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 - document the gras' timing macro
16 - add a real implementation (even if simplistic) to gras_os_myname
23 * Update the main page and the FAQ. Adding references to gforge.
26 * Add a gras_os_getpid function.
29 * Add MSG_task_get_compute_duration() and MSG_task_get_data_size()
30 * Extend the logs so that they also print PID, hostname, date, ... if
32 * Convert the MSG example to the use of xbt_logs instead of PRINT_MESSAGE,
33 and kill the old version which were in testsuite/
34 * Rewrite tools/MSG_visualization/colorize.pl for using with logs instead
38 * Add xbt_os_time(). As the rest of xbt/portability, this is not public
39 for users. Instead, each programming environment (GRAS, MSG,...) use it
40 when needed to provide such a feature to users.
41 Don't shortcut the mecanism or you will also shortcut the virtualization
42 you need on the simulator.
46 * Cleanups in configury with regard to compile optimization/warning flags.
47 Also add -fno-loop-optimize to any powerpc since it's the optimization
48 killing gcc (< 3.4.0).
49 * Doxygen cleanups: move MSG examples, kill the second Doxygen phase
50 needed by MSG examples complications
51 * Borrow configury beautifications from PHP
54 * Bugfix: XBT_LOG_NEW_DEFAULT_CATEGORY now compiles without compiler
55 warning (thanks loris for stumbling into this one).
58 * Change SIMGRID_INSTALL_PATH to GRAS_ROOT in Makefiles generated for user.
59 * Rename gras_get_my_fqdn to gras_os_myname and implement it in the simulator
60 RL would imply a DNS resolver, which is *hard* to do in a portable way
61 (and therefore delayed).
65 SimGrid (2.93) unstable; urgency=low
67 Alpha 4 on the path to SimGrid 3 (aka the "neuf-trois" version)
70 - Use Paje properly where used. Still to be sanitized properly.
71 - Portability fix: Add an implementation of the contexts using pthread
75 - Add xbt_procname(): returns the name of the current process.
76 Use it to show the current process's name in all logging.
78 - fix detection of older flex version and the reaction, since we do
79 depend on modern ones (we use lex_destroy)
80 - Better separation of SG and RL in the libs: remove all simulation code
81 from libgras. As a result, this lib is now only 200k when stripped.
82 Some of the xbt modules may also be duplicated (two sets and such) and
83 should be cleaned/killed before SG3.
84 - Insist on using xlC on AIX because of weird problems involving gcc there.
85 - Cleanup the make remote stuff. This is now done by scripts
86 tools/graspe-{master,slave} (GRAS Platform Expender). This is still
87 mainly for our private use, but we're working on changing them to user
90 - Bugfix: flush the socket on close only if there is some *output*.
91 - Bugfix: flush idempotent when there's nothing to send (don't send size=0)
93 - Add MSG_task_get_name. The task names are mainly for debugging purpose,
96 -- SimGrid team <simgrid2-users@listes.ens-lyon.fr> Fri, 4 Mar 2005 14:32:37 -0800
98 SimGrid (2.92) unstable; urgency=low
100 Alpha 3 on the path to SimGrid 3
104 - New! First try of benchmarking macros.
105 - New! First try so that gras_stub_generator generate deployment and
106 remote compilation helpers.
108 - Bugfix: Initialization fix in msg_test.
112 - Bugfix: applied patch to lexer so that it doesn't need a huge heap.
114 - Bugfix: let dicts work with NULL content (_foreach didn't) and cleanups
116 - API Change: gras_os_sleep to take the amount of seconds as a double.
117 Accepting an int was error prone since it was the only location where
118 seconds were coded as such. It leaded to damn rounding errors.
119 - Bugfix: Hard to belive that timers ever worked before this.
121 -- SimGrid team <simgrid2-users@listes.ens-lyon.fr> Wed, 23 Feb 2005 22:09:21 +0100
123 SimGrid (2.91) unstable; urgency=low
125 Alpha 2 on the path to SimGrid 3
129 - Bug fix in the lmm_solver.
131 - New! Interface to Paje (see http://www-id.imag.fr/Logiciels/paje/)
132 through the function MSG_paje_output.
133 - New! Introducing two new functions MSG_process_kill() and MSG_process_killall().
134 - It is possible to bound the rate of a communication in MSG with
135 MSG_task_put_bounded() (was already in the previous version but I had forgotten
136 to write it in the changelog).
137 - Bug fix to let GRAS run on top of MSG until we move it directly on top
142 - Various cleanups to the autotools stuff
143 - Begin to move Gras examples to examples/gras/
144 - Let make distcheck work again (yeah!)
146 - documentation overhauled using doxygen.
147 gtk-doc-tools is dead in SimGrid now.
148 - Automatically extract all existing logging categories, and add the list
149 to the documentation (long standing one, to say the less)
151 - Cleanup the known architecture table. Reorder the entries to group what
152 should be, and use a more consistent naming scheme.
153 (some of the test dataset are still to be regenerated)
154 - New! Allow library to register globals on each process just as userdata
156 This is implemented using a xbt_dict and not a xbt_set, so we loose the
157 lookup time (for now).
158 Use it in msg and trp.
159 This cleans a lot the internals and helps enforcing privacy of the
160 headers between the gras components.
161 - New! Add a timer mechanism, not unlike cron(8) and at(1).
162 - Bugfix: gras_os_time was delirious in RL.
163 - Bugfix: gras_trp_select/RL don't run into the wall when asked to select
165 - Reenable GRAS now that it works.
167 -- Arnaud Legrand <Arnaud.Legrand@imag.fr> Mon, 14 Feb 2005 14:02:13 -0800
169 SimGrid (2.90) unstable; urgency=low
171 Alpha 1 on the path to SimGrid 3
173 * It is a long time since the last release of SimGrid. I'm sorry about
174 that but as I had told you, I was rewriting a lot of things. I apologize
175 to those who had been reporting bugs to me and that I had not answered.
176 If your bug is still in the new version, please tell me. Here is a
177 summary of the main changes.
179 * REVOLUTION 1: The SimGrid project has merged with the GRAS project
180 lead by Martin Quinson. As a consequence SimGrid gains a lot in
181 portability, speed, and a lot more but you'll figure it out later.
182 SimGrid now comprises 3 different projects : MSG, GRAS and SMPI.
183 I wanted to release the new MSG as soon as possible and I have
184 broken GRAS, which is the reason why, for now, only MSG is fully
185 functional. A laconic description of these projects is available
186 in the documentation.
188 * REVOLUTION 2: I have removed SG and I am now using a new simulation
189 kernel optimized for our needs (called SURF but only the developers
190 should use it). Hence, MSG is now roughly 30 times faster and I think
191 that by rewriting a little bit MSG, I could event speed it up a little
192 bit more. Beside the gain in speed, it is also much easier to encode a
193 new platform model with SURF than it was with SG. More to come...
195 * REVOLUTION 3: I have tried to change a little as possible the API of
196 MSG but a few things really had to disappear. The main differences
197 with the previous version are :
198 1) no more m_links_t and the corresponding functions. Platforms are
199 directly read from a XML description and cannot be hard-coded
200 anymore. The same format is used for application deployment
201 description. The new format is described in the documentation.
202 Have a look in tools/platform_generation. There is a tiny script
203 that converts from the old platform format to the new one. Concerning
204 the application deployment format, parsing the old one is tricky.
205 I think most of you should however be able to convert your files. If
206 it is really an issue, I can write a C code that does the conversion.
208 2) the toolbox tbx does not exist anymore. We now have a library
209 with much more data-structures but without the hash-tables (we have
210 dictionaries that are much faster).
212 -- Arnaud Legrand <Arnaud.Legrand@imag.fr> Mon, 31 Jan 2005 10:45:53 -0800
214 *****************************************************************************
215 * Follows the old GRAS changelog. It does not follow the same syntax, but I *
216 * don't feel like converting the oldies. (Mt) *
217 *****************************************************************************
220 Version 2.90: "the long awaited one"
221 - Finished rewriting and debugging MSG. Rewrote the documentation.
222 - disable GRAS for now since it needs to be ported to the newest SG
225 - Finish the port to windows (using mingw32 for cross-compile)
228 - Main loop and datastructures of SURF. A cpu resource object is
229 functional. Surf can thus be used to create cpu's with variable
230 performance on which you can execute some actions.
232 2004-11-15 Martin Quinson
233 - Port to ARM. Simply added the alignment and size descriptions. Should
234 work, but the ARM machines are so slow that I didn't had the opportunity
235 to 'make check' over there yet.
237 2004-11-15 Arnaud Legrand
238 - Trace manager now written. It uses a heap structure and is therefore
239 expected to be efficient. It may however be speeded up (particularly
240 when many events occur at the same date) by using red and black
241 trees. One day maybe...
242 - Max-min linear system solver written. It uses a sparse matrix
243 structure taking advantage of its expected use. Most operations are
244 O(1) and free/calloc are called as few as possible. The computation of
245 the minimum could however be improved by using a red and black tree
248 2004-11-03 Arnaud Legrand
249 - Rename every gras_* function that was in xbt/ to its xbt_
251 - Add a heap and a doubly-linked list to xbt
252 - Added a dichotomy to the dictionaries. make check works as well before
253 so I assume that the patch is correct. I do not know however if things
254 run effectively faster than before now. :)
256 Inclusion of the SimGrid tree in the GRAS one. The archive is renamed to
257 SimGrid, and the version number is bumped to 2.x
259 2004-10-29 Martin Quinson
260 - Introduction of the remote errors.
261 They are the result of a RMI/RPC on the remote machine.
262 ErrCodes being scalar values, you can't get the host on which those
263 errors did happen. Extending the error mechanism as in Gnome is possible.
264 No idea yet whether it is a good idea.
266 2004-10-28 Martin Quinson
267 - Interface revolution: the Starred Structure Eradication.
268 I used to do typedef struct {} toto_t; and then handle *toto_t.
269 Arnaud (and Oli) didn't like it, and I surrendered. Now, you have:
270 - ???_t is a valid type (builded with typedef)
271 - s_toto_t is a structure (access to fields with .)
272 - s_toto is a structure needing 'struct' keyword to be used
273 - e_toto_t is an enum
274 - toto_t is an 'object' (struct*)
276 typedef struct s_toto {} s_toto_t, *toto_t;
277 typedef enum {} e_toto_t;
278 Moreover, only toto_t (and e_toto_t) are public. The rest (mainly
279 s_toto_t) is private.
281 - While I was at it, all gras_<obj>_free() functions want a gras_<obj>_t*
282 so that it can set the variable to NULL. It was so for dicts and sets,
283 it changed for dynars.
285 - Fix a bunch of memleaks in dict_remove
286 - Fix a bug in sg/server_socket opening: it failed all the time.
288 2004-10-07 Martin Quinson
289 - Speed up dynar lookup operation a bit.
291 gras_dynar_get is dead.
293 Now, you can choose between gras_dynar_get_cpy (the old gras_dynar_get
294 but should be avoided for efficiency reasons) and gras_dynar_get_ptr
295 (which gives you the address of the stored data).
297 gras_dynar_get_as is an helpful macro which allows you to retrieve a
298 copy of the data using an affectation to do the job and not a memcpy.
300 int toto = gras_dynar_get_as(dyn,0,int); rewrites itself to
301 int toto = *(int*)gras_dynar_get_ptr(dyn,0);
303 It does not really speedup the dynar test because they are
304 setting elements all the time (and look them seldom). But the dict does
305 far more lookup than setting.
307 So, this brings the dict_crash test from ~33s to ~25s (200000 elms).
309 2004-10-05 Martin Quinson
310 - Allow to (en/dis)able the cycle detection at run time.
312 Whether we should check for cycle or not is now a property of each
313 datatype. When you think there may be some cycle, use datadesc_cycle_set.
314 datadesc_cycle_unset allow to remove this property when previously set.
316 Note that the cycle detection is off by default since it impacts the
317 performance. Watch the data you feed GRAS with ;)
319 This property is hereditary. Any element embedded in a structure having it
320 set have it set for the time of this data exchange.
322 You should set it both on sender and receiver side. If you don't set it on
323 sender side, it will enter an endless loop. If you forget on receiver
324 side, the cycles won't be recreated after communication.
326 - Header reorganization.
327 Kill gras_private.h, each submodule must load the headers it needs.
329 2004-10-04 Martin Quinson
330 - Interface revolution: do not try to survive to malloc failure.
332 Now, gras_malloc and friends call gras_abort() on failure.
333 As a conclusion, malloc_error is not a valid error anymore, and all
334 functions for which it was the only gras_error_t return value are
335 changed. They now return void, or there result directly.
336 This simplify the API a lot.
338 2004-09-29 Martin Quinson
339 - Re-enable raw sockets.
340 Created by gras_socket_{client,server}_ext;
341 Used with gras_raw_{send,recv}
344 It should allow to kill the last bits of gras first version soon.
346 This is not completely satisfactory yet (duplicate code with
347 chunk_{send,recv}; a bit out of the plugin mechanism), but it should
350 - Simplify transport plugin (internal) interface by not passing any
351 argument to _server and _client, but embedding them in the socket
354 2004-09-28 Martin Quinson
355 - Finish the port to AIX.
356 autoconf was my problem (segfault within the malloc replacement
357 function. No idea why)
359 2004-09-16 Martin Quinson
360 - Fix some size_t madness on 64bit architectures.
362 2004-09-08 Martin Quinson
363 - Reduce the number of system headers loaded, overload some more system
364 calls (such as malloc to cast the result of the system one, and work
366 - Fix and reintroduce the config support
368 2004-09-07 Martin Quinson
369 - Source code reorganization to allow Arnaud to surf all over there.
370 - Allow to document the logging categories.
371 - Remove all uppercase from logging categories and useless cleanup in names.
373 2004-08-18 Martin Quinson
374 Version 0.6.2 (protocol not changed; API changed)
375 - Interface cleanup: gras_msgtype_by_name returns the type (instead of a
376 gras_error_t), and NULL when not found. Functions expecting a msgtype
377 as argument (msg_wait; msg_send) deal with NULL argument by providing a
378 hopefully usefull message.
379 - Portability to prehistoric sparcs again
381 2004-08-17 Martin Quinson
382 Version 0.6.1 (protocol not changed; ABI not changed)
383 - prealloc some buffers to speed things up
385 2004-08-11 Martin Quinson
386 Version 0.6 (protocol not changed; ABI expended)
387 - The parsing macro can deal with the references, provided that you add
388 the relevant annotations (using GRAS_ANNOTE(size,field_name))
390 2004-08-09 Martin Quinson
391 Version 0.5 (protocol not changed; ABI changed)
392 - Allow to off turn the cycle detection code in data exchange at
393 compilation time. It should be at run time, but I'm short of time (and
394 the config stuff is still broken). That way, we keep dict out of the
395 critical path, which is good because the performance is poor:
396 - search not dichotomial yet
397 - dynar give no way to access their content and memcpy everytime
398 - In composed data description (struct, ref and so on), stop foolness of
399 keeping the subtype's ID, but store the type itself. This keeps sets out
400 of the critical path, which is good since they rely on dynar and
401 dictionnaries. The only loose of that is that we cannot detect the
402 redeclaration of a structure/union with another content (but I'm not sure
403 the code detected well this error before anyway). We still can detect
404 the redefinition discrepancy for the other types.
405 - Use a whole bunch of optimisation flags (plus -fno-strict-aliasing since
406 it breaks the code because of type-punning used all over the place).
407 This breaks on all non-gcc architectures (for now).
409 All those changes (plus the buffer of last time) allow me to gain 2 order
410 of magnitude on cruel tests consisting of 800000 array of integers on two
411 level of a hierarchical structure (200 secondes -> 4 secondes)
414 - the selector of reference must now return the type it points to, not
417 2004-08-06 Martin Quinson
418 Version 0.4 (protocol changed; ABI not changed)
419 - Allow to pass --gras-log argument to processes in simulation mode. Really.
420 - New debugging level: trace (under debug) to see effect of GRAS_IN/OUT
421 - Implement a buffer transport, and use it by default (it relies on tcp in
422 real life and on sg in simulation).
423 That's a bit hackish since I had a new field to the structure to store
424 its data without interfering with the subtype ones. Inheritance
425 is tricky in C. And that's a kind of reverse inheritance with one class
426 derivating two classes. Or maybe a game with java interfaces. Anyway,
427 that's damn hard in C (at least).
428 Moreover, I got tired while trying to ensure plugin separation and
429 genericity in SG mode. MSG wants me to do weird things, so let's go for
430 cruel hacks (temporarily of course ;).
431 See comment in transport_private.h:71
432 - do not include all the _interface headers in private but in the files
433 which really need them (to cut the compilation time when they are
436 2004-07-26 Martin Quinson
437 Version 0.3 (protocol not changed; ABI changed)
438 - Major overhault of the datadesc interface to simplify it:
439 - shorted the function names:
440 s/gras_datadesc_declare_struct/gras_datadesc_struct/ and so on
441 - add a trivial way to push/pop integers into the cbps without malloc.
442 This allows to make really generic sub_type description, which simply
443 pop their size of the stack.
444 - add a function gras_datadesc_ref_pop_arr() which does what users want
445 most of the time: Declare a dynamic array (which pops its size of the
446 stack) and declare a reference to it. Poor name, but anyway.
447 - kill the post-send callback, add a post-receive one
449 2004-07-23 Martin Quinson
450 Version 0.2 (protocol changed; ABI changed)
451 - add some testing for cpbs in the test cases, and fix some more bugs.
452 This invalidate again the little64 data file, since I cannot regenerate
454 - remove an awfull optimization in the logging stuff, allowing me to:
455 - understand it again
456 - learn gcc how to check that the argument match the provided format
457 - fix all errors revealed by gcc after that
458 - internal keys of dict are not \0 terminated. Deal with it properly in
459 loggings instead of segfaulting when the user want to see the logs :-/
461 2004-07-22 Martin Quinson
462 - Fix some stupid bug preventing cbps (callback postit) from working
464 2004-07-21 Martin Quinson
465 - Some documentation cleanups
466 - remove the useless last argument of msgtype_declare
467 - rename the Virtu functions to fit into the 'os' namespace
468 - move headers src/include -> src/include/gras/ and stop fooling with
469 gras -> . symbolic link
470 - make distcheck is now successful
472 2004-07-19 Martin Quinson
474 - Build shared library also
475 - Install html doc to the right location
476 - stop removing maintainer files in make clean
477 - build tests only on make check
479 2004-07-13 Martin Quinson
481 - No major issue in previous version => change versionning schema
482 - Re-enable little64 convertion test now that Abdou kindly regenerated the
483 corresponding dataset.
485 2004-07-11 Martin Quinson
487 - Get it working with any kind of structure (we can compute the padding
488 bytes remotely for all the architectures I have access to)
489 - Implement the structure parsing macro (still not quite robust/complete)
490 - Improvement to the remote testing toysuite
492 2004-07-10 Martin Quinson
494 - get ride of a bunch of deprecated macros
495 - actually run the test for two-compliment, not only compile it :-/
496 - test whether the structures get packed (and bail out if yes. Damn.
497 Alignment is a serious matter)
498 - test whether the structures get compacted (but respecting the alignment
499 constraints of each types)
500 - test whether the array fields of structures can straddle alignment boundaries
502 - Damnit, double are bigger than float (typo in creation of 'double' datadesc)
503 (took me 2 hours to find that bug, looking at the wrong place)
504 - Add gras_datadesc_declare_{union,struct}_close(). They must be used
505 before sending/receiving and are used to compute the offsets of fields
506 - Given that padding size depend even on compiler options, keep track of
507 alignment and aligned_size only for the current architecture. Not a big
508 deal since we send structure fields one after the other (seems
510 - Add the datastructure used for IEEE paper by the PBIO guys to the test
511 program, let it work on linux/gcc/little32. portability todo.
513 2004-07-08 Martin Quinson
514 - import and improve remote compilation support from FAST
515 - make sure make check works on half a dozen of machines out there
517 2004-07-07 Martin Quinson
518 Let's say it's version 0.0.3 ;)
519 - Implement conversions (yuhu!)
520 - Let it work on solaris (beside conversion, of course)
521 - Stupid me, using rand() to generate the conversion datatests in not wise.
523 2004-07-06 Martin Quinson
524 - Let make dist work, since I'm gonna need it to compile on remote hosts
525 - Let Tests/datadesc_usage write the architecture on which the file was
526 generated as first byte.
527 - Add PowerPC (being also IRIX64), SPARC (also power4) and ALPHA
528 architecture descriptions.
529 - Add datadesc_usage.{i386,ppc,sparc} files being the result of execution
530 on those architectures.
531 - Optimization: send/recv array of scalar in one shoot
533 2004-07-05 Martin Quinson
534 - YEAH! GRAS/SG and GRAS/RL are both able to run the ping example !
536 - Plug a whole bunch of memleaks
537 - each process now have to call gras_{init,exit}. One day, their log
538 settings will be separated
539 - Continue the code factorisation between SG, RL and common in Transport.
541 2004-07-04 Martin Quinson
543 - Redistribution between SG and RL.
544 We wanna have to accept in SG, so move accepted related parts of RL in
545 the common part. (more precisely, the dynar of all known sockets is no
546 more a static in transport.c, but part of the process_data)
548 [gras_stub_generator]
549 - Bug fix: Do call gras_process_init from gras_init (wasnt called in RL).
551 2004-07-03 Martin Quinson
552 - Create a new log channel tbx containing dict, set, log, dynar (to shut
553 them all up in one shot)
555 - Fix the ugly case of reference to dynamic array.
556 - New (semi-public) function gras_datadesc_size to allow the messaging
557 layer to malloc the needed space for the buffer.
559 - gras_socket_close now expect the socket to close (and not its address to
560 put NULL in it after it). This is because the socket passed to handlers
561 is one of their argument (=> not writable).
563 - propagate the interface cleanup from last week in datadesc, ie remove a
564 superfluous level of indirection. User pass adress of variable
565 containing data (both when sending and receiving), and not of a variable
566 being a pointer to the data. Let's say that I like it better ;)
567 The price for that is constructs like "int msg=*(int*)payload" in
568 handlers, but it's a fine price, IMHO.
570 - Let it work in RL (yuhu)
572 2004-06-21 Martin Quinson
574 - porting SG plugin and SG select to new standards (works almost).
575 - plug memleaks and fix bugs around.
578 - cleanup the prototype of data recv and force users to specify when they
579 want to handle references to objects. Test case working even for cycles.
580 - plug memleaks. Valgrind is perfectly ok with this.
582 2004-06-12 Martin Quinson
584 - cleanup the separation between plugin and main code in plugin creation
586 2004-06-11 Martin Quinson
588 - Reput hook for raw sockets, needed for BW experiments
589 - kill a few lines of dead code
590 [Data description] Interface cleanup
591 - gras_datadesc_by_name returns the searched type or NULL.
592 That way, no variable is needed to use a type desc once, which makes
594 - gras_datadesc_declare_[struct|union]_append_name is removed. The last
595 two parameters were strings (field name, type name), leading to
597 [Dicos] Interface cleanup
598 - gras_dico_retrieve -> gras_dico_get ; gras_dico_insert -> gras_dico_set
599 This is consistant with the dynar API.
601 2004-04-21 Martin Quinson
603 - Porting to new standards.
606 There is no bag anymore, no need to take extra provision to mask the
607 pointers behind "ID".
608 Better splitup of functions between files create/exchange/convert.
609 This is still a bit artificial since convert and receive are so
610 interleaved, but anyway.
612 - add a queued message list to procdata (the ones not matching criteria
614 - factorize some more code between SG and RL wrt procdata
616 - use gras_exit in example to track memleaks
617 - get rid of gs_example now that GS is properly integrated into gras
618 - update run_test to integrate the lastest tests (datadesc)
620 - rename WARNINGn macros to WARNn since it prooved error-prone
622 2004-04-19 Martin Quinson
624 - register init/exit functions within gras module mechanism
625 - send/receive function.
626 Convertion is not implemented, but short-cutted if not needed.
627 struct/array elements are sent one by one (instead of block-wise), but
628 nobody really cares (yet). Get a prototype before optimizing.
629 - tests (using a file socket) for DD send/receive on:
630 - base types: int, float
631 - array: fixed size, string (ie ref to dynamic string)
632 - structure: homogeneous, heterogeneous
633 - chained list, graph with cycle
634 Believe it or not, valgrind is not too unhappy with the results. The
635 cycle happily segfaults, but the others are ok. And I'm sick of pointers
639 - Bugfix when using a filename explicitely (instead of '-')
641 2004-04-09 Martin Quinson
643 - factorize more code between RL and SG in socket creation
644 - Complete the implementation and tests of:
646 o file (only in RL, and mainly for debugging)
648 I lost 3 days to design a portable address resolver, and then decided
649 that the prototype mainly have to run on my box.
650 Addressing portability too early may be like optimizing too early :-/
652 - use gras_init in the Tests instead of the crappy parse_log_opt
653 (the latter function is removed)
654 [Conditional execution]
655 - New functions: gras_if_RL/gras_if_SG (basic support for this)
656 [Code reorganisation]
657 - Get rid of libgrasutils.a since it makes more trouble than it solves.
658 Build examples against the RL library, since there is no way to disable
659 its creation for now.