version)
104 - Use Paje properly where used. Still to be sanitized properly.
105 - Portability fix: Add an implementation of the contexts using pthread
109 - Add xbt_procname(): returns the name of the current process.
110 Use it to show the current process's name in all logging.
112 - fix detection of older flex version and the reaction, since we do
113 depend on modern ones (we use lex_destroy)
114 - Better separation of SG and RL in the libs: remove all simulation code
115 from libgras. As a result, this lib is now only 200k when stripped.
116 Some of the xbt modules may also be duplicated (two sets and such) and
117 should be cleaned/killed before SG3.
118 - Insist on using xlC on AIX because of weird problems involving gcc there.
119 - Cleanup the make remote stuff. This is now done by scripts
120 tools/graspe-{master,slave} (GRAS Platform Expender). This is still
121 mainly for our private use, but we're working on changing them to user
124 - Bugfix: flush the socket on close only if there is some *output*.
125 - Bugfix: flush idempotent when there's nothing to send (don't send size=0)
127 - Add MSG_task_get_name. The task names are mainly for debugging purpose,
130 -- SimGrid team <simgrid2-users@listes.ens-lyon.fr> Fri, 4 Mar 2005 14:32:37 -0800
132 SimGrid (2.92) unstable; urgency=low
134 Alpha 3 on the path to SimGrid 3
138 - New! First try of benchmarking macros.
139 - New! First try so that gras_stub_generator generate deployment and
140 remote compilation helpers.
142 - Bugfix: Initialization fix in msg_test.
146 - Bugfix: applied patch to lexer so that it doesn't need a huge heap.
148 - Bugfix: let dicts work with NULL content (_foreach didn't) and cleanups
150 - API Change: gras_os_sleep to take the amount of seconds as a double.
151 Accepting an int was error prone since it was the only location where
152 seconds were coded as such. It leaded to damn rounding errors.
153 - Bugfix: Hard to belive that timers ever worked before this.
155 -- SimGrid team <simgrid2-users@listes.ens-lyon.fr> Wed, 23 Feb 2005 22:09:21 +0100
157 SimGrid (2.91) unstable; urgency=low
159 Alpha 2 on the path to SimGrid 3
163 - Bug fix in the lmm_solver.
165 - New! Interface to Paje (see http://www-id.imag.fr/Logiciels/paje/)
166 through the function MSG_paje_output.
167 - New! Introducing two new functions MSG_process_kill() and MSG_process_killall().
168 - It is possible to bound the rate of a communication in MSG with
169 MSG_task_put_bounded() (was already in the previous version but I had forgotten
170 to write it in the changelog).
171 - Bug fix to let GRAS run on top of MSG until we move it directly on top
176 - Various cleanups to the autotools stuff
177 - Begin to move Gras examples to examples/gras/
178 - Let make distcheck work again (yeah!)
180 - documentation overhauled using doxygen.
181 gtk-doc-tools is dead in SimGrid now.
182 - Automatically extract all existing logging categories, and add the list
183 to the documentation (long standing one, to say the less)
185 - Cleanup the known architecture table. Reorder the entries to group what
186 should be, and use a more consistent naming scheme.
187 (some of the test dataset are still to be regenerated)
188 - New! Allow library to register globals on each process just as userdata
190 This is implemented using a xbt_dict and not a xbt_set, so we loose the
191 lookup time (for now).
192 Use it in msg and trp.
193 This cleans a lot the internals and helps enforcing privacy of the
194 headers between the gras components.
195 - New! Add a timer mechanism, not unlike cron(8) and at(1).
196 - Bugfix: gras_os_time was delirious in RL.
197 - Bugfix: gras_trp_select/RL don't run into the wall when asked to select
199 - Reenable GRAS now that it works.
201 -- Arnaud Legrand <Arnaud.Legrand@imag.fr> Mon, 14 Feb 2005 14:02:13 -0800
203 SimGrid (2.90) unstable; urgency=low
205 Alpha 1 on the path to SimGrid 3
207 * It is a long time since the last release of SimGrid. I'm sorry about
208 that but as I had told you, I was rewriting a lot of things. I apologize
209 to those who had been reporting bugs to me and that I had not answered.
210 If your bug is still in the new version, please tell me. Here is a
211 summary of the main changes.
213 * REVOLUTION 1: The SimGrid project has merged with the GRAS project
214 lead by Martin Quinson. As a consequence SimGrid gains a lot in
215 portability, speed, and a lot more but you'll figure it out later.
216 SimGrid now comprises 3 different projects : MSG, GRAS and SMPI.
217 I wanted to release the new MSG as soon as possible and I have
218 broken GRAS, which is the reason why, for now, only MSG is fully
219 functional. A laconic description of these projects is available
220 in the documentation.
222 * REVOLUTION 2: I have removed SG and I am now using a new simulation
223 kernel optimized for our needs (called SURF but only the developers
224 should use it). Hence, MSG is now roughly 30 times faster and I think
225 that by rewriting a little bit MSG, I could event speed it up a little
226 bit more. Beside the gain in speed, it is also much easier to encode a
227 new platform model with SURF than it was with SG. More to come...
229 * REVOLUTION 3: I have tried to change a little as possible the API of
230 MSG but a few things really had to disappear. The main differences
231 with the previous version are :
232 1) no more m_links_t and the corresponding functions. Platforms are
233 directly read from a XML description and cannot be hard-coded
234 anymore. The same format is used for application deployment
235 description. The new format is described in the documentation.
236 Have a look in tools/platform_generation. There is a tiny script
237 that converts from the old platform format to the new one. Concerning
238 the application deployment format, parsing the old one is tricky.
239 I think most of you should however be able to convert your files. If
240 it is really an issue, I can write a C code that does the conversion.
242 2) the toolbox tbx does not exist anymore. We now have a library
243 with much more data-structures but without the hash-tables (we have
244 dictionaries that are much faster).
246 -- Arnaud Legrand <Arnaud.Legrand@imag.fr> Mon, 31 Jan 2005 10:45:53 -0800
248 *****************************************************************************
249 * Follows the old GRAS changelog. It does not follow the same syntax, but I *
250 * don't feel like converting the oldies. (Mt) *
251 *****************************************************************************
254 Version 2.90: "the long awaited one"
255 - Finished rewriting and debugging MSG. Rewrote the documentation.
256 - disable GRAS for now since it needs to be ported to the newest SG
259 - Finish the port to windows (using mingw32 for cross-compile)
262 - Main loop and datastructures of SURF. A cpu resource object is
263 functional. Surf can thus be used to create cpu's with variable
264 performance on which you can execute some actions.
266 2004-11-15 Martin Quinson
267 - Port to ARM. Simply added the alignment and size descriptions. Should
268 work, but the ARM machines are so slow that I didn't had the opportunity
269 to 'make check' over there yet.
271 2004-11-15 Arnaud Legrand
272 - Trace manager now written. It uses a heap structure and is therefore
273 expected to be efficient. It may however be speeded up (particularly
274 when many events occur at the same date) by using red and black
275 trees. One day maybe...
276 - Max-min linear system solver written. It uses a sparse matrix
277 structure taking advantage of its expected use. Most operations are
278 O(1) and free/calloc are called as few as possible. The computation of
279 the minimum could however be improved by using a red and black tree
282 2004-11-03 Arnaud Legrand
283 - Rename every gras_* function that was in xbt/ to its xbt_
285 - Add a heap and a doubly-linked list to xbt
286 - Added a dichotomy to the dictionaries. make check works as well before
287 so I assume that the patch is correct. I do not know however if things
288 run effectively faster than before now. :)
290 Inclusion of the SimGrid tree in the GRAS one. The archive is renamed to
291 SimGrid, and the version number is bumped to 2.x
293 2004-10-29 Martin Quinson
294 - Introduction of the remote errors.
295 They are the result of a RMI/RPC on the remote machine.
296 ErrCodes being scalar values, you can't get the host on which those
297 errors did happen. Extending the error mechanism as in Gnome is possible.
298 No idea yet whether it is a good idea.
300 2004-10-28 Martin Quinson
301 - Interface revolution: the Starred Structure Eradication.
302 I used to do typedef struct {} toto_t; and then handle *toto_t.
303 Arnaud (and Oli) didn't like it, and I surrendered. Now, you have:
304 - ???_t is a valid type (builded with typedef)
305 - s_toto_t is a structure (access to fields with .)
306 - s_toto is a structure needing 'struct' keyword to be used
307 - e_toto_t is an enum
308 - toto_t is an 'object' (struct*)
310 typedef struct s_toto {} s_toto_t, *toto_t;
311 typedef enum {} e_toto_t;
312 Moreover, only toto_t (and e_toto_t) are public. The rest (mainly
313 s_toto_t) is private.
315 - While I was at it, all gras_<obj>_free() functions want a gras_<obj>_t*
316 so that it can set the variable to NULL. It was so for dicts and sets,
317 it changed for dynars.
319 - Fix a bunch of memleaks in dict_remove
320 - Fix a bug in sg/server_socket opening: it failed all the time.
322 2004-10-07 Martin Quinson
323 - Speed up dynar lookup operation a bit.
325 gras_dynar_get is dead.
327 Now, you can choose between gras_dynar_get_cpy (the old gras_dynar_get
328 but should be avoided for efficiency reasons) and gras_dynar_get_ptr
329 (which gives you the address of the stored data).
331 gras_dynar_get_as is an helpful macro which allows you to retrieve a
332 copy of the data using an affectation to do the job and not a memcpy.
334 int toto = gras_dynar_get_as(dyn,0,int); rewrites itself to
335 int toto = *(int*)gras_dynar_get_ptr(dyn,0);
337 It does not really speedup the dynar test because they are
338 setting elements all the time (and look them seldom). But the dict does
339 far more lookup than setting.
341 So, this brings the dict_crash test from ~33s to ~25s (200000 elms).
343 2004-10-05 Martin Quinson
344 - Allow to (en/dis)able the cycle detection at run time.
346 Whether we should check for cycle or not is now a property of each
347 datatype. When you think there may be some cycle, use datadesc_cycle_set.
348 datadesc_cycle_unset allow to remove this property when previously set.
350 Note that the cycle detection is off by default since it impacts the
351 performance. Watch the data you feed GRAS with ;)
353 This property is hereditary. Any element embedded in a structure having it
354 set have it set for the time of this data exchange.
356 You should set it both on sender and receiver side. If you don't set it on
357 sender side, it will enter an endless loop. If you forget on receiver
358 side, the cycles won't be recreated after communication.
360 - Header reorganization.
361 Kill gras_private.h, each submodule must load the headers it needs.
363 2004-10-04 Martin Quinson
364 - Interface revolution: do not try to survive to malloc failure.
366 Now, gras_malloc and friends call gras_abort() on failure.
367 As a conclusion, malloc_error is not a valid error anymore, and all
368 functions for which it was the only gras_error_t return value are
369 changed. They now return void, or there result directly.
370 This simplify the API a lot.
372 2004-09-29 Martin Quinson
373 - Re-enable raw sockets.
374 Created by gras_socket_{client,server}_ext;
375 Used with gras_raw_{send,recv}
378 It should allow to kill the last bits of gras first version soon.
380 This is not completely satisfactory yet (duplicate code with
381 chunk_{send,recv}; a bit out of the plugin mechanism), but it should
384 - Simplify transport plugin (internal) interface by not passing any
385 argument to _server and _client, but embedding them in the socket
388 2004-09-28 Martin Quinson
389 - Finish the port to AIX.
390 autoconf was my problem (segfault within the malloc replacement
391 function. No idea why)
393 2004-09-16 Martin Quinson
394 - Fix some size_t madness on 64bit architectures.
396 2004-09-08 Martin Quinson
397 - Reduce the number of system headers loaded, overload some more system
398 calls (such as malloc to cast the result of the system one, and work
400 - Fix and reintroduce the config support
402 2004-09-07 Martin Quinson
403 - Source code reorganization to allow Arnaud to surf all over there.
404 - Allow to document the logging categories.
405 - Remove all uppercase from logging categories and useless cleanup in names.
407 2004-08-18 Martin Quinson
408 Version 0.6.2 (protocol not changed; API changed)
409 - Interface cleanup: gras_msgtype_by_name returns the type (instead of a
410 gras_error_t), and NULL when not found. Functions expecting a msgtype
411 as argument (msg_wait; msg_send) deal with NULL argument by providing a
412 hopefully usefull message.
413 - Portability to prehistoric sparcs again
415 2004-08-17 Martin Quinson
416 Version 0.6.1 (protocol not changed; ABI not changed)
417 - prealloc some buffers to speed things up
419 2004-08-11 Martin Quinson
420 Version 0.6 (protocol not changed; ABI expended)
421 - The parsing macro can deal with the references, provided that you add
422 the relevant annotations (using GRAS_ANNOTE(size,field_name))
424 2004-08-09 Martin Quinson
425 Version 0.5 (protocol not changed; ABI changed)
426 - Allow to off turn the cycle detection code in data exchange at
427 compilation time. It should be at run time, but I'm short of time (and
428 the config stuff is still broken). That way, we keep dict out of the
429 critical path, which is good because the performance is poor:
430 - search not dichotomial yet
431 - dynar give no way to access their content and memcpy everytime
432 - In composed data description (struct, ref and so on), stop foolness of
433 keeping the subtype's ID, but store the type itself. This keeps sets out
434 of the critical path, which is good since they rely on dynar and
435 dictionnaries. The only loose of that is that we cannot detect the
436 redeclaration of a structure/union with another content (but I'm not sure
437 the code detected well this error before anyway). We still can detect
438 the redefinition discrepancy for the other types.
439 - Use a whole bunch of optimisation flags (plus -fno-strict-aliasing since
440 it breaks the code because of type-punning used all over the place).
441 This breaks on all non-gcc architectures (for now).
443 All those changes (plus the buffer of last time) allow me to gain 2 order
444 of magnitude on cruel tests consisting of 800000 array of integers on two
445 level of a hierarchical structure (200 secondes -> 4 secondes)
448 - the selector of reference must now return the type it points to, not
451 2004-08-06 Martin Quinson
452 Version 0.4 (protocol changed; ABI not changed)
453 - Allow to pass --gras-log argument to processes in simulation mode. Really.
454 - New debugging level: trace (under debug) to see effect of GRAS_IN/OUT
455 - Implement a buffer transport, and use it by default (it relies on tcp in
456 real life and on sg in simulation).
457 That's a bit hackish since I had a new field to the structure to store
458 its data without interfering with the subtype ones. Inheritance
459 is tricky in C. And that's a kind of reverse inheritance with one class
460 derivating two classes. Or maybe a game with java interfaces. Anyway,
461 that's damn hard in C (at least).
462 Moreover, I got tired while trying to ensure plugin separation and
463 genericity in SG mode. MSG wants me to do weird things, so let's go for
464 cruel hacks (temporarily of course ;).
465 See comment in transport_private.h:71
466 - do not include all the _interface headers in private but in the files
467 which really need them (to cut the compilation time when they are
470 2004-07-26 Martin Quinson
471 Version 0.3 (protocol not changed; ABI changed)
472 - Major overhault of the datadesc interface to simplify it:
473 - shorted the function names:
474 s/gras_datadesc_declare_struct/gras_datadesc_struct/ and so on
475 - add a trivial way to push/pop integers into the cbps without malloc.
476 This allows to make really generic sub_type description, which simply
477 pop their size of the stack.
478 - add a function gras_datadesc_ref_pop_arr() which does what users want
479 most of the time: Declare a dynamic array (which pops its size of the
480 stack) and declare a reference to it. Poor name, but anyway.
481 - kill the post-send callback, add a post-receive one
483 2004-07-23 Martin Quinson
484 Version 0.2 (protocol changed; ABI changed)
485 - add some testing for cpbs in the test cases, and fix some more bugs.
486 This invalidate again the little64 data file, since I cannot regenerate
488 - remove an awfull optimization in the logging stuff, allowing me to:
489 - understand it again
490 - learn gcc how to check that the argument match the provided format
491 - fix all errors revealed by gcc after that
492 - internal keys of dict are not \0 terminated. Deal with it properly in
493 loggings instead of segfaulting when the user want to see the logs :-/
495 2004-07-22 Martin Quinson
496 - Fix some stupid bug preventing cbps (callback postit) from working
498 2004-07-21 Martin Quinson
499 - Some documentation cleanups
500 - remove the useless last argument of msgtype_declare
501 - rename the Virtu functions to fit into the 'os' namespace
502 - move headers src/include -> src/include/gras/ and stop fooling with
503 gras -> . symbolic link
504 - make distcheck is now successful
506 2004-07-19 Martin Quinson
508 - Build shared library also
509 - Install html doc to the right location
510 - stop removing maintainer files in make clean
511 - build tests only on make check
513 2004-07-13 Martin Quinson
515 - No major issue in previous version => change versionning schema
516 - Re-enable little64 convertion test now that Abdou kindly regenerated the
517 corresponding dataset.
519 2004-07-11 Martin Quinson
521 - Get it working with any kind of structure (we can compute the padding
522 bytes remotely for all the architectures I have access to)
523 - Implement the structure parsing macro (still not quite robust/complete)
524 - Improvement to the remote testing toysuite
526 2004-07-10 Martin Quinson
528 - get ride of a bunch of deprecated macros
529 - actually run the test for two-compliment, not only compile it :-/
530 - test whether the structures get packed (and bail out if yes. Damn.
531 Alignment is a serious matter)
532 - test whether the structures get compacted (but respecting the alignment
533 constraints of each types)
534 - test whether the array fields of structures can straddle alignment boundaries
536 - Damnit, double are bigger than float (typo in creation of 'double' datadesc)
537 (took me 2 hours to find that bug, looking at the wrong place)
538 - Add gras_datadesc_declare_{union,struct}_close(). They must be used
539 before sending/receiving and are used to compute the offsets of fields
540 - Given that padding size depend even on compiler options, keep track of
541 alignment and aligned_size only for the current architecture. Not a big
542 deal since we send structure fields one after the other (seems
544 - Add the datastructure used for IEEE paper by the PBIO guys to the test
545 program, let it work on linux/gcc/little32. portability todo.
547 2004-07-08 Martin Quinson
548 - import and improve remote compilation support from FAST
549 - make sure make check works on half a dozen of machines out there
551 2004-07-07 Martin Quinson
552 Let's say it's version 0.0.3 ;)
553 - Implement conversions (yuhu!)
554 - Let it work on solaris (beside conversion, of course)
555 - Stupid me, using rand() to generate the conversion datatests in not wise.
557 2004-07-06 Martin Quinson
558 - Let make dist work, since I'm gonna need it to compile on remote hosts
559 - Let Tests/datadesc_usage write the architecture on which the file was
560 generated as first byte.
561 - Add PowerPC (being also IRIX64), SPARC (also power4) and ALPHA
562 architecture descriptions.
563 - Add datadesc_usage.{i386,ppc,sparc} files being the result of execution
564 on those architectures.
565 - Optimization: send/recv array of scalar in one shoot
567 2004-07-05 Martin Quinson
568 - YEAH! GRAS/SG and GRAS/RL are both able to run the ping example !
570 - Plug a whole bunch of memleaks
571 - each process now have to call gras_{init,exit}. One day, their log
572 settings will be separated
573 - Continue the code factorisation between SG, RL and common in Transport.
575 2004-07-04 Martin Quinson
577 - Redistribution between SG and RL.
578 We wanna have to accept in SG, so move accepted related parts of RL in
579 the common part. (more precisely, the dynar of all known sockets is no
580 more a static in transport.c, but part of the process_data)
582 [gras_stub_generator]
583 - Bug fix: Do call gras_process_init from gras_init (wasnt called in RL).
585 2004-07-03 Martin Quinson
586 - Create a new log channel tbx containing dict, set, log, dynar (to shut
587 them all up in one shot)
589 - Fix the ugly case of reference to dynamic array.
590 - New (semi-public) function gras_datadesc_size to allow the messaging
591 layer to malloc the needed space for the buffer.
593 - gras_socket_close now expect the socket to close (and not its address to
594 put NULL in it after it). This is because the socket passed to handlers
595 is one of their argument (=> not writable).
597 - propagate the interface cleanup from last week in datadesc, ie remove a
598 superfluous level of indirection. User pass adress of variable
599 containing data (both when sending and receiving), and not of a variable
600 being a pointer to the data. Let's say that I like it better ;)
601 The price for that is constructs like "int msg=*(int*)payload" in
602 handlers, but it's a fine price, IMHO.
604 - Let it work in RL (yuhu)
606 2004-06-21 Martin Quinson
608 - porting SG plugin and SG select to new standards (works almost).
609 - plug memleaks and fix bugs around.
612 - cleanup the prototype of data recv and force users to specify when they
613 want to handle references to objects. Test case working even for cycles.
614 - plug memleaks. Valgrind is perfectly ok with this.
616 2004-06-12 Martin Quinson
618 - cleanup the separation between plugin and main code in plugin creation
620 2004-06-11 Martin Quinson
622 - Reput hook for raw sockets, needed for BW experiments
623 - kill a few lines of dead code
624 [Data description] Interface cleanup
625 - gras_datadesc_by_name returns the searched type or NULL.
626 That way, no variable is needed to use a type desc once, which makes
628 - gras_datadesc_declare_[struct|union]_append_name is removed. The last
629 two parameters were strings (field name, type name), leading to
631 [Dicos] Interface cleanup
632 - gras_dico_retrieve -> gras_dico_get ; gras_dico_insert -> gras_dico_set
633 This is consistant with the dynar API.
635 2004-04-21 Martin Quinson
637 - Porting to new standards.
640 There is no bag anymore, no need to take extra provision to mask the
641 pointers behind "ID".
642 Better splitup of functions between files create/exchange/convert.
643 This is still a bit artificial since convert and receive are so
644 interleaved, but anyway.
646 - add a queued message list to procdata (the ones not matching criteria
648 - factorize some more code between SG and RL wrt procdata
650 - use gras_exit in example to track memleaks
651 - get rid of gs_example now that GS is properly integrated into gras
652 - update run_test to integrate the lastest tests (datadesc)
654 - rename WARNINGn macros to WARNn since it prooved error-prone
656 2004-04-19 Martin Quinson
658 - register init/exit functions within gras module mechanism
659 - send/receive function.
660 Convertion is not implemented, but short-cutted if not needed.
661 struct/array elements are sent one by one (instead of block-wise), but
662 nobody really cares (yet). Get a prototype before optimizing.
663 - tests (using a file socket) for DD send/receive on:
664 - base types: int, float
665 - array: fixed size, string (ie ref to dynamic string)
666 - structure: homogeneous, heterogeneous
667 - chained list, graph with cycle
668 Believe it or not, valgrind is not too unhappy with the results. The
669 cycle happily segfaults, but the others are ok. And I'm sick of pointers
673 - Bugfix when using a filename explicitely (instead of '-')
675 2004-04-09 Martin Quinson
677 - factorize more code between RL and SG in socket creation
678 - Complete the implementation and tests of:
680 o file (only in RL, and mainly for debugging)
682 I lost 3 days to design a portable address resolver, and then decided
683 that the prototype mainly have to run on my box.
684 Addressing portability too early may be like optimizing too early :-/
686 - use gras_init in the Tests instead of the crappy parse_log_opt
687 (the latter function is removed)
688 [Conditional execution]
689 - New functions: gras_if_RL/gras_if_SG (basic support for this)
690 [Code reorganisation]
691 - Get rid of libgrasutils.a since it makes more trouble than it solves.
692 Build examples against the RL library, since there is no way to disable
693 its creation for now.