Logo AND Algorithmique Numérique Distribuée

Public GIT Repository
simgrid.git
8 years agoBe verbose about errors encountered
Martin Quinson [Thu, 24 Sep 2015 23:10:56 +0000 (01:10 +0200)]
Be verbose about errors encountered

8 years agotry to please clang. No idea of why it fails on mac but works on linux...
Martin Quinson [Thu, 24 Sep 2015 20:48:10 +0000 (22:48 +0200)]
try to please clang. No idea of why it fails on mac but works on linux...

8 years agodon't fail miserably when raising an exception after simulation shutdown
Martin Quinson [Thu, 24 Sep 2015 20:23:26 +0000 (22:23 +0200)]
don't fail miserably when raising an exception after simulation shutdown

Fix #26

8 years agocosmetics: Use a C++ syntax that I know :-3
Martin Quinson [Thu, 24 Sep 2015 20:22:55 +0000 (22:22 +0200)]
cosmetics: Use a C++ syntax that I know :-3

8 years ago[tesh] kill unused variables
Martin Quinson [Tue, 22 Sep 2015 19:30:53 +0000 (21:30 +0200)]
[tesh] kill unused variables

8 years ago[tesh] cleanup: don't start useless process instead of killing it right away
Martin Quinson [Tue, 22 Sep 2015 19:23:00 +0000 (21:23 +0200)]
[tesh] cleanup: don't start useless process instead of killing it right away

8 years ago[tesh] speak of SIGKILL in a portable manner (hopefully)
Martin Quinson [Tue, 22 Sep 2015 13:49:58 +0000 (15:49 +0200)]
[tesh] speak of SIGKILL in a portable manner (hopefully)

8 years ago[tesh] exec the killer process after fork to make it work on windows
Martin Quinson [Tue, 22 Sep 2015 13:18:15 +0000 (15:18 +0200)]
[tesh] exec the killer process after fork to make it work on windows

fork() is only emulated on windows, using a thread of the current
process. So if we don't exec away the killer process, the tesh process
will not end before the timeout, leading to loooooong testing phases.

This script is still in a rather sorry state, more cleanups would be
welcome.

8 years ago[tesh] try to port to windows
Martin Quinson [Tue, 22 Sep 2015 12:36:56 +0000 (14:36 +0200)]
[tesh] try to port to windows

8 years ago[appveyor] disable MSVC for now. mingw-w64 is almost there
Martin Quinson [Tue, 22 Sep 2015 11:40:21 +0000 (13:40 +0200)]
[appveyor] disable MSVC for now. mingw-w64 is almost there

8 years agoMac OSX announces itself as 'darwin' in Perl
Martin Quinson [Tue, 22 Sep 2015 11:10:33 +0000 (13:10 +0200)]
Mac OSX announces itself as 'darwin' in Perl

8 years ago[tesh] try to load a handy perl module for the windows port
Martin Quinson [Tue, 22 Sep 2015 09:52:12 +0000 (11:52 +0200)]
[tesh] try to load a handy perl module for the windows port

This is just to check if it's installed on appveyor.

8 years ago[tesh] kill processes in a portable way
Martin Quinson [Tue, 22 Sep 2015 06:53:12 +0000 (08:53 +0200)]
[tesh] kill processes in a portable way

8 years ago[appveyor] don't install before running the tests
Martin Quinson [Tue, 22 Sep 2015 05:26:51 +0000 (07:26 +0200)]
[appveyor] don't install before running the tests

8 years agoavoid a name clash on windows
Martin Quinson [Tue, 22 Sep 2015 00:54:04 +0000 (02:54 +0200)]
avoid a name clash on windows

8 years ago[cmake] don't use path globbing: it breaks on windows
Martin Quinson [Tue, 22 Sep 2015 00:33:56 +0000 (02:33 +0200)]
[cmake] don't use path globbing: it breaks on windows

8 years agoFurther simplify the full.jar generation (the hard portability way)
Martin Quinson [Mon, 21 Sep 2015 23:54:22 +0000 (01:54 +0200)]
Further simplify the full.jar generation (the hard portability way)

Don't run the jar at build time to ensure that the java searches the
libs where the cmake puts them. It was nice but it's hard to get that
script running on windows.

Instead, manually determine in both systems (java and cmake) where the
libs should be. Of course both definitions should match.

 - Allows to remove the java_bundle.sh pimple (that breaks on windows)
 - Users compiling simgrid will use the installed libs instead
 - None of the solutions will work for users using a faulty jar
 - Our testing infrastructure will catch the glitches:
   Each system displays the libs location in the build logs.

8 years agoGreatly simplify the full jar generation
Martin Quinson [Mon, 21 Sep 2015 23:19:12 +0000 (01:19 +0200)]
Greatly simplify the full jar generation

8 years agothat condition will not happen very often, for sure :)
Martin Quinson [Mon, 21 Sep 2015 22:26:59 +0000 (00:26 +0200)]
that condition will not happen very often, for sure :)

8 years agoDon't trust the shebang to work [on windows]
Martin Quinson [Mon, 21 Sep 2015 22:08:47 +0000 (00:08 +0200)]
Don't trust the shebang to work [on windows]

8 years ago[appveyor] MSVC seems to freeze in 64 bits
Martin Quinson [Mon, 21 Sep 2015 22:07:43 +0000 (00:07 +0200)]
[appveyor] MSVC seems to freeze in 64 bits

8 years agocosmetics: mv tools/cmake/Scripts tools/cmake/scripts
Martin Quinson [Mon, 21 Sep 2015 21:57:06 +0000 (23:57 +0200)]
cosmetics: mv tools/cmake/Scripts tools/cmake/scripts

8 years agosilence some more MSVC warnings
Martin Quinson [Mon, 21 Sep 2015 21:53:10 +0000 (23:53 +0200)]
silence some more MSVC warnings

8 years agodon't mess with the name of the dll
Martin Quinson [Mon, 21 Sep 2015 15:23:11 +0000 (17:23 +0200)]
don't mess with the name of the dll

8 years ago[Doc] Fixed links to tutorial section.
Christian Heinrich [Mon, 21 Sep 2015 13:20:43 +0000 (15:20 +0200)]
[Doc] Fixed links to tutorial section.

8 years ago[Doc] Fixed spelling of smpi/simulate_computation
Christian Heinrich [Fri, 18 Sep 2015 14:17:44 +0000 (16:17 +0200)]
[Doc] Fixed spelling of smpi/simulate_computation

8 years ago[SMPI] Grammar...
Christian Heinrich [Tue, 15 Sep 2015 19:43:03 +0000 (21:43 +0200)]
[SMPI] Grammar...

8 years ago[mc] Add missing #include
Gabriel Corona [Mon, 21 Sep 2015 10:14:43 +0000 (12:14 +0200)]
[mc] Add missing #include

8 years ago[appveyor] switch to x64. Who still uses 32bits?
Martin Quinson [Mon, 21 Sep 2015 08:08:20 +0000 (10:08 +0200)]
[appveyor] switch to x64. Who still uses 32bits?

8 years agoone more strange casting to fit into LLP64 data model
Martin Quinson [Mon, 21 Sep 2015 08:05:16 +0000 (10:05 +0200)]
one more strange casting to fit into LLP64 data model

8 years agoCorrectly cast pointers into uintptr_t to fit into the (stupid) LLP64 data model
Martin Quinson [Mon, 21 Sep 2015 07:21:49 +0000 (09:21 +0200)]
Correctly cast pointers into uintptr_t to fit into the (stupid) LLP64 data model

8 years agoRevert "bump our dependency on cmake and stop trying to survive older versions"
Martin Quinson [Sun, 20 Sep 2015 19:51:23 +0000 (21:51 +0200)]
Revert "bump our dependency on cmake and stop trying to survive older versions"

It breaks on fedora, for some reason.

This reverts commit ae3968bcb65dde4ba2ad2d276a98fe665825eae0.

8 years agodisplay more info about the local data model
Martin Quinson [Sun, 20 Sep 2015 19:34:18 +0000 (21:34 +0200)]
display more info about the local data model

8 years ago[MSVC] try to shut some warnings in the right file
Martin Quinson [Sun, 20 Sep 2015 12:41:52 +0000 (14:41 +0200)]
[MSVC] try to shut some warnings in the right file

8 years ago[travis] get a recent cmake on linux
Martin Quinson [Sun, 20 Sep 2015 09:36:31 +0000 (11:36 +0200)]
[travis] get a recent cmake on linux

8 years ago[java/win] don't mess with the simgrid-java.dll prefix on windows
Martin Quinson [Sun, 20 Sep 2015 08:45:15 +0000 (10:45 +0200)]
[java/win] don't mess with the simgrid-java.dll prefix on windows

8 years agobump our dependency on cmake and stop trying to survive older versions
Martin Quinson [Sun, 20 Sep 2015 08:41:46 +0000 (10:41 +0200)]
bump our dependency on cmake and stop trying to survive older versions

8 years ago[appveyor] disable LTO as symengine's mingw does not provide it
Martin Quinson [Sun, 20 Sep 2015 08:09:54 +0000 (10:09 +0200)]
[appveyor] disable LTO as symengine's mingw does not provide it

8 years agoallow to specify on cmake command line whether we want LTO
Martin Quinson [Sun, 20 Sep 2015 08:08:38 +0000 (10:08 +0200)]
allow to specify on cmake command line whether we want LTO

8 years ago[appveyor] Enable Java. That's what we need on windows after all
Martin Quinson [Sun, 20 Sep 2015 00:57:12 +0000 (02:57 +0200)]
[appveyor] Enable Java. That's what we need on windows after all

8 years ago[windows] fix the detection of 32/64 bits
Martin Quinson [Sun, 20 Sep 2015 00:42:19 +0000 (02:42 +0200)]
[windows] fix the detection of 32/64 bits

Funny fact: _WIN32 is also defined in 64 bits mode

8 years agomore documentation
Martin Quinson [Sun, 20 Sep 2015 00:16:28 +0000 (02:16 +0200)]
more documentation

8 years ago[appveyor] restrict MinGW to Win32 as win_ucontext does not work on 64bits
Martin Quinson [Sun, 20 Sep 2015 00:15:38 +0000 (02:15 +0200)]
[appveyor] restrict MinGW to Win32 as win_ucontext does not work on 64bits

8 years ago[MinGW] bypass a broken detection: we do have asprintf here
Martin Quinson [Sat, 19 Sep 2015 23:54:27 +0000 (01:54 +0200)]
[MinGW] bypass a broken detection: we do have asprintf here

8 years agodisable a MSVC warning
Martin Quinson [Sat, 19 Sep 2015 23:40:34 +0000 (01:40 +0200)]
disable a MSVC warning

8 years agofix some 'signed/unsigned comparison' warnings
Martin Quinson [Sat, 19 Sep 2015 22:51:51 +0000 (00:51 +0200)]
fix some 'signed/unsigned comparison' warnings

8 years agodon't play cleaver games with MSVC. Build simgrid.dll and that's all
Martin Quinson [Sat, 19 Sep 2015 22:42:53 +0000 (00:42 +0200)]
don't play cleaver games with MSVC. Build simgrid.dll and that's all

8 years agodon't compare signed and unsigned
Martin Quinson [Sat, 19 Sep 2015 22:28:11 +0000 (00:28 +0200)]
don't compare signed and unsigned

8 years ago[MSVC] convert unsigned to signed before applying unary minus
Martin Quinson [Sat, 19 Sep 2015 22:25:00 +0000 (00:25 +0200)]
[MSVC] convert unsigned to signed before applying unary minus

This should avoid the following:
  warning C4146: unary minus operator applied to unsigned type,
                 result still unsigned

8 years agodon't pass gcc link options to MSVC
Martin Quinson [Sat, 19 Sep 2015 22:20:26 +0000 (00:20 +0200)]
don't pass gcc link options to MSVC

8 years ago[windows] only force the definition of HAVE_VA_COPY if using MSVC, MinGW defines...
Martin Quinson [Sat, 19 Sep 2015 21:42:09 +0000 (23:42 +0200)]
[windows] only force the definition of HAVE_VA_COPY if using MSVC, MinGW defines it on its own

8 years agostop being foolish and deal with public data the regular way
Martin Quinson [Sat, 19 Sep 2015 21:30:49 +0000 (23:30 +0200)]
stop being foolish and deal with public data the regular way

8 years agoreuse the mingw 'install' from symengine to our benefit
Martin Quinson [Sat, 19 Sep 2015 21:30:06 +0000 (23:30 +0200)]
reuse the mingw 'install' from symengine to our benefit

8 years ago[MSVC] export these variables in a windows way
Martin Quinson [Sat, 19 Sep 2015 06:22:04 +0000 (08:22 +0200)]
[MSVC] export these variables in a windows way

8 years ago[appveyor] give a spin to MinGW, just to see
Martin Quinson [Fri, 18 Sep 2015 18:43:15 +0000 (20:43 +0200)]
[appveyor] give a spin to MinGW, just to see

8 years agofix the detection of MSVC
Martin Quinson [Fri, 18 Sep 2015 13:36:22 +0000 (15:36 +0200)]
fix the detection of MSVC

8 years ago[MSVC] Give up: disable busy waiting on windows as it uses gcc synchronizations
Martin Quinson [Fri, 18 Sep 2015 12:55:36 +0000 (14:55 +0200)]
[MSVC] Give up: disable busy waiting on windows as it uses gcc synchronizations

A proper OOP implementation would be welcome, but not today

8 years agotry to refresh mingw before using it
Martin Quinson [Fri, 18 Sep 2015 12:47:47 +0000 (14:47 +0200)]
try to refresh mingw before using it

8 years ago[mc] Fix missing #include in Windows
Gabriel Corona [Fri, 18 Sep 2015 12:22:37 +0000 (14:22 +0200)]
[mc] Fix missing #include in Windows

We already have a polyfill in xbt/str.h

8 years agostupid bummer
Martin Quinson [Fri, 18 Sep 2015 09:38:11 +0000 (11:38 +0200)]
stupid bummer

8 years agoMerge branch 'master' of github.com:mquinson/simgrid
Martin Quinson [Fri, 18 Sep 2015 07:48:48 +0000 (09:48 +0200)]
Merge branch 'master' of github.com:mquinson/simgrid

8 years agoI think that MSVC wants the attrib before the symbol name
Martin Quinson [Fri, 18 Sep 2015 07:48:33 +0000 (09:48 +0200)]
I think that MSVC wants the attrib before the symbol name

8 years agoport XBT_ATTRIB_NORETURN to MSVC
Martin Quinson [Fri, 18 Sep 2015 07:48:06 +0000 (09:48 +0200)]
port XBT_ATTRIB_NORETURN to MSVC

8 years agos/_XBT_GNUC/XBT_ATTRIB/ as we will port them to MSVC
Martin Quinson [Thu, 17 Sep 2015 15:49:59 +0000 (17:49 +0200)]
s/_XBT_GNUC/XBT_ATTRIB/ as we will port them to MSVC

8 years ago[mc] Fix missing #include
Gabriel Corona [Thu, 17 Sep 2015 14:02:55 +0000 (16:02 +0200)]
[mc] Fix missing #include

8 years ago[mc] Use member hooks for surf_interface
Gabriel Corona [Tue, 15 Sep 2015 09:41:44 +0000 (11:41 +0200)]
[mc] Use member hooks for surf_interface

8 years ago[mc] Use member hooks for VM
Gabriel Corona [Tue, 15 Sep 2015 09:07:14 +0000 (11:07 +0200)]
[mc] Use member hooks for VM

8 years ago[mc] Use member hooks for CpuTi
Gabriel Corona [Tue, 15 Sep 2015 09:01:19 +0000 (11:01 +0200)]
[mc] Use member hooks for CpuTi

8 years ago[mc] Remove remaining bits of libdw outside of mc_libdw.{h,c}pp
Gabriel Corona [Thu, 17 Sep 2015 12:29:52 +0000 (14:29 +0200)]
[mc] Remove remaining bits of libdw outside of mc_libdw.{h,c}pp

8 years ago[mc] Use string::compare instead of strcmp
Gabriel Corona [Thu, 17 Sep 2015 12:26:05 +0000 (14:26 +0200)]
[mc] Use string::compare instead of strcmp

8 years ago[mc] Fix initialisation of dwarf expression
Gabriel Corona [Thu, 17 Sep 2015 12:20:10 +0000 (14:20 +0200)]
[mc] Fix initialisation of dwarf expression

8 years ago[mc] Move libdw binding code in its own file
Gabriel Corona [Thu, 17 Sep 2015 11:39:57 +0000 (13:39 +0200)]
[mc] Move libdw binding code in its own file

8 years agoMerge branch 'master' of github.com:mquinson/simgrid
Martin Quinson [Thu, 17 Sep 2015 12:12:12 +0000 (14:12 +0200)]
Merge branch 'master' of github.com:mquinson/simgrid

8 years ago[appveyor] disable x64 for now, it's borken
Martin Quinson [Thu, 17 Sep 2015 12:10:14 +0000 (14:10 +0200)]
[appveyor] disable x64 for now, it's borken

8 years ago[appveyor] try MinGW-w64 instead of MSys2
Martin Quinson [Tue, 15 Sep 2015 08:21:03 +0000 (10:21 +0200)]
[appveyor] try MinGW-w64 instead of MSys2

8 years agoDon't add git-only files in git-archive
Martin Quinson [Mon, 14 Sep 2015 19:20:08 +0000 (21:20 +0200)]
Don't add git-only files in git-archive

So, don't ignore them during the distcheck

8 years ago[appveyor] now that it does (not) compile with MSVC, try with MSYS2
Martin Quinson [Mon, 14 Sep 2015 14:34:45 +0000 (16:34 +0200)]
[appveyor] now that it does (not) compile with MSVC, try with MSYS2

8 years ago[appveyor] try to use boost 1.59 as the older ones don't recognize MSVC 2015
Martin Quinson [Mon, 14 Sep 2015 13:06:55 +0000 (15:06 +0200)]
[appveyor] try to use boost 1.59 as the older ones don't recognize MSVC 2015

8 years agoMSCV also define _THROW, so pick another name for our internal macro
Martin Quinson [Mon, 14 Sep 2015 12:18:02 +0000 (14:18 +0200)]
MSCV also define _THROW, so pick another name for our internal macro

8 years ago[MSVC] help that stupid parser: we are not closing any comment here...
Martin Quinson [Mon, 14 Sep 2015 10:17:00 +0000 (12:17 +0200)]
[MSVC] help that stupid parser: we are not closing any comment here...

-void NetworkConstantAction::setCategory(const char */*category*/)
+void NetworkConstantAction::setCategory(const char * /*category*/)

When seing char >>> */* blah */ <<<, MSVC spits the following message:
  warning C4138: '*/' found outside of comment

8 years ago[MSVC] strcasecmp has a weird name on non-POSIX systems
Martin Quinson [Mon, 14 Sep 2015 10:11:52 +0000 (12:11 +0200)]
[MSVC] strcasecmp has a weird name on non-POSIX systems

8 years agodelay the inclusion of standard headers so that _CRT_*_NO_WARNINGS gets defined on...
Martin Quinson [Mon, 14 Sep 2015 09:55:07 +0000 (11:55 +0200)]
delay the inclusion of standard headers so that _CRT_*_NO_WARNINGS gets defined on time

8 years ago[MSVC] cleaner solution to the lack of unistd for flex parsers
Martin Quinson [Mon, 14 Sep 2015 09:06:06 +0000 (11:06 +0200)]
[MSVC] cleaner solution to the lack of unistd for flex parsers

8 years agouse new/delete instead of xbt_new0/xbt_free on C++ objects
Martin Quinson [Mon, 14 Sep 2015 08:28:14 +0000 (10:28 +0200)]
use new/delete instead of xbt_new0/xbt_free on C++ objects

8 years ago[appveyor] disable mallocators on windows: they need __sync_lock_test_and_set
Martin Quinson [Mon, 14 Sep 2015 08:07:57 +0000 (10:07 +0200)]
[appveyor] disable mallocators on windows: they need __sync_lock_test_and_set

8 years agodefine isatty() on architectures that lack it
Martin Quinson [Mon, 14 Sep 2015 08:06:22 +0000 (10:06 +0200)]
define isatty() on architectures that lack it

8 years agoload system headers after simgrid ones
Martin Quinson [Mon, 14 Sep 2015 08:01:35 +0000 (10:01 +0200)]
load system headers after simgrid ones

This is an attempt to get xbt/base.h loaded before stdio.h, so that
_CRT_SECURE_NO_WARNINGS gets defined before the declaration of fopen

8 years ago[MSVC] 32bits/64bits is declared that way on this arch
Martin Quinson [Mon, 14 Sep 2015 07:24:14 +0000 (09:24 +0200)]
[MSVC] 32bits/64bits is declared that way on this arch

8 years agoregenerate maintainer-mode files
Martin Quinson [Mon, 14 Sep 2015 05:20:52 +0000 (07:20 +0200)]
regenerate maintainer-mode files

8 years agobummer
Martin Quinson [Mon, 14 Sep 2015 05:17:59 +0000 (07:17 +0200)]
bummer

8 years ago[MSCV] kill unused variable
Martin Quinson [Sun, 13 Sep 2015 22:51:44 +0000 (00:51 +0200)]
[MSCV] kill unused variable

8 years agoUse std::atomic instead of __sync_fetch_and_add for portability
Martin Quinson [Sun, 13 Sep 2015 22:50:06 +0000 (00:50 +0200)]
Use std::atomic instead of __sync_fetch_and_add for portability

8 years agotypo
Martin Quinson [Sun, 13 Sep 2015 22:38:10 +0000 (00:38 +0200)]
typo

8 years ago[MSVC] don't declare the 'xbt' log channel both extern and local in xbt_main.c
Martin Quinson [Sun, 13 Sep 2015 21:20:24 +0000 (23:20 +0200)]
[MSVC] don't declare the 'xbt' log channel both extern and local in xbt_main.c

8 years ago[MSCV] no UNISTD here
Martin Quinson [Sun, 13 Sep 2015 21:16:55 +0000 (23:16 +0200)]
[MSCV] no UNISTD here

8 years agodont load smpi.h when not needed: it breaks on MSVC
Martin Quinson [Sun, 13 Sep 2015 20:32:10 +0000 (22:32 +0200)]
dont load smpi.h when not needed: it breaks on MSVC

8 years ago[MSVC] silence some warnings
Martin Quinson [Sun, 13 Sep 2015 20:26:47 +0000 (22:26 +0200)]
[MSVC] silence some warnings

8 years agofix make distcheck
Martin Quinson [Sun, 13 Sep 2015 20:26:29 +0000 (22:26 +0200)]
fix make distcheck

8 years agobummer
Martin Quinson [Sun, 13 Sep 2015 19:59:38 +0000 (21:59 +0200)]
bummer

8 years agoCut the dirname in error messages so that the tests remain reproducible
Martin Quinson [Sun, 13 Sep 2015 19:52:20 +0000 (21:52 +0200)]
Cut the dirname in error messages so that the tests remain reproducible

This was already the case 2 hours ago, but now, do it in a portable way.