1 SimGrid (2.94-cvs) UNRELEASED
7 - Faire descendre gras de msg?
8 - Compilation de data_send/recv?
9 - La révolution des exceptions ?
10 Je pense que c'est plutot post SG3, tout ca.
12 Plutot finaliser ce qui est commencé et finir le Grand Nettoyage:
13 - regarder ce qu'on peut couper de la xbt
14 - diffusion "pull" dans graspe
21 * Update the main page and the FAQ. Adding references to gforge.
24 * Add a gras_os_getpid function.
27 * Add MSG_task_get_compute_duration() and MSG_task_get_data_size()
28 * Extend the logs so that they also print PID, hostname, date, ... if
30 * Convert the MSG example to the use of xbt_logs instead of PRINT_MESSAGE,
31 and kill the old version which were in testsuite/
32 * Rewrite tools/MSG_visualization/colorize.pl for using with logs instead
36 * Add xbt_os_time(). As the rest of xbt/portability, this is not public
37 for users. Instead, each programming environment (GRAS, MSG,...) use it
38 when needed to provide such a feature to users.
39 Don't shortcut the mecanism or you will also shortcut the virtualization
40 you need on the simulator.
44 * Cleanups in configury with regard to compile optimization/warning flags.
45 Also add -fno-loop-optimize to any powerpc since it's the optimization
46 killing gcc (< 3.4.0).
47 * Doxygen cleanups: move MSG examples, kill the second Doxygen phase
48 needed by MSG examples complications
51 * Bugfix: XBT_LOG_NEW_DEFAULT_CATEGORY now compiles without compiler
52 warning (thanks loris for stumbling into this one).
55 * Change SIMGRID_INSTALL_PATH to GRAS_ROOT in Makefiles generated for user.
56 * Rename gras_get_my_fqdn to gras_os_myname and implement it in the simulator
57 RL would imply a DNS resolver, which is *hard* to do in a portable way
58 (and therefore delayed).
62 SimGrid (2.93) unstable; urgency=low
64 Alpha 4 on the path to SimGrid 3 (aka the "neuf-trois" version)
67 - Use Paje properly where used. Still to be sanitized properly.
68 - Portability fix: Add an implementation of the contexts using pthread
72 - Add xbt_procname(): returns the name of the current process.
73 Use it to show the current process's name in all logging.
75 - fix detection of older flex version and the reaction, since we do
76 depend on modern ones (we use lex_destroy)
77 - Better separation of SG and RL in the libs: remove all simulation code
78 from libgras. As a result, this lib is now only 200k when stripped.
79 Some of the xbt modules may also be duplicated (two sets and such) and
80 should be cleaned/killed before SG3.
81 - Insist on using xlC on AIX because of weird problems involving gcc there.
82 - Cleanup the make remote stuff. This is now done by scripts
83 tools/graspe-{master,slave} (GRAS Platform Expender). This is still
84 mainly for our private use, but we're working on changing them to user
87 - Bugfix: flush the socket on close only if there is some *output*.
88 - Bugfix: flush idempotent when there's nothing to send (don't send size=0)
90 - Add MSG_task_get_name. The task names are mainly for debugging purpose,
93 -- SimGrid team <simgrid2-users@listes.ens-lyon.fr> Fri, 4 Mar 2005 14:32:37 -0800
95 SimGrid (2.92) unstable; urgency=low
97 Alpha 3 on the path to SimGrid 3
101 - New! First try of benchmarking macros.
102 - New! First try so that gras_stub_generator generate deployment and
103 remote compilation helpers.
105 - Bugfix: Initialization fix in msg_test.
109 - Bugfix: applied patch to lexer so that it doesn't need a huge heap.
111 - Bugfix: let dicts work with NULL content (_foreach didn't) and cleanups
113 - API Change: gras_os_sleep to take the amount of seconds as a double.
114 Accepting an int was error prone since it was the only location where
115 seconds were coded as such. It leaded to damn rounding errors.
116 - Bugfix: Hard to belive that timers ever worked before this.
118 -- SimGrid team <simgrid2-users@listes.ens-lyon.fr> Wed, 23 Feb 2005 22:09:21 +0100
120 SimGrid (2.91) unstable; urgency=low
122 Alpha 2 on the path to SimGrid 3
126 - Bug fix in the lmm_solver.
128 - New! Interface to Paje (see http://www-id.imag.fr/Logiciels/paje/)
129 through the function MSG_paje_output.
130 - New! Introducing two new functions MSG_process_kill() and MSG_process_killall().
131 - It is possible to bound the rate of a communication in MSG with
132 MSG_task_put_bounded() (was already in the previous version but I had forgotten
133 to write it in the changelog).
134 - Bug fix to let GRAS run on top of MSG until we move it directly on top
139 - Various cleanups to the autotools stuff
140 - Begin to move Gras examples to examples/gras/
141 - Let make distcheck work again (yeah!)
143 - documentation overhauled using doxygen.
144 gtk-doc-tools is dead in SimGrid now.
145 - Automatically extract all existing logging categories, and add the list
146 to the documentation (long standing one, to say the less)
148 - Cleanup the known architecture table. Reorder the entries to group what
149 should be, and use a more consistent naming scheme.
150 (some of the test dataset are still to be regenerated)
151 - New! Allow library to register globals on each process just as userdata
153 This is implemented using a xbt_dict and not a xbt_set, so we loose the
154 lookup time (for now).
155 Use it in msg and trp.
156 This cleans a lot the internals and helps enforcing privacy of the
157 headers between the gras components.
158 - New! Add a timer mechanism, not unlike cron(8) and at(1).
159 - Bugfix: gras_os_time was delirious in RL.
160 - Bugfix: gras_trp_select/RL don't run into the wall when asked to select
162 - Reenable GRAS now that it works.
164 -- Arnaud Legrand <Arnaud.Legrand@imag.fr> Mon, 14 Feb 2005 14:02:13 -0800
166 SimGrid (2.90) unstable; urgency=low
168 Alpha 1 on the path to SimGrid 3
170 * It is a long time since the last release of SimGrid. I'm sorry about
171 that but as I had told you, I was rewriting a lot of things. I apologize
172 to those who had been reporting bugs to me and that I had not answered.
173 If your bug is still in the new version, please tell me. Here is a
174 summary of the main changes.
176 * REVOLUTION 1: The SimGrid project has merged with the GRAS project
177 lead by Martin Quinson. As a consequence SimGrid gains a lot in
178 portability, speed, and a lot more but you'll figure it out later.
179 SimGrid now comprises 3 different projects : MSG, GRAS and SMPI.
180 I wanted to release the new MSG as soon as possible and I have
181 broken GRAS, which is the reason why, for now, only MSG is fully
182 functional. A laconic description of these projects is available
183 in the documentation.
185 * REVOLUTION 2: I have removed SG and I am now using a new simulation
186 kernel optimized for our needs (called SURF but only the developers
187 should use it). Hence, MSG is now roughly 30 times faster and I think
188 that by rewriting a little bit MSG, I could event speed it up a little
189 bit more. Beside the gain in speed, it is also much easier to encode a
190 new platform model with SURF than it was with SG. More to come...
192 * REVOLUTION 3: I have tried to change a little as possible the API of
193 MSG but a few things really had to disappear. The main differences
194 with the previous version are :
195 1) no more m_links_t and the corresponding functions. Platforms are
196 directly read from a XML description and cannot be hard-coded
197 anymore. The same format is used for application deployment
198 description. The new format is described in the documentation.
199 Have a look in tools/platform_generation. There is a tiny script
200 that converts from the old platform format to the new one. Concerning
201 the application deployment format, parsing the old one is tricky.
202 I think most of you should however be able to convert your files. If
203 it is really an issue, I can write a C code that does the conversion.
205 2) the toolbox tbx does not exist anymore. We now have a library
206 with much more data-structures but without the hash-tables (we have
207 dictionaries that are much faster).
209 -- Arnaud Legrand <Arnaud.Legrand@imag.fr> Mon, 31 Jan 2005 10:45:53 -0800
211 *****************************************************************************
212 * Follows the old GRAS changelog. It does not follow the same syntax, but I *
213 * don't feel like converting the oldies. (Mt) *
214 *****************************************************************************
217 Version 2.90: "the long awaited one"
218 - Finished rewriting and debugging MSG. Rewrote the documentation.
219 - disable GRAS for now since it needs to be ported to the newest SG
222 - Finish the port to windows (using mingw32 for cross-compile)
225 - Main loop and datastructures of SURF. A cpu resource object is
226 functional. Surf can thus be used to create cpu's with variable
227 performance on which you can execute some actions.
229 2004-11-15 Martin Quinson
230 - Port to ARM. Simply added the alignment and size descriptions. Should
231 work, but the ARM machines are so slow that I didn't had the opportunity
232 to 'make check' over there yet.
234 2004-11-15 Arnaud Legrand
235 - Trace manager now written. It uses a heap structure and is therefore
236 expected to be efficient. It may however be speeded up (particularly
237 when many events occur at the same date) by using red and black
238 trees. One day maybe...
239 - Max-min linear system solver written. It uses a sparse matrix
240 structure taking advantage of its expected use. Most operations are
241 O(1) and free/calloc are called as few as possible. The computation of
242 the minimum could however be improved by using a red and black tree
245 2004-11-03 Arnaud Legrand
246 - Rename every gras_* function that was in xbt/ to its xbt_
248 - Add a heap and a doubly-linked list to xbt
249 - Added a dichotomy to the dictionaries. make check works as well before
250 so I assume that the patch is correct. I do not know however if things
251 run effectively faster than before now. :)
253 Inclusion of the SimGrid tree in the GRAS one. The archive is renamed to
254 SimGrid, and the version number is bumped to 2.x
256 2004-10-29 Martin Quinson
257 - Introduction of the remote errors.
258 They are the result of a RMI/RPC on the remote machine.
259 ErrCodes being scalar values, you can't get the host on which those
260 errors did happen. Extending the error mechanism as in Gnome is possible.
261 No idea yet whether it is a good idea.
263 2004-10-28 Martin Quinson
264 - Interface revolution: the Starred Structure Eradication.
265 I used to do typedef struct {} toto_t; and then handle *toto_t.
266 Arnaud (and Oli) didn't like it, and I surrendered. Now, you have:
267 - ???_t is a valid type (builded with typedef)
268 - s_toto_t is a structure (access to fields with .)
269 - s_toto is a structure needing 'struct' keyword to be used
270 - e_toto_t is an enum
271 - toto_t is an 'object' (struct*)
273 typedef struct s_toto {} s_toto_t, *toto_t;
274 typedef enum {} e_toto_t;
275 Moreover, only toto_t (and e_toto_t) are public. The rest (mainly
276 s_toto_t) is private.
278 - While I was at it, all gras_<obj>_free() functions want a gras_<obj>_t*
279 so that it can set the variable to NULL. It was so for dicts and sets,
280 it changed for dynars.
282 - Fix a bunch of memleaks in dict_remove
283 - Fix a bug in sg/server_socket opening: it failed all the time.
285 2004-10-07 Martin Quinson
286 - Speed up dynar lookup operation a bit.
288 gras_dynar_get is dead.
290 Now, you can choose between gras_dynar_get_cpy (the old gras_dynar_get
291 but should be avoided for efficiency reasons) and gras_dynar_get_ptr
292 (which gives you the address of the stored data).
294 gras_dynar_get_as is an helpful macro which allows you to retrieve a
295 copy of the data using an affectation to do the job and not a memcpy.
297 int toto = gras_dynar_get_as(dyn,0,int); rewrites itself to
298 int toto = *(int*)gras_dynar_get_ptr(dyn,0);
300 It does not really speedup the dynar test because they are
301 setting elements all the time (and look them seldom). But the dict does
302 far more lookup than setting.
304 So, this brings the dict_crash test from ~33s to ~25s (200000 elms).
306 2004-10-05 Martin Quinson
307 - Allow to (en/dis)able the cycle detection at run time.
309 Whether we should check for cycle or not is now a property of each
310 datatype. When you think there may be some cycle, use datadesc_cycle_set.
311 datadesc_cycle_unset allow to remove this property when previously set.
313 Note that the cycle detection is off by default since it impacts the
314 performance. Watch the data you feed GRAS with ;)
316 This property is hereditary. Any element embedded in a structure having it
317 set have it set for the time of this data exchange.
319 You should set it both on sender and receiver side. If you don't set it on
320 sender side, it will enter an endless loop. If you forget on receiver
321 side, the cycles won't be recreated after communication.
323 - Header reorganization.
324 Kill gras_private.h, each submodule must load the headers it needs.
326 2004-10-04 Martin Quinson
327 - Interface revolution: do not try to survive to malloc failure.
329 Now, gras_malloc and friends call gras_abort() on failure.
330 As a conclusion, malloc_error is not a valid error anymore, and all
331 functions for which it was the only gras_error_t return value are
332 changed. They now return void, or there result directly.
333 This simplify the API a lot.
335 2004-09-29 Martin Quinson
336 - Re-enable raw sockets.
337 Created by gras_socket_{client,server}_ext;
338 Used with gras_raw_{send,recv}
341 It should allow to kill the last bits of gras first version soon.
343 This is not completely satisfactory yet (duplicate code with
344 chunk_{send,recv}; a bit out of the plugin mechanism), but it should
347 - Simplify transport plugin (internal) interface by not passing any
348 argument to _server and _client, but embedding them in the socket
351 2004-09-28 Martin Quinson
352 - Finish the port to AIX.
353 autoconf was my problem (segfault within the malloc replacement
354 function. No idea why)
356 2004-09-16 Martin Quinson
357 - Fix some size_t madness on 64bit architectures.
359 2004-09-08 Martin Quinson
360 - Reduce the number of system headers loaded, overload some more system
361 calls (such as malloc to cast the result of the system one, and work
363 - Fix and reintroduce the config support
365 2004-09-07 Martin Quinson
366 - Source code reorganization to allow Arnaud to surf all over there.
367 - Allow to document the logging categories.
368 - Remove all uppercase from logging categories and useless cleanup in names.
370 2004-08-18 Martin Quinson
371 Version 0.6.2 (protocol not changed; API changed)
372 - Interface cleanup: gras_msgtype_by_name returns the type (instead of a
373 gras_error_t), and NULL when not found. Functions expecting a msgtype
374 as argument (msg_wait; msg_send) deal with NULL argument by providing a
375 hopefully usefull message.
376 - Portability to prehistoric sparcs again
378 2004-08-17 Martin Quinson
379 Version 0.6.1 (protocol not changed; ABI not changed)
380 - prealloc some buffers to speed things up
382 2004-08-11 Martin Quinson
383 Version 0.6 (protocol not changed; ABI expended)
384 - The parsing macro can deal with the references, provided that you add
385 the relevant annotations (using GRAS_ANNOTE(size,field_name))
387 2004-08-09 Martin Quinson
388 Version 0.5 (protocol not changed; ABI changed)
389 - Allow to off turn the cycle detection code in data exchange at
390 compilation time. It should be at run time, but I'm short of time (and
391 the config stuff is still broken). That way, we keep dict out of the
392 critical path, which is good because the performance is poor:
393 - search not dichotomial yet
394 - dynar give no way to access their content and memcpy everytime
395 - In composed data description (struct, ref and so on), stop foolness of
396 keeping the subtype's ID, but store the type itself. This keeps sets out
397 of the critical path, which is good since they rely on dynar and
398 dictionnaries. The only loose of that is that we cannot detect the
399 redeclaration of a structure/union with another content (but I'm not sure
400 the code detected well this error before anyway). We still can detect
401 the redefinition discrepancy for the other types.
402 - Use a whole bunch of optimisation flags (plus -fno-strict-aliasing since
403 it breaks the code because of type-punning used all over the place).
404 This breaks on all non-gcc architectures (for now).
406 All those changes (plus the buffer of last time) allow me to gain 2 order
407 of magnitude on cruel tests consisting of 800000 array of integers on two
408 level of a hierarchical structure (200 secondes -> 4 secondes)
411 - the selector of reference must now return the type it points to, not
414 2004-08-06 Martin Quinson
415 Version 0.4 (protocol changed; ABI not changed)
416 - Allow to pass --gras-log argument to processes in simulation mode. Really.
417 - New debugging level: trace (under debug) to see effect of GRAS_IN/OUT
418 - Implement a buffer transport, and use it by default (it relies on tcp in
419 real life and on sg in simulation).
420 That's a bit hackish since I had a new field to the structure to store
421 its data without interfering with the subtype ones. Inheritance
422 is tricky in C. And that's a kind of reverse inheritance with one class
423 derivating two classes. Or maybe a game with java interfaces. Anyway,
424 that's damn hard in C (at least).
425 Moreover, I got tired while trying to ensure plugin separation and
426 genericity in SG mode. MSG wants me to do weird things, so let's go for
427 cruel hacks (temporarily of course ;).
428 See comment in transport_private.h:71
429 - do not include all the _interface headers in private but in the files
430 which really need them (to cut the compilation time when they are
433 2004-07-26 Martin Quinson
434 Version 0.3 (protocol not changed; ABI changed)
435 - Major overhault of the datadesc interface to simplify it:
436 - shorted the function names:
437 s/gras_datadesc_declare_struct/gras_datadesc_struct/ and so on
438 - add a trivial way to push/pop integers into the cbps without malloc.
439 This allows to make really generic sub_type description, which simply
440 pop their size of the stack.
441 - add a function gras_datadesc_ref_pop_arr() which does what users want
442 most of the time: Declare a dynamic array (which pops its size of the
443 stack) and declare a reference to it. Poor name, but anyway.
444 - kill the post-send callback, add a post-receive one
446 2004-07-23 Martin Quinson
447 Version 0.2 (protocol changed; ABI changed)
448 - add some testing for cpbs in the test cases, and fix some more bugs.
449 This invalidate again the little64 data file, since I cannot regenerate
451 - remove an awfull optimization in the logging stuff, allowing me to:
452 - understand it again
453 - learn gcc how to check that the argument match the provided format
454 - fix all errors revealed by gcc after that
455 - internal keys of dict are not \0 terminated. Deal with it properly in
456 loggings instead of segfaulting when the user want to see the logs :-/
458 2004-07-22 Martin Quinson
459 - Fix some stupid bug preventing cbps (callback postit) from working
461 2004-07-21 Martin Quinson
462 - Some documentation cleanups
463 - remove the useless last argument of msgtype_declare
464 - rename the Virtu functions to fit into the 'os' namespace
465 - move headers src/include -> src/include/gras/ and stop fooling with
466 gras -> . symbolic link
467 - make distcheck is now successful
469 2004-07-19 Martin Quinson
471 - Build shared library also
472 - Install html doc to the right location
473 - stop removing maintainer files in make clean
474 - build tests only on make check
476 2004-07-13 Martin Quinson
478 - No major issue in previous version => change versionning schema
479 - Re-enable little64 convertion test now that Abdou kindly regenerated the
480 corresponding dataset.
482 2004-07-11 Martin Quinson
484 - Get it working with any kind of structure (we can compute the padding
485 bytes remotely for all the architectures I have access to)
486 - Implement the structure parsing macro (still not quite robust/complete)
487 - Improvement to the remote testing toysuite
489 2004-07-10 Martin Quinson
491 - get ride of a bunch of deprecated macros
492 - actually run the test for two-compliment, not only compile it :-/
493 - test whether the structures get packed (and bail out if yes. Damn.
494 Alignment is a serious matter)
495 - test whether the structures get compacted (but respecting the alignment
496 constraints of each types)
497 - test whether the array fields of structures can straddle alignment boundaries
499 - Damnit, double are bigger than float (typo in creation of 'double' datadesc)
500 (took me 2 hours to find that bug, looking at the wrong place)
501 - Add gras_datadesc_declare_{union,struct}_close(). They must be used
502 before sending/receiving and are used to compute the offsets of fields
503 - Given that padding size depend even on compiler options, keep track of
504 alignment and aligned_size only for the current architecture. Not a big
505 deal since we send structure fields one after the other (seems
507 - Add the datastructure used for IEEE paper by the PBIO guys to the test
508 program, let it work on linux/gcc/little32. portability todo.
510 2004-07-08 Martin Quinson
511 - import and improve remote compilation support from FAST
512 - make sure make check works on half a dozen of machines out there
514 2004-07-07 Martin Quinson
515 Let's say it's version 0.0.3 ;)
516 - Implement conversions (yuhu!)
517 - Let it work on solaris (beside conversion, of course)
518 - Stupid me, using rand() to generate the conversion datatests in not wise.
520 2004-07-06 Martin Quinson
521 - Let make dist work, since I'm gonna need it to compile on remote hosts
522 - Let Tests/datadesc_usage write the architecture on which the file was
523 generated as first byte.
524 - Add PowerPC (being also IRIX64), SPARC (also power4) and ALPHA
525 architecture descriptions.
526 - Add datadesc_usage.{i386,ppc,sparc} files being the result of execution
527 on those architectures.
528 - Optimization: send/recv array of scalar in one shoot
530 2004-07-05 Martin Quinson
531 - YEAH! GRAS/SG and GRAS/RL are both able to run the ping example !
533 - Plug a whole bunch of memleaks
534 - each process now have to call gras_{init,exit}. One day, their log
535 settings will be separated
536 - Continue the code factorisation between SG, RL and common in Transport.
538 2004-07-04 Martin Quinson
540 - Redistribution between SG and RL.
541 We wanna have to accept in SG, so move accepted related parts of RL in
542 the common part. (more precisely, the dynar of all known sockets is no
543 more a static in transport.c, but part of the process_data)
545 [gras_stub_generator]
546 - Bug fix: Do call gras_process_init from gras_init (wasnt called in RL).
548 2004-07-03 Martin Quinson
549 - Create a new log channel tbx containing dict, set, log, dynar (to shut
550 them all up in one shot)
552 - Fix the ugly case of reference to dynamic array.
553 - New (semi-public) function gras_datadesc_size to allow the messaging
554 layer to malloc the needed space for the buffer.
556 - gras_socket_close now expect the socket to close (and not its address to
557 put NULL in it after it). This is because the socket passed to handlers
558 is one of their argument (=> not writable).
560 - propagate the interface cleanup from last week in datadesc, ie remove a
561 superfluous level of indirection. User pass adress of variable
562 containing data (both when sending and receiving), and not of a variable
563 being a pointer to the data. Let's say that I like it better ;)
564 The price for that is constructs like "int msg=*(int*)payload" in
565 handlers, but it's a fine price, IMHO.
567 - Let it work in RL (yuhu)
569 2004-06-21 Martin Quinson
571 - porting SG plugin and SG select to new standards (works almost).
572 - plug memleaks and fix bugs around.
575 - cleanup the prototype of data recv and force users to specify when they
576 want to handle references to objects. Test case working even for cycles.
577 - plug memleaks. Valgrind is perfectly ok with this.
579 2004-06-12 Martin Quinson
581 - cleanup the separation between plugin and main code in plugin creation
583 2004-06-11 Martin Quinson
585 - Reput hook for raw sockets, needed for BW experiments
586 - kill a few lines of dead code
587 [Data description] Interface cleanup
588 - gras_datadesc_by_name returns the searched type or NULL.
589 That way, no variable is needed to use a type desc once, which makes
591 - gras_datadesc_declare_[struct|union]_append_name is removed. The last
592 two parameters were strings (field name, type name), leading to
594 [Dicos] Interface cleanup
595 - gras_dico_retrieve -> gras_dico_get ; gras_dico_insert -> gras_dico_set
596 This is consistant with the dynar API.
598 2004-04-21 Martin Quinson
600 - Porting to new standards.
603 There is no bag anymore, no need to take extra provision to mask the
604 pointers behind "ID".
605 Better splitup of functions between files create/exchange/convert.
606 This is still a bit artificial since convert and receive are so
607 interleaved, but anyway.
609 - add a queued message list to procdata (the ones not matching criteria
611 - factorize some more code between SG and RL wrt procdata
613 - use gras_exit in example to track memleaks
614 - get rid of gs_example now that GS is properly integrated into gras
615 - update run_test to integrate the lastest tests (datadesc)
617 - rename WARNINGn macros to WARNn since it prooved error-prone
619 2004-04-19 Martin Quinson
621 - register init/exit functions within gras module mechanism
622 - send/receive function.
623 Convertion is not implemented, but short-cutted if not needed.
624 struct/array elements are sent one by one (instead of block-wise), but
625 nobody really cares (yet). Get a prototype before optimizing.
626 - tests (using a file socket) for DD send/receive on:
627 - base types: int, float
628 - array: fixed size, string (ie ref to dynamic string)
629 - structure: homogeneous, heterogeneous
630 - chained list, graph with cycle
631 Believe it or not, valgrind is not too unhappy with the results. The
632 cycle happily segfaults, but the others are ok. And I'm sick of pointers
636 - Bugfix when using a filename explicitely (instead of '-')
638 2004-04-09 Martin Quinson
640 - factorize more code between RL and SG in socket creation
641 - Complete the implementation and tests of:
643 o file (only in RL, and mainly for debugging)
645 I lost 3 days to design a portable address resolver, and then decided
646 that the prototype mainly have to run on my box.
647 Addressing portability too early may be like optimizing too early :-/
649 - use gras_init in the Tests instead of the crappy parse_log_opt
650 (the latter function is removed)
651 [Conditional execution]
652 - New functions: gras_if_RL/gras_if_SG (basic support for this)
653 [Code reorganisation]
654 - Get rid of libgrasutils.a since it makes more trouble than it solves.
655 Build examples against the RL library, since there is no way to disable
656 its creation for now.