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 - virer le test msg soit de examples/ soit de testsuite/
13 - déplacer les exemples msg pour faire une seule passe doxygen
14 - regarder ce qu'on peut couper de la xbt
15 - diffusion "pull" dans graspe
16 - certains tests n'utilisent pas les loggeries, et/ou sont trop bavards
23 * Add a gras_os_getpid function.
26 * Add MSG_task_get_compute_duration() and MSG_task_get_data_size()
27 * Extend the logs so that they also print PID, hostname, date, ... if
29 * Convert the MSG example to the use of xbt_logs instead of PRINT_MESSAGE
30 * Rewrite tools/MSG_visualization/colorize.pl for using with logs instead
36 * Cleanups in configury with regard to compile optimization/warning flags.
37 Also add -fno-loop-optimize to any powerpc since it's the optimization
38 killing gcc (< 3.4.0).
40 * Bugfix: XBT_LOG_NEW_DEFAULT_CATEGORY now compiles without compiler
41 warning (thanks loris for stumbling into this one).
43 * Change SIMGRID_INSTALL_PATH to GRAS_ROOT in Makefiles generated for user.
44 * Rename gras_get_my_fqdn to gras_os_myname and implement it in the simulator
45 RL would imply a DNS resolver, which is *hard* to do in a portable way.
49 SimGrid (2.93) unstable; urgency=low
51 Alpha 4 on the path to SimGrid 3 (aka the "neuf-trois" version)
54 - Use Paje properly where used. Still to be sanitized properly.
55 - Portability fix: Add an implementation of the contexts using pthread
59 - Add xbt_procname(): returns the name of the current process.
60 Use it to show the current process's name in all logging.
62 - fix detection of older flex version and the reaction, since we do
63 depend on modern ones (we use lex_destroy)
64 - Better separation of SG and RL in the libs: remove all simulation code
65 from libgras. As a result, this lib is now only 200k when stripped.
66 Some of the xbt modules may also be duplicated (two sets and such) and
67 should be cleaned/killed before SG3.
68 - Insist on using xlC on AIX because of weird problems involving gcc there.
69 - Cleanup the make remote stuff. This is now done by scripts
70 tools/graspe-{master,slave} (GRAS Platform Expender). This is still
71 mainly for our private use, but we're working on changing them to user
74 - Bugfix: flush the socket on close only if there is some *output*.
75 - Bugfix: flush idempotent when there's nothing to send (don't send size=0)
77 - Add MSG_task_get_name. The task names are mainly for debugging purpose,
80 -- SimGrid team <simgrid2-users@listes.ens-lyon.fr> Fri, 4 Mar 2005 14:32:37 -0800
82 SimGrid (2.92) unstable; urgency=low
84 Alpha 3 on the path to SimGrid 3
88 - New! First try of benchmarking macros.
89 - New! First try so that gras_stub_generator generate deployment and
90 remote compilation helpers.
92 - Bugfix: Initialization fix in msg_test.
96 - Bugfix: applied patch to lexer so that it doesn't need a huge heap.
98 - Bugfix: let dicts work with NULL content (_foreach didn't) and cleanups
100 - API Change: gras_os_sleep to take the amount of seconds as a double.
101 Accepting an int was error prone since it was the only location where
102 seconds were coded as such. It leaded to damn rounding errors.
103 - Bugfix: Hard to belive that timers ever worked before this.
105 -- SimGrid team <simgrid2-users@listes.ens-lyon.fr> Wed, 23 Feb 2005 22:09:21 +0100
107 SimGrid (2.91) unstable; urgency=low
109 Alpha 2 on the path to SimGrid 3
113 - Bug fix in the lmm_solver.
115 - New! Interface to Paje (see http://www-id.imag.fr/Logiciels/paje/)
116 through the function MSG_paje_output.
117 - New! Introducing two new functions MSG_process_kill() and MSG_process_killall().
118 - It is possible to bound the rate of a communication in MSG with
119 MSG_task_put_bounded() (was already in the previous version but I had forgotten
120 to write it in the changelog).
121 - Bug fix to let GRAS run on top of MSG until we move it directly on top
126 - Various cleanups to the autotools stuff
127 - Begin to move Gras examples to examples/gras/
128 - Let make distcheck work again (yeah!)
130 - documentation overhauled using doxygen.
131 gtk-doc-tools is dead in SimGrid now.
132 - Automatically extract all existing logging categories, and add the list
133 to the documentation (long standing one, to say the less)
135 - Cleanup the known architecture table. Reorder the entries to group what
136 should be, and use a more consistent naming scheme.
137 (some of the test dataset are still to be regenerated)
138 - New! Allow library to register globals on each process just as userdata
140 This is implemented using a xbt_dict and not a xbt_set, so we loose the
141 lookup time (for now).
142 Use it in msg and trp.
143 This cleans a lot the internals and helps enforcing privacy of the
144 headers between the gras components.
145 - New! Add a timer mechanism, not unlike cron(8) and at(1).
146 - Bugfix: gras_os_time was delirious in RL.
147 - Bugfix: gras_trp_select/RL don't run into the wall when asked to select
149 - Reenable GRAS now that it works.
151 -- Arnaud Legrand <Arnaud.Legrand@imag.fr> Mon, 14 Feb 2005 14:02:13 -0800
153 SimGrid (2.90) unstable; urgency=low
155 Alpha 1 on the path to SimGrid 3
157 * It is a long time since the last release of SimGrid. I'm sorry about
158 that but as I had told you, I was rewriting a lot of things. I apologize
159 to those who had been reporting bugs to me and that I had not answered.
160 If your bug is still in the new version, please tell me. Here is a
161 summary of the main changes.
163 * REVOLUTION 1: The SimGrid project has merged with the GRAS project
164 lead by Martin Quinson. As a consequence SimGrid gains a lot in
165 portability, speed, and a lot more but you'll figure it out later.
166 SimGrid now comprises 3 different projects : MSG, GRAS and SMPI.
167 I wanted to release the new MSG as soon as possible and I have
168 broken GRAS, which is the reason why, for now, only MSG is fully
169 functional. A laconic description of these projects is available
170 in the documentation.
172 * REVOLUTION 2: I have removed SG and I am now using a new simulation
173 kernel optimized for our needs (called SURF but only the developers
174 should use it). Hence, MSG is now roughly 30 times faster and I think
175 that by rewriting a little bit MSG, I could event speed it up a little
176 bit more. Beside the gain in speed, it is also much easier to encode a
177 new platform model with SURF than it was with SG. More to come...
179 * REVOLUTION 3: I have tried to change a little as possible the API of
180 MSG but a few things really had to disappear. The main differences
181 with the previous version are :
182 1) no more m_links_t and the corresponding functions. Platforms are
183 directly read from a XML description and cannot be hard-coded
184 anymore. The same format is used for application deployment
185 description. The new format is described in the documentation.
186 Have a look in tools/platform_generation. There is a tiny script
187 that converts from the old platform format to the new one. Concerning
188 the application deployment format, parsing the old one is tricky.
189 I think most of you should however be able to convert your files. If
190 it is really an issue, I can write a C code that does the conversion.
192 2) the toolbox tbx does not exist anymore. We now have a library
193 with much more data-structures but without the hash-tables (we have
194 dictionaries that are much faster).
196 -- Arnaud Legrand <Arnaud.Legrand@imag.fr> Mon, 31 Jan 2005 10:45:53 -0800
198 *****************************************************************************
199 * Follows the old GRAS changelog. It does not follow the same syntax, but I *
200 * don't feel like converting the oldies. (Mt) *
201 *****************************************************************************
204 Version 2.90: "the long awaited one"
205 - Finished rewriting and debugging MSG. Rewrote the documentation.
206 - disable GRAS for now since it needs to be ported to the newest SG
209 - Finish the port to windows (using mingw32 for cross-compile)
212 - Main loop and datastructures of SURF. A cpu resource object is
213 functional. Surf can thus be used to create cpu's with variable
214 performance on which you can execute some actions.
216 2004-11-15 Martin Quinson
217 - Port to ARM. Simply added the alignment and size descriptions. Should
218 work, but the ARM machines are so slow that I didn't had the opportunity
219 to 'make check' over there yet.
221 2004-11-15 Arnaud Legrand
222 - Trace manager now written. It uses a heap structure and is therefore
223 expected to be efficient. It may however be speeded up (particularly
224 when many events occur at the same date) by using red and black
225 trees. One day maybe...
226 - Max-min linear system solver written. It uses a sparse matrix
227 structure taking advantage of its expected use. Most operations are
228 O(1) and free/calloc are called as few as possible. The computation of
229 the minimum could however be improved by using a red and black tree
232 2004-11-03 Arnaud Legrand
233 - Rename every gras_* function that was in xbt/ to its xbt_
235 - Add a heap and a doubly-linked list to xbt
236 - Added a dichotomy to the dictionaries. make check works as well before
237 so I assume that the patch is correct. I do not know however if things
238 run effectively faster than before now. :)
240 Inclusion of the SimGrid tree in the GRAS one. The archive is renamed to
241 SimGrid, and the version number is bumped to 2.x
243 2004-10-29 Martin Quinson
244 - Introduction of the remote errors.
245 They are the result of a RMI/RPC on the remote machine.
246 ErrCodes being scalar values, you can't get the host on which those
247 errors did happen. Extending the error mechanism as in Gnome is possible.
248 No idea yet whether it is a good idea.
250 2004-10-28 Martin Quinson
251 - Interface revolution: the Starred Structure Eradication.
252 I used to do typedef struct {} toto_t; and then handle *toto_t.
253 Arnaud (and Oli) didn't like it, and I surrendered. Now, you have:
254 - ???_t is a valid type (builded with typedef)
255 - s_toto_t is a structure (access to fields with .)
256 - s_toto is a structure needing 'struct' keyword to be used
257 - e_toto_t is an enum
258 - toto_t is an 'object' (struct*)
260 typedef struct s_toto {} s_toto_t, *toto_t;
261 typedef enum {} e_toto_t;
262 Moreover, only toto_t (and e_toto_t) are public. The rest (mainly
263 s_toto_t) is private.
265 - While I was at it, all gras_<obj>_free() functions want a gras_<obj>_t*
266 so that it can set the variable to NULL. It was so for dicts and sets,
267 it changed for dynars.
269 - Fix a bunch of memleaks in dict_remove
270 - Fix a bug in sg/server_socket opening: it failed all the time.
272 2004-10-07 Martin Quinson
273 - Speed up dynar lookup operation a bit.
275 gras_dynar_get is dead.
277 Now, you can choose between gras_dynar_get_cpy (the old gras_dynar_get
278 but should be avoided for efficiency reasons) and gras_dynar_get_ptr
279 (which gives you the address of the stored data).
281 gras_dynar_get_as is an helpful macro which allows you to retrieve a
282 copy of the data using an affectation to do the job and not a memcpy.
284 int toto = gras_dynar_get_as(dyn,0,int); rewrites itself to
285 int toto = *(int*)gras_dynar_get_ptr(dyn,0);
287 It does not really speedup the dynar test because they are
288 setting elements all the time (and look them seldom). But the dict does
289 far more lookup than setting.
291 So, this brings the dict_crash test from ~33s to ~25s (200000 elms).
293 2004-10-05 Martin Quinson
294 - Allow to (en/dis)able the cycle detection at run time.
296 Whether we should check for cycle or not is now a property of each
297 datatype. When you think there may be some cycle, use datadesc_cycle_set.
298 datadesc_cycle_unset allow to remove this property when previously set.
300 Note that the cycle detection is off by default since it impacts the
301 performance. Watch the data you feed GRAS with ;)
303 This property is hereditary. Any element embedded in a structure having it
304 set have it set for the time of this data exchange.
306 You should set it both on sender and receiver side. If you don't set it on
307 sender side, it will enter an endless loop. If you forget on receiver
308 side, the cycles won't be recreated after communication.
310 - Header reorganization.
311 Kill gras_private.h, each submodule must load the headers it needs.
313 2004-10-04 Martin Quinson
314 - Interface revolution: do not try to survive to malloc failure.
316 Now, gras_malloc and friends call gras_abort() on failure.
317 As a conclusion, malloc_error is not a valid error anymore, and all
318 functions for which it was the only gras_error_t return value are
319 changed. They now return void, or there result directly.
320 This simplify the API a lot.
322 2004-09-29 Martin Quinson
323 - Re-enable raw sockets.
324 Created by gras_socket_{client,server}_ext;
325 Used with gras_raw_{send,recv}
328 It should allow to kill the last bits of gras first version soon.
330 This is not completely satisfactory yet (duplicate code with
331 chunk_{send,recv}; a bit out of the plugin mechanism), but it should
334 - Simplify transport plugin (internal) interface by not passing any
335 argument to _server and _client, but embedding them in the socket
338 2004-09-28 Martin Quinson
339 - Finish the port to AIX.
340 autoconf was my problem (segfault within the malloc replacement
341 function. No idea why)
343 2004-09-16 Martin Quinson
344 - Fix some size_t madness on 64bit architectures.
346 2004-09-08 Martin Quinson
347 - Reduce the number of system headers loaded, overload some more system
348 calls (such as malloc to cast the result of the system one, and work
350 - Fix and reintroduce the config support
352 2004-09-07 Martin Quinson
353 - Source code reorganization to allow Arnaud to surf all over there.
354 - Allow to document the logging categories.
355 - Remove all uppercase from logging categories and useless cleanup in names.
357 2004-08-18 Martin Quinson
358 Version 0.6.2 (protocol not changed; API changed)
359 - Interface cleanup: gras_msgtype_by_name returns the type (instead of a
360 gras_error_t), and NULL when not found. Functions expecting a msgtype
361 as argument (msg_wait; msg_send) deal with NULL argument by providing a
362 hopefully usefull message.
363 - Portability to prehistoric sparcs again
365 2004-08-17 Martin Quinson
366 Version 0.6.1 (protocol not changed; ABI not changed)
367 - prealloc some buffers to speed things up
369 2004-08-11 Martin Quinson
370 Version 0.6 (protocol not changed; ABI expended)
371 - The parsing macro can deal with the references, provided that you add
372 the relevant annotations (using GRAS_ANNOTE(size,field_name))
374 2004-08-09 Martin Quinson
375 Version 0.5 (protocol not changed; ABI changed)
376 - Allow to off turn the cycle detection code in data exchange at
377 compilation time. It should be at run time, but I'm short of time (and
378 the config stuff is still broken). That way, we keep dict out of the
379 critical path, which is good because the performance is poor:
380 - search not dichotomial yet
381 - dynar give no way to access their content and memcpy everytime
382 - In composed data description (struct, ref and so on), stop foolness of
383 keeping the subtype's ID, but store the type itself. This keeps sets out
384 of the critical path, which is good since they rely on dynar and
385 dictionnaries. The only loose of that is that we cannot detect the
386 redeclaration of a structure/union with another content (but I'm not sure
387 the code detected well this error before anyway). We still can detect
388 the redefinition discrepancy for the other types.
389 - Use a whole bunch of optimisation flags (plus -fno-strict-aliasing since
390 it breaks the code because of type-punning used all over the place).
391 This breaks on all non-gcc architectures (for now).
393 All those changes (plus the buffer of last time) allow me to gain 2 order
394 of magnitude on cruel tests consisting of 800000 array of integers on two
395 level of a hierarchical structure (200 secondes -> 4 secondes)
398 - the selector of reference must now return the type it points to, not
401 2004-08-06 Martin Quinson
402 Version 0.4 (protocol changed; ABI not changed)
403 - Allow to pass --gras-log argument to processes in simulation mode. Really.
404 - New debugging level: trace (under debug) to see effect of GRAS_IN/OUT
405 - Implement a buffer transport, and use it by default (it relies on tcp in
406 real life and on sg in simulation).
407 That's a bit hackish since I had a new field to the structure to store
408 its data without interfering with the subtype ones. Inheritance
409 is tricky in C. And that's a kind of reverse inheritance with one class
410 derivating two classes. Or maybe a game with java interfaces. Anyway,
411 that's damn hard in C (at least).
412 Moreover, I got tired while trying to ensure plugin separation and
413 genericity in SG mode. MSG wants me to do weird things, so let's go for
414 cruel hacks (temporarily of course ;).
415 See comment in transport_private.h:71
416 - do not include all the _interface headers in private but in the files
417 which really need them (to cut the compilation time when they are
420 2004-07-26 Martin Quinson
421 Version 0.3 (protocol not changed; ABI changed)
422 - Major overhault of the datadesc interface to simplify it:
423 - shorted the function names:
424 s/gras_datadesc_declare_struct/gras_datadesc_struct/ and so on
425 - add a trivial way to push/pop integers into the cbps without malloc.
426 This allows to make really generic sub_type description, which simply
427 pop their size of the stack.
428 - add a function gras_datadesc_ref_pop_arr() which does what users want
429 most of the time: Declare a dynamic array (which pops its size of the
430 stack) and declare a reference to it. Poor name, but anyway.
431 - kill the post-send callback, add a post-receive one
433 2004-07-23 Martin Quinson
434 Version 0.2 (protocol changed; ABI changed)
435 - add some testing for cpbs in the test cases, and fix some more bugs.
436 This invalidate again the little64 data file, since I cannot regenerate
438 - remove an awfull optimization in the logging stuff, allowing me to:
439 - understand it again
440 - learn gcc how to check that the argument match the provided format
441 - fix all errors revealed by gcc after that
442 - internal keys of dict are not \0 terminated. Deal with it properly in
443 loggings instead of segfaulting when the user want to see the logs :-/
445 2004-07-22 Martin Quinson
446 - Fix some stupid bug preventing cbps (callback postit) from working
448 2004-07-21 Martin Quinson
449 - Some documentation cleanups
450 - remove the useless last argument of msgtype_declare
451 - rename the Virtu functions to fit into the 'os' namespace
452 - move headers src/include -> src/include/gras/ and stop fooling with
453 gras -> . symbolic link
454 - make distcheck is now successful
456 2004-07-19 Martin Quinson
458 - Build shared library also
459 - Install html doc to the right location
460 - stop removing maintainer files in make clean
461 - build tests only on make check
463 2004-07-13 Martin Quinson
465 - No major issue in previous version => change versionning schema
466 - Re-enable little64 convertion test now that Abdou kindly regenerated the
467 corresponding dataset.
469 2004-07-11 Martin Quinson
471 - Get it working with any kind of structure (we can compute the padding
472 bytes remotely for all the architectures I have access to)
473 - Implement the structure parsing macro (still not quite robust/complete)
474 - Improvement to the remote testing toysuite
476 2004-07-10 Martin Quinson
478 - get ride of a bunch of deprecated macros
479 - actually run the test for two-compliment, not only compile it :-/
480 - test whether the structures get packed (and bail out if yes. Damn.
481 Alignment is a serious matter)
482 - test whether the structures get compacted (but respecting the alignment
483 constraints of each types)
484 - test whether the array fields of structures can straddle alignment boundaries
486 - Damnit, double are bigger than float (typo in creation of 'double' datadesc)
487 (took me 2 hours to find that bug, looking at the wrong place)
488 - Add gras_datadesc_declare_{union,struct}_close(). They must be used
489 before sending/receiving and are used to compute the offsets of fields
490 - Given that padding size depend even on compiler options, keep track of
491 alignment and aligned_size only for the current architecture. Not a big
492 deal since we send structure fields one after the other (seems
494 - Add the datastructure used for IEEE paper by the PBIO guys to the test
495 program, let it work on linux/gcc/little32. portability todo.
497 2004-07-08 Martin Quinson
498 - import and improve remote compilation support from FAST
499 - make sure make check works on half a dozen of machines out there
501 2004-07-07 Martin Quinson
502 Let's say it's version 0.0.3 ;)
503 - Implement conversions (yuhu!)
504 - Let it work on solaris (beside conversion, of course)
505 - Stupid me, using rand() to generate the conversion datatests in not wise.
507 2004-07-06 Martin Quinson
508 - Let make dist work, since I'm gonna need it to compile on remote hosts
509 - Let Tests/datadesc_usage write the architecture on which the file was
510 generated as first byte.
511 - Add PowerPC (being also IRIX64), SPARC (also power4) and ALPHA
512 architecture descriptions.
513 - Add datadesc_usage.{i386,ppc,sparc} files being the result of execution
514 on those architectures.
515 - Optimization: send/recv array of scalar in one shoot
517 2004-07-05 Martin Quinson
518 - YEAH! GRAS/SG and GRAS/RL are both able to run the ping example !
520 - Plug a whole bunch of memleaks
521 - each process now have to call gras_{init,exit}. One day, their log
522 settings will be separated
523 - Continue the code factorisation between SG, RL and common in Transport.
525 2004-07-04 Martin Quinson
527 - Redistribution between SG and RL.
528 We wanna have to accept in SG, so move accepted related parts of RL in
529 the common part. (more precisely, the dynar of all known sockets is no
530 more a static in transport.c, but part of the process_data)
532 [gras_stub_generator]
533 - Bug fix: Do call gras_process_init from gras_init (wasnt called in RL).
535 2004-07-03 Martin Quinson
536 - Create a new log channel tbx containing dict, set, log, dynar (to shut
537 them all up in one shot)
539 - Fix the ugly case of reference to dynamic array.
540 - New (semi-public) function gras_datadesc_size to allow the messaging
541 layer to malloc the needed space for the buffer.
543 - gras_socket_close now expect the socket to close (and not its address to
544 put NULL in it after it). This is because the socket passed to handlers
545 is one of their argument (=> not writable).
547 - propagate the interface cleanup from last week in datadesc, ie remove a
548 superfluous level of indirection. User pass adress of variable
549 containing data (both when sending and receiving), and not of a variable
550 being a pointer to the data. Let's say that I like it better ;)
551 The price for that is constructs like "int msg=*(int*)payload" in
552 handlers, but it's a fine price, IMHO.
554 - Let it work in RL (yuhu)
556 2004-06-21 Martin Quinson
558 - porting SG plugin and SG select to new standards (works almost).
559 - plug memleaks and fix bugs around.
562 - cleanup the prototype of data recv and force users to specify when they
563 want to handle references to objects. Test case working even for cycles.
564 - plug memleaks. Valgrind is perfectly ok with this.
566 2004-06-12 Martin Quinson
568 - cleanup the separation between plugin and main code in plugin creation
570 2004-06-11 Martin Quinson
572 - Reput hook for raw sockets, needed for BW experiments
573 - kill a few lines of dead code
574 [Data description] Interface cleanup
575 - gras_datadesc_by_name returns the searched type or NULL.
576 That way, no variable is needed to use a type desc once, which makes
578 - gras_datadesc_declare_[struct|union]_append_name is removed. The last
579 two parameters were strings (field name, type name), leading to
581 [Dicos] Interface cleanup
582 - gras_dico_retrieve -> gras_dico_get ; gras_dico_insert -> gras_dico_set
583 This is consistant with the dynar API.
585 2004-04-21 Martin Quinson
587 - Porting to new standards.
590 There is no bag anymore, no need to take extra provision to mask the
591 pointers behind "ID".
592 Better splitup of functions between files create/exchange/convert.
593 This is still a bit artificial since convert and receive are so
594 interleaved, but anyway.
596 - add a queued message list to procdata (the ones not matching criteria
598 - factorize some more code between SG and RL wrt procdata
600 - use gras_exit in example to track memleaks
601 - get rid of gs_example now that GS is properly integrated into gras
602 - update run_test to integrate the lastest tests (datadesc)
604 - rename WARNINGn macros to WARNn since it prooved error-prone
606 2004-04-19 Martin Quinson
608 - register init/exit functions within gras module mechanism
609 - send/receive function.
610 Convertion is not implemented, but short-cutted if not needed.
611 struct/array elements are sent one by one (instead of block-wise), but
612 nobody really cares (yet). Get a prototype before optimizing.
613 - tests (using a file socket) for DD send/receive on:
614 - base types: int, float
615 - array: fixed size, string (ie ref to dynamic string)
616 - structure: homogeneous, heterogeneous
617 - chained list, graph with cycle
618 Believe it or not, valgrind is not too unhappy with the results. The
619 cycle happily segfaults, but the others are ok. And I'm sick of pointers
623 - Bugfix when using a filename explicitely (instead of '-')
625 2004-04-09 Martin Quinson
627 - factorize more code between RL and SG in socket creation
628 - Complete the implementation and tests of:
630 o file (only in RL, and mainly for debugging)
632 I lost 3 days to design a portable address resolver, and then decided
633 that the prototype mainly have to run on my box.
634 Addressing portability too early may be like optimizing too early :-/
636 - use gras_init in the Tests instead of the crappy parse_log_opt
637 (the latter function is removed)
638 [Conditional execution]
639 - New functions: gras_if_RL/gras_if_SG (basic support for this)
640 [Code reorganisation]
641 - Get rid of libgrasutils.a since it makes more trouble than it solves.
642 Build examples against the RL library, since there is no way to disable
643 its creation for now.