1 SimGrid (2.94-cvs) UNRELEASED
7 - Faire descendre gras de msg?
8 - Compilation de data_send/recv?
9 Je pense que c'est plutot post SG3, tout ca.
11 La révolution des exceptions !!
13 Plutot finaliser ce qui est commencé et finir le Grand Nettoyage:
14 - déplacer les exemples msg pour faire une seule passe doxygen
15 - regarder ce qu'on peut couper de la xbt
16 - diffusion "pull" dans graspe
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).
50 * Bugfix: XBT_LOG_NEW_DEFAULT_CATEGORY now compiles without compiler
51 warning (thanks loris for stumbling into this one).
53 * Change SIMGRID_INSTALL_PATH to GRAS_ROOT in Makefiles generated for user.
54 * Rename gras_get_my_fqdn to gras_os_myname and implement it in the simulator
55 RL would imply a DNS resolver, which is *hard* to do in a portable way
56 (and therefore delayed).
60 SimGrid (2.93) unstable; urgency=low
62 Alpha 4 on the path to SimGrid 3 (aka the "neuf-trois" version)
65 - Use Paje properly where used. Still to be sanitized properly.
66 - Portability fix: Add an implementation of the contexts using pthread
70 - Add xbt_procname(): returns the name of the current process.
71 Use it to show the current process's name in all logging.
73 - fix detection of older flex version and the reaction, since we do
74 depend on modern ones (we use lex_destroy)
75 - Better separation of SG and RL in the libs: remove all simulation code
76 from libgras. As a result, this lib is now only 200k when stripped.
77 Some of the xbt modules may also be duplicated (two sets and such) and
78 should be cleaned/killed before SG3.
79 - Insist on using xlC on AIX because of weird problems involving gcc there.
80 - Cleanup the make remote stuff. This is now done by scripts
81 tools/graspe-{master,slave} (GRAS Platform Expender). This is still
82 mainly for our private use, but we're working on changing them to user
85 - Bugfix: flush the socket on close only if there is some *output*.
86 - Bugfix: flush idempotent when there's nothing to send (don't send size=0)
88 - Add MSG_task_get_name. The task names are mainly for debugging purpose,
91 -- SimGrid team <simgrid2-users@listes.ens-lyon.fr> Fri, 4 Mar 2005 14:32:37 -0800
93 SimGrid (2.92) unstable; urgency=low
95 Alpha 3 on the path to SimGrid 3
99 - New! First try of benchmarking macros.
100 - New! First try so that gras_stub_generator generate deployment and
101 remote compilation helpers.
103 - Bugfix: Initialization fix in msg_test.
107 - Bugfix: applied patch to lexer so that it doesn't need a huge heap.
109 - Bugfix: let dicts work with NULL content (_foreach didn't) and cleanups
111 - API Change: gras_os_sleep to take the amount of seconds as a double.
112 Accepting an int was error prone since it was the only location where
113 seconds were coded as such. It leaded to damn rounding errors.
114 - Bugfix: Hard to belive that timers ever worked before this.
116 -- SimGrid team <simgrid2-users@listes.ens-lyon.fr> Wed, 23 Feb 2005 22:09:21 +0100
118 SimGrid (2.91) unstable; urgency=low
120 Alpha 2 on the path to SimGrid 3
124 - Bug fix in the lmm_solver.
126 - New! Interface to Paje (see http://www-id.imag.fr/Logiciels/paje/)
127 through the function MSG_paje_output.
128 - New! Introducing two new functions MSG_process_kill() and MSG_process_killall().
129 - It is possible to bound the rate of a communication in MSG with
130 MSG_task_put_bounded() (was already in the previous version but I had forgotten
131 to write it in the changelog).
132 - Bug fix to let GRAS run on top of MSG until we move it directly on top
137 - Various cleanups to the autotools stuff
138 - Begin to move Gras examples to examples/gras/
139 - Let make distcheck work again (yeah!)
141 - documentation overhauled using doxygen.
142 gtk-doc-tools is dead in SimGrid now.
143 - Automatically extract all existing logging categories, and add the list
144 to the documentation (long standing one, to say the less)
146 - Cleanup the known architecture table. Reorder the entries to group what
147 should be, and use a more consistent naming scheme.
148 (some of the test dataset are still to be regenerated)
149 - New! Allow library to register globals on each process just as userdata
151 This is implemented using a xbt_dict and not a xbt_set, so we loose the
152 lookup time (for now).
153 Use it in msg and trp.
154 This cleans a lot the internals and helps enforcing privacy of the
155 headers between the gras components.
156 - New! Add a timer mechanism, not unlike cron(8) and at(1).
157 - Bugfix: gras_os_time was delirious in RL.
158 - Bugfix: gras_trp_select/RL don't run into the wall when asked to select
160 - Reenable GRAS now that it works.
162 -- Arnaud Legrand <Arnaud.Legrand@imag.fr> Mon, 14 Feb 2005 14:02:13 -0800
164 SimGrid (2.90) unstable; urgency=low
166 Alpha 1 on the path to SimGrid 3
168 * It is a long time since the last release of SimGrid. I'm sorry about
169 that but as I had told you, I was rewriting a lot of things. I apologize
170 to those who had been reporting bugs to me and that I had not answered.
171 If your bug is still in the new version, please tell me. Here is a
172 summary of the main changes.
174 * REVOLUTION 1: The SimGrid project has merged with the GRAS project
175 lead by Martin Quinson. As a consequence SimGrid gains a lot in
176 portability, speed, and a lot more but you'll figure it out later.
177 SimGrid now comprises 3 different projects : MSG, GRAS and SMPI.
178 I wanted to release the new MSG as soon as possible and I have
179 broken GRAS, which is the reason why, for now, only MSG is fully
180 functional. A laconic description of these projects is available
181 in the documentation.
183 * REVOLUTION 2: I have removed SG and I am now using a new simulation
184 kernel optimized for our needs (called SURF but only the developers
185 should use it). Hence, MSG is now roughly 30 times faster and I think
186 that by rewriting a little bit MSG, I could event speed it up a little
187 bit more. Beside the gain in speed, it is also much easier to encode a
188 new platform model with SURF than it was with SG. More to come...
190 * REVOLUTION 3: I have tried to change a little as possible the API of
191 MSG but a few things really had to disappear. The main differences
192 with the previous version are :
193 1) no more m_links_t and the corresponding functions. Platforms are
194 directly read from a XML description and cannot be hard-coded
195 anymore. The same format is used for application deployment
196 description. The new format is described in the documentation.
197 Have a look in tools/platform_generation. There is a tiny script
198 that converts from the old platform format to the new one. Concerning
199 the application deployment format, parsing the old one is tricky.
200 I think most of you should however be able to convert your files. If
201 it is really an issue, I can write a C code that does the conversion.
203 2) the toolbox tbx does not exist anymore. We now have a library
204 with much more data-structures but without the hash-tables (we have
205 dictionaries that are much faster).
207 -- Arnaud Legrand <Arnaud.Legrand@imag.fr> Mon, 31 Jan 2005 10:45:53 -0800
209 *****************************************************************************
210 * Follows the old GRAS changelog. It does not follow the same syntax, but I *
211 * don't feel like converting the oldies. (Mt) *
212 *****************************************************************************
215 Version 2.90: "the long awaited one"
216 - Finished rewriting and debugging MSG. Rewrote the documentation.
217 - disable GRAS for now since it needs to be ported to the newest SG
220 - Finish the port to windows (using mingw32 for cross-compile)
223 - Main loop and datastructures of SURF. A cpu resource object is
224 functional. Surf can thus be used to create cpu's with variable
225 performance on which you can execute some actions.
227 2004-11-15 Martin Quinson
228 - Port to ARM. Simply added the alignment and size descriptions. Should
229 work, but the ARM machines are so slow that I didn't had the opportunity
230 to 'make check' over there yet.
232 2004-11-15 Arnaud Legrand
233 - Trace manager now written. It uses a heap structure and is therefore
234 expected to be efficient. It may however be speeded up (particularly
235 when many events occur at the same date) by using red and black
236 trees. One day maybe...
237 - Max-min linear system solver written. It uses a sparse matrix
238 structure taking advantage of its expected use. Most operations are
239 O(1) and free/calloc are called as few as possible. The computation of
240 the minimum could however be improved by using a red and black tree
243 2004-11-03 Arnaud Legrand
244 - Rename every gras_* function that was in xbt/ to its xbt_
246 - Add a heap and a doubly-linked list to xbt
247 - Added a dichotomy to the dictionaries. make check works as well before
248 so I assume that the patch is correct. I do not know however if things
249 run effectively faster than before now. :)
251 Inclusion of the SimGrid tree in the GRAS one. The archive is renamed to
252 SimGrid, and the version number is bumped to 2.x
254 2004-10-29 Martin Quinson
255 - Introduction of the remote errors.
256 They are the result of a RMI/RPC on the remote machine.
257 ErrCodes being scalar values, you can't get the host on which those
258 errors did happen. Extending the error mechanism as in Gnome is possible.
259 No idea yet whether it is a good idea.
261 2004-10-28 Martin Quinson
262 - Interface revolution: the Starred Structure Eradication.
263 I used to do typedef struct {} toto_t; and then handle *toto_t.
264 Arnaud (and Oli) didn't like it, and I surrendered. Now, you have:
265 - ???_t is a valid type (builded with typedef)
266 - s_toto_t is a structure (access to fields with .)
267 - s_toto is a structure needing 'struct' keyword to be used
268 - e_toto_t is an enum
269 - toto_t is an 'object' (struct*)
271 typedef struct s_toto {} s_toto_t, *toto_t;
272 typedef enum {} e_toto_t;
273 Moreover, only toto_t (and e_toto_t) are public. The rest (mainly
274 s_toto_t) is private.
276 - While I was at it, all gras_<obj>_free() functions want a gras_<obj>_t*
277 so that it can set the variable to NULL. It was so for dicts and sets,
278 it changed for dynars.
280 - Fix a bunch of memleaks in dict_remove
281 - Fix a bug in sg/server_socket opening: it failed all the time.
283 2004-10-07 Martin Quinson
284 - Speed up dynar lookup operation a bit.
286 gras_dynar_get is dead.
288 Now, you can choose between gras_dynar_get_cpy (the old gras_dynar_get
289 but should be avoided for efficiency reasons) and gras_dynar_get_ptr
290 (which gives you the address of the stored data).
292 gras_dynar_get_as is an helpful macro which allows you to retrieve a
293 copy of the data using an affectation to do the job and not a memcpy.
295 int toto = gras_dynar_get_as(dyn,0,int); rewrites itself to
296 int toto = *(int*)gras_dynar_get_ptr(dyn,0);
298 It does not really speedup the dynar test because they are
299 setting elements all the time (and look them seldom). But the dict does
300 far more lookup than setting.
302 So, this brings the dict_crash test from ~33s to ~25s (200000 elms).
304 2004-10-05 Martin Quinson
305 - Allow to (en/dis)able the cycle detection at run time.
307 Whether we should check for cycle or not is now a property of each
308 datatype. When you think there may be some cycle, use datadesc_cycle_set.
309 datadesc_cycle_unset allow to remove this property when previously set.
311 Note that the cycle detection is off by default since it impacts the
312 performance. Watch the data you feed GRAS with ;)
314 This property is hereditary. Any element embedded in a structure having it
315 set have it set for the time of this data exchange.
317 You should set it both on sender and receiver side. If you don't set it on
318 sender side, it will enter an endless loop. If you forget on receiver
319 side, the cycles won't be recreated after communication.
321 - Header reorganization.
322 Kill gras_private.h, each submodule must load the headers it needs.
324 2004-10-04 Martin Quinson
325 - Interface revolution: do not try to survive to malloc failure.
327 Now, gras_malloc and friends call gras_abort() on failure.
328 As a conclusion, malloc_error is not a valid error anymore, and all
329 functions for which it was the only gras_error_t return value are
330 changed. They now return void, or there result directly.
331 This simplify the API a lot.
333 2004-09-29 Martin Quinson
334 - Re-enable raw sockets.
335 Created by gras_socket_{client,server}_ext;
336 Used with gras_raw_{send,recv}
339 It should allow to kill the last bits of gras first version soon.
341 This is not completely satisfactory yet (duplicate code with
342 chunk_{send,recv}; a bit out of the plugin mechanism), but it should
345 - Simplify transport plugin (internal) interface by not passing any
346 argument to _server and _client, but embedding them in the socket
349 2004-09-28 Martin Quinson
350 - Finish the port to AIX.
351 autoconf was my problem (segfault within the malloc replacement
352 function. No idea why)
354 2004-09-16 Martin Quinson
355 - Fix some size_t madness on 64bit architectures.
357 2004-09-08 Martin Quinson
358 - Reduce the number of system headers loaded, overload some more system
359 calls (such as malloc to cast the result of the system one, and work
361 - Fix and reintroduce the config support
363 2004-09-07 Martin Quinson
364 - Source code reorganization to allow Arnaud to surf all over there.
365 - Allow to document the logging categories.
366 - Remove all uppercase from logging categories and useless cleanup in names.
368 2004-08-18 Martin Quinson
369 Version 0.6.2 (protocol not changed; API changed)
370 - Interface cleanup: gras_msgtype_by_name returns the type (instead of a
371 gras_error_t), and NULL when not found. Functions expecting a msgtype
372 as argument (msg_wait; msg_send) deal with NULL argument by providing a
373 hopefully usefull message.
374 - Portability to prehistoric sparcs again
376 2004-08-17 Martin Quinson
377 Version 0.6.1 (protocol not changed; ABI not changed)
378 - prealloc some buffers to speed things up
380 2004-08-11 Martin Quinson
381 Version 0.6 (protocol not changed; ABI expended)
382 - The parsing macro can deal with the references, provided that you add
383 the relevant annotations (using GRAS_ANNOTE(size,field_name))
385 2004-08-09 Martin Quinson
386 Version 0.5 (protocol not changed; ABI changed)
387 - Allow to off turn the cycle detection code in data exchange at
388 compilation time. It should be at run time, but I'm short of time (and
389 the config stuff is still broken). That way, we keep dict out of the
390 critical path, which is good because the performance is poor:
391 - search not dichotomial yet
392 - dynar give no way to access their content and memcpy everytime
393 - In composed data description (struct, ref and so on), stop foolness of
394 keeping the subtype's ID, but store the type itself. This keeps sets out
395 of the critical path, which is good since they rely on dynar and
396 dictionnaries. The only loose of that is that we cannot detect the
397 redeclaration of a structure/union with another content (but I'm not sure
398 the code detected well this error before anyway). We still can detect
399 the redefinition discrepancy for the other types.
400 - Use a whole bunch of optimisation flags (plus -fno-strict-aliasing since
401 it breaks the code because of type-punning used all over the place).
402 This breaks on all non-gcc architectures (for now).
404 All those changes (plus the buffer of last time) allow me to gain 2 order
405 of magnitude on cruel tests consisting of 800000 array of integers on two
406 level of a hierarchical structure (200 secondes -> 4 secondes)
409 - the selector of reference must now return the type it points to, not
412 2004-08-06 Martin Quinson
413 Version 0.4 (protocol changed; ABI not changed)
414 - Allow to pass --gras-log argument to processes in simulation mode. Really.
415 - New debugging level: trace (under debug) to see effect of GRAS_IN/OUT
416 - Implement a buffer transport, and use it by default (it relies on tcp in
417 real life and on sg in simulation).
418 That's a bit hackish since I had a new field to the structure to store
419 its data without interfering with the subtype ones. Inheritance
420 is tricky in C. And that's a kind of reverse inheritance with one class
421 derivating two classes. Or maybe a game with java interfaces. Anyway,
422 that's damn hard in C (at least).
423 Moreover, I got tired while trying to ensure plugin separation and
424 genericity in SG mode. MSG wants me to do weird things, so let's go for
425 cruel hacks (temporarily of course ;).
426 See comment in transport_private.h:71
427 - do not include all the _interface headers in private but in the files
428 which really need them (to cut the compilation time when they are
431 2004-07-26 Martin Quinson
432 Version 0.3 (protocol not changed; ABI changed)
433 - Major overhault of the datadesc interface to simplify it:
434 - shorted the function names:
435 s/gras_datadesc_declare_struct/gras_datadesc_struct/ and so on
436 - add a trivial way to push/pop integers into the cbps without malloc.
437 This allows to make really generic sub_type description, which simply
438 pop their size of the stack.
439 - add a function gras_datadesc_ref_pop_arr() which does what users want
440 most of the time: Declare a dynamic array (which pops its size of the
441 stack) and declare a reference to it. Poor name, but anyway.
442 - kill the post-send callback, add a post-receive one
444 2004-07-23 Martin Quinson
445 Version 0.2 (protocol changed; ABI changed)
446 - add some testing for cpbs in the test cases, and fix some more bugs.
447 This invalidate again the little64 data file, since I cannot regenerate
449 - remove an awfull optimization in the logging stuff, allowing me to:
450 - understand it again
451 - learn gcc how to check that the argument match the provided format
452 - fix all errors revealed by gcc after that
453 - internal keys of dict are not \0 terminated. Deal with it properly in
454 loggings instead of segfaulting when the user want to see the logs :-/
456 2004-07-22 Martin Quinson
457 - Fix some stupid bug preventing cbps (callback postit) from working
459 2004-07-21 Martin Quinson
460 - Some documentation cleanups
461 - remove the useless last argument of msgtype_declare
462 - rename the Virtu functions to fit into the 'os' namespace
463 - move headers src/include -> src/include/gras/ and stop fooling with
464 gras -> . symbolic link
465 - make distcheck is now successful
467 2004-07-19 Martin Quinson
469 - Build shared library also
470 - Install html doc to the right location
471 - stop removing maintainer files in make clean
472 - build tests only on make check
474 2004-07-13 Martin Quinson
476 - No major issue in previous version => change versionning schema
477 - Re-enable little64 convertion test now that Abdou kindly regenerated the
478 corresponding dataset.
480 2004-07-11 Martin Quinson
482 - Get it working with any kind of structure (we can compute the padding
483 bytes remotely for all the architectures I have access to)
484 - Implement the structure parsing macro (still not quite robust/complete)
485 - Improvement to the remote testing toysuite
487 2004-07-10 Martin Quinson
489 - get ride of a bunch of deprecated macros
490 - actually run the test for two-compliment, not only compile it :-/
491 - test whether the structures get packed (and bail out if yes. Damn.
492 Alignment is a serious matter)
493 - test whether the structures get compacted (but respecting the alignment
494 constraints of each types)
495 - test whether the array fields of structures can straddle alignment boundaries
497 - Damnit, double are bigger than float (typo in creation of 'double' datadesc)
498 (took me 2 hours to find that bug, looking at the wrong place)
499 - Add gras_datadesc_declare_{union,struct}_close(). They must be used
500 before sending/receiving and are used to compute the offsets of fields
501 - Given that padding size depend even on compiler options, keep track of
502 alignment and aligned_size only for the current architecture. Not a big
503 deal since we send structure fields one after the other (seems
505 - Add the datastructure used for IEEE paper by the PBIO guys to the test
506 program, let it work on linux/gcc/little32. portability todo.
508 2004-07-08 Martin Quinson
509 - import and improve remote compilation support from FAST
510 - make sure make check works on half a dozen of machines out there
512 2004-07-07 Martin Quinson
513 Let's say it's version 0.0.3 ;)
514 - Implement conversions (yuhu!)
515 - Let it work on solaris (beside conversion, of course)
516 - Stupid me, using rand() to generate the conversion datatests in not wise.
518 2004-07-06 Martin Quinson
519 - Let make dist work, since I'm gonna need it to compile on remote hosts
520 - Let Tests/datadesc_usage write the architecture on which the file was
521 generated as first byte.
522 - Add PowerPC (being also IRIX64), SPARC (also power4) and ALPHA
523 architecture descriptions.
524 - Add datadesc_usage.{i386,ppc,sparc} files being the result of execution
525 on those architectures.
526 - Optimization: send/recv array of scalar in one shoot
528 2004-07-05 Martin Quinson
529 - YEAH! GRAS/SG and GRAS/RL are both able to run the ping example !
531 - Plug a whole bunch of memleaks
532 - each process now have to call gras_{init,exit}. One day, their log
533 settings will be separated
534 - Continue the code factorisation between SG, RL and common in Transport.
536 2004-07-04 Martin Quinson
538 - Redistribution between SG and RL.
539 We wanna have to accept in SG, so move accepted related parts of RL in
540 the common part. (more precisely, the dynar of all known sockets is no
541 more a static in transport.c, but part of the process_data)
543 [gras_stub_generator]
544 - Bug fix: Do call gras_process_init from gras_init (wasnt called in RL).
546 2004-07-03 Martin Quinson
547 - Create a new log channel tbx containing dict, set, log, dynar (to shut
548 them all up in one shot)
550 - Fix the ugly case of reference to dynamic array.
551 - New (semi-public) function gras_datadesc_size to allow the messaging
552 layer to malloc the needed space for the buffer.
554 - gras_socket_close now expect the socket to close (and not its address to
555 put NULL in it after it). This is because the socket passed to handlers
556 is one of their argument (=> not writable).
558 - propagate the interface cleanup from last week in datadesc, ie remove a
559 superfluous level of indirection. User pass adress of variable
560 containing data (both when sending and receiving), and not of a variable
561 being a pointer to the data. Let's say that I like it better ;)
562 The price for that is constructs like "int msg=*(int*)payload" in
563 handlers, but it's a fine price, IMHO.
565 - Let it work in RL (yuhu)
567 2004-06-21 Martin Quinson
569 - porting SG plugin and SG select to new standards (works almost).
570 - plug memleaks and fix bugs around.
573 - cleanup the prototype of data recv and force users to specify when they
574 want to handle references to objects. Test case working even for cycles.
575 - plug memleaks. Valgrind is perfectly ok with this.
577 2004-06-12 Martin Quinson
579 - cleanup the separation between plugin and main code in plugin creation
581 2004-06-11 Martin Quinson
583 - Reput hook for raw sockets, needed for BW experiments
584 - kill a few lines of dead code
585 [Data description] Interface cleanup
586 - gras_datadesc_by_name returns the searched type or NULL.
587 That way, no variable is needed to use a type desc once, which makes
589 - gras_datadesc_declare_[struct|union]_append_name is removed. The last
590 two parameters were strings (field name, type name), leading to
592 [Dicos] Interface cleanup
593 - gras_dico_retrieve -> gras_dico_get ; gras_dico_insert -> gras_dico_set
594 This is consistant with the dynar API.
596 2004-04-21 Martin Quinson
598 - Porting to new standards.
601 There is no bag anymore, no need to take extra provision to mask the
602 pointers behind "ID".
603 Better splitup of functions between files create/exchange/convert.
604 This is still a bit artificial since convert and receive are so
605 interleaved, but anyway.
607 - add a queued message list to procdata (the ones not matching criteria
609 - factorize some more code between SG and RL wrt procdata
611 - use gras_exit in example to track memleaks
612 - get rid of gs_example now that GS is properly integrated into gras
613 - update run_test to integrate the lastest tests (datadesc)
615 - rename WARNINGn macros to WARNn since it prooved error-prone
617 2004-04-19 Martin Quinson
619 - register init/exit functions within gras module mechanism
620 - send/receive function.
621 Convertion is not implemented, but short-cutted if not needed.
622 struct/array elements are sent one by one (instead of block-wise), but
623 nobody really cares (yet). Get a prototype before optimizing.
624 - tests (using a file socket) for DD send/receive on:
625 - base types: int, float
626 - array: fixed size, string (ie ref to dynamic string)
627 - structure: homogeneous, heterogeneous
628 - chained list, graph with cycle
629 Believe it or not, valgrind is not too unhappy with the results. The
630 cycle happily segfaults, but the others are ok. And I'm sick of pointers
634 - Bugfix when using a filename explicitely (instead of '-')
636 2004-04-09 Martin Quinson
638 - factorize more code between RL and SG in socket creation
639 - Complete the implementation and tests of:
641 o file (only in RL, and mainly for debugging)
643 I lost 3 days to design a portable address resolver, and then decided
644 that the prototype mainly have to run on my box.
645 Addressing portability too early may be like optimizing too early :-/
647 - use gras_init in the Tests instead of the crappy parse_log_opt
648 (the latter function is removed)
649 [Conditional execution]
650 - New functions: gras_if_RL/gras_if_SG (basic support for this)
651 [Code reorganisation]
652 - Get rid of libgrasutils.a since it makes more trouble than it solves.
653 Build examples against the RL library, since there is no way to disable
654 its creation for now.