1 SimGrid (2.96) unstable; urgency=low
4 * New! Add a FATPIPE model. [AL]
5 * New! Add a parallel task model. [AL]
6 * New! Add automatically a loopback interface (in the default
7 network model) if none was precised.
10 * Bugfix: MSG_process_resume now works with the current running process.
12 * New! Add MSG_parallel_task_create and MSG_parallel_task_execute. [AL]
13 * Modification of MSG_task_get_compute_duration. Once a task has been
14 processed, the value returned by this function is now equal to 0. [AL]
15 * New! Add double MSG_task_get_remaining_computation(m_task_t task) and
16 MSG_error_t MSG_task_cancel(m_task_t task). Add a state
17 (MSG_TASK_CANCELLED) to MSG_error_t corresponding to the cancelation
18 of a m_task. For now, MSG_task_cancel only works with computation
20 * New! Add double MSG_get_host_speed(m_host_t h) that returns the speed
21 of the processor (in Mflop/s) regardless of the current load on the
26 SimGrid (2.95) unstable; urgency=low
29 * Steal some nice code to GNU pth to fix context detection and usage [AL]
30 * Cleanup in the xbt_config API; add configuration callbacks. [MQ]
31 * Cleanup in the initialization API: the unused "defaultlog" is dead. [MQ]
34 * Bugfix: Allow absolute paths for platform description files [MQ]
35 * Bugfix: do free the variables after use. Leads to drastic performance
37 * Implement max_duration (ie, timeouts) on resources [AL]
40 * Implement MSG_config to configure MSG at runtime. xbt_cfg test on a real
42 * Implement MSG_channel_select_from() to help GRAS now that SURF provide
43 the needed support (timeouts) [AL]
46 * Implement measurement sockets. You can now get the bandwidth between two
47 hosts thanks to AMOK (see below). [MQ]
48 * gras_datadesc_dynar() builds a dynar type descriptor, allowing to send
49 dynar over the network (yeah) [MQ]
50 * Real (even if simplistic) implementation of gras_os_myname() on RL [MQ]
51 * simple/static token-ring example. [Alexandre Colucci and MQ]
52 * Use MSG_channel_select_from() and remove the *slow* hack we had to put
56 * Differentiate the types "char[22]" and "unsigned char[22]" in automatic
57 type parsing. "short" and "long" modifiers were also ignored; other
58 modifier (such as reference level) are still ignored. [MQ]
59 * Embeed the buffer size within the buffer itself on SG. [MQ]
60 That way, send() are atomic and cannot get intermixed anymore (at least
61 the ones which are less than 100k; bigger messages still have the issue)
62 * Array size pushed by the field, not by the field type (or each
63 and every long int will push stuff to the cbps) [MQ]
64 * use select() to sleep since it allows to portably sleep less than one
68 * <project>.Makefile.local (generated from gras_stub_generator) |MQ]:
71 * Type Callbacks now receive the gras_datadesc_type_t they work on as argument.
72 * type category 'ignored' killed as it was never used and were difficult
74 * whether a type can cycle or not is now a flag, leaving room for more
75 flags (as "ignored", if we feel the need one day ;)
76 * Rename raw sockets to measurement sockets since "raw" has another
77 meaning in networking community.
80 * Advanced Metacomputing Overlay Kit introduction. It is based over GRAS
81 and offers features not belonging to GRAS but that most applications
82 need. One day, it may be a set of plugins loadable at runtime.
83 * New module: bandwidth
84 bandwidth measurement between arbitrary nodes running this module. [MQ]
86 -- Da SimGrid team <simgrid-devel@lists.gforge.inria.fr> Thu, 30 Jun 2005 16:29:20 -0700
88 SimGrid (2.94) unstable; urgency=low
90 The first beta release of SimGrid 3 !
94 * Update the main page and the FAQ. Adding references to gforge.
97 * Add a gras_os_getpid function.
100 * Add MSG_task_get_compute_duration() and MSG_task_get_data_size()
101 * Extend the logs so that they also print PID, hostname, date, ... if
103 * Convert the MSG example to the use of xbt_logs instead of PRINT_MESSAGE,
104 and kill the old version which were in testsuite/
105 * Rewrite tools/MSG_visualization/colorize.pl for using with logs instead
109 * Add xbt_os_time(). As the rest of xbt/portability, this is not public
110 for users. Instead, each programming environment (GRAS, MSG,...) use it
111 when needed to provide such a feature to users.
112 Don't shortcut the mecanism or you will also shortcut the virtualization
113 you need on the simulator.
117 * Cleanups in configury with regard to compile optimization/warning flags.
118 Also add -fno-loop-optimize to any powerpc since it's the optimization
119 killing gcc (< 3.4.0).
120 * Doxygen cleanups: move MSG examples, kill the second Doxygen phase
121 needed by MSG examples complications
122 * Borrow configury beautifications from PHP
125 * Bugfix: XBT_LOG_NEW_DEFAULT_CATEGORY now compiles without compiler
126 warning (thanks loris for stumbling into this one).
127 * Bugfix: stop loading private headers (gras_config.h) from the public
131 * Change SIMGRID_INSTALL_PATH to GRAS_ROOT in Makefiles generated for user.
132 * Rename gras_get_my_fqdn to gras_os_myname and implement it in the simulator
133 RL would imply a DNS resolver, which is *hard* to do in a portable way
134 (and therefore delayed).
135 * Implement a real timer mecanism and use it in timing macros. This allows
136 to avoid rounding errors and get a 0.000005 sec precision in timing
137 macros. While I was at it, various cleanups:
138 - allow to declare more than one timed section per file (fix a stupid bug)
139 - move some private declaration to the right place
140 - merge conditional execution and timing macros into emulation module
141 - document the module
142 - make sure the module cleanups its mess on gras_exit
143 * Documentation improvements:
144 - (new) how to compile applications using GRAS
145 - (new) emulation support (timing macros)
147 -- Da SimGrid team <simgrid-devel@lists.gforge.inria.fr> Fri, 13 May 2005 10:49:31 +0200
149 SimGrid (2.93) unstable; urgency=low
151 Alpha 4 on the path to SimGrid 3 (aka the "neuf-trois" version)
154 - Use Paje properly where used. Still to be sanitized properly.
155 - Portability fix: Add an implementation of the contexts using pthread
159 - Add xbt_procname(): returns the name of the current process.
160 Use it to show the current process's name in all logging.
162 - fix detection of older flex version and the reaction, since we do
163 depend on modern ones (we use lex_destroy)
164 - Better separation of SG and RL in the libs: remove all simulation code
165 from libgras. As a result, this lib is now only 200k when stripped.
166 Some of the xbt modules may also be duplicated (two sets and such) and
167 should be cleaned/killed before SG3.
168 - Insist on using xlC on AIX because of weird problems involving gcc there.
169 - Cleanup the make remote stuff. This is now done by scripts
170 tools/graspe-{master,slave} (GRAS Platform Expender). This is still
171 mainly for our private use, but we're working on changing them to user
174 - Bugfix: flush the socket on close only if there is some *output*.
175 - Bugfix: flush idempotent when there's nothing to send (don't send size=0)
177 - Add MSG_task_get_name. The task names are mainly for debugging purpose,
180 -- SimGrid team <simgrid2-users@listes.ens-lyon.fr> Fri, 4 Mar 2005 14:32:37 -0800
182 SimGrid (2.92) unstable; urgency=low
184 Alpha 3 on the path to SimGrid 3
188 - New! First try of benchmarking macros.
189 - New! First try so that gras_stub_generator generate deployment and
190 remote compilation helpers.
192 - Bugfix: Initialization fix in msg_test.
196 - Bugfix: applied patch to lexer so that it doesn't need a huge heap.
198 - Bugfix: let dicts work with NULL content (_foreach didn't) and cleanups
200 - API Change: gras_os_sleep to take the amount of seconds as a double.
201 Accepting an int was error prone since it was the only location where
202 seconds were coded as such. It leaded to damn rounding errors.
203 - Bugfix: Hard to belive that timers ever worked before this.
205 -- SimGrid team <simgrid2-users@listes.ens-lyon.fr> Wed, 23 Feb 2005 22:09:21 +0100
207 SimGrid (2.91) unstable; urgency=low
209 Alpha 2 on the path to SimGrid 3
213 - Bug fix in the lmm_solver.
215 - New! Interface to Paje (see http://www-id.imag.fr/Logiciels/paje/)
216 through the function MSG_paje_output.
217 - New! Introducing two new functions MSG_process_kill() and MSG_process_killall().
218 - It is possible to bound the rate of a communication in MSG with
219 MSG_task_put_bounded() (was already in the previous version but I had forgotten
220 to write it in the changelog).
221 - Bug fix to let GRAS run on top of MSG until we move it directly on top
226 - Various cleanups to the autotools stuff
227 - Begin to move Gras examples to examples/gras/
228 - Let make distcheck work again (yeah!)
230 - documentation overhauled using doxygen.
231 gtk-doc-tools is dead in SimGrid now.
232 - Automatically extract all existing logging categories, and add the list
233 to the documentation (long standing one, to say the less)
235 - Cleanup the known architecture table. Reorder the entries to group what
236 should be, and use a more consistent naming scheme.
237 (some of the test dataset are still to be regenerated)
238 - New! Allow library to register globals on each process just as userdata
240 This is implemented using a xbt_dict and not a xbt_set, so we loose the
241 lookup time (for now).
242 Use it in msg and trp.
243 This cleans a lot the internals and helps enforcing privacy of the
244 headers between the gras components.
245 - New! Add a timer mechanism, not unlike cron(8) and at(1).
246 - Bugfix: gras_os_time was delirious in RL.
247 - Bugfix: gras_trp_select/RL don't run into the wall when asked to select
249 - Reenable GRAS now that it works.
251 -- Arnaud Legrand <Arnaud.Legrand@imag.fr> Mon, 14 Feb 2005 14:02:13 -0800
253 SimGrid (2.90) unstable; urgency=low
255 Alpha 1 on the path to SimGrid 3
257 * It is a long time since the last release of SimGrid. I'm sorry about
258 that but as I had told you, I was rewriting a lot of things. I apologize
259 to those who had been reporting bugs to me and that I had not answered.
260 If your bug is still in the new version, please tell me. Here is a
261 summary of the main changes.
263 * REVOLUTION 1: The SimGrid project has merged with the GRAS project
264 lead by Martin Quinson. As a consequence SimGrid gains a lot in
265 portability, speed, and a lot more but you'll figure it out later.
266 SimGrid now comprises 3 different projects : MSG, GRAS and SMPI.
267 I wanted to release the new MSG as soon as possible and I have
268 broken GRAS, which is the reason why, for now, only MSG is fully
269 functional. A laconic description of these projects is available
270 in the documentation.
272 * REVOLUTION 2: I have removed SG and I am now using a new simulation
273 kernel optimized for our needs (called SURF but only the developers
274 should use it). Hence, MSG is now roughly 30 times faster and I think
275 that by rewriting a little bit MSG, I could event speed it up a little
276 bit more. Beside the gain in speed, it is also much easier to encode a
277 new platform model with SURF than it was with SG. More to come...
279 * REVOLUTION 3: I have tried to change a little as possible the API of
280 MSG but a few things really had to disappear. The main differences
281 with the previous version are :
282 1) no more m_links_t and the corresponding functions. Platforms are
283 directly read from a XML description and cannot be hard-coded
284 anymore. The same format is used for application deployment
285 description. The new format is described in the documentation.
286 Have a look in tools/platform_generation. There is a tiny script
287 that converts from the old platform format to the new one. Concerning
288 the application deployment format, parsing the old one is tricky.
289 I think most of you should however be able to convert your files. If
290 it is really an issue, I can write a C code that does the conversion.
292 2) the toolbox tbx does not exist anymore. We now have a library
293 with much more data-structures but without the hash-tables (we have
294 dictionaries that are much faster).
296 -- Arnaud Legrand <Arnaud.Legrand@imag.fr> Mon, 31 Jan 2005 10:45:53 -0800
298 *****************************************************************************
299 * Follows the old GRAS changelog. It does not follow the same syntax, but I *
300 * don't feel like converting the oldies. (Mt) *
301 *****************************************************************************
304 Version 2.90: "the long awaited one"
305 - Finished rewriting and debugging MSG. Rewrote the documentation.
306 - disable GRAS for now since it needs to be ported to the newest SG
309 - Finish the port to windows (using mingw32 for cross-compile)
312 - Main loop and datastructures of SURF. A cpu resource object is
313 functional. Surf can thus be used to create cpu's with variable
314 performance on which you can execute some actions.
316 2004-11-15 Martin Quinson
317 - Port to ARM. Simply added the alignment and size descriptions. Should
318 work, but the ARM machines are so slow that I didn't had the opportunity
319 to 'make check' over there yet.
321 2004-11-15 Arnaud Legrand
322 - Trace manager now written. It uses a heap structure and is therefore
323 expected to be efficient. It may however be speeded up (particularly
324 when many events occur at the same date) by using red and black
325 trees. One day maybe...
326 - Max-min linear system solver written. It uses a sparse matrix
327 structure taking advantage of its expected use. Most operations are
328 O(1) and free/calloc are called as few as possible. The computation of
329 the minimum could however be improved by using a red and black tree
332 2004-11-03 Arnaud Legrand
333 - Rename every gras_* function that was in xbt/ to its xbt_
335 - Add a heap and a doubly-linked list to xbt
336 - Added a dichotomy to the dictionaries. make check works as well before
337 so I assume that the patch is correct. I do not know however if things
338 run effectively faster than before now. :)
340 Inclusion of the SimGrid tree in the GRAS one. The archive is renamed to
341 SimGrid, and the version number is bumped to 2.x
343 2004-10-29 Martin Quinson
344 - Introduction of the remote errors.
345 They are the result of a RMI/RPC on the remote machine.
346 ErrCodes being scalar values, you can't get the host on which those
347 errors did happen. Extending the error mechanism as in Gnome is possible.
348 No idea yet whether it is a good idea.
350 2004-10-28 Martin Quinson
351 - Interface revolution: the Starred Structure Eradication.
352 I used to do typedef struct {} toto_t; and then handle *toto_t.
353 Arnaud (and Oli) didn't like it, and I surrendered. Now, you have:
354 - ???_t is a valid type (builded with typedef)
355 - s_toto_t is a structure (access to fields with .)
356 - s_toto is a structure needing 'struct' keyword to be used
357 - e_toto_t is an enum
358 - toto_t is an 'object' (struct*)
360 typedef struct s_toto {} s_toto_t, *toto_t;
361 typedef enum {} e_toto_t;
362 Moreover, only toto_t (and e_toto_t) are public. The rest (mainly
363 s_toto_t) is private.
365 - While I was at it, all gras_<obj>_free() functions want a gras_<obj>_t*
366 so that it can set the variable to NULL. It was so for dicts and sets,
367 it changed for dynars.
369 - Fix a bunch of memleaks in dict_remove
370 - Fix a bug in sg/server_socket opening: it failed all the time.
372 2004-10-07 Martin Quinson
373 - Speed up dynar lookup operation a bit.
375 gras_dynar_get is dead.
377 Now, you can choose between gras_dynar_get_cpy (the old gras_dynar_get
378 but should be avoided for efficiency reasons) and gras_dynar_get_ptr
379 (which gives you the address of the stored data).
381 gras_dynar_get_as is an helpful macro which allows you to retrieve a
382 copy of the data using an affectation to do the job and not a memcpy.
384 int toto = gras_dynar_get_as(dyn,0,int); rewrites itself to
385 int toto = *(int*)gras_dynar_get_ptr(dyn,0);
387 It does not really speedup the dynar test because they are
388 setting elements all the time (and look them seldom). But the dict does
389 far more lookup than setting.
391 So, this brings the dict_crash test from ~33s to ~25s (200000 elms).
393 2004-10-05 Martin Quinson
394 - Allow to (en/dis)able the cycle detection at run time.
396 Whether we should check for cycle or not is now a property of each
397 datatype. When you think there may be some cycle, use datadesc_cycle_set.
398 datadesc_cycle_unset allow to remove this property when previously set.
400 Note that the cycle detection is off by default since it impacts the
401 performance. Watch the data you feed GRAS with ;)
403 This property is hereditary. Any element embedded in a structure having it
404 set have it set for the time of this data exchange.
406 You should set it both on sender and receiver side. If you don't set it on
407 sender side, it will enter an endless loop. If you forget on receiver
408 side, the cycles won't be recreated after communication.
410 - Header reorganization.
411 Kill gras_private.h, each submodule must load the headers it needs.
413 2004-10-04 Martin Quinson
414 - Interface revolution: do not try to survive to malloc failure.
416 Now, gras_malloc and friends call gras_abort() on failure.
417 As a conclusion, malloc_error is not a valid error anymore, and all
418 functions for which it was the only gras_error_t return value are
419 changed. They now return void, or there result directly.
420 This simplify the API a lot.
422 2004-09-29 Martin Quinson
423 - Re-enable raw sockets.
424 Created by gras_socket_{client,server}_ext;
425 Used with gras_raw_{send,recv}
428 It should allow to kill the last bits of gras first version soon.
430 This is not completely satisfactory yet (duplicate code with
431 chunk_{send,recv}; a bit out of the plugin mechanism), but it should
434 - Simplify transport plugin (internal) interface by not passing any
435 argument to _server and _client, but embedding them in the socket
438 2004-09-28 Martin Quinson
439 - Finish the port to AIX.
440 autoconf was my problem (segfault within the malloc replacement
441 function. No idea why)
443 2004-09-16 Martin Quinson
444 - Fix some size_t madness on 64bit architectures.
446 2004-09-08 Martin Quinson
447 - Reduce the number of system headers loaded, overload some more system
448 calls (such as malloc to cast the result of the system one, and work
450 - Fix and reintroduce the config support
452 2004-09-07 Martin Quinson
453 - Source code reorganization to allow Arnaud to surf all over there.
454 - Allow to document the logging categories.
455 - Remove all uppercase from logging categories and useless cleanup in names.
457 2004-08-18 Martin Quinson
458 Version 0.6.2 (protocol not changed; API changed)
459 - Interface cleanup: gras_msgtype_by_name returns the type (instead of a
460 gras_error_t), and NULL when not found. Functions expecting a msgtype
461 as argument (msg_wait; msg_send) deal with NULL argument by providing a
462 hopefully usefull message.
463 - Portability to prehistoric sparcs again
465 2004-08-17 Martin Quinson
466 Version 0.6.1 (protocol not changed; ABI not changed)
467 - prealloc some buffers to speed things up
469 2004-08-11 Martin Quinson
470 Version 0.6 (protocol not changed; ABI expended)
471 - The parsing macro can deal with the references, provided that you add
472 the relevant annotations (using GRAS_ANNOTE(size,field_name))
474 2004-08-09 Martin Quinson
475 Version 0.5 (protocol not changed; ABI changed)
476 - Allow to off turn the cycle detection code in data exchange at
477 compilation time. It should be at run time, but I'm short of time (and
478 the config stuff is still broken). That way, we keep dict out of the
479 critical path, which is good because the performance is poor:
480 - search not dichotomial yet
481 - dynar give no way to access their content and memcpy everytime
482 - In composed data description (struct, ref and so on), stop foolness of
483 keeping the subtype's ID, but store the type itself. This keeps sets out
484 of the critical path, which is good since they rely on dynar and
485 dictionnaries. The only loose of that is that we cannot detect the
486 redeclaration of a structure/union with another content (but I'm not sure
487 the code detected well this error before anyway). We still can detect
488 the redefinition discrepancy for the other types.
489 - Use a whole bunch of optimisation flags (plus -fno-strict-aliasing since
490 it breaks the code because of type-punning used all over the place).
491 This breaks on all non-gcc architectures (for now).
493 All those changes (plus the buffer of last time) allow me to gain 2 order
494 of magnitude on cruel tests consisting of 800000 array of integers on two
495 level of a hierarchical structure (200 secondes -> 4 secondes)
498 - the selector of reference must now return the type it points to, not
501 2004-08-06 Martin Quinson
502 Version 0.4 (protocol changed; ABI not changed)
503 - Allow to pass --gras-log argument to processes in simulation mode. Really.
504 - New debugging level: trace (under debug) to see effect of GRAS_IN/OUT
505 - Implement a buffer transport, and use it by default (it relies on tcp in
506 real life and on sg in simulation).
507 That's a bit hackish since I had a new field to the structure to store
508 its data without interfering with the subtype ones. Inheritance
509 is tricky in C. And that's a kind of reverse inheritance with one class
510 derivating two classes. Or maybe a game with java interfaces. Anyway,
511 that's damn hard in C (at least).
512 Moreover, I got tired while trying to ensure plugin separation and
513 genericity in SG mode. MSG wants me to do weird things, so let's go for
514 cruel hacks (temporarily of course ;).
515 See comment in transport_private.h:71
516 - do not include all the _interface headers in private but in the files
517 which really need them (to cut the compilation time when they are
520 2004-07-26 Martin Quinson
521 Version 0.3 (protocol not changed; ABI changed)
522 - Major overhault of the datadesc interface to simplify it:
523 - shorted the function names:
524 s/gras_datadesc_declare_struct/gras_datadesc_struct/ and so on
525 - add a trivial way to push/pop integers into the cbps without malloc.
526 This allows to make really generic sub_type description, which simply
527 pop their size of the stack.
528 - add a function gras_datadesc_ref_pop_arr() which does what users want
529 most of the time: Declare a dynamic array (which pops its size of the
530 stack) and declare a reference to it. Poor name, but anyway.
531 - kill the post-send callback, add a post-receive one
533 2004-07-23 Martin Quinson
534 Version 0.2 (protocol changed; ABI changed)
535 - add some testing for cpbs in the test cases, and fix some more bugs.
536 This invalidate again the little64 data file, since I cannot regenerate
538 - remove an awfull optimization in the logging stuff, allowing me to:
539 - understand it again
540 - learn gcc how to check that the argument match the provided format
541 - fix all errors revealed by gcc after that
542 - internal keys of dict are not \0 terminated. Deal with it properly in
543 loggings instead of segfaulting when the user want to see the logs :-/
545 2004-07-22 Martin Quinson
546 - Fix some stupid bug preventing cbps (callback postit) from working
548 2004-07-21 Martin Quinson
549 - Some documentation cleanups
550 - remove the useless last argument of msgtype_declare
551 - rename the Virtu functions to fit into the 'os' namespace
552 - move headers src/include -> src/include/gras/ and stop fooling with
553 gras -> . symbolic link
554 - make distcheck is now successful
556 2004-07-19 Martin Quinson
558 - Build shared library also
559 - Install html doc to the right location
560 - stop removing maintainer files in make clean
561 - build tests only on make check
563 2004-07-13 Martin Quinson
565 - No major issue in previous version => change versionning schema
566 - Re-enable little64 convertion test now that Abdou kindly regenerated the
567 corresponding dataset.
569 2004-07-11 Martin Quinson
571 - Get it working with any kind of structure (we can compute the padding
572 bytes remotely for all the architectures I have access to)
573 - Implement the structure parsing macro (still not quite robust/complete)
574 - Improvement to the remote testing toysuite
576 2004-07-10 Martin Quinson
578 - get ride of a bunch of deprecated macros
579 - actually run the test for two-compliment, not only compile it :-/
580 - test whether the structures get packed (and bail out if yes. Damn.
581 Alignment is a serious matter)
582 - test whether the structures get compacted (but respecting the alignment
583 constraints of each types)
584 - test whether the array fields of structures can straddle alignment boundaries
586 - Damnit, double are bigger than float (typo in creation of 'double' datadesc)
587 (took me 2 hours to find that bug, looking at the wrong place)
588 - Add gras_datadesc_declare_{union,struct}_close(). They must be used
589 before sending/receiving and are used to compute the offsets of fields
590 - Given that padding size depend even on compiler options, keep track of
591 alignment and aligned_size only for the current architecture. Not a big
592 deal since we send structure fields one after the other (seems
594 - Add the datastructure used for IEEE paper by the PBIO guys to the test
595 program, let it work on linux/gcc/little32. portability todo.
597 2004-07-08 Martin Quinson
598 - import and improve remote compilation support from FAST
599 - make sure make check works on half a dozen of machines out there
601 2004-07-07 Martin Quinson
602 Let's say it's version 0.0.3 ;)
603 - Implement conversions (yuhu!)
604 - Let it work on solaris (beside conversion, of course)
605 - Stupid me, using rand() to generate the conversion datatests in not wise.
607 2004-07-06 Martin Quinson
608 - Let make dist work, since I'm gonna need it to compile on remote hosts
609 - Let Tests/datadesc_usage write the architecture on which the file was
610 generated as first byte.
611 - Add PowerPC (being also IRIX64), SPARC (also power4) and ALPHA
612 architecture descriptions.
613 - Add datadesc_usage.{i386,ppc,sparc} files being the result of execution
614 on those architectures.
615 - Optimization: send/recv array of scalar in one shoot
617 2004-07-05 Martin Quinson
618 - YEAH! GRAS/SG and GRAS/RL are both able to run the ping example !
620 - Plug a whole bunch of memleaks
621 - each process now have to call gras_{init,exit}. One day, their log
622 settings will be separated
623 - Continue the code factorisation between SG, RL and common in Transport.
625 2004-07-04 Martin Quinson
627 - Redistribution between SG and RL.
628 We wanna have to accept in SG, so move accepted related parts of RL in
629 the common part. (more precisely, the dynar of all known sockets is no
630 more a static in transport.c, but part of the process_data)
632 [gras_stub_generator]
633 - Bug fix: Do call gras_process_init from gras_init (wasnt called in RL).
635 2004-07-03 Martin Quinson
636 - Create a new log channel tbx containing dict, set, log, dynar (to shut
637 them all up in one shot)
639 - Fix the ugly case of reference to dynamic array.
640 - New (semi-public) function gras_datadesc_size to allow the messaging
641 layer to malloc the needed space for the buffer.
643 - gras_socket_close now expect the socket to close (and not its address to
644 put NULL in it after it). This is because the socket passed to handlers
645 is one of their argument (=> not writable).
647 - propagate the interface cleanup from last week in datadesc, ie remove a
648 superfluous level of indirection. User pass adress of variable
649 containing data (both when sending and receiving), and not of a variable
650 being a pointer to the data. Let's say that I like it better ;)
651 The price for that is constructs like "int msg=*(int*)payload" in
652 handlers, but it's a fine price, IMHO.
654 - Let it work in RL (yuhu)
656 2004-06-21 Martin Quinson
658 - porting SG plugin and SG select to new standards (works almost).
659 - plug memleaks and fix bugs around.
662 - cleanup the prototype of data recv and force users to specify when they
663 want to handle references to objects. Test case working even for cycles.
664 - plug memleaks. Valgrind is perfectly ok with this.
666 2004-06-12 Martin Quinson
668 - cleanup the separation between plugin and main code in plugin creation
670 2004-06-11 Martin Quinson
672 - Reput hook for raw sockets, needed for BW experiments
673 - kill a few lines of dead code
674 [Data description] Interface cleanup
675 - gras_datadesc_by_name returns the searched type or NULL.
676 That way, no variable is needed to use a type desc once, which makes
678 - gras_datadesc_declare_[struct|union]_append_name is removed. The last
679 two parameters were strings (field name, type name), leading to
681 [Dicos] Interface cleanup
682 - gras_dico_retrieve -> gras_dico_get ; gras_dico_insert -> gras_dico_set
683 This is consistant with the dynar API.
685 2004-04-21 Martin Quinson
687 - Porting to new standards.
690 There is no bag anymore, no need to take extra provision to mask the
691 pointers behind "ID".
692 Better splitup of functions between files create/exchange/convert.
693 This is still a bit artificial since convert and receive are so
694 interleaved, but anyway.
696 - add a queued message list to procdata (the ones not matching criteria
698 - factorize some more code between SG and RL wrt procdata
700 - use gras_exit in example to track memleaks
701 - get rid of gs_example now that GS is properly integrated into gras
702 - update run_test to integrate the lastest tests (datadesc)
704 - rename WARNINGn macros to WARNn since it prooved error-prone
706 2004-04-19 Martin Quinson
708 - register init/exit functions within gras module mechanism
709 - send/receive function.
710 Convertion is not implemented, but short-cutted if not needed.
711 struct/array elements are sent one by one (instead of block-wise), but
712 nobody really cares (yet). Get a prototype before optimizing.
713 - tests (using a file socket) for DD send/receive on:
714 - base types: int, float
715 - array: fixed size, string (ie ref to dynamic string)
716 - structure: homogeneous, heterogeneous
717 - chained list, graph with cycle
718 Believe it or not, valgrind is not too unhappy with the results. The
719 cycle happily segfaults, but the others are ok. And I'm sick of pointers
723 - Bugfix when using a filename explicitely (instead of '-')
725 2004-04-09 Martin Quinson
727 - factorize more code between RL and SG in socket creation
728 - Complete the implementation and tests of:
730 o file (only in RL, and mainly for debugging)
732 I lost 3 days to design a portable address resolver, and then decided
733 that the prototype mainly have to run on my box.
734 Addressing portability too early may be like optimizing too early :-/
736 - use gras_init in the Tests instead of the crappy parse_log_opt
737 (the latter function is removed)
738 [Conditional execution]
739 - New functions: gras_if_RL/gras_if_SG (basic support for this)
740 [Code reorganisation]
741 - Get rid of libgrasutils.a since it makes more trouble than it solves.
742 Build examples against the RL library, since there is no way to disable
743 its creation for now.