1 SimGrid (3.0.1) unstable; urgency=low
4 * Unfortunately, I had missed 5 misnamed functions:
5 xbt_fifo_item_t xbt_fifo_newitem(void);
6 void xbt_fifo_freeitem(xbt_fifo_item_t);
7 xbt_fifo_item_t xbt_fifo_getFirstItem(xbt_fifo_t l);
8 xbt_fifo_item_t xbt_fifo_getNextItem(xbt_fifo_item_t i);
9 xbt_fifo_item_t xbt_fifo_getPrevItem(xbt_fifo_item_t i);
10 They're now deprecated. Please use their new versions:
11 xbt_fifo_item_t xbt_fifo_new_item(void);
12 void xbt_fifo_free_item(xbt_fifo_item_t);
13 xbt_fifo_item_t xbt_fifo_get_first_item(xbt_fifo_t l);
14 xbt_fifo_item_t xbt_fifo_get_next_item(xbt_fifo_item_t i);
15 xbt_fifo_item_t xbt_fifo_get_prev_item(xbt_fifo_item_t i);
17 * Bugfix: really disconnect fifo items which are remove_item()ed [AL]
18 * Doc: xbt_log module unmercifully reworked [MQ]
21 * Add addtionnal checkings on channel values in communicating functions.
24 * Actually implement gras_datadesc_copy() so that we don't have to mimick
25 RL communication on top of SG since it's so uneffective.
26 It may also be used for inter-thread communication in RL, one day. [MQ]
27 * Use gras_datadesc_copy() to exchange messages on top of SG [MQ]
29 - improve message exchange performance on top of SG
30 - deprecate transport_plugin_sg.c:gras_trp_sg_chunk_send() & recv()
34 SimGrid (3.00) stable; urgency=low
37 * New! Give the possibility to hijack the surf parser and thus bypass
38 MSG_create_environment and MSG_launch_application. Have a look at
39 examples/msg/msg_test_surfxml_bypassed.c to see how it can be done.
41 -- Arnaud Legrand <simgrid-devel@lists.gforge.inria.fr> Sat, 20 Aug 2005 23:25:25 -0700
43 SimGrid (2.96) unstable; urgency=low
48 * New! Exception handling with setjmp or such (code from OSSP ex) [MQ]
49 This deprecates the xbt_error_t mecanisms.
50 It modifies (simplifies) all XBT and GRAS API.
51 MSG API keeps unchanged (exceptions raised by XBT are catched from
52 within MSG and masked with existing error handling facilities)
55 * New! Add a FATPIPE model. [AL]
56 * New! Add a parallel task model. [AL]
57 * New! Add automatically a loopback interface (in the default
58 network model) if none was precised.
61 * Bugfix: MSG_process_resume now works with the current running process.
63 * New! Add MSG_parallel_task_create and MSG_parallel_task_execute. [AL]
64 * Modification of MSG_task_get_compute_duration. Once a task has been
65 processed, the value returned by this function is now equal to 0. [AL]
66 * New! Add double MSG_task_get_remaining_computation(m_task_t task) and
67 MSG_error_t MSG_task_cancel(m_task_t task). Add a state
68 (MSG_TASK_CANCELLED) to MSG_error_t corresponding to the cancelation
69 of a m_task. For now, MSG_task_cancel only works with computation
71 * New! Add double MSG_get_host_speed(m_host_t h) that returns the speed
72 of the processor (in Mflop/s) regardless of the current load on the
74 * API Change: use proper naming convention for MSG_getClock and
75 MSG_process_isSuspended: MSG_get_clock and MSG_process_is_suspended.
77 * New! Add void MSG_task_set_priority(m_task_t task, double priority).
78 This function changes the priority of a computation task. This priority
79 doesn't affect the transfer rate. A priority of 2 will make a task
80 receive two times more cpu power than the other ones. This function
81 has been added to suit the needs of Nguyen The Loc and hasn't been that
82 much tested yet. So if it fails, please report it and send me your code.
84 * API Change: removed all functions and types that were marked "deprecated"
85 since many months. Renamed MSG_global_init_args to MSG_global_init.
87 -- Da SimGrid team <simgrid-devel@lists.gforge.inria.fr> Mon, 8 Aug 2005 17:58:47 -0700
89 SimGrid (2.95) unstable; urgency=low
92 * Steal some nice code to GNU pth to fix context detection and usage [AL]
93 * Cleanup in the xbt_config API; add configuration callbacks. [MQ]
94 * Cleanup in the initialization API: the unused "defaultlog" is dead. [MQ]
97 * Bugfix: Allow absolute paths for platform description files [MQ]
98 * Bugfix: do free the variables after use. Leads to drastic performance
100 * Implement max_duration (ie, timeouts) on resources [AL]
103 * Implement MSG_config to configure MSG at runtime. xbt_cfg test on a real
105 * Implement MSG_channel_select_from() to help GRAS now that SURF provide
106 the needed support (timeouts) [AL]
109 * Implement measurement sockets. You can now get the bandwidth between two
110 hosts thanks to AMOK (see below). [MQ]
111 * gras_datadesc_dynar() builds a dynar type descriptor, allowing to send
112 dynar over the network (yeah) [MQ]
113 * Real (even if simplistic) implementation of gras_os_myname() on RL [MQ]
114 * simple/static token-ring example. [Alexandre Colucci and MQ]
115 * Use MSG_channel_select_from() and remove the *slow* hack we had to put
119 * Differentiate the types "char[22]" and "unsigned char[22]" in automatic
120 type parsing. "short" and "long" modifiers were also ignored; other
121 modifier (such as reference level) are still ignored. [MQ]
122 * Embeed the buffer size within the buffer itself on SG. [MQ]
123 That way, send() are atomic and cannot get intermixed anymore (at least
124 the ones which are less than 100k; bigger messages still have the issue)
125 * Array size pushed by the field, not by the field type (or each
126 and every long int will push stuff to the cbps) [MQ]
127 * use select() to sleep since it allows to portably sleep less than one
130 GRAS (minor cleanups)
131 * <project>.Makefile.local (generated from gras_stub_generator) |MQ]:
134 * Type Callbacks now receive the gras_datadesc_type_t they work on as argument.
135 * type category 'ignored' killed as it was never used and were difficult
137 * whether a type can cycle or not is now a flag, leaving room for more
138 flags (as "ignored", if we feel the need one day ;)
139 * Rename raw sockets to measurement sockets since "raw" has another
140 meaning in networking community.
143 * Advanced Metacomputing Overlay Kit introduction. It is based over GRAS
144 and offers features not belonging to GRAS but that most applications
145 need. One day, it may be a set of plugins loadable at runtime.
146 * New module: bandwidth
147 bandwidth measurement between arbitrary nodes running this module. [MQ]
149 -- Da SimGrid team <simgrid-devel@lists.gforge.inria.fr> Thu, 30 Jun 2005 16:29:20 -0700
151 SimGrid (2.94) unstable; urgency=low
153 The first beta release of SimGrid 3 !
157 * Update the main page and the FAQ. Adding references to gforge.
160 * Add a gras_os_getpid function.
163 * Add MSG_task_get_compute_duration() and MSG_task_get_data_size()
164 * Extend the logs so that they also print PID, hostname, date, ... if
166 * Convert the MSG example to the use of xbt_logs instead of PRINT_MESSAGE,
167 and kill the old version which were in testsuite/
168 * Rewrite tools/MSG_visualization/colorize.pl for using with logs instead
172 * Add xbt_os_time(). As the rest of xbt/portability, this is not public
173 for users. Instead, each programming environment (GRAS, MSG,...) use it
174 when needed to provide such a feature to users.
175 Don't shortcut the mecanism or you will also shortcut the virtualization
176 you need on the simulator.
180 * Cleanups in configury with regard to compile optimization/warning flags.
181 Also add -fno-loop-optimize to any powerpc since it's the optimization
182 killing gcc (< 3.4.0).
183 * Doxygen cleanups: move MSG examples, kill the second Doxygen phase
184 needed by MSG examples complications
185 * Borrow configury beautifications from PHP
188 * Bugfix: XBT_LOG_NEW_DEFAULT_CATEGORY now compiles without compiler
189 warning (thanks loris for stumbling into this one).
190 * Bugfix: stop loading private headers (gras_config.h) from the public
194 * Change SIMGRID_INSTALL_PATH to GRAS_ROOT in Makefiles generated for user.
195 * Rename gras_get_my_fqdn to gras_os_myname and implement it in the simulator
196 RL would imply a DNS resolver, which is *hard* to do in a portable way
197 (and therefore delayed).
198 * Implement a real timer mecanism and use it in timing macros. This allows
199 to avoid rounding errors and get a 0.000005 sec precision in timing
200 macros. While I was at it, various cleanups:
201 - allow to declare more than one timed section per file (fix a stupid bug)
202 - move some private declaration to the right place
203 - merge conditional execution and timing macros into emulation module
204 - document the module
205 - make sure the module cleanups its mess on gras_exit
206 * Documentation improvements:
207 - (new) how to compile applications using GRAS
208 - (new) emulation support (timing macros)
210 -- Da SimGrid team <simgrid-devel@lists.gforge.inria.fr> Fri, 13 May 2005 10:49:31 +0200
212 SimGrid (2.93) unstable; urgency=low
214 Alpha 4 on the path to SimGrid 3 (aka the "neuf-trois" version)
217 - Use Paje properly where used. Still to be sanitized properly.
218 - Portability fix: Add an implementation of the contexts using pthread
222 - Add xbt_procname(): returns the name of the current process.
223 Use it to show the current process's name in all logging.
225 - fix detection of older flex version and the reaction, since we do
226 depend on modern ones (we use lex_destroy)
227 - Better separation of SG and RL in the libs: remove all simulation code
228 from libgras. As a result, this lib is now only 200k when stripped.
229 Some of the xbt modules may also be duplicated (two sets and such) and
230 should be cleaned/killed before SG3.
231 - Insist on using xlC on AIX because of weird problems involving gcc there.
232 - Cleanup the make remote stuff. This is now done by scripts
233 tools/graspe-{master,slave} (GRAS Platform Expender). This is still
234 mainly for our private use, but we're working on changing them to user
237 - Bugfix: flush the socket on close only if there is some *output*.
238 - Bugfix: flush idempotent when there's nothing to send (don't send size=0)
240 - Add MSG_task_get_name. The task names are mainly for debugging purpose,
243 -- SimGrid team <simgrid2-users@listes.ens-lyon.fr> Fri, 4 Mar 2005 14:32:37 -0800
245 SimGrid (2.92) unstable; urgency=low
247 Alpha 3 on the path to SimGrid 3
251 - New! First try of benchmarking macros.
252 - New! First try so that gras_stub_generator generate deployment and
253 remote compilation helpers.
255 - Bugfix: Initialization fix in msg_test.
259 - Bugfix: applied patch to lexer so that it doesn't need a huge heap.
261 - Bugfix: let dicts work with NULL content (_foreach didn't) and cleanups
263 - API Change: gras_os_sleep to take the amount of seconds as a double.
264 Accepting an int was error prone since it was the only location where
265 seconds were coded as such. It leaded to damn rounding errors.
266 - Bugfix: Hard to belive that timers ever worked before this.
268 -- SimGrid team <simgrid2-users@listes.ens-lyon.fr> Wed, 23 Feb 2005 22:09:21 +0100
270 SimGrid (2.91) unstable; urgency=low
272 Alpha 2 on the path to SimGrid 3
276 - Bug fix in the lmm_solver.
278 - New! Interface to Paje (see http://www-id.imag.fr/Logiciels/paje/)
279 through the function MSG_paje_output.
280 - New! Introducing two new functions MSG_process_kill() and MSG_process_killall().
281 - It is possible to bound the rate of a communication in MSG with
282 MSG_task_put_bounded() (was already in the previous version but I had forgotten
283 to write it in the changelog).
284 - Bug fix to let GRAS run on top of MSG until we move it directly on top
289 - Various cleanups to the autotools stuff
290 - Begin to move Gras examples to examples/gras/
291 - Let make distcheck work again (yeah!)
293 - documentation overhauled using doxygen.
294 gtk-doc-tools is dead in SimGrid now.
295 - Automatically extract all existing logging categories, and add the list
296 to the documentation (long standing one, to say the less)
298 - Cleanup the known architecture table. Reorder the entries to group what
299 should be, and use a more consistent naming scheme.
300 (some of the test dataset are still to be regenerated)
301 - New! Allow library to register globals on each process just as userdata
303 This is implemented using a xbt_dict and not a xbt_set, so we loose the
304 lookup time (for now).
305 Use it in msg and trp.
306 This cleans a lot the internals and helps enforcing privacy of the
307 headers between the gras components.
308 - New! Add a timer mechanism, not unlike cron(8) and at(1).
309 - Bugfix: gras_os_time was delirious in RL.
310 - Bugfix: gras_trp_select/RL don't run into the wall when asked to select
312 - Reenable GRAS now that it works.
314 -- Arnaud Legrand <Arnaud.Legrand@imag.fr> Mon, 14 Feb 2005 14:02:13 -0800
316 SimGrid (2.90) unstable; urgency=low
318 Alpha 1 on the path to SimGrid 3
320 * It is a long time since the last release of SimGrid. I'm sorry about
321 that but as I had told you, I was rewriting a lot of things. I apologize
322 to those who had been reporting bugs to me and that I had not answered.
323 If your bug is still in the new version, please tell me. Here is a
324 summary of the main changes.
326 * REVOLUTION 1: The SimGrid project has merged with the GRAS project
327 lead by Martin Quinson. As a consequence SimGrid gains a lot in
328 portability, speed, and a lot more but you'll figure it out later.
329 SimGrid now comprises 3 different projects : MSG, GRAS and SMPI.
330 I wanted to release the new MSG as soon as possible and I have
331 broken GRAS, which is the reason why, for now, only MSG is fully
332 functional. A laconic description of these projects is available
333 in the documentation.
335 * REVOLUTION 2: I have removed SG and I am now using a new simulation
336 kernel optimized for our needs (called SURF but only the developers
337 should use it). Hence, MSG is now roughly 30 times faster and I think
338 that by rewriting a little bit MSG, I could event speed it up a little
339 bit more. Beside the gain in speed, it is also much easier to encode a
340 new platform model with SURF than it was with SG. More to come...
342 * REVOLUTION 3: I have tried to change a little as possible the API of
343 MSG but a few things really had to disappear. The main differences
344 with the previous version are :
345 1) no more m_links_t and the corresponding functions. Platforms are
346 directly read from a XML description and cannot be hard-coded
347 anymore. The same format is used for application deployment
348 description. The new format is described in the documentation.
349 Have a look in tools/platform_generation. There is a tiny script
350 that converts from the old platform format to the new one. Concerning
351 the application deployment format, parsing the old one is tricky.
352 I think most of you should however be able to convert your files. If
353 it is really an issue, I can write a C code that does the conversion.
355 2) the toolbox tbx does not exist anymore. We now have a library
356 with much more data-structures but without the hash-tables (we have
357 dictionaries that are much faster).
359 -- Arnaud Legrand <Arnaud.Legrand@imag.fr> Mon, 31 Jan 2005 10:45:53 -0800
361 *****************************************************************************
362 * Follows the old GRAS changelog. It does not follow the same syntax, but I *
363 * don't feel like converting the oldies. (Mt) *
364 *****************************************************************************
367 Version 2.90: "the long awaited one"
368 - Finished rewriting and debugging MSG. Rewrote the documentation.
369 - disable GRAS for now since it needs to be ported to the newest SG
372 - Finish the port to windows (using mingw32 for cross-compile)
375 - Main loop and datastructures of SURF. A cpu resource object is
376 functional. Surf can thus be used to create cpu's with variable
377 performance on which you can execute some actions.
379 2004-11-15 Martin Quinson
380 - Port to ARM. Simply added the alignment and size descriptions. Should
381 work, but the ARM machines are so slow that I didn't had the opportunity
382 to 'make check' over there yet.
384 2004-11-15 Arnaud Legrand
385 - Trace manager now written. It uses a heap structure and is therefore
386 expected to be efficient. It may however be speeded up (particularly
387 when many events occur at the same date) by using red and black
388 trees. One day maybe...
389 - Max-min linear system solver written. It uses a sparse matrix
390 structure taking advantage of its expected use. Most operations are
391 O(1) and free/calloc are called as few as possible. The computation of
392 the minimum could however be improved by using a red and black tree
395 2004-11-03 Arnaud Legrand
396 - Rename every gras_* function that was in xbt/ to its xbt_
398 - Add a heap and a doubly-linked list to xbt
399 - Added a dichotomy to the dictionaries. make check works as well before
400 so I assume that the patch is correct. I do not know however if things
401 run effectively faster than before now. :)
403 Inclusion of the SimGrid tree in the GRAS one. The archive is renamed to
404 SimGrid, and the version number is bumped to 2.x
406 2004-10-29 Martin Quinson
407 - Introduction of the remote errors.
408 They are the result of a RMI/RPC on the remote machine.
409 ErrCodes being scalar values, you can't get the host on which those
410 errors did happen. Extending the error mechanism as in Gnome is possible.
411 No idea yet whether it is a good idea.
413 2004-10-28 Martin Quinson
414 - Interface revolution: the Starred Structure Eradication.
415 I used to do typedef struct {} toto_t; and then handle *toto_t.
416 Arnaud (and Oli) didn't like it, and I surrendered. Now, you have:
417 - ???_t is a valid type (builded with typedef)
418 - s_toto_t is a structure (access to fields with .)
419 - s_toto is a structure needing 'struct' keyword to be used
420 - e_toto_t is an enum
421 - toto_t is an 'object' (struct*)
423 typedef struct s_toto {} s_toto_t, *toto_t;
424 typedef enum {} e_toto_t;
425 Moreover, only toto_t (and e_toto_t) are public. The rest (mainly
426 s_toto_t) is private.
428 - While I was at it, all gras_<obj>_free() functions want a gras_<obj>_t*
429 so that it can set the variable to NULL. It was so for dicts and sets,
430 it changed for dynars.
432 - Fix a bunch of memleaks in dict_remove
433 - Fix a bug in sg/server_socket opening: it failed all the time.
435 2004-10-07 Martin Quinson
436 - Speed up dynar lookup operation a bit.
438 gras_dynar_get is dead.
440 Now, you can choose between gras_dynar_get_cpy (the old gras_dynar_get
441 but should be avoided for efficiency reasons) and gras_dynar_get_ptr
442 (which gives you the address of the stored data).
444 gras_dynar_get_as is an helpful macro which allows you to retrieve a
445 copy of the data using an affectation to do the job and not a memcpy.
447 int toto = gras_dynar_get_as(dyn,0,int); rewrites itself to
448 int toto = *(int*)gras_dynar_get_ptr(dyn,0);
450 It does not really speedup the dynar test because they are
451 setting elements all the time (and look them seldom). But the dict does
452 far more lookup than setting.
454 So, this brings the dict_crash test from ~33s to ~25s (200000 elms).
456 2004-10-05 Martin Quinson
457 - Allow to (en/dis)able the cycle detection at run time.
459 Whether we should check for cycle or not is now a property of each
460 datatype. When you think there may be some cycle, use datadesc_cycle_set.
461 datadesc_cycle_unset allow to remove this property when previously set.
463 Note that the cycle detection is off by default since it impacts the
464 performance. Watch the data you feed GRAS with ;)
466 This property is hereditary. Any element embedded in a structure having it
467 set have it set for the time of this data exchange.
469 You should set it both on sender and receiver side. If you don't set it on
470 sender side, it will enter an endless loop. If you forget on receiver
471 side, the cycles won't be recreated after communication.
473 - Header reorganization.
474 Kill gras_private.h, each submodule must load the headers it needs.
476 2004-10-04 Martin Quinson
477 - Interface revolution: do not try to survive to malloc failure.
479 Now, gras_malloc and friends call gras_abort() on failure.
480 As a conclusion, malloc_error is not a valid error anymore, and all
481 functions for which it was the only gras_error_t return value are
482 changed. They now return void, or there result directly.
483 This simplify the API a lot.
485 2004-09-29 Martin Quinson
486 - Re-enable raw sockets.
487 Created by gras_socket_{client,server}_ext;
488 Used with gras_raw_{send,recv}
491 It should allow to kill the last bits of gras first version soon.
493 This is not completely satisfactory yet (duplicate code with
494 chunk_{send,recv}; a bit out of the plugin mechanism), but it should
497 - Simplify transport plugin (internal) interface by not passing any
498 argument to _server and _client, but embedding them in the socket
501 2004-09-28 Martin Quinson
502 - Finish the port to AIX.
503 autoconf was my problem (segfault within the malloc replacement
504 function. No idea why)
506 2004-09-16 Martin Quinson
507 - Fix some size_t madness on 64bit architectures.
509 2004-09-08 Martin Quinson
510 - Reduce the number of system headers loaded, overload some more system
511 calls (such as malloc to cast the result of the system one, and work
513 - Fix and reintroduce the config support
515 2004-09-07 Martin Quinson
516 - Source code reorganization to allow Arnaud to surf all over there.
517 - Allow to document the logging categories.
518 - Remove all uppercase from logging categories and useless cleanup in names.
520 2004-08-18 Martin Quinson
521 Version 0.6.2 (protocol not changed; API changed)
522 - Interface cleanup: gras_msgtype_by_name returns the type (instead of a
523 gras_error_t), and NULL when not found. Functions expecting a msgtype
524 as argument (msg_wait; msg_send) deal with NULL argument by providing a
525 hopefully usefull message.
526 - Portability to prehistoric sparcs again
528 2004-08-17 Martin Quinson
529 Version 0.6.1 (protocol not changed; ABI not changed)
530 - prealloc some buffers to speed things up
532 2004-08-11 Martin Quinson
533 Version 0.6 (protocol not changed; ABI expended)
534 - The parsing macro can deal with the references, provided that you add
535 the relevant annotations (using GRAS_ANNOTE(size,field_name))
537 2004-08-09 Martin Quinson
538 Version 0.5 (protocol not changed; ABI changed)
539 - Allow to off turn the cycle detection code in data exchange at
540 compilation time. It should be at run time, but I'm short of time (and
541 the config stuff is still broken). That way, we keep dict out of the
542 critical path, which is good because the performance is poor:
543 - search not dichotomial yet
544 - dynar give no way to access their content and memcpy everytime
545 - In composed data description (struct, ref and so on), stop foolness of
546 keeping the subtype's ID, but store the type itself. This keeps sets out
547 of the critical path, which is good since they rely on dynar and
548 dictionnaries. The only loose of that is that we cannot detect the
549 redeclaration of a structure/union with another content (but I'm not sure
550 the code detected well this error before anyway). We still can detect
551 the redefinition discrepancy for the other types.
552 - Use a whole bunch of optimisation flags (plus -fno-strict-aliasing since
553 it breaks the code because of type-punning used all over the place).
554 This breaks on all non-gcc architectures (for now).
556 All those changes (plus the buffer of last time) allow me to gain 2 order
557 of magnitude on cruel tests consisting of 800000 array of integers on two
558 level of a hierarchical structure (200 secondes -> 4 secondes)
561 - the selector of reference must now return the type it points to, not
564 2004-08-06 Martin Quinson
565 Version 0.4 (protocol changed; ABI not changed)
566 - Allow to pass --gras-log argument to processes in simulation mode. Really.
567 - New debugging level: trace (under debug) to see effect of GRAS_IN/OUT
568 - Implement a buffer transport, and use it by default (it relies on tcp in
569 real life and on sg in simulation).
570 That's a bit hackish since I had a new field to the structure to store
571 its data without interfering with the subtype ones. Inheritance
572 is tricky in C. And that's a kind of reverse inheritance with one class
573 derivating two classes. Or maybe a game with java interfaces. Anyway,
574 that's damn hard in C (at least).
575 Moreover, I got tired while trying to ensure plugin separation and
576 genericity in SG mode. MSG wants me to do weird things, so let's go for
577 cruel hacks (temporarily of course ;).
578 See comment in transport_private.h:71
579 - do not include all the _interface headers in private but in the files
580 which really need them (to cut the compilation time when they are
583 2004-07-26 Martin Quinson
584 Version 0.3 (protocol not changed; ABI changed)
585 - Major overhault of the datadesc interface to simplify it:
586 - shorted the function names:
587 s/gras_datadesc_declare_struct/gras_datadesc_struct/ and so on
588 - add a trivial way to push/pop integers into the cbps without malloc.
589 This allows to make really generic sub_type description, which simply
590 pop their size of the stack.
591 - add a function gras_datadesc_ref_pop_arr() which does what users want
592 most of the time: Declare a dynamic array (which pops its size of the
593 stack) and declare a reference to it. Poor name, but anyway.
594 - kill the post-send callback, add a post-receive one
596 2004-07-23 Martin Quinson
597 Version 0.2 (protocol changed; ABI changed)
598 - add some testing for cpbs in the test cases, and fix some more bugs.
599 This invalidate again the little64 data file, since I cannot regenerate
601 - remove an awfull optimization in the logging stuff, allowing me to:
602 - understand it again
603 - learn gcc how to check that the argument match the provided format
604 - fix all errors revealed by gcc after that
605 - internal keys of dict are not \0 terminated. Deal with it properly in
606 loggings instead of segfaulting when the user want to see the logs :-/
608 2004-07-22 Martin Quinson
609 - Fix some stupid bug preventing cbps (callback postit) from working
611 2004-07-21 Martin Quinson
612 - Some documentation cleanups
613 - remove the useless last argument of msgtype_declare
614 - rename the Virtu functions to fit into the 'os' namespace
615 - move headers src/include -> src/include/gras/ and stop fooling with
616 gras -> . symbolic link
617 - make distcheck is now successful
619 2004-07-19 Martin Quinson
621 - Build shared library also
622 - Install html doc to the right location
623 - stop removing maintainer files in make clean
624 - build tests only on make check
626 2004-07-13 Martin Quinson
628 - No major issue in previous version => change versionning schema
629 - Re-enable little64 convertion test now that Abdou kindly regenerated the
630 corresponding dataset.
632 2004-07-11 Martin Quinson
634 - Get it working with any kind of structure (we can compute the padding
635 bytes remotely for all the architectures I have access to)
636 - Implement the structure parsing macro (still not quite robust/complete)
637 - Improvement to the remote testing toysuite
639 2004-07-10 Martin Quinson
641 - get ride of a bunch of deprecated macros
642 - actually run the test for two-compliment, not only compile it :-/
643 - test whether the structures get packed (and bail out if yes. Damn.
644 Alignment is a serious matter)
645 - test whether the structures get compacted (but respecting the alignment
646 constraints of each types)
647 - test whether the array fields of structures can straddle alignment boundaries
649 - Damnit, double are bigger than float (typo in creation of 'double' datadesc)
650 (took me 2 hours to find that bug, looking at the wrong place)
651 - Add gras_datadesc_declare_{union,struct}_close(). They must be used
652 before sending/receiving and are used to compute the offsets of fields
653 - Given that padding size depend even on compiler options, keep track of
654 alignment and aligned_size only for the current architecture. Not a big
655 deal since we send structure fields one after the other (seems
657 - Add the datastructure used for IEEE paper by the PBIO guys to the test
658 program, let it work on linux/gcc/little32. portability todo.
660 2004-07-08 Martin Quinson
661 - import and improve remote compilation support from FAST
662 - make sure make check works on half a dozen of machines out there
664 2004-07-07 Martin Quinson
665 Let's say it's version 0.0.3 ;)
666 - Implement conversions (yuhu!)
667 - Let it work on solaris (beside conversion, of course)
668 - Stupid me, using rand() to generate the conversion datatests in not wise.
670 2004-07-06 Martin Quinson
671 - Let make dist work, since I'm gonna need it to compile on remote hosts
672 - Let Tests/datadesc_usage write the architecture on which the file was
673 generated as first byte.
674 - Add PowerPC (being also IRIX64), SPARC (also power4) and ALPHA
675 architecture descriptions.
676 - Add datadesc_usage.{i386,ppc,sparc} files being the result of execution
677 on those architectures.
678 - Optimization: send/recv array of scalar in one shoot
680 2004-07-05 Martin Quinson
681 - YEAH! GRAS/SG and GRAS/RL are both able to run the ping example !
683 - Plug a whole bunch of memleaks
684 - each process now have to call gras_{init,exit}. One day, their log
685 settings will be separated
686 - Continue the code factorisation between SG, RL and common in Transport.
688 2004-07-04 Martin Quinson
690 - Redistribution between SG and RL.
691 We wanna have to accept in SG, so move accepted related parts of RL in
692 the common part. (more precisely, the dynar of all known sockets is no
693 more a static in transport.c, but part of the process_data)
695 [gras_stub_generator]
696 - Bug fix: Do call gras_process_init from gras_init (wasnt called in RL).
698 2004-07-03 Martin Quinson
699 - Create a new log channel tbx containing dict, set, log, dynar (to shut
700 them all up in one shot)
702 - Fix the ugly case of reference to dynamic array.
703 - New (semi-public) function gras_datadesc_size to allow the messaging
704 layer to malloc the needed space for the buffer.
706 - gras_socket_close now expect the socket to close (and not its address to
707 put NULL in it after it). This is because the socket passed to handlers
708 is one of their argument (=> not writable).
710 - propagate the interface cleanup from last week in datadesc, ie remove a
711 superfluous level of indirection. User pass adress of variable
712 containing data (both when sending and receiving), and not of a variable
713 being a pointer to the data. Let's say that I like it better ;)
714 The price for that is constructs like "int msg=*(int*)payload" in
715 handlers, but it's a fine price, IMHO.
717 - Let it work in RL (yuhu)
719 2004-06-21 Martin Quinson
721 - porting SG plugin and SG select to new standards (works almost).
722 - plug memleaks and fix bugs around.
725 - cleanup the prototype of data recv and force users to specify when they
726 want to handle references to objects. Test case working even for cycles.
727 - plug memleaks. Valgrind is perfectly ok with this.
729 2004-06-12 Martin Quinson
731 - cleanup the separation between plugin and main code in plugin creation
733 2004-06-11 Martin Quinson
735 - Reput hook for raw sockets, needed for BW experiments
736 - kill a few lines of dead code
737 [Data description] Interface cleanup
738 - gras_datadesc_by_name returns the searched type or NULL.
739 That way, no variable is needed to use a type desc once, which makes
741 - gras_datadesc_declare_[struct|union]_append_name is removed. The last
742 two parameters were strings (field name, type name), leading to
744 [Dicos] Interface cleanup
745 - gras_dico_retrieve -> gras_dico_get ; gras_dico_insert -> gras_dico_set
746 This is consistant with the dynar API.
748 2004-04-21 Martin Quinson
750 - Porting to new standards.
753 There is no bag anymore, no need to take extra provision to mask the
754 pointers behind "ID".
755 Better splitup of functions between files create/exchange/convert.
756 This is still a bit artificial since convert and receive are so
757 interleaved, but anyway.
759 - add a queued message list to procdata (the ones not matching criteria
761 - factorize some more code between SG and RL wrt procdata
763 - use gras_exit in example to track memleaks
764 - get rid of gs_example now that GS is properly integrated into gras
765 - update run_test to integrate the lastest tests (datadesc)
767 - rename WARNINGn macros to WARNn since it prooved error-prone
769 2004-04-19 Martin Quinson
771 - register init/exit functions within gras module mechanism
772 - send/receive function.
773 Convertion is not implemented, but short-cutted if not needed.
774 struct/array elements are sent one by one (instead of block-wise), but
775 nobody really cares (yet). Get a prototype before optimizing.
776 - tests (using a file socket) for DD send/receive on:
777 - base types: int, float
778 - array: fixed size, string (ie ref to dynamic string)
779 - structure: homogeneous, heterogeneous
780 - chained list, graph with cycle
781 Believe it or not, valgrind is not too unhappy with the results. The
782 cycle happily segfaults, but the others are ok. And I'm sick of pointers
786 - Bugfix when using a filename explicitely (instead of '-')
788 2004-04-09 Martin Quinson
790 - factorize more code between RL and SG in socket creation
791 - Complete the implementation and tests of:
793 o file (only in RL, and mainly for debugging)
795 I lost 3 days to design a portable address resolver, and then decided
796 that the prototype mainly have to run on my box.
797 Addressing portability too early may be like optimizing too early :-/
799 - use gras_init in the Tests instead of the crappy parse_log_opt
800 (the latter function is removed)
801 [Conditional execution]
802 - New functions: gras_if_RL/gras_if_SG (basic support for this)
803 [Code reorganisation]
804 - Get rid of libgrasutils.a since it makes more trouble than it solves.
805 Build examples against the RL library, since there is no way to disable
806 its creation for now.