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 Plutot finaliser ce qui est commencé et finir le Grand Nettoyage:
12 - déplacer les exemples msg pour faire une seule passe doxygen
13 - regarder ce qu'on peut couper de la xbt
14 - diffusion "pull" dans graspe
15 - certains tests n'utilisent pas les loggeries, et/ou sont trop bavards
22 * Update the documentation 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 * Rewrite tools/MSG_visualization/colorize.pl for using with logs instead
34 * Removed msg_test from testsuite. Using the newer version of examples as
38 * Add xbt_os_time(). This functions returns the number of seconds
39 since the Epoch (00:00:00 UTC, January 1, 1970). Most users should
40 use gras_os_time and should not use this function unless they
41 really know what they are doing.
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).
49 * Bugfix: XBT_LOG_NEW_DEFAULT_CATEGORY now compiles without compiler
50 warning (thanks loris for stumbling into this one).
52 * Change SIMGRID_INSTALL_PATH to GRAS_ROOT in Makefiles generated for user.
53 * Rename gras_get_my_fqdn to gras_os_myname and implement it in the simulator
54 RL would imply a DNS resolver, which is *hard* to do in a portable way.
58 SimGrid (2.93) unstable; urgency=low
60 Alpha 4 on the path to SimGrid 3 (aka the "neuf-trois" version)
63 - Use Paje properly where used. Still to be sanitized properly.
64 - Portability fix: Add an implementation of the contexts using pthread
68 - Add xbt_procname(): returns the name of the current process.
69 Use it to show the current process's name in all logging.
71 - fix detection of older flex version and the reaction, since we do
72 depend on modern ones (we use lex_destroy)
73 - Better separation of SG and RL in the libs: remove all simulation code
74 from libgras. As a result, this lib is now only 200k when stripped.
75 Some of the xbt modules may also be duplicated (two sets and such) and
76 should be cleaned/killed before SG3.
77 - Insist on using xlC on AIX because of weird problems involving gcc there.
78 - Cleanup the make remote stuff. This is now done by scripts
79 tools/graspe-{master,slave} (GRAS Platform Expender). This is still
80 mainly for our private use, but we're working on changing them to user
83 - Bugfix: flush the socket on close only if there is some *output*.
84 - Bugfix: flush idempotent when there's nothing to send (don't send size=0)
86 - Add MSG_task_get_name. The task names are mainly for debugging purpose,
89 -- SimGrid team <simgrid2-users@listes.ens-lyon.fr> Fri, 4 Mar 2005 14:32:37 -0800
91 SimGrid (2.92) unstable; urgency=low
93 Alpha 3 on the path to SimGrid 3
97 - New! First try of benchmarking macros.
98 - New! First try so that gras_stub_generator generate deployment and
99 remote compilation helpers.
101 - Bugfix: Initialization fix in msg_test.
105 - Bugfix: applied patch to lexer so that it doesn't need a huge heap.
107 - Bugfix: let dicts work with NULL content (_foreach didn't) and cleanups
109 - API Change: gras_os_sleep to take the amount of seconds as a double.
110 Accepting an int was error prone since it was the only location where
111 seconds were coded as such. It leaded to damn rounding errors.
112 - Bugfix: Hard to belive that timers ever worked before this.
114 -- SimGrid team <simgrid2-users@listes.ens-lyon.fr> Wed, 23 Feb 2005 22:09:21 +0100
116 SimGrid (2.91) unstable; urgency=low
118 Alpha 2 on the path to SimGrid 3
122 - Bug fix in the lmm_solver.
124 - New! Interface to Paje (see http://www-id.imag.fr/Logiciels/paje/)
125 through the function MSG_paje_output.
126 - New! Introducing two new functions MSG_process_kill() and MSG_process_killall().
127 - It is possible to bound the rate of a communication in MSG with
128 MSG_task_put_bounded() (was already in the previous version but I had forgotten
129 to write it in the changelog).
130 - Bug fix to let GRAS run on top of MSG until we move it directly on top
135 - Various cleanups to the autotools stuff
136 - Begin to move Gras examples to examples/gras/
137 - Let make distcheck work again (yeah!)
139 - documentation overhauled using doxygen.
140 gtk-doc-tools is dead in SimGrid now.
141 - Automatically extract all existing logging categories, and add the list
142 to the documentation (long standing one, to say the less)
144 - Cleanup the known architecture table. Reorder the entries to group what
145 should be, and use a more consistent naming scheme.
146 (some of the test dataset are still to be regenerated)
147 - New! Allow library to register globals on each process just as userdata
149 This is implemented using a xbt_dict and not a xbt_set, so we loose the
150 lookup time (for now).
151 Use it in msg and trp.
152 This cleans a lot the internals and helps enforcing privacy of the
153 headers between the gras components.
154 - New! Add a timer mechanism, not unlike cron(8) and at(1).
155 - Bugfix: gras_os_time was delirious in RL.
156 - Bugfix: gras_trp_select/RL don't run into the wall when asked to select
158 - Reenable GRAS now that it works.
160 -- Arnaud Legrand <Arnaud.Legrand@imag.fr> Mon, 14 Feb 2005 14:02:13 -0800
162 SimGrid (2.90) unstable; urgency=low
164 Alpha 1 on the path to SimGrid 3
166 * It is a long time since the last release of SimGrid. I'm sorry about
167 that but as I had told you, I was rewriting a lot of things. I apologize
168 to those who had been reporting bugs to me and that I had not answered.
169 If your bug is still in the new version, please tell me. Here is a
170 summary of the main changes.
172 * REVOLUTION 1: The SimGrid project has merged with the GRAS project
173 lead by Martin Quinson. As a consequence SimGrid gains a lot in
174 portability, speed, and a lot more but you'll figure it out later.
175 SimGrid now comprises 3 different projects : MSG, GRAS and SMPI.
176 I wanted to release the new MSG as soon as possible and I have
177 broken GRAS, which is the reason why, for now, only MSG is fully
178 functional. A laconic description of these projects is available
179 in the documentation.
181 * REVOLUTION 2: I have removed SG and I am now using a new simulation
182 kernel optimized for our needs (called SURF but only the developers
183 should use it). Hence, MSG is now roughly 30 times faster and I think
184 that by rewriting a little bit MSG, I could event speed it up a little
185 bit more. Beside the gain in speed, it is also much easier to encode a
186 new platform model with SURF than it was with SG. More to come...
188 * REVOLUTION 3: I have tried to change a little as possible the API of
189 MSG but a few things really had to disappear. The main differences
190 with the previous version are :
191 1) no more m_links_t and the corresponding functions. Platforms are
192 directly read from a XML description and cannot be hard-coded
193 anymore. The same format is used for application deployment
194 description. The new format is described in the documentation.
195 Have a look in tools/platform_generation. There is a tiny script
196 that converts from the old platform format to the new one. Concerning
197 the application deployment format, parsing the old one is tricky.
198 I think most of you should however be able to convert your files. If
199 it is really an issue, I can write a C code that does the conversion.
201 2) the toolbox tbx does not exist anymore. We now have a library
202 with much more data-structures but without the hash-tables (we have
203 dictionaries that are much faster).
205 -- Arnaud Legrand <Arnaud.Legrand@imag.fr> Mon, 31 Jan 2005 10:45:53 -0800
207 *****************************************************************************
208 * Follows the old GRAS changelog. It does not follow the same syntax, but I *
209 * don't feel like converting the oldies. (Mt) *
210 *****************************************************************************
213 Version 2.90: "the long awaited one"
214 - Finished rewriting and debugging MSG. Rewrote the documentation.
215 - disable GRAS for now since it needs to be ported to the newest SG
218 - Finish the port to windows (using mingw32 for cross-compile)
221 - Main loop and datastructures of SURF. A cpu resource object is
222 functional. Surf can thus be used to create cpu's with variable
223 performance on which you can execute some actions.
225 2004-11-15 Martin Quinson
226 - Port to ARM. Simply added the alignment and size descriptions. Should
227 work, but the ARM machines are so slow that I didn't had the opportunity
228 to 'make check' over there yet.
230 2004-11-15 Arnaud Legrand
231 - Trace manager now written. It uses a heap structure and is therefore
232 expected to be efficient. It may however be speeded up (particularly
233 when many events occur at the same date) by using red and black
234 trees. One day maybe...
235 - Max-min linear system solver written. It uses a sparse matrix
236 structure taking advantage of its expected use. Most operations are
237 O(1) and free/calloc are called as few as possible. The computation of
238 the minimum could however be improved by using a red and black tree
241 2004-11-03 Arnaud Legrand
242 - Rename every gras_* function that was in xbt/ to its xbt_
244 - Add a heap and a doubly-linked list to xbt
245 - Added a dichotomy to the dictionaries. make check works as well before
246 so I assume that the patch is correct. I do not know however if things
247 run effectively faster than before now. :)
249 Inclusion of the SimGrid tree in the GRAS one. The archive is renamed to
250 SimGrid, and the version number is bumped to 2.x
252 2004-10-29 Martin Quinson
253 - Introduction of the remote errors.
254 They are the result of a RMI/RPC on the remote machine.
255 ErrCodes being scalar values, you can't get the host on which those
256 errors did happen. Extending the error mechanism as in Gnome is possible.
257 No idea yet whether it is a good idea.
259 2004-10-28 Martin Quinson
260 - Interface revolution: the Starred Structure Eradication.
261 I used to do typedef struct {} toto_t; and then handle *toto_t.
262 Arnaud (and Oli) didn't like it, and I surrendered. Now, you have:
263 - ???_t is a valid type (builded with typedef)
264 - s_toto_t is a structure (access to fields with .)
265 - s_toto is a structure needing 'struct' keyword to be used
266 - e_toto_t is an enum
267 - toto_t is an 'object' (struct*)
269 typedef struct s_toto {} s_toto_t, *toto_t;
270 typedef enum {} e_toto_t;
271 Moreover, only toto_t (and e_toto_t) are public. The rest (mainly
272 s_toto_t) is private.
274 - While I was at it, all gras_<obj>_free() functions want a gras_<obj>_t*
275 so that it can set the variable to NULL. It was so for dicts and sets,
276 it changed for dynars.
278 - Fix a bunch of memleaks in dict_remove
279 - Fix a bug in sg/server_socket opening: it failed all the time.
281 2004-10-07 Martin Quinson
282 - Speed up dynar lookup operation a bit.
284 gras_dynar_get is dead.
286 Now, you can choose between gras_dynar_get_cpy (the old gras_dynar_get
287 but should be avoided for efficiency reasons) and gras_dynar_get_ptr
288 (which gives you the address of the stored data).
290 gras_dynar_get_as is an helpful macro which allows you to retrieve a
291 copy of the data using an affectation to do the job and not a memcpy.
293 int toto = gras_dynar_get_as(dyn,0,int); rewrites itself to
294 int toto = *(int*)gras_dynar_get_ptr(dyn,0);
296 It does not really speedup the dynar test because they are
297 setting elements all the time (and look them seldom). But the dict does
298 far more lookup than setting.
300 So, this brings the dict_crash test from ~33s to ~25s (200000 elms).
302 2004-10-05 Martin Quinson
303 - Allow to (en/dis)able the cycle detection at run time.
305 Whether we should check for cycle or not is now a property of each
306 datatype. When you think there may be some cycle, use datadesc_cycle_set.
307 datadesc_cycle_unset allow to remove this property when previously set.
309 Note that the cycle detection is off by default since it impacts the
310 performance. Watch the data you feed GRAS with ;)
312 This property is hereditary. Any element embedded in a structure having it
313 set have it set for the time of this data exchange.
315 You should set it both on sender and receiver side. If you don't set it on
316 sender side, it will enter an endless loop. If you forget on receiver
317 side, the cycles won't be recreated after communication.
319 - Header reorganization.
320 Kill gras_private.h, each submodule must load the headers it needs.
322 2004-10-04 Martin Quinson
323 - Interface revolution: do not try to survive to malloc failure.
325 Now, gras_malloc and friends call gras_abort() on failure.
326 As a conclusion, malloc_error is not a valid error anymore, and all
327 functions for which it was the only gras_error_t return value are
328 changed. They now return void, or there result directly.
329 This simplify the API a lot.
331 2004-09-29 Martin Quinson
332 - Re-enable raw sockets.
333 Created by gras_socket_{client,server}_ext;
334 Used with gras_raw_{send,recv}
337 It should allow to kill the last bits of gras first version soon.
339 This is not completely satisfactory yet (duplicate code with
340 chunk_{send,recv}; a bit out of the plugin mechanism), but it should
343 - Simplify transport plugin (internal) interface by not passing any
344 argument to _server and _client, but embedding them in the socket
347 2004-09-28 Martin Quinson
348 - Finish the port to AIX.
349 autoconf was my problem (segfault within the malloc replacement
350 function. No idea why)
352 2004-09-16 Martin Quinson
353 - Fix some size_t madness on 64bit architectures.
355 2004-09-08 Martin Quinson
356 - Reduce the number of system headers loaded, overload some more system
357 calls (such as malloc to cast the result of the system one, and work
359 - Fix and reintroduce the config support
361 2004-09-07 Martin Quinson
362 - Source code reorganization to allow Arnaud to surf all over there.
363 - Allow to document the logging categories.
364 - Remove all uppercase from logging categories and useless cleanup in names.
366 2004-08-18 Martin Quinson
367 Version 0.6.2 (protocol not changed; API changed)
368 - Interface cleanup: gras_msgtype_by_name returns the type (instead of a
369 gras_error_t), and NULL when not found. Functions expecting a msgtype
370 as argument (msg_wait; msg_send) deal with NULL argument by providing a
371 hopefully usefull message.
372 - Portability to prehistoric sparcs again
374 2004-08-17 Martin Quinson
375 Version 0.6.1 (protocol not changed; ABI not changed)
376 - prealloc some buffers to speed things up
378 2004-08-11 Martin Quinson
379 Version 0.6 (protocol not changed; ABI expended)
380 - The parsing macro can deal with the references, provided that you add
381 the relevant annotations (using GRAS_ANNOTE(size,field_name))
383 2004-08-09 Martin Quinson
384 Version 0.5 (protocol not changed; ABI changed)
385 - Allow to off turn the cycle detection code in data exchange at
386 compilation time. It should be at run time, but I'm short of time (and
387 the config stuff is still broken). That way, we keep dict out of the
388 critical path, which is good because the performance is poor:
389 - search not dichotomial yet
390 - dynar give no way to access their content and memcpy everytime
391 - In composed data description (struct, ref and so on), stop foolness of
392 keeping the subtype's ID, but store the type itself. This keeps sets out
393 of the critical path, which is good since they rely on dynar and
394 dictionnaries. The only loose of that is that we cannot detect the
395 redeclaration of a structure/union with another content (but I'm not sure
396 the code detected well this error before anyway). We still can detect
397 the redefinition discrepancy for the other types.
398 - Use a whole bunch of optimisation flags (plus -fno-strict-aliasing since
399 it breaks the code because of type-punning used all over the place).
400 This breaks on all non-gcc architectures (for now).
402 All those changes (plus the buffer of last time) allow me to gain 2 order
403 of magnitude on cruel tests consisting of 800000 array of integers on two
404 level of a hierarchical structure (200 secondes -> 4 secondes)
407 - the selector of reference must now return the type it points to, not
410 2004-08-06 Martin Quinson
411 Version 0.4 (protocol changed; ABI not changed)
412 - Allow to pass --gras-log argument to processes in simulation mode. Really.
413 - New debugging level: trace (under debug) to see effect of GRAS_IN/OUT
414 - Implement a buffer transport, and use it by default (it relies on tcp in
415 real life and on sg in simulation).
416 That's a bit hackish since I had a new field to the structure to store
417 its data without interfering with the subtype ones. Inheritance
418 is tricky in C. And that's a kind of reverse inheritance with one class
419 derivating two classes. Or maybe a game with java interfaces. Anyway,
420 that's damn hard in C (at least).
421 Moreover, I got tired while trying to ensure plugin separation and
422 genericity in SG mode. MSG wants me to do weird things, so let's go for
423 cruel hacks (temporarily of course ;).
424 See comment in transport_private.h:71
425 - do not include all the _interface headers in private but in the files
426 which really need them (to cut the compilation time when they are
429 2004-07-26 Martin Quinson
430 Version 0.3 (protocol not changed; ABI changed)
431 - Major overhault of the datadesc interface to simplify it:
432 - shorted the function names:
433 s/gras_datadesc_declare_struct/gras_datadesc_struct/ and so on
434 - add a trivial way to push/pop integers into the cbps without malloc.
435 This allows to make really generic sub_type description, which simply
436 pop their size of the stack.
437 - add a function gras_datadesc_ref_pop_arr() which does what users want
438 most of the time: Declare a dynamic array (which pops its size of the
439 stack) and declare a reference to it. Poor name, but anyway.
440 - kill the post-send callback, add a post-receive one
442 2004-07-23 Martin Quinson
443 Version 0.2 (protocol changed; ABI changed)
444 - add some testing for cpbs in the test cases, and fix some more bugs.
445 This invalidate again the little64 data file, since I cannot regenerate
447 - remove an awfull optimization in the logging stuff, allowing me to:
448 - understand it again
449 - learn gcc how to check that the argument match the provided format
450 - fix all errors revealed by gcc after that
451 - internal keys of dict are not \0 terminated. Deal with it properly in
452 loggings instead of segfaulting when the user want to see the logs :-/
454 2004-07-22 Martin Quinson
455 - Fix some stupid bug preventing cbps (callback postit) from working
457 2004-07-21 Martin Quinson
458 - Some documentation cleanups
459 - remove the useless last argument of msgtype_declare
460 - rename the Virtu functions to fit into the 'os' namespace
461 - move headers src/include -> src/include/gras/ and stop fooling with
462 gras -> . symbolic link
463 - make distcheck is now successful
465 2004-07-19 Martin Quinson
467 - Build shared library also
468 - Install html doc to the right location
469 - stop removing maintainer files in make clean
470 - build tests only on make check
472 2004-07-13 Martin Quinson
474 - No major issue in previous version => change versionning schema
475 - Re-enable little64 convertion test now that Abdou kindly regenerated the
476 corresponding dataset.
478 2004-07-11 Martin Quinson
480 - Get it working with any kind of structure (we can compute the padding
481 bytes remotely for all the architectures I have access to)
482 - Implement the structure parsing macro (still not quite robust/complete)
483 - Improvement to the remote testing toysuite
485 2004-07-10 Martin Quinson
487 - get ride of a bunch of deprecated macros
488 - actually run the test for two-compliment, not only compile it :-/
489 - test whether the structures get packed (and bail out if yes. Damn.
490 Alignment is a serious matter)
491 - test whether the structures get compacted (but respecting the alignment
492 constraints of each types)
493 - test whether the array fields of structures can straddle alignment boundaries
495 - Damnit, double are bigger than float (typo in creation of 'double' datadesc)
496 (took me 2 hours to find that bug, looking at the wrong place)
497 - Add gras_datadesc_declare_{union,struct}_close(). They must be used
498 before sending/receiving and are used to compute the offsets of fields
499 - Given that padding size depend even on compiler options, keep track of
500 alignment and aligned_size only for the current architecture. Not a big
501 deal since we send structure fields one after the other (seems
503 - Add the datastructure used for IEEE paper by the PBIO guys to the test
504 program, let it work on linux/gcc/little32. portability todo.
506 2004-07-08 Martin Quinson
507 - import and improve remote compilation support from FAST
508 - make sure make check works on half a dozen of machines out there
510 2004-07-07 Martin Quinson
511 Let's say it's version 0.0.3 ;)
512 - Implement conversions (yuhu!)
513 - Let it work on solaris (beside conversion, of course)
514 - Stupid me, using rand() to generate the conversion datatests in not wise.
516 2004-07-06 Martin Quinson
517 - Let make dist work, since I'm gonna need it to compile on remote hosts
518 - Let Tests/datadesc_usage write the architecture on which the file was
519 generated as first byte.
520 - Add PowerPC (being also IRIX64), SPARC (also power4) and ALPHA
521 architecture descriptions.
522 - Add datadesc_usage.{i386,ppc,sparc} files being the result of execution
523 on those architectures.
524 - Optimization: send/recv array of scalar in one shoot
526 2004-07-05 Martin Quinson
527 - YEAH! GRAS/SG and GRAS/RL are both able to run the ping example !
529 - Plug a whole bunch of memleaks
530 - each process now have to call gras_{init,exit}. One day, their log
531 settings will be separated
532 - Continue the code factorisation between SG, RL and common in Transport.
534 2004-07-04 Martin Quinson
536 - Redistribution between SG and RL.
537 We wanna have to accept in SG, so move accepted related parts of RL in
538 the common part. (more precisely, the dynar of all known sockets is no
539 more a static in transport.c, but part of the process_data)
541 [gras_stub_generator]
542 - Bug fix: Do call gras_process_init from gras_init (wasnt called in RL).
544 2004-07-03 Martin Quinson
545 - Create a new log channel tbx containing dict, set, log, dynar (to shut
546 them all up in one shot)
548 - Fix the ugly case of reference to dynamic array.
549 - New (semi-public) function gras_datadesc_size to allow the messaging
550 layer to malloc the needed space for the buffer.
552 - gras_socket_close now expect the socket to close (and not its address to
553 put NULL in it after it). This is because the socket passed to handlers
554 is one of their argument (=> not writable).
556 - propagate the interface cleanup from last week in datadesc, ie remove a
557 superfluous level of indirection. User pass adress of variable
558 containing data (both when sending and receiving), and not of a variable
559 being a pointer to the data. Let's say that I like it better ;)
560 The price for that is constructs like "int msg=*(int*)payload" in
561 handlers, but it's a fine price, IMHO.
563 - Let it work in RL (yuhu)
565 2004-06-21 Martin Quinson
567 - porting SG plugin and SG select to new standards (works almost).
568 - plug memleaks and fix bugs around.
571 - cleanup the prototype of data recv and force users to specify when they
572 want to handle references to objects. Test case working even for cycles.
573 - plug memleaks. Valgrind is perfectly ok with this.
575 2004-06-12 Martin Quinson
577 - cleanup the separation between plugin and main code in plugin creation
579 2004-06-11 Martin Quinson
581 - Reput hook for raw sockets, needed for BW experiments
582 - kill a few lines of dead code
583 [Data description] Interface cleanup
584 - gras_datadesc_by_name returns the searched type or NULL.
585 That way, no variable is needed to use a type desc once, which makes
587 - gras_datadesc_declare_[struct|union]_append_name is removed. The last
588 two parameters were strings (field name, type name), leading to
590 [Dicos] Interface cleanup
591 - gras_dico_retrieve -> gras_dico_get ; gras_dico_insert -> gras_dico_set
592 This is consistant with the dynar API.
594 2004-04-21 Martin Quinson
596 - Porting to new standards.
599 There is no bag anymore, no need to take extra provision to mask the
600 pointers behind "ID".
601 Better splitup of functions between files create/exchange/convert.
602 This is still a bit artificial since convert and receive are so
603 interleaved, but anyway.
605 - add a queued message list to procdata (the ones not matching criteria
607 - factorize some more code between SG and RL wrt procdata
609 - use gras_exit in example to track memleaks
610 - get rid of gs_example now that GS is properly integrated into gras
611 - update run_test to integrate the lastest tests (datadesc)
613 - rename WARNINGn macros to WARNn since it prooved error-prone
615 2004-04-19 Martin Quinson
617 - register init/exit functions within gras module mechanism
618 - send/receive function.
619 Convertion is not implemented, but short-cutted if not needed.
620 struct/array elements are sent one by one (instead of block-wise), but
621 nobody really cares (yet). Get a prototype before optimizing.
622 - tests (using a file socket) for DD send/receive on:
623 - base types: int, float
624 - array: fixed size, string (ie ref to dynamic string)
625 - structure: homogeneous, heterogeneous
626 - chained list, graph with cycle
627 Believe it or not, valgrind is not too unhappy with the results. The
628 cycle happily segfaults, but the others are ok. And I'm sick of pointers
632 - Bugfix when using a filename explicitely (instead of '-')
634 2004-04-09 Martin Quinson
636 - factorize more code between RL and SG in socket creation
637 - Complete the implementation and tests of:
639 o file (only in RL, and mainly for debugging)
641 I lost 3 days to design a portable address resolver, and then decided
642 that the prototype mainly have to run on my box.
643 Addressing portability too early may be like optimizing too early :-/
645 - use gras_init in the Tests instead of the crappy parse_log_opt
646 (the latter function is removed)
647 [Conditional execution]
648 - New functions: gras_if_RL/gras_if_SG (basic support for this)
649 [Code reorganisation]
650 - Get rid of libgrasutils.a since it makes more trouble than it solves.
651 Build examples against the RL library, since there is no way to disable
652 its creation for now.