1 SimGrid (2.95) unstable; urgency=low
4 * Implement MSG_get_msgload [MQ]
8 SimGrid (2.94) unstable; urgency=low
10 The first release candidate for SimGrid 3 !
14 * Update the main page and the FAQ. Adding references to gforge.
17 * Add a gras_os_getpid function.
20 * Add MSG_task_get_compute_duration() and MSG_task_get_data_size()
21 * Extend the logs so that they also print PID, hostname, date, ... if
23 * Convert the MSG example to the use of xbt_logs instead of PRINT_MESSAGE,
24 and kill the old version which were in testsuite/
25 * Rewrite tools/MSG_visualization/colorize.pl for using with logs instead
29 * Add xbt_os_time(). As the rest of xbt/portability, this is not public
30 for users. Instead, each programming environment (GRAS, MSG,...) use it
31 when needed to provide such a feature to users.
32 Don't shortcut the mecanism or you will also shortcut the virtualization
33 you need on the simulator.
37 * Cleanups in configury with regard to compile optimization/warning flags.
38 Also add -fno-loop-optimize to any powerpc since it's the optimization
39 killing gcc (< 3.4.0).
40 * Doxygen cleanups: move MSG examples, kill the second Doxygen phase
41 needed by MSG examples complications
42 * Borrow configury beautifications from PHP
45 * Bugfix: XBT_LOG_NEW_DEFAULT_CATEGORY now compiles without compiler
46 warning (thanks loris for stumbling into this one).
47 * Bugfix: stop loading private headers (gras_config.h) from the public
51 * Change SIMGRID_INSTALL_PATH to GRAS_ROOT in Makefiles generated for user.
52 * Rename gras_get_my_fqdn to gras_os_myname and implement it in the simulator
53 RL would imply a DNS resolver, which is *hard* to do in a portable way
54 (and therefore delayed).
55 * Implement a real timer mecanism and use it in timing macros. This allows
56 to avoid rounding errors and get a 0.000005 sec precision in timing
57 macros. While I was at it, various cleanups:
58 - allow to declare more than one timed section per file (fix a stupid bug)
59 - move some private declaration to the right place
60 - merge conditional execution and timing macros into emulation module
62 - make sure the module cleanups its mess on gras_exit
63 * Documentation improvements:
64 - (new) how to compile applications using GRAS
65 - (new) emulation support (timing macros)
67 -- Da SimGrid team <simgrid-devel@lists.gforge.inria.fr> Fri, 13 May 2005 10:49:31 +0200
69 SimGrid (2.93) unstable; urgency=low
71 Alpha 4 on the path to SimGrid 3 (aka the "neuf-trois" version)
74 - Use Paje properly where used. Still to be sanitized properly.
75 - Portability fix: Add an implementation of the contexts using pthread
79 - Add xbt_procname(): returns the name of the current process.
80 Use it to show the current process's name in all logging.
82 - fix detection of older flex version and the reaction, since we do
83 depend on modern ones (we use lex_destroy)
84 - Better separation of SG and RL in the libs: remove all simulation code
85 from libgras. As a result, this lib is now only 200k when stripped.
86 Some of the xbt modules may also be duplicated (two sets and such) and
87 should be cleaned/killed before SG3.
88 - Insist on using xlC on AIX because of weird problems involving gcc there.
89 - Cleanup the make remote stuff. This is now done by scripts
90 tools/graspe-{master,slave} (GRAS Platform Expender). This is still
91 mainly for our private use, but we're working on changing them to user
94 - Bugfix: flush the socket on close only if there is some *output*.
95 - Bugfix: flush idempotent when there's nothing to send (don't send size=0)
97 - Add MSG_task_get_name. The task names are mainly for debugging purpose,
100 -- SimGrid team <simgrid2-users@listes.ens-lyon.fr> Fri, 4 Mar 2005 14:32:37 -0800
102 SimGrid (2.92) unstable; urgency=low
104 Alpha 3 on the path to SimGrid 3
108 - New! First try of benchmarking macros.
109 - New! First try so that gras_stub_generator generate deployment and
110 remote compilation helpers.
112 - Bugfix: Initialization fix in msg_test.
116 - Bugfix: applied patch to lexer so that it doesn't need a huge heap.
118 - Bugfix: let dicts work with NULL content (_foreach didn't) and cleanups
120 - API Change: gras_os_sleep to take the amount of seconds as a double.
121 Accepting an int was error prone since it was the only location where
122 seconds were coded as such. It leaded to damn rounding errors.
123 - Bugfix: Hard to belive that timers ever worked before this.
125 -- SimGrid team <simgrid2-users@listes.ens-lyon.fr> Wed, 23 Feb 2005 22:09:21 +0100
127 SimGrid (2.91) unstable; urgency=low
129 Alpha 2 on the path to SimGrid 3
133 - Bug fix in the lmm_solver.
135 - New! Interface to Paje (see http://www-id.imag.fr/Logiciels/paje/)
136 through the function MSG_paje_output.
137 - New! Introducing two new functions MSG_process_kill() and MSG_process_killall().
138 - It is possible to bound the rate of a communication in MSG with
139 MSG_task_put_bounded() (was already in the previous version but I had forgotten
140 to write it in the changelog).
141 - Bug fix to let GRAS run on top of MSG until we move it directly on top
146 - Various cleanups to the autotools stuff
147 - Begin to move Gras examples to examples/gras/
148 - Let make distcheck work again (yeah!)
150 - documentation overhauled using doxygen.
151 gtk-doc-tools is dead in SimGrid now.
152 - Automatically extract all existing logging categories, and add the list
153 to the documentation (long standing one, to say the less)
155 - Cleanup the known architecture table. Reorder the entries to group what
156 should be, and use a more consistent naming scheme.
157 (some of the test dataset are still to be regenerated)
158 - New! Allow library to register globals on each process just as userdata
160 This is implemented using a xbt_dict and not a xbt_set, so we loose the
161 lookup time (for now).
162 Use it in msg and trp.
163 This cleans a lot the internals and helps enforcing privacy of the
164 headers between the gras components.
165 - New! Add a timer mechanism, not unlike cron(8) and at(1).
166 - Bugfix: gras_os_time was delirious in RL.
167 - Bugfix: gras_trp_select/RL don't run into the wall when asked to select
169 - Reenable GRAS now that it works.
171 -- Arnaud Legrand <Arnaud.Legrand@imag.fr> Mon, 14 Feb 2005 14:02:13 -0800
173 SimGrid (2.90) unstable; urgency=low
175 Alpha 1 on the path to SimGrid 3
177 * It is a long time since the last release of SimGrid. I'm sorry about
178 that but as I had told you, I was rewriting a lot of things. I apologize
179 to those who had been reporting bugs to me and that I had not answered.
180 If your bug is still in the new version, please tell me. Here is a
181 summary of the main changes.
183 * REVOLUTION 1: The SimGrid project has merged with the GRAS project
184 lead by Martin Quinson. As a consequence SimGrid gains a lot in
185 portability, speed, and a lot more but you'll figure it out later.
186 SimGrid now comprises 3 different projects : MSG, GRAS and SMPI.
187 I wanted to release the new MSG as soon as possible and I have
188 broken GRAS, which is the reason why, for now, only MSG is fully
189 functional. A laconic description of these projects is available
190 in the documentation.
192 * REVOLUTION 2: I have removed SG and I am now using a new simulation
193 kernel optimized for our needs (called SURF but only the developers
194 should use it). Hence, MSG is now roughly 30 times faster and I think
195 that by rewriting a little bit MSG, I could event speed it up a little
196 bit more. Beside the gain in speed, it is also much easier to encode a
197 new platform model with SURF than it was with SG. More to come...
199 * REVOLUTION 3: I have tried to change a little as possible the API of
200 MSG but a few things really had to disappear. The main differences
201 with the previous version are :
202 1) no more m_links_t and the corresponding functions. Platforms are
203 directly read from a XML description and cannot be hard-coded
204 anymore. The same format is used for application deployment
205 description. The new format is described in the documentation.
206 Have a look in tools/platform_generation. There is a tiny script
207 that converts from the old platform format to the new one. Concerning
208 the application deployment format, parsing the old one is tricky.
209 I think most of you should however be able to convert your files. If
210 it is really an issue, I can write a C code that does the conversion.
212 2) the toolbox tbx does not exist anymore. We now have a library
213 with much more data-structures but without the hash-tables (we have
214 dictionaries that are much faster).
216 -- Arnaud Legrand <Arnaud.Legrand@imag.fr> Mon, 31 Jan 2005 10:45:53 -0800
218 *****************************************************************************
219 * Follows the old GRAS changelog. It does not follow the same syntax, but I *
220 * don't feel like converting the oldies. (Mt) *
221 *****************************************************************************
224 Version 2.90: "the long awaited one"
225 - Finished rewriting and debugging MSG. Rewrote the documentation.
226 - disable GRAS for now since it needs to be ported to the newest SG
229 - Finish the port to windows (using mingw32 for cross-compile)
232 - Main loop and datastructures of SURF. A cpu resource object is
233 functional. Surf can thus be used to create cpu's with variable
234 performance on which you can execute some actions.
236 2004-11-15 Martin Quinson
237 - Port to ARM. Simply added the alignment and size descriptions. Should
238 work, but the ARM machines are so slow that I didn't had the opportunity
239 to 'make check' over there yet.
241 2004-11-15 Arnaud Legrand
242 - Trace manager now written. It uses a heap structure and is therefore
243 expected to be efficient. It may however be speeded up (particularly
244 when many events occur at the same date) by using red and black
245 trees. One day maybe...
246 - Max-min linear system solver written. It uses a sparse matrix
247 structure taking advantage of its expected use. Most operations are
248 O(1) and free/calloc are called as few as possible. The computation of
249 the minimum could however be improved by using a red and black tree
252 2004-11-03 Arnaud Legrand
253 - Rename every gras_* function that was in xbt/ to its xbt_
255 - Add a heap and a doubly-linked list to xbt
256 - Added a dichotomy to the dictionaries. make check works as well before
257 so I assume that the patch is correct. I do not know however if things
258 run effectively faster than before now. :)
260 Inclusion of the SimGrid tree in the GRAS one. The archive is renamed to
261 SimGrid, and the version number is bumped to 2.x
263 2004-10-29 Martin Quinson
264 - Introduction of the remote errors.
265 They are the result of a RMI/RPC on the remote machine.
266 ErrCodes being scalar values, you can't get the host on which those
267 errors did happen. Extending the error mechanism as in Gnome is possible.
268 No idea yet whether it is a good idea.
270 2004-10-28 Martin Quinson
271 - Interface revolution: the Starred Structure Eradication.
272 I used to do typedef struct {} toto_t; and then handle *toto_t.
273 Arnaud (and Oli) didn't like it, and I surrendered. Now, you have:
274 - ???_t is a valid type (builded with typedef)
275 - s_toto_t is a structure (access to fields with .)
276 - s_toto is a structure needing 'struct' keyword to be used
277 - e_toto_t is an enum
278 - toto_t is an 'object' (struct*)
280 typedef struct s_toto {} s_toto_t, *toto_t;
281 typedef enum {} e_toto_t;
282 Moreover, only toto_t (and e_toto_t) are public. The rest (mainly
283 s_toto_t) is private.
285 - While I was at it, all gras_<obj>_free() functions want a gras_<obj>_t*
286 so that it can set the variable to NULL. It was so for dicts and sets,
287 it changed for dynars.
289 - Fix a bunch of memleaks in dict_remove
290 - Fix a bug in sg/server_socket opening: it failed all the time.
292 2004-10-07 Martin Quinson
293 - Speed up dynar lookup operation a bit.
295 gras_dynar_get is dead.
297 Now, you can choose between gras_dynar_get_cpy (the old gras_dynar_get
298 but should be avoided for efficiency reasons) and gras_dynar_get_ptr
299 (which gives you the address of the stored data).
301 gras_dynar_get_as is an helpful macro which allows you to retrieve a
302 copy of the data using an affectation to do the job and not a memcpy.
304 int toto = gras_dynar_get_as(dyn,0,int); rewrites itself to
305 int toto = *(int*)gras_dynar_get_ptr(dyn,0);
307 It does not really speedup the dynar test because they are
308 setting elements all the time (and look them seldom). But the dict does
309 far more lookup than setting.
311 So, this brings the dict_crash test from ~33s to ~25s (200000 elms).
313 2004-10-05 Martin Quinson
314 - Allow to (en/dis)able the cycle detection at run time.
316 Whether we should check for cycle or not is now a property of each
317 datatype. When you think there may be some cycle, use datadesc_cycle_set.
318 datadesc_cycle_unset allow to remove this property when previously set.
320 Note that the cycle detection is off by default since it impacts the
321 performance. Watch the data you feed GRAS with ;)
323 This property is hereditary. Any element embedded in a structure having it
324 set have it set for the time of this data exchange.
326 You should set it both on sender and receiver side. If you don't set it on
327 sender side, it will enter an endless loop. If you forget on receiver
328 side, the cycles won't be recreated after communication.
330 - Header reorganization.
331 Kill gras_private.h, each submodule must load the headers it needs.
333 2004-10-04 Martin Quinson
334 - Interface revolution: do not try to survive to malloc failure.
336 Now, gras_malloc and friends call gras_abort() on failure.
337 As a conclusion, malloc_error is not a valid error anymore, and all
338 functions for which it was the only gras_error_t return value are
339 changed. They now return void, or there result directly.
340 This simplify the API a lot.
342 2004-09-29 Martin Quinson
343 - Re-enable raw sockets.
344 Created by gras_socket_{client,server}_ext;
345 Used with gras_raw_{send,recv}
348 It should allow to kill the last bits of gras first version soon.
350 This is not completely satisfactory yet (duplicate code with
351 chunk_{send,recv}; a bit out of the plugin mechanism), but it should
354 - Simplify transport plugin (internal) interface by not passing any
355 argument to _server and _client, but embedding them in the socket
358 2004-09-28 Martin Quinson
359 - Finish the port to AIX.
360 autoconf was my problem (segfault within the malloc replacement
361 function. No idea why)
363 2004-09-16 Martin Quinson
364 - Fix some size_t madness on 64bit architectures.
366 2004-09-08 Martin Quinson
367 - Reduce the number of system headers loaded, overload some more system
368 calls (such as malloc to cast the result of the system one, and work
370 - Fix and reintroduce the config support
372 2004-09-07 Martin Quinson
373 - Source code reorganization to allow Arnaud to surf all over there.
374 - Allow to document the logging categories.
375 - Remove all uppercase from logging categories and useless cleanup in names.
377 2004-08-18 Martin Quinson
378 Version 0.6.2 (protocol not changed; API changed)
379 - Interface cleanup: gras_msgtype_by_name returns the type (instead of a
380 gras_error_t), and NULL when not found. Functions expecting a msgtype
381 as argument (msg_wait; msg_send) deal with NULL argument by providing a
382 hopefully usefull message.
383 - Portability to prehistoric sparcs again
385 2004-08-17 Martin Quinson
386 Version 0.6.1 (protocol not changed; ABI not changed)
387 - prealloc some buffers to speed things up
389 2004-08-11 Martin Quinson
390 Version 0.6 (protocol not changed; ABI expended)
391 - The parsing macro can deal with the references, provided that you add
392 the relevant annotations (using GRAS_ANNOTE(size,field_name))
394 2004-08-09 Martin Quinson
395 Version 0.5 (protocol not changed; ABI changed)
396 - Allow to off turn the cycle detection code in data exchange at
397 compilation time. It should be at run time, but I'm short of time (and
398 the config stuff is still broken). That way, we keep dict out of the
399 critical path, which is good because the performance is poor:
400 - search not dichotomial yet
401 - dynar give no way to access their content and memcpy everytime
402 - In composed data description (struct, ref and so on), stop foolness of
403 keeping the subtype's ID, but store the type itself. This keeps sets out
404 of the critical path, which is good since they rely on dynar and
405 dictionnaries. The only loose of that is that we cannot detect the
406 redeclaration of a structure/union with another content (but I'm not sure
407 the code detected well this error before anyway). We still can detect
408 the redefinition discrepancy for the other types.
409 - Use a whole bunch of optimisation flags (plus -fno-strict-aliasing since
410 it breaks the code because of type-punning used all over the place).
411 This breaks on all non-gcc architectures (for now).
413 All those changes (plus the buffer of last time) allow me to gain 2 order
414 of magnitude on cruel tests consisting of 800000 array of integers on two
415 level of a hierarchical structure (200 secondes -> 4 secondes)
418 - the selector of reference must now return the type it points to, not
421 2004-08-06 Martin Quinson
422 Version 0.4 (protocol changed; ABI not changed)
423 - Allow to pass --gras-log argument to processes in simulation mode. Really.
424 - New debugging level: trace (under debug) to see effect of GRAS_IN/OUT
425 - Implement a buffer transport, and use it by default (it relies on tcp in
426 real life and on sg in simulation).
427 That's a bit hackish since I had a new field to the structure to store
428 its data without interfering with the subtype ones. Inheritance
429 is tricky in C. And that's a kind of reverse inheritance with one class
430 derivating two classes. Or maybe a game with java interfaces. Anyway,
431 that's damn hard in C (at least).
432 Moreover, I got tired while trying to ensure plugin separation and
433 genericity in SG mode. MSG wants me to do weird things, so let's go for
434 cruel hacks (temporarily of course ;).
435 See comment in transport_private.h:71
436 - do not include all the _interface headers in private but in the files
437 which really need them (to cut the compilation time when they are
440 2004-07-26 Martin Quinson
441 Version 0.3 (protocol not changed; ABI changed)
442 - Major overhault of the datadesc interface to simplify it:
443 - shorted the function names:
444 s/gras_datadesc_declare_struct/gras_datadesc_struct/ and so on
445 - add a trivial way to push/pop integers into the cbps without malloc.
446 This allows to make really generic sub_type description, which simply
447 pop their size of the stack.
448 - add a function gras_datadesc_ref_pop_arr() which does what users want
449 most of the time: Declare a dynamic array (which pops its size of the
450 stack) and declare a reference to it. Poor name, but anyway.
451 - kill the post-send callback, add a post-receive one
453 2004-07-23 Martin Quinson
454 Version 0.2 (protocol changed; ABI changed)
455 - add some testing for cpbs in the test cases, and fix some more bugs.
456 This invalidate again the little64 data file, since I cannot regenerate
458 - remove an awfull optimization in the logging stuff, allowing me to:
459 - understand it again
460 - learn gcc how to check that the argument match the provided format
461 - fix all errors revealed by gcc after that
462 - internal keys of dict are not \0 terminated. Deal with it properly in
463 loggings instead of segfaulting when the user want to see the logs :-/
465 2004-07-22 Martin Quinson
466 - Fix some stupid bug preventing cbps (callback postit) from working
468 2004-07-21 Martin Quinson
469 - Some documentation cleanups
470 - remove the useless last argument of msgtype_declare
471 - rename the Virtu functions to fit into the 'os' namespace
472 - move headers src/include -> src/include/gras/ and stop fooling with
473 gras -> . symbolic link
474 - make distcheck is now successful
476 2004-07-19 Martin Quinson
478 - Build shared library also
479 - Install html doc to the right location
480 - stop removing maintainer files in make clean
481 - build tests only on make check
483 2004-07-13 Martin Quinson
485 - No major issue in previous version => change versionning schema
486 - Re-enable little64 convertion test now that Abdou kindly regenerated the
487 corresponding dataset.
489 2004-07-11 Martin Quinson
491 - Get it working with any kind of structure (we can compute the padding
492 bytes remotely for all the architectures I have access to)
493 - Implement the structure parsing macro (still not quite robust/complete)
494 - Improvement to the remote testing toysuite
496 2004-07-10 Martin Quinson
498 - get ride of a bunch of deprecated macros
499 - actually run the test for two-compliment, not only compile it :-/
500 - test whether the structures get packed (and bail out if yes. Damn.
501 Alignment is a serious matter)
502 - test whether the structures get compacted (but respecting the alignment
503 constraints of each types)
504 - test whether the array fields of structures can straddle alignment boundaries
506 - Damnit, double are bigger than float (typo in creation of 'double' datadesc)
507 (took me 2 hours to find that bug, looking at the wrong place)
508 - Add gras_datadesc_declare_{union,struct}_close(). They must be used
509 before sending/receiving and are used to compute the offsets of fields
510 - Given that padding size depend even on compiler options, keep track of
511 alignment and aligned_size only for the current architecture. Not a big
512 deal since we send structure fields one after the other (seems
514 - Add the datastructure used for IEEE paper by the PBIO guys to the test
515 program, let it work on linux/gcc/little32. portability todo.
517 2004-07-08 Martin Quinson
518 - import and improve remote compilation support from FAST
519 - make sure make check works on half a dozen of machines out there
521 2004-07-07 Martin Quinson
522 Let's say it's version 0.0.3 ;)
523 - Implement conversions (yuhu!)
524 - Let it work on solaris (beside conversion, of course)
525 - Stupid me, using rand() to generate the conversion datatests in not wise.
527 2004-07-06 Martin Quinson
528 - Let make dist work, since I'm gonna need it to compile on remote hosts
529 - Let Tests/datadesc_usage write the architecture on which the file was
530 generated as first byte.
531 - Add PowerPC (being also IRIX64), SPARC (also power4) and ALPHA
532 architecture descriptions.
533 - Add datadesc_usage.{i386,ppc,sparc} files being the result of execution
534 on those architectures.
535 - Optimization: send/recv array of scalar in one shoot
537 2004-07-05 Martin Quinson
538 - YEAH! GRAS/SG and GRAS/RL are both able to run the ping example !
540 - Plug a whole bunch of memleaks
541 - each process now have to call gras_{init,exit}. One day, their log
542 settings will be separated
543 - Continue the code factorisation between SG, RL and common in Transport.
545 2004-07-04 Martin Quinson
547 - Redistribution between SG and RL.
548 We wanna have to accept in SG, so move accepted related parts of RL in
549 the common part. (more precisely, the dynar of all known sockets is no
550 more a static in transport.c, but part of the process_data)
552 [gras_stub_generator]
553 - Bug fix: Do call gras_process_init from gras_init (wasnt called in RL).
555 2004-07-03 Martin Quinson
556 - Create a new log channel tbx containing dict, set, log, dynar (to shut
557 them all up in one shot)
559 - Fix the ugly case of reference to dynamic array.
560 - New (semi-public) function gras_datadesc_size to allow the messaging
561 layer to malloc the needed space for the buffer.
563 - gras_socket_close now expect the socket to close (and not its address to
564 put NULL in it after it). This is because the socket passed to handlers
565 is one of their argument (=> not writable).
567 - propagate the interface cleanup from last week in datadesc, ie remove a
568 superfluous level of indirection. User pass adress of variable
569 containing data (both when sending and receiving), and not of a variable
570 being a pointer to the data. Let's say that I like it better ;)
571 The price for that is constructs like "int msg=*(int*)payload" in
572 handlers, but it's a fine price, IMHO.
574 - Let it work in RL (yuhu)
576 2004-06-21 Martin Quinson
578 - porting SG plugin and SG select to new standards (works almost).
579 - plug memleaks and fix bugs around.
582 - cleanup the prototype of data recv and force users to specify when they
583 want to handle references to objects. Test case working even for cycles.
584 - plug memleaks. Valgrind is perfectly ok with this.
586 2004-06-12 Martin Quinson
588 - cleanup the separation between plugin and main code in plugin creation
590 2004-06-11 Martin Quinson
592 - Reput hook for raw sockets, needed for BW experiments
593 - kill a few lines of dead code
594 [Data description] Interface cleanup
595 - gras_datadesc_by_name returns the searched type or NULL.
596 That way, no variable is needed to use a type desc once, which makes
598 - gras_datadesc_declare_[struct|union]_append_name is removed. The last
599 two parameters were strings (field name, type name), leading to
601 [Dicos] Interface cleanup
602 - gras_dico_retrieve -> gras_dico_get ; gras_dico_insert -> gras_dico_set
603 This is consistant with the dynar API.
605 2004-04-21 Martin Quinson
607 - Porting to new standards.
610 There is no bag anymore, no need to take extra provision to mask the
611 pointers behind "ID".
612 Better splitup of functions between files create/exchange/convert.
613 This is still a bit artificial since convert and receive are so
614 interleaved, but anyway.
616 - add a queued message list to procdata (the ones not matching criteria
618 - factorize some more code between SG and RL wrt procdata
620 - use gras_exit in example to track memleaks
621 - get rid of gs_example now that GS is properly integrated into gras
622 - update run_test to integrate the lastest tests (datadesc)
624 - rename WARNINGn macros to WARNn since it prooved error-prone
626 2004-04-19 Martin Quinson
628 - register init/exit functions within gras module mechanism
629 - send/receive function.
630 Convertion is not implemented, but short-cutted if not needed.
631 struct/array elements are sent one by one (instead of block-wise), but
632 nobody really cares (yet). Get a prototype before optimizing.
633 - tests (using a file socket) for DD send/receive on:
634 - base types: int, float
635 - array: fixed size, string (ie ref to dynamic string)
636 - structure: homogeneous, heterogeneous
637 - chained list, graph with cycle
638 Believe it or not, valgrind is not too unhappy with the results. The
639 cycle happily segfaults, but the others are ok. And I'm sick of pointers
643 - Bugfix when using a filename explicitely (instead of '-')
645 2004-04-09 Martin Quinson
647 - factorize more code between RL and SG in socket creation
648 - Complete the implementation and tests of:
650 o file (only in RL, and mainly for debugging)
652 I lost 3 days to design a portable address resolver, and then decided
653 that the prototype mainly have to run on my box.
654 Addressing portability too early may be like optimizing too early :-/
656 - use gras_init in the Tests instead of the crappy parse_log_opt
657 (the latter function is removed)
658 [Conditional execution]
659 - New functions: gras_if_RL/gras_if_SG (basic support for this)
660 [Code reorganisation]
661 - Get rid of libgrasutils.a since it makes more trouble than it solves.
662 Build examples against the RL library, since there is no way to disable
663 its creation for now.