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);
19 * Add addtionnal checkings on channel values in communicating functions.
22 Documentation improvements:
23 * xbt_log module unmercifully reworked [MQ]
27 SimGrid (3.00) stable; urgency=low
30 * New! Give the possibility to hijack the surf parser and thus bypass
31 MSG_create_environment and MSG_launch_application. Have a look at
32 examples/msg/msg_test_surfxml_bypassed.c to see how it can be done.
34 -- Arnaud Legrand <simgrid-devel@lists.gforge.inria.fr> Sat, 20 Aug 2005 23:25:25 -0700
36 SimGrid (2.96) unstable; urgency=low
41 * New! Exception handling with setjmp or such (code from OSSP ex) [MQ]
42 This deprecates the xbt_error_t mecanisms.
43 It modifies (simplifies) all XBT and GRAS API.
44 MSG API keeps unchanged (exceptions raised by XBT are catched from
45 within MSG and masked with existing error handling facilities)
48 * New! Add a FATPIPE model. [AL]
49 * New! Add a parallel task model. [AL]
50 * New! Add automatically a loopback interface (in the default
51 network model) if none was precised.
54 * Bugfix: MSG_process_resume now works with the current running process.
56 * New! Add MSG_parallel_task_create and MSG_parallel_task_execute. [AL]
57 * Modification of MSG_task_get_compute_duration. Once a task has been
58 processed, the value returned by this function is now equal to 0. [AL]
59 * New! Add double MSG_task_get_remaining_computation(m_task_t task) and
60 MSG_error_t MSG_task_cancel(m_task_t task). Add a state
61 (MSG_TASK_CANCELLED) to MSG_error_t corresponding to the cancelation
62 of a m_task. For now, MSG_task_cancel only works with computation
64 * New! Add double MSG_get_host_speed(m_host_t h) that returns the speed
65 of the processor (in Mflop/s) regardless of the current load on the
67 * API Change: use proper naming convention for MSG_getClock and
68 MSG_process_isSuspended: MSG_get_clock and MSG_process_is_suspended.
70 * New! Add void MSG_task_set_priority(m_task_t task, double priority).
71 This function changes the priority of a computation task. This priority
72 doesn't affect the transfer rate. A priority of 2 will make a task
73 receive two times more cpu power than the other ones. This function
74 has been added to suit the needs of Nguyen The Loc and hasn't been that
75 much tested yet. So if it fails, please report it and send me your code.
77 * API Change: removed all functions and types that were marked "deprecated"
78 since many months. Renamed MSG_global_init_args to MSG_global_init.
80 -- Da SimGrid team <simgrid-devel@lists.gforge.inria.fr> Mon, 8 Aug 2005 17:58:47 -0700
82 SimGrid (2.95) unstable; urgency=low
85 * Steal some nice code to GNU pth to fix context detection and usage [AL]
86 * Cleanup in the xbt_config API; add configuration callbacks. [MQ]
87 * Cleanup in the initialization API: the unused "defaultlog" is dead. [MQ]
90 * Bugfix: Allow absolute paths for platform description files [MQ]
91 * Bugfix: do free the variables after use. Leads to drastic performance
93 * Implement max_duration (ie, timeouts) on resources [AL]
96 * Implement MSG_config to configure MSG at runtime. xbt_cfg test on a real
98 * Implement MSG_channel_select_from() to help GRAS now that SURF provide
99 the needed support (timeouts) [AL]
102 * Implement measurement sockets. You can now get the bandwidth between two
103 hosts thanks to AMOK (see below). [MQ]
104 * gras_datadesc_dynar() builds a dynar type descriptor, allowing to send
105 dynar over the network (yeah) [MQ]
106 * Real (even if simplistic) implementation of gras_os_myname() on RL [MQ]
107 * simple/static token-ring example. [Alexandre Colucci and MQ]
108 * Use MSG_channel_select_from() and remove the *slow* hack we had to put
112 * Differentiate the types "char[22]" and "unsigned char[22]" in automatic
113 type parsing. "short" and "long" modifiers were also ignored; other
114 modifier (such as reference level) are still ignored. [MQ]
115 * Embeed the buffer size within the buffer itself on SG. [MQ]
116 That way, send() are atomic and cannot get intermixed anymore (at least
117 the ones which are less than 100k; bigger messages still have the issue)
118 * Array size pushed by the field, not by the field type (or each
119 and every long int will push stuff to the cbps) [MQ]
120 * use select() to sleep since it allows to portably sleep less than one
123 GRAS (minor cleanups)
124 * <project>.Makefile.local (generated from gras_stub_generator) |MQ]:
127 * Type Callbacks now receive the gras_datadesc_type_t they work on as argument.
128 * type category 'ignored' killed as it was never used and were difficult
130 * whether a type can cycle or not is now a flag, leaving room for more
131 flags (as "ignored", if we feel the need one day ;)
132 * Rename raw sockets to measurement sockets since "raw" has another
133 meaning in networking community.
136 * Advanced Metacomputing Overlay Kit introduction. It is based over GRAS
137 and offers features not belonging to GRAS but that most applications
138 need. One day, it may be a set of plugins loadable at runtime.
139 * New module: bandwidth
140 bandwidth measurement between arbitrary nodes running this module. [MQ]
142 -- Da SimGrid team <simgrid-devel@lists.gforge.inria.fr> Thu, 30 Jun 2005 16:29:20 -0700
144 SimGrid (2.94) unstable; urgency=low
146 The first beta release of SimGrid 3 !
150 * Update the main page and the FAQ. Adding references to gforge.
153 * Add a gras_os_getpid function.
156 * Add MSG_task_get_compute_duration() and MSG_task_get_data_size()
157 * Extend the logs so that they also print PID, hostname, date, ... if
159 * Convert the MSG example to the use of xbt_logs instead of PRINT_MESSAGE,
160 and kill the old version which were in testsuite/
161 * Rewrite tools/MSG_visualization/colorize.pl for using with logs instead
165 * Add xbt_os_time(). As the rest of xbt/portability, this is not public
166 for users. Instead, each programming environment (GRAS, MSG,...) use it
167 when needed to provide such a feature to users.
168 Don't shortcut the mecanism or you will also shortcut the virtualization
169 you need on the simulator.
173 * Cleanups in configury with regard to compile optimization/warning flags.
174 Also add -fno-loop-optimize to any powerpc since it's the optimization
175 killing gcc (< 3.4.0).
176 * Doxygen cleanups: move MSG examples, kill the second Doxygen phase
177 needed by MSG examples complications
178 * Borrow configury beautifications from PHP
181 * Bugfix: XBT_LOG_NEW_DEFAULT_CATEGORY now compiles without compiler
182 warning (thanks loris for stumbling into this one).
183 * Bugfix: stop loading private headers (gras_config.h) from the public
187 * Change SIMGRID_INSTALL_PATH to GRAS_ROOT in Makefiles generated for user.
188 * Rename gras_get_my_fqdn to gras_os_myname and implement it in the simulator
189 RL would imply a DNS resolver, which is *hard* to do in a portable way
190 (and therefore delayed).
191 * Implement a real timer mecanism and use it in timing macros. This allows
192 to avoid rounding errors and get a 0.000005 sec precision in timing
193 macros. While I was at it, various cleanups:
194 - allow to declare more than one timed section per file (fix a stupid bug)
195 - move some private declaration to the right place
196 - merge conditional execution and timing macros into emulation module
197 - document the module
198 - make sure the module cleanups its mess on gras_exit
199 * Documentation improvements:
200 - (new) how to compile applications using GRAS
201 - (new) emulation support (timing macros)
203 -- Da SimGrid team <simgrid-devel@lists.gforge.inria.fr> Fri, 13 May 2005 10:49:31 +0200
205 SimGrid (2.93) unstable; urgency=low
207 Alpha 4 on the path to SimGrid 3 (aka the "neuf-trois" version)
210 - Use Paje properly where used. Still to be sanitized properly.
211 - Portability fix: Add an implementation of the contexts using pthread
215 - Add xbt_procname(): returns the name of the current process.
216 Use it to show the current process's name in all logging.
218 - fix detection of older flex version and the reaction, since we do
219 depend on modern ones (we use lex_destroy)
220 - Better separation of SG and RL in the libs: remove all simulation code
221 from libgras. As a result, this lib is now only 200k when stripped.
222 Some of the xbt modules may also be duplicated (two sets and such) and
223 should be cleaned/killed before SG3.
224 - Insist on using xlC on AIX because of weird problems involving gcc there.
225 - Cleanup the make remote stuff. This is now done by scripts
226 tools/graspe-{master,slave} (GRAS Platform Expender). This is still
227 mainly for our private use, but we're working on changing them to user
230 - Bugfix: flush the socket on close only if there is some *output*.
231 - Bugfix: flush idempotent when there's nothing to send (don't send size=0)
233 - Add MSG_task_get_name. The task names are mainly for debugging purpose,
236 -- SimGrid team <simgrid2-users@listes.ens-lyon.fr> Fri, 4 Mar 2005 14:32:37 -0800
238 SimGrid (2.92) unstable; urgency=low
240 Alpha 3 on the path to SimGrid 3
244 - New! First try of benchmarking macros.
245 - New! First try so that gras_stub_generator generate deployment and
246 remote compilation helpers.
248 - Bugfix: Initialization fix in msg_test.
252 - Bugfix: applied patch to lexer so that it doesn't need a huge heap.
254 - Bugfix: let dicts work with NULL content (_foreach didn't) and cleanups
256 - API Change: gras_os_sleep to take the amount of seconds as a double.
257 Accepting an int was error prone since it was the only location where
258 seconds were coded as such. It leaded to damn rounding errors.
259 - Bugfix: Hard to belive that timers ever worked before this.
261 -- SimGrid team <simgrid2-users@listes.ens-lyon.fr> Wed, 23 Feb 2005 22:09:21 +0100
263 SimGrid (2.91) unstable; urgency=low
265 Alpha 2 on the path to SimGrid 3
269 - Bug fix in the lmm_solver.
271 - New! Interface to Paje (see http://www-id.imag.fr/Logiciels/paje/)
272 through the function MSG_paje_output.
273 - New! Introducing two new functions MSG_process_kill() and MSG_process_killall().
274 - It is possible to bound the rate of a communication in MSG with
275 MSG_task_put_bounded() (was already in the previous version but I had forgotten
276 to write it in the changelog).
277 - Bug fix to let GRAS run on top of MSG until we move it directly on top
282 - Various cleanups to the autotools stuff
283 - Begin to move Gras examples to examples/gras/
284 - Let make distcheck work again (yeah!)
286 - documentation overhauled using doxygen.
287 gtk-doc-tools is dead in SimGrid now.
288 - Automatically extract all existing logging categories, and add the list
289 to the documentation (long standing one, to say the less)
291 - Cleanup the known architecture table. Reorder the entries to group what
292 should be, and use a more consistent naming scheme.
293 (some of the test dataset are still to be regenerated)
294 - New! Allow library to register globals on each process just as userdata
296 This is implemented using a xbt_dict and not a xbt_set, so we loose the
297 lookup time (for now).
298 Use it in msg and trp.
299 This cleans a lot the internals and helps enforcing privacy of the
300 headers between the gras components.
301 - New! Add a timer mechanism, not unlike cron(8) and at(1).
302 - Bugfix: gras_os_time was delirious in RL.
303 - Bugfix: gras_trp_select/RL don't run into the wall when asked to select
305 - Reenable GRAS now that it works.
307 -- Arnaud Legrand <Arnaud.Legrand@imag.fr> Mon, 14 Feb 2005 14:02:13 -0800
309 SimGrid (2.90) unstable; urgency=low
311 Alpha 1 on the path to SimGrid 3
313 * It is a long time since the last release of SimGrid. I'm sorry about
314 that but as I had told you, I was rewriting a lot of things. I apologize
315 to those who had been reporting bugs to me and that I had not answered.
316 If your bug is still in the new version, please tell me. Here is a
317 summary of the main changes.
319 * REVOLUTION 1: The SimGrid project has merged with the GRAS project
320 lead by Martin Quinson. As a consequence SimGrid gains a lot in
321 portability, speed, and a lot more but you'll figure it out later.
322 SimGrid now comprises 3 different projects : MSG, GRAS and SMPI.
323 I wanted to release the new MSG as soon as possible and I have
324 broken GRAS, which is the reason why, for now, only MSG is fully
325 functional. A laconic description of these projects is available
326 in the documentation.
328 * REVOLUTION 2: I have removed SG and I am now using a new simulation
329 kernel optimized for our needs (called SURF but only the developers
330 should use it). Hence, MSG is now roughly 30 times faster and I think
331 that by rewriting a little bit MSG, I could event speed it up a little
332 bit more. Beside the gain in speed, it is also much easier to encode a
333 new platform model with SURF than it was with SG. More to come...
335 * REVOLUTION 3: I have tried to change a little as possible the API of
336 MSG but a few things really had to disappear. The main differences
337 with the previous version are :
338 1) no more m_links_t and the corresponding functions. Platforms are
339 directly read from a XML description and cannot be hard-coded
340 anymore. The same format is used for application deployment
341 description. The new format is described in the documentation.
342 Have a look in tools/platform_generation. There is a tiny script
343 that converts from the old platform format to the new one. Concerning
344 the application deployment format, parsing the old one is tricky.
345 I think most of you should however be able to convert your files. If
346 it is really an issue, I can write a C code that does the conversion.
348 2) the toolbox tbx does not exist anymore. We now have a library
349 with much more data-structures but without the hash-tables (we have
350 dictionaries that are much faster).
352 -- Arnaud Legrand <Arnaud.Legrand@imag.fr> Mon, 31 Jan 2005 10:45:53 -0800
354 *****************************************************************************
355 * Follows the old GRAS changelog. It does not follow the same syntax, but I *
356 * don't feel like converting the oldies. (Mt) *
357 *****************************************************************************
360 Version 2.90: "the long awaited one"
361 - Finished rewriting and debugging MSG. Rewrote the documentation.
362 - disable GRAS for now since it needs to be ported to the newest SG
365 - Finish the port to windows (using mingw32 for cross-compile)
368 - Main loop and datastructures of SURF. A cpu resource object is
369 functional. Surf can thus be used to create cpu's with variable
370 performance on which you can execute some actions.
372 2004-11-15 Martin Quinson
373 - Port to ARM. Simply added the alignment and size descriptions. Should
374 work, but the ARM machines are so slow that I didn't had the opportunity
375 to 'make check' over there yet.
377 2004-11-15 Arnaud Legrand
378 - Trace manager now written. It uses a heap structure and is therefore
379 expected to be efficient. It may however be speeded up (particularly
380 when many events occur at the same date) by using red and black
381 trees. One day maybe...
382 - Max-min linear system solver written. It uses a sparse matrix
383 structure taking advantage of its expected use. Most operations are
384 O(1) and free/calloc are called as few as possible. The computation of
385 the minimum could however be improved by using a red and black tree
388 2004-11-03 Arnaud Legrand
389 - Rename every gras_* function that was in xbt/ to its xbt_
391 - Add a heap and a doubly-linked list to xbt
392 - Added a dichotomy to the dictionaries. make check works as well before
393 so I assume that the patch is correct. I do not know however if things
394 run effectively faster than before now. :)
396 Inclusion of the SimGrid tree in the GRAS one. The archive is renamed to
397 SimGrid, and the version number is bumped to 2.x
399 2004-10-29 Martin Quinson
400 - Introduction of the remote errors.
401 They are the result of a RMI/RPC on the remote machine.
402 ErrCodes being scalar values, you can't get the host on which those
403 errors did happen. Extending the error mechanism as in Gnome is possible.
404 No idea yet whether it is a good idea.
406 2004-10-28 Martin Quinson
407 - Interface revolution: the Starred Structure Eradication.
408 I used to do typedef struct {} toto_t; and then handle *toto_t.
409 Arnaud (and Oli) didn't like it, and I surrendered. Now, you have:
410 - ???_t is a valid type (builded with typedef)
411 - s_toto_t is a structure (access to fields with .)
412 - s_toto is a structure needing 'struct' keyword to be used
413 - e_toto_t is an enum
414 - toto_t is an 'object' (struct*)
416 typedef struct s_toto {} s_toto_t, *toto_t;
417 typedef enum {} e_toto_t;
418 Moreover, only toto_t (and e_toto_t) are public. The rest (mainly
419 s_toto_t) is private.
421 - While I was at it, all gras_<obj>_free() functions want a gras_<obj>_t*
422 so that it can set the variable to NULL. It was so for dicts and sets,
423 it changed for dynars.
425 - Fix a bunch of memleaks in dict_remove
426 - Fix a bug in sg/server_socket opening: it failed all the time.
428 2004-10-07 Martin Quinson
429 - Speed up dynar lookup operation a bit.
431 gras_dynar_get is dead.
433 Now, you can choose between gras_dynar_get_cpy (the old gras_dynar_get
434 but should be avoided for efficiency reasons) and gras_dynar_get_ptr
435 (which gives you the address of the stored data).
437 gras_dynar_get_as is an helpful macro which allows you to retrieve a
438 copy of the data using an affectation to do the job and not a memcpy.
440 int toto = gras_dynar_get_as(dyn,0,int); rewrites itself to
441 int toto = *(int*)gras_dynar_get_ptr(dyn,0);
443 It does not really speedup the dynar test because they are
444 setting elements all the time (and look them seldom). But the dict does
445 far more lookup than setting.
447 So, this brings the dict_crash test from ~33s to ~25s (200000 elms).
449 2004-10-05 Martin Quinson
450 - Allow to (en/dis)able the cycle detection at run time.
452 Whether we should check for cycle or not is now a property of each
453 datatype. When you think there may be some cycle, use datadesc_cycle_set.
454 datadesc_cycle_unset allow to remove this property when previously set.
456 Note that the cycle detection is off by default since it impacts the
457 performance. Watch the data you feed GRAS with ;)
459 This property is hereditary. Any element embedded in a structure having it
460 set have it set for the time of this data exchange.
462 You should set it both on sender and receiver side. If you don't set it on
463 sender side, it will enter an endless loop. If you forget on receiver
464 side, the cycles won't be recreated after communication.
466 - Header reorganization.
467 Kill gras_private.h, each submodule must load the headers it needs.
469 2004-10-04 Martin Quinson
470 - Interface revolution: do not try to survive to malloc failure.
472 Now, gras_malloc and friends call gras_abort() on failure.
473 As a conclusion, malloc_error is not a valid error anymore, and all
474 functions for which it was the only gras_error_t return value are
475 changed. They now return void, or there result directly.
476 This simplify the API a lot.
478 2004-09-29 Martin Quinson
479 - Re-enable raw sockets.
480 Created by gras_socket_{client,server}_ext;
481 Used with gras_raw_{send,recv}
484 It should allow to kill the last bits of gras first version soon.
486 This is not completely satisfactory yet (duplicate code with
487 chunk_{send,recv}; a bit out of the plugin mechanism), but it should
490 - Simplify transport plugin (internal) interface by not passing any
491 argument to _server and _client, but embedding them in the socket
494 2004-09-28 Martin Quinson
495 - Finish the port to AIX.
496 autoconf was my problem (segfault within the malloc replacement
497 function. No idea why)
499 2004-09-16 Martin Quinson
500 - Fix some size_t madness on 64bit architectures.
502 2004-09-08 Martin Quinson
503 - Reduce the number of system headers loaded, overload some more system
504 calls (such as malloc to cast the result of the system one, and work
506 - Fix and reintroduce the config support
508 2004-09-07 Martin Quinson
509 - Source code reorganization to allow Arnaud to surf all over there.
510 - Allow to document the logging categories.
511 - Remove all uppercase from logging categories and useless cleanup in names.
513 2004-08-18 Martin Quinson
514 Version 0.6.2 (protocol not changed; API changed)
515 - Interface cleanup: gras_msgtype_by_name returns the type (instead of a
516 gras_error_t), and NULL when not found. Functions expecting a msgtype
517 as argument (msg_wait; msg_send) deal with NULL argument by providing a
518 hopefully usefull message.
519 - Portability to prehistoric sparcs again
521 2004-08-17 Martin Quinson
522 Version 0.6.1 (protocol not changed; ABI not changed)
523 - prealloc some buffers to speed things up
525 2004-08-11 Martin Quinson
526 Version 0.6 (protocol not changed; ABI expended)
527 - The parsing macro can deal with the references, provided that you add
528 the relevant annotations (using GRAS_ANNOTE(size,field_name))
530 2004-08-09 Martin Quinson
531 Version 0.5 (protocol not changed; ABI changed)
532 - Allow to off turn the cycle detection code in data exchange at
533 compilation time. It should be at run time, but I'm short of time (and
534 the config stuff is still broken). That way, we keep dict out of the
535 critical path, which is good because the performance is poor:
536 - search not dichotomial yet
537 - dynar give no way to access their content and memcpy everytime
538 - In composed data description (struct, ref and so on), stop foolness of
539 keeping the subtype's ID, but store the type itself. This keeps sets out
540 of the critical path, which is good since they rely on dynar and
541 dictionnaries. The only loose of that is that we cannot detect the
542 redeclaration of a structure/union with another content (but I'm not sure
543 the code detected well this error before anyway). We still can detect
544 the redefinition discrepancy for the other types.
545 - Use a whole bunch of optimisation flags (plus -fno-strict-aliasing since
546 it breaks the code because of type-punning used all over the place).
547 This breaks on all non-gcc architectures (for now).
549 All those changes (plus the buffer of last time) allow me to gain 2 order
550 of magnitude on cruel tests consisting of 800000 array of integers on two
551 level of a hierarchical structure (200 secondes -> 4 secondes)
554 - the selector of reference must now return the type it points to, not
557 2004-08-06 Martin Quinson
558 Version 0.4 (protocol changed; ABI not changed)
559 - Allow to pass --gras-log argument to processes in simulation mode. Really.
560 - New debugging level: trace (under debug) to see effect of GRAS_IN/OUT
561 - Implement a buffer transport, and use it by default (it relies on tcp in
562 real life and on sg in simulation).
563 That's a bit hackish since I had a new field to the structure to store
564 its data without interfering with the subtype ones. Inheritance
565 is tricky in C. And that's a kind of reverse inheritance with one class
566 derivating two classes. Or maybe a game with java interfaces. Anyway,
567 that's damn hard in C (at least).
568 Moreover, I got tired while trying to ensure plugin separation and
569 genericity in SG mode. MSG wants me to do weird things, so let's go for
570 cruel hacks (temporarily of course ;).
571 See comment in transport_private.h:71
572 - do not include all the _interface headers in private but in the files
573 which really need them (to cut the compilation time when they are
576 2004-07-26 Martin Quinson
577 Version 0.3 (protocol not changed; ABI changed)
578 - Major overhault of the datadesc interface to simplify it:
579 - shorted the function names:
580 s/gras_datadesc_declare_struct/gras_datadesc_struct/ and so on
581 - add a trivial way to push/pop integers into the cbps without malloc.
582 This allows to make really generic sub_type description, which simply
583 pop their size of the stack.
584 - add a function gras_datadesc_ref_pop_arr() which does what users want
585 most of the time: Declare a dynamic array (which pops its size of the
586 stack) and declare a reference to it. Poor name, but anyway.
587 - kill the post-send callback, add a post-receive one
589 2004-07-23 Martin Quinson
590 Version 0.2 (protocol changed; ABI changed)
591 - add some testing for cpbs in the test cases, and fix some more bugs.
592 This invalidate again the little64 data file, since I cannot regenerate
594 - remove an awfull optimization in the logging stuff, allowing me to:
595 - understand it again
596 - learn gcc how to check that the argument match the provided format
597 - fix all errors revealed by gcc after that
598 - internal keys of dict are not \0 terminated. Deal with it properly in
599 loggings instead of segfaulting when the user want to see the logs :-/
601 2004-07-22 Martin Quinson
602 - Fix some stupid bug preventing cbps (callback postit) from working
604 2004-07-21 Martin Quinson
605 - Some documentation cleanups
606 - remove the useless last argument of msgtype_declare
607 - rename the Virtu functions to fit into the 'os' namespace
608 - move headers src/include -> src/include/gras/ and stop fooling with
609 gras -> . symbolic link
610 - make distcheck is now successful
612 2004-07-19 Martin Quinson
614 - Build shared library also
615 - Install html doc to the right location
616 - stop removing maintainer files in make clean
617 - build tests only on make check
619 2004-07-13 Martin Quinson
621 - No major issue in previous version => change versionning schema
622 - Re-enable little64 convertion test now that Abdou kindly regenerated the
623 corresponding dataset.
625 2004-07-11 Martin Quinson
627 - Get it working with any kind of structure (we can compute the padding
628 bytes remotely for all the architectures I have access to)
629 - Implement the structure parsing macro (still not quite robust/complete)
630 - Improvement to the remote testing toysuite
632 2004-07-10 Martin Quinson
634 - get ride of a bunch of deprecated macros
635 - actually run the test for two-compliment, not only compile it :-/
636 - test whether the structures get packed (and bail out if yes. Damn.
637 Alignment is a serious matter)
638 - test whether the structures get compacted (but respecting the alignment
639 constraints of each types)
640 - test whether the array fields of structures can straddle alignment boundaries
642 - Damnit, double are bigger than float (typo in creation of 'double' datadesc)
643 (took me 2 hours to find that bug, looking at the wrong place)
644 - Add gras_datadesc_declare_{union,struct}_close(). They must be used
645 before sending/receiving and are used to compute the offsets of fields
646 - Given that padding size depend even on compiler options, keep track of
647 alignment and aligned_size only for the current architecture. Not a big
648 deal since we send structure fields one after the other (seems
650 - Add the datastructure used for IEEE paper by the PBIO guys to the test
651 program, let it work on linux/gcc/little32. portability todo.
653 2004-07-08 Martin Quinson
654 - import and improve remote compilation support from FAST
655 - make sure make check works on half a dozen of machines out there
657 2004-07-07 Martin Quinson
658 Let's say it's version 0.0.3 ;)
659 - Implement conversions (yuhu!)
660 - Let it work on solaris (beside conversion, of course)
661 - Stupid me, using rand() to generate the conversion datatests in not wise.
663 2004-07-06 Martin Quinson
664 - Let make dist work, since I'm gonna need it to compile on remote hosts
665 - Let Tests/datadesc_usage write the architecture on which the file was
666 generated as first byte.
667 - Add PowerPC (being also IRIX64), SPARC (also power4) and ALPHA
668 architecture descriptions.
669 - Add datadesc_usage.{i386,ppc,sparc} files being the result of execution
670 on those architectures.
671 - Optimization: send/recv array of scalar in one shoot
673 2004-07-05 Martin Quinson
674 - YEAH! GRAS/SG and GRAS/RL are both able to run the ping example !
676 - Plug a whole bunch of memleaks
677 - each process now have to call gras_{init,exit}. One day, their log
678 settings will be separated
679 - Continue the code factorisation between SG, RL and common in Transport.
681 2004-07-04 Martin Quinson
683 - Redistribution between SG and RL.
684 We wanna have to accept in SG, so move accepted related parts of RL in
685 the common part. (more precisely, the dynar of all known sockets is no
686 more a static in transport.c, but part of the process_data)
688 [gras_stub_generator]
689 - Bug fix: Do call gras_process_init from gras_init (wasnt called in RL).
691 2004-07-03 Martin Quinson
692 - Create a new log channel tbx containing dict, set, log, dynar (to shut
693 them all up in one shot)
695 - Fix the ugly case of reference to dynamic array.
696 - New (semi-public) function gras_datadesc_size to allow the messaging
697 layer to malloc the needed space for the buffer.
699 - gras_socket_close now expect the socket to close (and not its address to
700 put NULL in it after it). This is because the socket passed to handlers
701 is one of their argument (=> not writable).
703 - propagate the interface cleanup from last week in datadesc, ie remove a
704 superfluous level of indirection. User pass adress of variable
705 containing data (both when sending and receiving), and not of a variable
706 being a pointer to the data. Let's say that I like it better ;)
707 The price for that is constructs like "int msg=*(int*)payload" in
708 handlers, but it's a fine price, IMHO.
710 - Let it work in RL (yuhu)
712 2004-06-21 Martin Quinson
714 - porting SG plugin and SG select to new standards (works almost).
715 - plug memleaks and fix bugs around.
718 - cleanup the prototype of data recv and force users to specify when they
719 want to handle references to objects. Test case working even for cycles.
720 - plug memleaks. Valgrind is perfectly ok with this.
722 2004-06-12 Martin Quinson
724 - cleanup the separation between plugin and main code in plugin creation
726 2004-06-11 Martin Quinson
728 - Reput hook for raw sockets, needed for BW experiments
729 - kill a few lines of dead code
730 [Data description] Interface cleanup
731 - gras_datadesc_by_name returns the searched type or NULL.
732 That way, no variable is needed to use a type desc once, which makes
734 - gras_datadesc_declare_[struct|union]_append_name is removed. The last
735 two parameters were strings (field name, type name), leading to
737 [Dicos] Interface cleanup
738 - gras_dico_retrieve -> gras_dico_get ; gras_dico_insert -> gras_dico_set
739 This is consistant with the dynar API.
741 2004-04-21 Martin Quinson
743 - Porting to new standards.
746 There is no bag anymore, no need to take extra provision to mask the
747 pointers behind "ID".
748 Better splitup of functions between files create/exchange/convert.
749 This is still a bit artificial since convert and receive are so
750 interleaved, but anyway.
752 - add a queued message list to procdata (the ones not matching criteria
754 - factorize some more code between SG and RL wrt procdata
756 - use gras_exit in example to track memleaks
757 - get rid of gs_example now that GS is properly integrated into gras
758 - update run_test to integrate the lastest tests (datadesc)
760 - rename WARNINGn macros to WARNn since it prooved error-prone
762 2004-04-19 Martin Quinson
764 - register init/exit functions within gras module mechanism
765 - send/receive function.
766 Convertion is not implemented, but short-cutted if not needed.
767 struct/array elements are sent one by one (instead of block-wise), but
768 nobody really cares (yet). Get a prototype before optimizing.
769 - tests (using a file socket) for DD send/receive on:
770 - base types: int, float
771 - array: fixed size, string (ie ref to dynamic string)
772 - structure: homogeneous, heterogeneous
773 - chained list, graph with cycle
774 Believe it or not, valgrind is not too unhappy with the results. The
775 cycle happily segfaults, but the others are ok. And I'm sick of pointers
779 - Bugfix when using a filename explicitely (instead of '-')
781 2004-04-09 Martin Quinson
783 - factorize more code between RL and SG in socket creation
784 - Complete the implementation and tests of:
786 o file (only in RL, and mainly for debugging)
788 I lost 3 days to design a portable address resolver, and then decided
789 that the prototype mainly have to run on my box.
790 Addressing portability too early may be like optimizing too early :-/
792 - use gras_init in the Tests instead of the crappy parse_log_opt
793 (the latter function is removed)
794 [Conditional execution]
795 - New functions: gras_if_RL/gras_if_SG (basic support for this)
796 [Code reorganisation]
797 - Get rid of libgrasutils.a since it makes more trouble than it solves.
798 Build examples against the RL library, since there is no way to disable
799 its creation for now.