1 SimGrid (2.94-cvs) UNRELEASED
7 - Stop layering gras over msg?
8 - Pre-compilation of data_send/recv?
9 - Exception revolution?
10 All of them seem to be post-SG3 issues.
12 Rather finalize what's underway and finish the Big Cleanup:
13 - Check what can be cut from the xbt
14 - "pull" code diffusion in graspe
15 - add a real implementation (even if simplistic) to gras_os_myname
22 * Update the main page and the FAQ. Adding references to gforge.
25 * Add a gras_os_getpid function.
28 * Add MSG_task_get_compute_duration() and MSG_task_get_data_size()
29 * Extend the logs so that they also print PID, hostname, date, ... if
31 * Convert the MSG example to the use of xbt_logs instead of PRINT_MESSAGE,
32 and kill the old version which were in testsuite/
33 * Rewrite tools/MSG_visualization/colorize.pl for using with logs instead
37 * Add xbt_os_time(). As the rest of xbt/portability, this is not public
38 for users. Instead, each programming environment (GRAS, MSG,...) use it
39 when needed to provide such a feature to users.
40 Don't shortcut the mecanism or you will also shortcut the virtualization
41 you need on the simulator.
45 * Cleanups in configury with regard to compile optimization/warning flags.
46 Also add -fno-loop-optimize to any powerpc since it's the optimization
47 killing gcc (< 3.4.0).
48 * Doxygen cleanups: move MSG examples, kill the second Doxygen phase
49 needed by MSG examples complications
50 * Borrow configury beautifications from PHP
53 * Bugfix: XBT_LOG_NEW_DEFAULT_CATEGORY now compiles without compiler
54 warning (thanks loris for stumbling into this one).
57 * Change SIMGRID_INSTALL_PATH to GRAS_ROOT in Makefiles generated for user.
58 * Rename gras_get_my_fqdn to gras_os_myname and implement it in the simulator
59 RL would imply a DNS resolver, which is *hard* to do in a portable way
60 (and therefore delayed).
61 * Implement a real timer mecanism and use it in timing macros. This allows
62 to avoid rounding errors and get a 0.000005 sec precision in timing
63 macros. While I was at it, various cleanups:
64 - allow to declare more than one timed section per file (fix a stupid bug)
65 - move some private declaration to the right place
66 - merge conditional execution and timing macros into emulation module
68 - make sure the module cleanups its mess on gras_exit
72 SimGrid (2.93) unstable; urgency=low
74 Alpha 4 on the path to SimGrid 3 (aka the "neuf-trois" version)
77 - Use Paje properly where used. Still to be sanitized properly.
78 - Portability fix: Add an implementation of the contexts using pthread
82 - Add xbt_procname(): returns the name of the current process.
83 Use it to show the current process's name in all logging.
85 - fix detection of older flex version and the reaction, since we do
86 depend on modern ones (we use lex_destroy)
87 - Better separation of SG and RL in the libs: remove all simulation code
88 from libgras. As a result, this lib is now only 200k when stripped.
89 Some of the xbt modules may also be duplicated (two sets and such) and
90 should be cleaned/killed before SG3.
91 - Insist on using xlC on AIX because of weird problems involving gcc there.
92 - Cleanup the make remote stuff. This is now done by scripts
93 tools/graspe-{master,slave} (GRAS Platform Expender). This is still
94 mainly for our private use, but we're working on changing them to user
97 - Bugfix: flush the socket on close only if there is some *output*.
98 - Bugfix: flush idempotent when there's nothing to send (don't send size=0)
100 - Add MSG_task_get_name. The task names are mainly for debugging purpose,
103 -- SimGrid team <simgrid2-users@listes.ens-lyon.fr> Fri, 4 Mar 2005 14:32:37 -0800
105 SimGrid (2.92) unstable; urgency=low
107 Alpha 3 on the path to SimGrid 3
111 - New! First try of benchmarking macros.
112 - New! First try so that gras_stub_generator generate deployment and
113 remote compilation helpers.
115 - Bugfix: Initialization fix in msg_test.
119 - Bugfix: applied patch to lexer so that it doesn't need a huge heap.
121 - Bugfix: let dicts work with NULL content (_foreach didn't) and cleanups
123 - API Change: gras_os_sleep to take the amount of seconds as a double.
124 Accepting an int was error prone since it was the only location where
125 seconds were coded as such. It leaded to damn rounding errors.
126 - Bugfix: Hard to belive that timers ever worked before this.
128 -- SimGrid team <simgrid2-users@listes.ens-lyon.fr> Wed, 23 Feb 2005 22:09:21 +0100
130 SimGrid (2.91) unstable; urgency=low
132 Alpha 2 on the path to SimGrid 3
136 - Bug fix in the lmm_solver.
138 - New! Interface to Paje (see http://www-id.imag.fr/Logiciels/paje/)
139 through the function MSG_paje_output.
140 - New! Introducing two new functions MSG_process_kill() and MSG_process_killall().
141 - It is possible to bound the rate of a communication in MSG with
142 MSG_task_put_bounded() (was already in the previous version but I had forgotten
143 to write it in the changelog).
144 - Bug fix to let GRAS run on top of MSG until we move it directly on top
149 - Various cleanups to the autotools stuff
150 - Begin to move Gras examples to examples/gras/
151 - Let make distcheck work again (yeah!)
153 - documentation overhauled using doxygen.
154 gtk-doc-tools is dead in SimGrid now.
155 - Automatically extract all existing logging categories, and add the list
156 to the documentation (long standing one, to say the less)
158 - Cleanup the known architecture table. Reorder the entries to group what
159 should be, and use a more consistent naming scheme.
160 (some of the test dataset are still to be regenerated)
161 - New! Allow library to register globals on each process just as userdata
163 This is implemented using a xbt_dict and not a xbt_set, so we loose the
164 lookup time (for now).
165 Use it in msg and trp.
166 This cleans a lot the internals and helps enforcing privacy of the
167 headers between the gras components.
168 - New! Add a timer mechanism, not unlike cron(8) and at(1).
169 - Bugfix: gras_os_time was delirious in RL.
170 - Bugfix: gras_trp_select/RL don't run into the wall when asked to select
172 - Reenable GRAS now that it works.
174 -- Arnaud Legrand <Arnaud.Legrand@imag.fr> Mon, 14 Feb 2005 14:02:13 -0800
176 SimGrid (2.90) unstable; urgency=low
178 Alpha 1 on the path to SimGrid 3
180 * It is a long time since the last release of SimGrid. I'm sorry about
181 that but as I had told you, I was rewriting a lot of things. I apologize
182 to those who had been reporting bugs to me and that I had not answered.
183 If your bug is still in the new version, please tell me. Here is a
184 summary of the main changes.
186 * REVOLUTION 1: The SimGrid project has merged with the GRAS project
187 lead by Martin Quinson. As a consequence SimGrid gains a lot in
188 portability, speed, and a lot more but you'll figure it out later.
189 SimGrid now comprises 3 different projects : MSG, GRAS and SMPI.
190 I wanted to release the new MSG as soon as possible and I have
191 broken GRAS, which is the reason why, for now, only MSG is fully
192 functional. A laconic description of these projects is available
193 in the documentation.
195 * REVOLUTION 2: I have removed SG and I am now using a new simulation
196 kernel optimized for our needs (called SURF but only the developers
197 should use it). Hence, MSG is now roughly 30 times faster and I think
198 that by rewriting a little bit MSG, I could event speed it up a little
199 bit more. Beside the gain in speed, it is also much easier to encode a
200 new platform model with SURF than it was with SG. More to come...
202 * REVOLUTION 3: I have tried to change a little as possible the API of
203 MSG but a few things really had to disappear. The main differences
204 with the previous version are :
205 1) no more m_links_t and the corresponding functions. Platforms are
206 directly read from a XML description and cannot be hard-coded
207 anymore. The same format is used for application deployment
208 description. The new format is described in the documentation.
209 Have a look in tools/platform_generation. There is a tiny script
210 that converts from the old platform format to the new one. Concerning
211 the application deployment format, parsing the old one is tricky.
212 I think most of you should however be able to convert your files. If
213 it is really an issue, I can write a C code that does the conversion.
215 2) the toolbox tbx does not exist anymore. We now have a library
216 with much more data-structures but without the hash-tables (we have
217 dictionaries that are much faster).
219 -- Arnaud Legrand <Arnaud.Legrand@imag.fr> Mon, 31 Jan 2005 10:45:53 -0800
221 *****************************************************************************
222 * Follows the old GRAS changelog. It does not follow the same syntax, but I *
223 * don't feel like converting the oldies. (Mt) *
224 *****************************************************************************
227 Version 2.90: "the long awaited one"
228 - Finished rewriting and debugging MSG. Rewrote the documentation.
229 - disable GRAS for now since it needs to be ported to the newest SG
232 - Finish the port to windows (using mingw32 for cross-compile)
235 - Main loop and datastructures of SURF. A cpu resource object is
236 functional. Surf can thus be used to create cpu's with variable
237 performance on which you can execute some actions.
239 2004-11-15 Martin Quinson
240 - Port to ARM. Simply added the alignment and size descriptions. Should
241 work, but the ARM machines are so slow that I didn't had the opportunity
242 to 'make check' over there yet.
244 2004-11-15 Arnaud Legrand
245 - Trace manager now written. It uses a heap structure and is therefore
246 expected to be efficient. It may however be speeded up (particularly
247 when many events occur at the same date) by using red and black
248 trees. One day maybe...
249 - Max-min linear system solver written. It uses a sparse matrix
250 structure taking advantage of its expected use. Most operations are
251 O(1) and free/calloc are called as few as possible. The computation of
252 the minimum could however be improved by using a red and black tree
255 2004-11-03 Arnaud Legrand
256 - Rename every gras_* function that was in xbt/ to its xbt_
258 - Add a heap and a doubly-linked list to xbt
259 - Added a dichotomy to the dictionaries. make check works as well before
260 so I assume that the patch is correct. I do not know however if things
261 run effectively faster than before now. :)
263 Inclusion of the SimGrid tree in the GRAS one. The archive is renamed to
264 SimGrid, and the version number is bumped to 2.x
266 2004-10-29 Martin Quinson
267 - Introduction of the remote errors.
268 They are the result of a RMI/RPC on the remote machine.
269 ErrCodes being scalar values, you can't get the host on which those
270 errors did happen. Extending the error mechanism as in Gnome is possible.
271 No idea yet whether it is a good idea.
273 2004-10-28 Martin Quinson
274 - Interface revolution: the Starred Structure Eradication.
275 I used to do typedef struct {} toto_t; and then handle *toto_t.
276 Arnaud (and Oli) didn't like it, and I surrendered. Now, you have:
277 - ???_t is a valid type (builded with typedef)
278 - s_toto_t is a structure (access to fields with .)
279 - s_toto is a structure needing 'struct' keyword to be used
280 - e_toto_t is an enum
281 - toto_t is an 'object' (struct*)
283 typedef struct s_toto {} s_toto_t, *toto_t;
284 typedef enum {} e_toto_t;
285 Moreover, only toto_t (and e_toto_t) are public. The rest (mainly
286 s_toto_t) is private.
288 - While I was at it, all gras_<obj>_free() functions want a gras_<obj>_t*
289 so that it can set the variable to NULL. It was so for dicts and sets,
290 it changed for dynars.
292 - Fix a bunch of memleaks in dict_remove
293 - Fix a bug in sg/server_socket opening: it failed all the time.
295 2004-10-07 Martin Quinson
296 - Speed up dynar lookup operation a bit.
298 gras_dynar_get is dead.
300 Now, you can choose between gras_dynar_get_cpy (the old gras_dynar_get
301 but should be avoided for efficiency reasons) and gras_dynar_get_ptr
302 (which gives you the address of the stored data).
304 gras_dynar_get_as is an helpful macro which allows you to retrieve a
305 copy of the data using an affectation to do the job and not a memcpy.
307 int toto = gras_dynar_get_as(dyn,0,int); rewrites itself to
308 int toto = *(int*)gras_dynar_get_ptr(dyn,0);
310 It does not really speedup the dynar test because they are
311 setting elements all the time (and look them seldom). But the dict does
312 far more lookup than setting.
314 So, this brings the dict_crash test from ~33s to ~25s (200000 elms).
316 2004-10-05 Martin Quinson
317 - Allow to (en/dis)able the cycle detection at run time.
319 Whether we should check for cycle or not is now a property of each
320 datatype. When you think there may be some cycle, use datadesc_cycle_set.
321 datadesc_cycle_unset allow to remove this property when previously set.
323 Note that the cycle detection is off by default since it impacts the
324 performance. Watch the data you feed GRAS with ;)
326 This property is hereditary. Any element embedded in a structure having it
327 set have it set for the time of this data exchange.
329 You should set it both on sender and receiver side. If you don't set it on
330 sender side, it will enter an endless loop. If you forget on receiver
331 side, the cycles won't be recreated after communication.
333 - Header reorganization.
334 Kill gras_private.h, each submodule must load the headers it needs.
336 2004-10-04 Martin Quinson
337 - Interface revolution: do not try to survive to malloc failure.
339 Now, gras_malloc and friends call gras_abort() on failure.
340 As a conclusion, malloc_error is not a valid error anymore, and all
341 functions for which it was the only gras_error_t return value are
342 changed. They now return void, or there result directly.
343 This simplify the API a lot.
345 2004-09-29 Martin Quinson
346 - Re-enable raw sockets.
347 Created by gras_socket_{client,server}_ext;
348 Used with gras_raw_{send,recv}
351 It should allow to kill the last bits of gras first version soon.
353 This is not completely satisfactory yet (duplicate code with
354 chunk_{send,recv}; a bit out of the plugin mechanism), but it should
357 - Simplify transport plugin (internal) interface by not passing any
358 argument to _server and _client, but embedding them in the socket
361 2004-09-28 Martin Quinson
362 - Finish the port to AIX.
363 autoconf was my problem (segfault within the malloc replacement
364 function. No idea why)
366 2004-09-16 Martin Quinson
367 - Fix some size_t madness on 64bit architectures.
369 2004-09-08 Martin Quinson
370 - Reduce the number of system headers loaded, overload some more system
371 calls (such as malloc to cast the result of the system one, and work
373 - Fix and reintroduce the config support
375 2004-09-07 Martin Quinson
376 - Source code reorganization to allow Arnaud to surf all over there.
377 - Allow to document the logging categories.
378 - Remove all uppercase from logging categories and useless cleanup in names.
380 2004-08-18 Martin Quinson
381 Version 0.6.2 (protocol not changed; API changed)
382 - Interface cleanup: gras_msgtype_by_name returns the type (instead of a
383 gras_error_t), and NULL when not found. Functions expecting a msgtype
384 as argument (msg_wait; msg_send) deal with NULL argument by providing a
385 hopefully usefull message.
386 - Portability to prehistoric sparcs again
388 2004-08-17 Martin Quinson
389 Version 0.6.1 (protocol not changed; ABI not changed)
390 - prealloc some buffers to speed things up
392 2004-08-11 Martin Quinson
393 Version 0.6 (protocol not changed; ABI expended)
394 - The parsing macro can deal with the references, provided that you add
395 the relevant annotations (using GRAS_ANNOTE(size,field_name))
397 2004-08-09 Martin Quinson
398 Version 0.5 (protocol not changed; ABI changed)
399 - Allow to off turn the cycle detection code in data exchange at
400 compilation time. It should be at run time, but I'm short of time (and
401 the config stuff is still broken). That way, we keep dict out of the
402 critical path, which is good because the performance is poor:
403 - search not dichotomial yet
404 - dynar give no way to access their content and memcpy everytime
405 - In composed data description (struct, ref and so on), stop foolness of
406 keeping the subtype's ID, but store the type itself. This keeps sets out
407 of the critical path, which is good since they rely on dynar and
408 dictionnaries. The only loose of that is that we cannot detect the
409 redeclaration of a structure/union with another content (but I'm not sure
410 the code detected well this error before anyway). We still can detect
411 the redefinition discrepancy for the other types.
412 - Use a whole bunch of optimisation flags (plus -fno-strict-aliasing since
413 it breaks the code because of type-punning used all over the place).
414 This breaks on all non-gcc architectures (for now).
416 All those changes (plus the buffer of last time) allow me to gain 2 order
417 of magnitude on cruel tests consisting of 800000 array of integers on two
418 level of a hierarchical structure (200 secondes -> 4 secondes)
421 - the selector of reference must now return the type it points to, not
424 2004-08-06 Martin Quinson
425 Version 0.4 (protocol changed; ABI not changed)
426 - Allow to pass --gras-log argument to processes in simulation mode. Really.
427 - New debugging level: trace (under debug) to see effect of GRAS_IN/OUT
428 - Implement a buffer transport, and use it by default (it relies on tcp in
429 real life and on sg in simulation).
430 That's a bit hackish since I had a new field to the structure to store
431 its data without interfering with the subtype ones. Inheritance
432 is tricky in C. And that's a kind of reverse inheritance with one class
433 derivating two classes. Or maybe a game with java interfaces. Anyway,
434 that's damn hard in C (at least).
435 Moreover, I got tired while trying to ensure plugin separation and
436 genericity in SG mode. MSG wants me to do weird things, so let's go for
437 cruel hacks (temporarily of course ;).
438 See comment in transport_private.h:71
439 - do not include all the _interface headers in private but in the files
440 which really need them (to cut the compilation time when they are
443 2004-07-26 Martin Quinson
444 Version 0.3 (protocol not changed; ABI changed)
445 - Major overhault of the datadesc interface to simplify it:
446 - shorted the function names:
447 s/gras_datadesc_declare_struct/gras_datadesc_struct/ and so on
448 - add a trivial way to push/pop integers into the cbps without malloc.
449 This allows to make really generic sub_type description, which simply
450 pop their size of the stack.
451 - add a function gras_datadesc_ref_pop_arr() which does what users want
452 most of the time: Declare a dynamic array (which pops its size of the
453 stack) and declare a reference to it. Poor name, but anyway.
454 - kill the post-send callback, add a post-receive one
456 2004-07-23 Martin Quinson
457 Version 0.2 (protocol changed; ABI changed)
458 - add some testing for cpbs in the test cases, and fix some more bugs.
459 This invalidate again the little64 data file, since I cannot regenerate
461 - remove an awfull optimization in the logging stuff, allowing me to:
462 - understand it again
463 - learn gcc how to check that the argument match the provided format
464 - fix all errors revealed by gcc after that
465 - internal keys of dict are not \0 terminated. Deal with it properly in
466 loggings instead of segfaulting when the user want to see the logs :-/
468 2004-07-22 Martin Quinson
469 - Fix some stupid bug preventing cbps (callback postit) from working
471 2004-07-21 Martin Quinson
472 - Some documentation cleanups
473 - remove the useless last argument of msgtype_declare
474 - rename the Virtu functions to fit into the 'os' namespace
475 - move headers src/include -> src/include/gras/ and stop fooling with
476 gras -> . symbolic link
477 - make distcheck is now successful
479 2004-07-19 Martin Quinson
481 - Build shared library also
482 - Install html doc to the right location
483 - stop removing maintainer files in make clean
484 - build tests only on make check
486 2004-07-13 Martin Quinson
488 - No major issue in previous version => change versionning schema
489 - Re-enable little64 convertion test now that Abdou kindly regenerated the
490 corresponding dataset.
492 2004-07-11 Martin Quinson
494 - Get it working with any kind of structure (we can compute the padding
495 bytes remotely for all the architectures I have access to)
496 - Implement the structure parsing macro (still not quite robust/complete)
497 - Improvement to the remote testing toysuite
499 2004-07-10 Martin Quinson
501 - get ride of a bunch of deprecated macros
502 - actually run the test for two-compliment, not only compile it :-/
503 - test whether the structures get packed (and bail out if yes. Damn.
504 Alignment is a serious matter)
505 - test whether the structures get compacted (but respecting the alignment
506 constraints of each types)
507 - test whether the array fields of structures can straddle alignment boundaries
509 - Damnit, double are bigger than float (typo in creation of 'double' datadesc)
510 (took me 2 hours to find that bug, looking at the wrong place)
511 - Add gras_datadesc_declare_{union,struct}_close(). They must be used
512 before sending/receiving and are used to compute the offsets of fields
513 - Given that padding size depend even on compiler options, keep track of
514 alignment and aligned_size only for the current architecture. Not a big
515 deal since we send structure fields one after the other (seems
517 - Add the datastructure used for IEEE paper by the PBIO guys to the test
518 program, let it work on linux/gcc/little32. portability todo.
520 2004-07-08 Martin Quinson
521 - import and improve remote compilation support from FAST
522 - make sure make check works on half a dozen of machines out there
524 2004-07-07 Martin Quinson
525 Let's say it's version 0.0.3 ;)
526 - Implement conversions (yuhu!)
527 - Let it work on solaris (beside conversion, of course)
528 - Stupid me, using rand() to generate the conversion datatests in not wise.
530 2004-07-06 Martin Quinson
531 - Let make dist work, since I'm gonna need it to compile on remote hosts
532 - Let Tests/datadesc_usage write the architecture on which the file was
533 generated as first byte.
534 - Add PowerPC (being also IRIX64), SPARC (also power4) and ALPHA
535 architecture descriptions.
536 - Add datadesc_usage.{i386,ppc,sparc} files being the result of execution
537 on those architectures.
538 - Optimization: send/recv array of scalar in one shoot
540 2004-07-05 Martin Quinson
541 - YEAH! GRAS/SG and GRAS/RL are both able to run the ping example !
543 - Plug a whole bunch of memleaks
544 - each process now have to call gras_{init,exit}. One day, their log
545 settings will be separated
546 - Continue the code factorisation between SG, RL and common in Transport.
548 2004-07-04 Martin Quinson
550 - Redistribution between SG and RL.
551 We wanna have to accept in SG, so move accepted related parts of RL in
552 the common part. (more precisely, the dynar of all known sockets is no
553 more a static in transport.c, but part of the process_data)
555 [gras_stub_generator]
556 - Bug fix: Do call gras_process_init from gras_init (wasnt called in RL).
558 2004-07-03 Martin Quinson
559 - Create a new log channel tbx containing dict, set, log, dynar (to shut
560 them all up in one shot)
562 - Fix the ugly case of reference to dynamic array.
563 - New (semi-public) function gras_datadesc_size to allow the messaging
564 layer to malloc the needed space for the buffer.
566 - gras_socket_close now expect the socket to close (and not its address to
567 put NULL in it after it). This is because the socket passed to handlers
568 is one of their argument (=> not writable).
570 - propagate the interface cleanup from last week in datadesc, ie remove a
571 superfluous level of indirection. User pass adress of variable
572 containing data (both when sending and receiving), and not of a variable
573 being a pointer to the data. Let's say that I like it better ;)
574 The price for that is constructs like "int msg=*(int*)payload" in
575 handlers, but it's a fine price, IMHO.
577 - Let it work in RL (yuhu)
579 2004-06-21 Martin Quinson
581 - porting SG plugin and SG select to new standards (works almost).
582 - plug memleaks and fix bugs around.
585 - cleanup the prototype of data recv and force users to specify when they
586 want to handle references to objects. Test case working even for cycles.
587 - plug memleaks. Valgrind is perfectly ok with this.
589 2004-06-12 Martin Quinson
591 - cleanup the separation between plugin and main code in plugin creation
593 2004-06-11 Martin Quinson
595 - Reput hook for raw sockets, needed for BW experiments
596 - kill a few lines of dead code
597 [Data description] Interface cleanup
598 - gras_datadesc_by_name returns the searched type or NULL.
599 That way, no variable is needed to use a type desc once, which makes
601 - gras_datadesc_declare_[struct|union]_append_name is removed. The last
602 two parameters were strings (field name, type name), leading to
604 [Dicos] Interface cleanup
605 - gras_dico_retrieve -> gras_dico_get ; gras_dico_insert -> gras_dico_set
606 This is consistant with the dynar API.
608 2004-04-21 Martin Quinson
610 - Porting to new standards.
613 There is no bag anymore, no need to take extra provision to mask the
614 pointers behind "ID".
615 Better splitup of functions between files create/exchange/convert.
616 This is still a bit artificial since convert and receive are so
617 interleaved, but anyway.
619 - add a queued message list to procdata (the ones not matching criteria
621 - factorize some more code between SG and RL wrt procdata
623 - use gras_exit in example to track memleaks
624 - get rid of gs_example now that GS is properly integrated into gras
625 - update run_test to integrate the lastest tests (datadesc)
627 - rename WARNINGn macros to WARNn since it prooved error-prone
629 2004-04-19 Martin Quinson
631 - register init/exit functions within gras module mechanism
632 - send/receive function.
633 Convertion is not implemented, but short-cutted if not needed.
634 struct/array elements are sent one by one (instead of block-wise), but
635 nobody really cares (yet). Get a prototype before optimizing.
636 - tests (using a file socket) for DD send/receive on:
637 - base types: int, float
638 - array: fixed size, string (ie ref to dynamic string)
639 - structure: homogeneous, heterogeneous
640 - chained list, graph with cycle
641 Believe it or not, valgrind is not too unhappy with the results. The
642 cycle happily segfaults, but the others are ok. And I'm sick of pointers
646 - Bugfix when using a filename explicitely (instead of '-')
648 2004-04-09 Martin Quinson
650 - factorize more code between RL and SG in socket creation
651 - Complete the implementation and tests of:
653 o file (only in RL, and mainly for debugging)
655 I lost 3 days to design a portable address resolver, and then decided
656 that the prototype mainly have to run on my box.
657 Addressing portability too early may be like optimizing too early :-/
659 - use gras_init in the Tests instead of the crappy parse_log_opt
660 (the latter function is removed)
661 [Conditional execution]
662 - New functions: gras_if_RL/gras_if_SG (basic support for this)
663 [Code reorganisation]
664 - Get rid of libgrasutils.a since it makes more trouble than it solves.
665 Build examples against the RL library, since there is no way to disable
666 its creation for now.