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 ;)
36 SimGrid (2.94) unstable; urgency=low
38 The first release candidate for SimGrid 3 !
42 * Update the main page and the FAQ. Adding references to gforge.
45 * Add a gras_os_getpid function.
48 * Add MSG_task_get_compute_duration() and MSG_task_get_data_size()
49 * Extend the logs so that they also print PID, hostname, date, ... if
51 * Convert the MSG example to the use of xbt_logs instead of PRINT_MESSAGE,
52 and kill the old version which were in testsuite/
53 * Rewrite tools/MSG_visualization/colorize.pl for using with logs instead
57 * Add xbt_os_time(). As the rest of xbt/portability, this is not public
58 for users. Instead, each programming environment (GRAS, MSG,...) use it
59 when needed to provide such a feature to users.
60 Don't shortcut the mecanism or you will also shortcut the virtualization
61 you need on the simulator.
65 * Cleanups in configury with regard to compile optimization/warning flags.
66 Also add -fno-loop-optimize to any powerpc since it's the optimization
67 killing gcc (< 3.4.0).
68 * Doxygen cleanups: move MSG examples, kill the second Doxygen phase
69 needed by MSG examples complications
70 * Borrow configury beautifications from PHP
73 * Bugfix: XBT_LOG_NEW_DEFAULT_CATEGORY now compiles without compiler
74 warning (thanks loris for stumbling into this one).
75 * Bugfix: stop loading private headers (gras_config.h) from the public
79 * Change SIMGRID_INSTALL_PATH to GRAS_ROOT in Makefiles generated for user.
80 * Rename gras_get_my_fqdn to gras_os_myname and implement it in the simulator
81 RL would imply a DNS resolver, which is *hard* to do in a portable way
82 (and therefore delayed).
83 * Implement a real timer mecanism and use it in timing macros. This allows
84 to avoid rounding errors and get a 0.000005 sec precision in timing
85 macros. While I was at it, various cleanups:
86 - allow to declare more than one timed section per file (fix a stupid bug)
87 - move some private declaration to the right place
88 - merge conditional execution and timing macros into emulation module
90 - make sure the module cleanups its mess on gras_exit
91 * Documentation improvements:
92 - (new) how to compile applications using GRAS
93 - (new) emulation support (timing macros)
95 -- Da SimGrid team <simgrid-devel@lists.gforge.inria.fr> Fri, 13 May 2005 10:49:31 +0200
97 SimGrid (2.93) unstable; urgency=low
99 Alpha 4 on the path to SimGrid 3 (aka the "neuf-trois" version)
102 - Use Paje properly where used. Still to be sanitized properly.
103 - Portability fix: Add an implementation of the contexts using pthread
107 - Add xbt_procname(): returns the name of the current process.
108 Use it to show the current process's name in all logging.
110 - fix detection of older flex version and the reaction, since we do
111 depend on modern ones (we use lex_destroy)
112 - Better separation of SG and RL in the libs: remove all simulation code
113 from libgras. As a result, this lib is now only 200k when stripped.
114 Some of the xbt modules may also be duplicated (two sets and such) and
115 should be cleaned/killed before SG3.
116 - Insist on using xlC on AIX because of weird problems involving gcc there.
117 - Cleanup the make remote stuff. This is now done by scripts
118 tools/graspe-{master,slave} (GRAS Platform Expender). This is still
119 mainly for our private use, but we're working on changing them to user
122 - Bugfix: flush the socket on close only if there is some *output*.
123 - Bugfix: flush idempotent when there's nothing to send (don't send size=0)
125 - Add MSG_task_get_name. The task names are mainly for debugging purpose,
128 -- SimGrid team <simgrid2-users@listes.ens-lyon.fr> Fri, 4 Mar 2005 14:32:37 -0800
130 SimGrid (2.92) unstable; urgency=low
132 Alpha 3 on the path to SimGrid 3
136 - New! First try of benchmarking macros.
137 - New! First try so that gras_stub_generator generate deployment and
138 remote compilation helpers.
140 - Bugfix: Initialization fix in msg_test.
144 - Bugfix: applied patch to lexer so that it doesn't need a huge heap.
146 - Bugfix: let dicts work with NULL content (_foreach didn't) and cleanups
148 - API Change: gras_os_sleep to take the amount of seconds as a double.
149 Accepting an int was error prone since it was the only location where
150 seconds were coded as such. It leaded to damn rounding errors.
151 - Bugfix: Hard to belive that timers ever worked before this.
153 -- SimGrid team <simgrid2-users@listes.ens-lyon.fr> Wed, 23 Feb 2005 22:09:21 +0100
155 SimGrid (2.91) unstable; urgency=low
157 Alpha 2 on the path to SimGrid 3
161 - Bug fix in the lmm_solver.
163 - New! Interface to Paje (see http://www-id.imag.fr/Logiciels/paje/)
164 through the function MSG_paje_output.
165 - New! Introducing two new functions MSG_process_kill() and MSG_process_killall().
166 - It is possible to bound the rate of a communication in MSG with
167 MSG_task_put_bounded() (was already in the previous version but I had forgotten
168 to write it in the changelog).
169 - Bug fix to let GRAS run on top of MSG until we move it directly on top
174 - Various cleanups to the autotools stuff
175 - Begin to move Gras examples to examples/gras/
176 - Let make distcheck work again (yeah!)
178 - documentation overhauled using doxygen.
179 gtk-doc-tools is dead in SimGrid now.
180 - Automatically extract all existing logging categories, and add the list
181 to the documentation (long standing one, to say the less)
183 - Cleanup the known architecture table. Reorder the entries to group what
184 should be, and use a more consistent naming scheme.
185 (some of the test dataset are still to be regenerated)
186 - New! Allow library to register globals on each process just as userdata
188 This is implemented using a xbt_dict and not a xbt_set, so we loose the
189 lookup time (for now).
190 Use it in msg and trp.
191 This cleans a lot the internals and helps enforcing privacy of the
192 headers between the gras components.
193 - New! Add a timer mechanism, not unlike cron(8) and at(1).
194 - Bugfix: gras_os_time was delirious in RL.
195 - Bugfix: gras_trp_select/RL don't run into the wall when asked to select
197 - Reenable GRAS now that it works.
199 -- Arnaud Legrand <Arnaud.Legrand@imag.fr> Mon, 14 Feb 2005 14:02:13 -0800
201 SimGrid (2.90) unstable; urgency=low
203 Alpha 1 on the path to SimGrid 3
205 * It is a long time since the last release of SimGrid. I'm sorry about
206 that but as I had told you, I was rewriting a lot of things. I apologize
207 to those who had been reporting bugs to me and that I had not answered.
208 If your bug is still in the new version, please tell me. Here is a
209 summary of the main changes.
211 * REVOLUTION 1: The SimGrid project has merged with the GRAS project
212 lead by Martin Quinson. As a consequence SimGrid gains a lot in
213 portability, speed, and a lot more but you'll figure it out later.
214 SimGrid now comprises 3 different projects : MSG, GRAS and SMPI.
215 I wanted to release the new MSG as soon as possible and I have
216 broken GRAS, which is the reason why, for now, only MSG is fully
217 functional. A laconic description of these projects is available
218 in the documentation.
220 * REVOLUTION 2: I have removed SG and I am now using a new simulation
221 kernel optimized for our needs (called SURF but only the developers
222 should use it). Hence, MSG is now roughly 30 times faster and I think
223 that by rewriting a little bit MSG, I could event speed it up a little
224 bit more. Beside the gain in speed, it is also much easier to encode a
225 new platform model with SURF than it was with SG. More to come...
227 * REVOLUTION 3: I have tried to change a little as possible the API of
228 MSG but a few things really had to disappear. The main differences
229 with the previous version are :
230 1) no more m_links_t and the corresponding functions. Platforms are
231 directly read from a XML description and cannot be hard-coded
232 anymore. The same format is used for application deployment
233 description. The new format is described in the documentation.
234 Have a look in tools/platform_generation. There is a tiny script
235 that converts from the old platform format to the new one. Concerning
236 the application deployment format, parsing the old one is tricky.
237 I think most of you should however be able to convert your files. If
238 it is really an issue, I can write a C code that does the conversion.
240 2) the toolbox tbx does not exist anymore. We now have a library
241 with much more data-structures but without the hash-tables (we have
242 dictionaries that are much faster).
244 -- Arnaud Legrand <Arnaud.Legrand@imag.fr> Mon, 31 Jan 2005 10:45:53 -0800
246 *****************************************************************************
247 * Follows the old GRAS changelog. It does not follow the same syntax, but I *
248 * don't feel like converting the oldies. (Mt) *
249 *****************************************************************************
252 Version 2.90: "the long awaited one"
253 - Finished rewriting and debugging MSG. Rewrote the documentation.
254 - disable GRAS for now since it needs to be ported to the newest SG
257 - Finish the port to windows (using mingw32 for cross-compile)
260 - Main loop and datastructures of SURF. A cpu resource object is
261 functional. Surf can thus be used to create cpu's with variable
262 performance on which you can execute some actions.
264 2004-11-15 Martin Quinson
265 - Port to ARM. Simply added the alignment and size descriptions. Should
266 work, but the ARM machines are so slow that I didn't had the opportunity
267 to 'make check' over there yet.
269 2004-11-15 Arnaud Legrand
270 - Trace manager now written. It uses a heap structure and is therefore
271 expected to be efficient. It may however be speeded up (particularly
272 when many events occur at the same date) by using red and black
273 trees. One day maybe...
274 - Max-min linear system solver written. It uses a sparse matrix
275 structure taking advantage of its expected use. Most operations are
276 O(1) and free/calloc are called as few as possible. The computation of
277 the minimum could however be improved by using a red and black tree
280 2004-11-03 Arnaud Legrand
281 - Rename every gras_* function that was in xbt/ to its xbt_
283 - Add a heap and a doubly-linked list to xbt
284 - Added a dichotomy to the dictionaries. make check works as well before
285 so I assume that the patch is correct. I do not know however if things
286 run effectively faster than before now. :)
288 Inclusion of the SimGrid tree in the GRAS one. The archive is renamed to
289 SimGrid, and the version number is bumped to 2.x
291 2004-10-29 Martin Quinson
292 - Introduction of the remote errors.
293 They are the result of a RMI/RPC on the remote machine.
294 ErrCodes being scalar values, you can't get the host on which those
295 errors did happen. Extending the error mechanism as in Gnome is possible.
296 No idea yet whether it is a good idea.
298 2004-10-28 Martin Quinson
299 - Interface revolution: the Starred Structure Eradication.
300 I used to do typedef struct {} toto_t; and then handle *toto_t.
301 Arnaud (and Oli) didn't like it, and I surrendered. Now, you have:
302 - ???_t is a valid type (builded with typedef)
303 - s_toto_t is a structure (access to fields with .)
304 - s_toto is a structure needing 'struct' keyword to be used
305 - e_toto_t is an enum
306 - toto_t is an 'object' (struct*)
308 typedef struct s_toto {} s_toto_t, *toto_t;
309 typedef enum {} e_toto_t;
310 Moreover, only toto_t (and e_toto_t) are public. The rest (mainly
311 s_toto_t) is private.
313 - While I was at it, all gras_<obj>_free() functions want a gras_<obj>_t*
314 so that it can set the variable to NULL. It was so for dicts and sets,
315 it changed for dynars.
317 - Fix a bunch of memleaks in dict_remove
318 - Fix a bug in sg/server_socket opening: it failed all the time.
320 2004-10-07 Martin Quinson
321 - Speed up dynar lookup operation a bit.
323 gras_dynar_get is dead.
325 Now, you can choose between gras_dynar_get_cpy (the old gras_dynar_get
326 but should be avoided for efficiency reasons) and gras_dynar_get_ptr
327 (which gives you the address of the stored data).
329 gras_dynar_get_as is an helpful macro which allows you to retrieve a
330 copy of the data using an affectation to do the job and not a memcpy.
332 int toto = gras_dynar_get_as(dyn,0,int); rewrites itself to
333 int toto = *(int*)gras_dynar_get_ptr(dyn,0);
335 It does not really speedup the dynar test because they are
336 setting elements all the time (and look them seldom). But the dict does
337 far more lookup than setting.
339 So, this brings the dict_crash test from ~33s to ~25s (200000 elms).
341 2004-10-05 Martin Quinson
342 - Allow to (en/dis)able the cycle detection at run time.
344 Whether we should check for cycle or not is now a property of each
345 datatype. When you think there may be some cycle, use datadesc_cycle_set.
346 datadesc_cycle_unset allow to remove this property when previously set.
348 Note that the cycle detection is off by default since it impacts the
349 performance. Watch the data you feed GRAS with ;)
351 This property is hereditary. Any element embedded in a structure having it
352 set have it set for the time of this data exchange.
354 You should set it both on sender and receiver side. If you don't set it on
355 sender side, it will enter an endless loop. If you forget on receiver
356 side, the cycles won't be recreated after communication.
358 - Header reorganization.
359 Kill gras_private.h, each submodule must load the headers it needs.
361 2004-10-04 Martin Quinson
362 - Interface revolution: do not try to survive to malloc failure.
364 Now, gras_malloc and friends call gras_abort() on failure.
365 As a conclusion, malloc_error is not a valid error anymore, and all
366 functions for which it was the only gras_error_t return value are
367 changed. They now return void, or there result directly.
368 This simplify the API a lot.
370 2004-09-29 Martin Quinson
371 - Re-enable raw sockets.
372 Created by gras_socket_{client,server}_ext;
373 Used with gras_raw_{send,recv}
376 It should allow to kill the last bits of gras first version soon.
378 This is not completely satisfactory yet (duplicate code with
379 chunk_{send,recv}; a bit out of the plugin mechanism), but it should
382 - Simplify transport plugin (internal) interface by not passing any
383 argument to _server and _client, but embedding them in the socket
386 2004-09-28 Martin Quinson
387 - Finish the port to AIX.
388 autoconf was my problem (segfault within the malloc replacement
389 function. No idea why)
391 2004-09-16 Martin Quinson
392 - Fix some size_t madness on 64bit architectures.
394 2004-09-08 Martin Quinson
395 - Reduce the number of system headers loaded, overload some more system
396 calls (such as malloc to cast the result of the system one, and work
398 - Fix and reintroduce the config support
400 2004-09-07 Martin Quinson
401 - Source code reorganization to allow Arnaud to surf all over there.
402 - Allow to document the logging categories.
403 - Remove all uppercase from logging categories and useless cleanup in names.
405 2004-08-18 Martin Quinson
406 Version 0.6.2 (protocol not changed; API changed)
407 - Interface cleanup: gras_msgtype_by_name returns the type (instead of a
408 gras_error_t), and NULL when not found. Functions expecting a msgtype
409 as argument (msg_wait; msg_send) deal with NULL argument by providing a
410 hopefully usefull message.
411 - Portability to prehistoric sparcs again
413 2004-08-17 Martin Quinson
414 Version 0.6.1 (protocol not changed; ABI not changed)
415 - prealloc some buffers to speed things up
417 2004-08-11 Martin Quinson
418 Version 0.6 (protocol not changed; ABI expended)
419 - The parsing macro can deal with the references, provided that you add
420 the relevant annotations (using GRAS_ANNOTE(size,field_name))
422 2004-08-09 Martin Quinson
423 Version 0.5 (protocol not changed; ABI changed)
424 - Allow to off turn the cycle detection code in data exchange at
425 compilation time. It should be at run time, but I'm short of time (and
426 the config stuff is still broken). That way, we keep dict out of the
427 critical path, which is good because the performance is poor:
428 - search not dichotomial yet
429 - dynar give no way to access their content and memcpy everytime
430 - In composed data description (struct, ref and so on), stop foolness of
431 keeping the subtype's ID, but store the type itself. This keeps sets out
432 of the critical path, which is good since they rely on dynar and
433 dictionnaries. The only loose of that is that we cannot detect the
434 redeclaration of a structure/union with another content (but I'm not sure
435 the code detected well this error before anyway). We still can detect
436 the redefinition discrepancy for the other types.
437 - Use a whole bunch of optimisation flags (plus -fno-strict-aliasing since
438 it breaks the code because of type-punning used all over the place).
439 This breaks on all non-gcc architectures (for now).
441 All those changes (plus the buffer of last time) allow me to gain 2 order
442 of magnitude on cruel tests consisting of 800000 array of integers on two
443 level of a hierarchical structure (200 secondes -> 4 secondes)
446 - the selector of reference must now return the type it points to, not
449 2004-08-06 Martin Quinson
450 Version 0.4 (protocol changed; ABI not changed)
451 - Allow to pass --gras-log argument to processes in simulation mode. Really.
452 - New debugging level: trace (under debug) to see effect of GRAS_IN/OUT
453 - Implement a buffer transport, and use it by default (it relies on tcp in
454 real life and on sg in simulation).
455 That's a bit hackish since I had a new field to the structure to store
456 its data without interfering with the subtype ones. Inheritance
457 is tricky in C. And that's a kind of reverse inheritance with one class
458 derivating two classes. Or maybe a game with java interfaces. Anyway,
459 that's damn hard in C (at least).
460 Moreover, I got tired while trying to ensure plugin separation and
461 genericity in SG mode. MSG wants me to do weird things, so let's go for
462 cruel hacks (temporarily of course ;).
463 See comment in transport_private.h:71
464 - do not include all the _interface headers in private but in the files
465 which really need them (to cut the compilation time when they are
468 2004-07-26 Martin Quinson
469 Version 0.3 (protocol not changed; ABI changed)
470 - Major overhault of the datadesc interface to simplify it:
471 - shorted the function names:
472 s/gras_datadesc_declare_struct/gras_datadesc_struct/ and so on
473 - add a trivial way to push/pop integers into the cbps without malloc.
474 This allows to make really generic sub_type description, which simply
475 pop their size of the stack.
476 - add a function gras_datadesc_ref_pop_arr() which does what users want
477 most of the time: Declare a dynamic array (which pops its size of the
478 stack) and declare a reference to it. Poor name, but anyway.
479 - kill the post-send callback, add a post-receive one
481 2004-07-23 Martin Quinson
482 Version 0.2 (protocol changed; ABI changed)
483 - add some testing for cpbs in the test cases, and fix some more bugs.
484 This invalidate again the little64 data file, since I cannot regenerate
486 - remove an awfull optimization in the logging stuff, allowing me to:
487 - understand it again
488 - learn gcc how to check that the argument match the provided format
489 - fix all errors revealed by gcc after that
490 - internal keys of dict are not \0 terminated. Deal with it properly in
491 loggings instead of segfaulting when the user want to see the logs :-/
493 2004-07-22 Martin Quinson
494 - Fix some stupid bug preventing cbps (callback postit) from working
496 2004-07-21 Martin Quinson
497 - Some documentation cleanups
498 - remove the useless last argument of msgtype_declare
499 - rename the Virtu functions to fit into the 'os' namespace
500 - move headers src/include -> src/include/gras/ and stop fooling with
501 gras -> . symbolic link
502 - make distcheck is now successful
504 2004-07-19 Martin Quinson
506 - Build shared library also
507 - Install html doc to the right location
508 - stop removing maintainer files in make clean
509 - build tests only on make check
511 2004-07-13 Martin Quinson
513 - No major issue in previous version => change versionning schema
514 - Re-enable little64 convertion test now that Abdou kindly regenerated the
515 corresponding dataset.
517 2004-07-11 Martin Quinson
519 - Get it working with any kind of structure (we can compute the padding
520 bytes remotely for all the architectures I have access to)
521 - Implement the structure parsing macro (still not quite robust/complete)
522 - Improvement to the remote testing toysuite
524 2004-07-10 Martin Quinson
526 - get ride of a bunch of deprecated macros
527 - actually run the test for two-compliment, not only compile it :-/
528 - test whether the structures get packed (and bail out if yes. Damn.
529 Alignment is a serious matter)
530 - test whether the structures get compacted (but respecting the alignment
531 constraints of each types)
532 - test whether the array fields of structures can straddle alignment boundaries
534 - Damnit, double are bigger than float (typo in creation of 'double' datadesc)
535 (took me 2 hours to find that bug, looking at the wrong place)
536 - Add gras_datadesc_declare_{union,struct}_close(). They must be used
537 before sending/receiving and are used to compute the offsets of fields
538 - Given that padding size depend even on compiler options, keep track of
539 alignment and aligned_size only for the current architecture. Not a big
540 deal since we send structure fields one after the other (seems
542 - Add the datastructure used for IEEE paper by the PBIO guys to the test
543 program, let it work on linux/gcc/little32. portability todo.
545 2004-07-08 Martin Quinson
546 - import and improve remote compilation support from FAST
547 - make sure make check works on half a dozen of machines out there
549 2004-07-07 Martin Quinson
550 Let's say it's version 0.0.3 ;)
551 - Implement conversions (yuhu!)
552 - Let it work on solaris (beside conversion, of course)
553 - Stupid me, using rand() to generate the conversion datatests in not wise.
555 2004-07-06 Martin Quinson
556 - Let make dist work, since I'm gonna need it to compile on remote hosts
557 - Let Tests/datadesc_usage write the architecture on which the file was
558 generated as first byte.
559 - Add PowerPC (being also IRIX64), SPARC (also power4) and ALPHA
560 architecture descriptions.
561 - Add datadesc_usage.{i386,ppc,sparc} files being the result of execution
562 on those architectures.
563 - Optimization: send/recv array of scalar in one shoot
565 2004-07-05 Martin Quinson
566 - YEAH! GRAS/SG and GRAS/RL are both able to run the ping example !
568 - Plug a whole bunch of memleaks
569 - each process now have to call gras_{init,exit}. One day, their log
570 settings will be separated
571 - Continue the code factorisation between SG, RL and common in Transport.
573 2004-07-04 Martin Quinson
575 - Redistribution between SG and RL.
576 We wanna have to accept in SG, so move accepted related parts of RL in
577 the common part. (more precisely, the dynar of all known sockets is no
578 more a static in transport.c, but part of the process_data)
580 [gras_stub_generator]
581 - Bug fix: Do call gras_process_init from gras_init (wasnt called in RL).
583 2004-07-03 Martin Quinson
584 - Create a new log channel tbx containing dict, set, log, dynar (to shut
585 them all up in one shot)
587 - Fix the ugly case of reference to dynamic array.
588 - New (semi-public) function gras_datadesc_size to allow the messaging
589 layer to malloc the needed space for the buffer.
591 - gras_socket_close now expect the socket to close (and not its address to
592 put NULL in it after it). This is because the socket passed to handlers
593 is one of their argument (=> not writable).
595 - propagate the interface cleanup from last week in datadesc, ie remove a
596 superfluous level of indirection. User pass adress of variable
597 containing data (both when sending and receiving), and not of a variable
598 being a pointer to the data. Let's say that I like it better ;)
599 The price for that is constructs like "int msg=*(int*)payload" in
600 handlers, but it's a fine price, IMHO.
602 - Let it work in RL (yuhu)
604 2004-06-21 Martin Quinson
606 - porting SG plugin and SG select to new standards (works almost).
607 - plug memleaks and fix bugs around.
610 - cleanup the prototype of data recv and force users to specify when they
611 want to handle references to objects. Test case working even for cycles.
612 - plug memleaks. Valgrind is perfectly ok with this.
614 2004-06-12 Martin Quinson
616 - cleanup the separation between plugin and main code in plugin creation
618 2004-06-11 Martin Quinson
620 - Reput hook for raw sockets, needed for BW experiments
621 - kill a few lines of dead code
622 [Data description] Interface cleanup
623 - gras_datadesc_by_name returns the searched type or NULL.
624 That way, no variable is needed to use a type desc once, which makes
626 - gras_datadesc_declare_[struct|union]_append_name is removed. The last
627 two parameters were strings (field name, type name), leading to
629 [Dicos] Interface cleanup
630 - gras_dico_retrieve -> gras_dico_get ; gras_dico_insert -> gras_dico_set
631 This is consistant with the dynar API.
633 2004-04-21 Martin Quinson
635 - Porting to new standards.
638 There is no bag anymore, no need to take extra provision to mask the
639 pointers behind "ID".
640 Better splitup of functions between files create/exchange/convert.
641 This is still a bit artificial since convert and receive are so
642 interleaved, but anyway.
644 - add a queued message list to procdata (the ones not matching criteria
646 - factorize some more code between SG and RL wrt procdata
648 - use gras_exit in example to track memleaks
649 - get rid of gs_example now that GS is properly integrated into gras
650 - update run_test to integrate the lastest tests (datadesc)
652 - rename WARNINGn macros to WARNn since it prooved error-prone
654 2004-04-19 Martin Quinson
656 - register init/exit functions within gras module mechanism
657 - send/receive function.
658 Convertion is not implemented, but short-cutted if not needed.
659 struct/array elements are sent one by one (instead of block-wise), but
660 nobody really cares (yet). Get a prototype before optimizing.
661 - tests (using a file socket) for DD send/receive on:
662 - base types: int, float
663 - array: fixed size, string (ie ref to dynamic string)
664 - structure: homogeneous, heterogeneous
665 - chained list, graph with cycle
666 Believe it or not, valgrind is not too unhappy with the results. The
667 cycle happily segfaults, but the others are ok. And I'm sick of pointers
671 - Bugfix when using a filename explicitely (instead of '-')
673 2004-04-09 Martin Quinson
675 - factorize more code between RL and SG in socket creation
676 - Complete the implementation and tests of:
678 o file (only in RL, and mainly for debugging)
680 I lost 3 days to design a portable address resolver, and then decided
681 that the prototype mainly have to run on my box.
682 Addressing portability too early may be like optimizing too early :-/
684 - use gras_init in the Tests instead of the crappy parse_log_opt
685 (the latter function is removed)
686 [Conditional execution]
687 - New functions: gras_if_RL/gras_if_SG (basic support for this)
688 [Code reorganisation]
689 - Get rid of libgrasutils.a since it makes more trouble than it solves.
690 Build examples against the RL library, since there is no way to disable
691 its creation for now.