Samuel Lepetit [Fri, 18 May 2012 12:03:57 +0000 (14:03 +0200)]
Fix the handling of errors in Task.send
Samuel Lepetit [Fri, 18 May 2012 12:01:22 +0000 (14:01 +0200)]
Changed the way current_context is handled in smx_context_java
Samuel Lepetit [Wed, 16 May 2012 14:59:56 +0000 (16:59 +0200)]
Removed "ApplicationHandler" and "japplication_handler" and use MSG_launch_application instead.
Also, removed the mandatory "Host, String, String[], double, double" constructor and replaced it with the old "Host, String, String[]" mandatory constructor
for the classes (updated the examples if you need an example), since we can't know from MSG the start_time and the kill_time if we launch it from
deployment file.
Samuel Lepetit [Wed, 16 May 2012 11:56:08 +0000 (13:56 +0200)]
Add an example to test starting_time and kill_time
Samuel Lepetit [Wed, 16 May 2012 11:47:13 +0000 (13:47 +0200)]
Change the way the process killing is handled, using the flag from the base context and not a
specific flag on smx_context_java.
It lets us now use a "process_kill" naturally :).
Samuel Lepetit [Wed, 16 May 2012 09:47:19 +0000 (11:47 +0200)]
Add a straightforward implementation of start_time and kill_time. This change the constructor you have to implement (sorry). Fix the examples for that purpose
Samuel Lepetit [Tue, 15 May 2012 14:47:48 +0000 (16:47 +0200)]
Change "ProcessKilledException" to "ProcessKilledError" -> make it not catchable by catch
(Exception e) statements (because someone's gonna do it).
Samuel Lepetit [Mon, 14 May 2012 15:19:37 +0000 (17:19 +0200)]
Fix platform file for mutual exclusion example. Fix the associated test
Samuel Lepetit [Mon, 14 May 2012 15:16:21 +0000 (17:16 +0200)]
Use jxbt_get_jfield instead of jxbt_get_sfield (faster)
Samuel Lepetit [Mon, 14 May 2012 15:12:25 +0000 (17:12 +0200)]
Removed jprocess_start (useless now), cached ids in jmsg_process
Samuel Lepetit [Mon, 14 May 2012 14:37:07 +0000 (16:37 +0200)]
Change the way the Process are killed, since we need the java stack to be empty (IE, the "run"
call must have returned).
Updated the examples tesh to match it.
Updated the changelog
Samuel Lepetit [Mon, 14 May 2012 11:43:50 +0000 (13:43 +0200)]
Change the way the threads are stopped, not killing the JVM at the end of the simulation.
Samuel Lepetit [Mon, 14 May 2012 08:07:36 +0000 (10:07 +0200)]
Change the include file for os_thread
Samuel Lepetit [Fri, 11 May 2012 15:15:57 +0000 (17:15 +0200)]
Change the way the threads are launched: they are now launched as
xbt_os_thread and then attached to the JVM, meaning that the C has
control over them.
This means that now, "Process" does not herit anymore from Thread, and
that you now need to call explicitely the method "start" if you bypass
the deployment file.
Updated the examples.
Samuel Lepetit [Fri, 11 May 2012 13:44:59 +0000 (15:44 +0200)]
Forgot to update migrate signature in Process
Samuel Lepetit [Thu, 10 May 2012 13:06:32 +0000 (15:06 +0200)]
Add setlocale in jmsg.c, to have traces not depending on the locale of the user
Samuel Lepetit [Thu, 10 May 2012 12:39:20 +0000 (14:39 +0200)]
Missing import in examples
Samuel Lepetit [Thu, 10 May 2012 10:00:29 +0000 (12:00 +0200)]
Change the way the objects are created in ApplicationHandler: they are
now created the same way than when you create the object manually,
using the (Host, String, String[]) constructor. It means that now your
classes need to have a (Host, String, String[]) constructor calling the
same constructor in Process.
Update the examples to reflect the change.
Update the changelog.
Samuel Lepetit [Wed, 9 May 2012 16:10:16 +0000 (18:10 +0200)]
Add sleep in Process, with a Thread.sleep like use (in milliseconds/nanoseconds).
Samuel Lepetit [Wed, 9 May 2012 15:38:23 +0000 (17:38 +0200)]
Add setName in Task
Samuel Lepetit [Wed, 9 May 2012 15:11:17 +0000 (17:11 +0200)]
Add setComputeDuration in Task
Samuel Lepetit [Wed, 9 May 2012 15:11:03 +0000 (17:11 +0200)]
Remove useless line in examples/async/Master.java
Samuel Lepetit [Wed, 9 May 2012 14:14:44 +0000 (16:14 +0200)]
Cache the task name in Task
Samuel Lepetit [Wed, 9 May 2012 13:58:40 +0000 (15:58 +0200)]
Removed the HostFailureException from waitFor, since it's never thrown
Samuel Lepetit [Wed, 9 May 2012 13:50:29 +0000 (15:50 +0200)]
Add cache on jmsg_syncro.c/Mutex.java
Samuel Lepetit [Wed, 9 May 2012 13:39:47 +0000 (15:39 +0200)]
Add a cache on Host.name. Add a cache on some jfieldID/jmethodID on Host/Process
Samuel Lepetit [Wed, 9 May 2012 11:10:32 +0000 (13:10 +0200)]
Deleted simulatedSleep, use waitFor instead. Moved everything from MsgNative to their respective classes. Some optimization work on Process.
Samuel Lepetit [Wed, 9 May 2012 08:24:48 +0000 (10:24 +0200)]
Add checking on getSender/getSource to prevent a SEGFAULT
Samuel Lepetit [Sat, 5 May 2012 11:00:00 +0000 (13:00 +0200)]
Check if the fields were found in Task.nativeInit and Comm.nativeInit
Samuel Lepetit [Fri, 4 May 2012 17:36:10 +0000 (19:36 +0200)]
Move functions from MsgNative to Host, and from jmsg to jmsg_host
Samuel Lepetit [Fri, 4 May 2012 16:58:53 +0000 (18:58 +0200)]
Move all the remaining stuff for Task from MsgNative to Task
Samuel Lepetit [Fri, 4 May 2012 15:07:20 +0000 (17:07 +0200)]
Move some functions from MsgNative to Task again,to depopulate MsgNative/jmsg.c
Samuel Lepetit [Fri, 4 May 2012 14:24:59 +0000 (16:24 +0200)]
Move some functions from MsgNative to Task, to make one less call
Samuel Lepetit [Fri, 4 May 2012 13:12:25 +0000 (15:12 +0200)]
Merge branch 'master' of scm.gforge.inria.fr:/gitroot/simgrid/simgrid-java
Samuel Lepetit [Fri, 4 May 2012 13:12:23 +0000 (15:12 +0200)]
Add cache for jfieldID/jmethodID on jtask/jcomm
navarro [Fri, 4 May 2012 12:22:50 +0000 (14:22 +0200)]
Used of tesh for test.
Samuel Lepetit [Fri, 4 May 2012 11:57:55 +0000 (13:57 +0200)]
Add waitCompletion to Comm as a binding of MSG_comm_wait
Samuel Lepetit [Fri, 4 May 2012 11:02:11 +0000 (13:02 +0200)]
Add isend in Task
Samuel Lepetit [Fri, 4 May 2012 08:21:23 +0000 (10:21 +0200)]
Bugfix in async example. maybe our code should be more defensive for that kind of case, since it can crash simgrid without the user understanding why.
Samuel Lepetit [Thu, 3 May 2012 17:23:33 +0000 (19:23 +0200)]
Add asynchronous API (except wait) for communications
Samuel Lepetit [Thu, 3 May 2012 13:00:06 +0000 (15:00 +0200)]
Properly finalize the simgrid mutexes
Martin Quinson [Wed, 2 May 2012 20:19:33 +0000 (22:19 +0200)]
Quick (and probably dirty) exportation of SimGrid simulated mutexes to Java
I have unfortunately to confess that I didn't test it at all. It
should work, but no promise here. Please people, break it and report
this.
Martin Quinson [Wed, 2 May 2012 20:19:13 +0000 (22:19 +0200)]
cosmetics
Martin Quinson [Wed, 2 May 2012 19:24:33 +0000 (21:24 +0200)]
Do not embeed our version of semaphores
Java 1.5 can be considered as sufficiently prevalent for us to not
dupplicate its features. Closes #14217.
Martin Quinson [Mon, 30 Apr 2012 23:42:26 +0000 (01:42 +0200)]
cosmetics
Martin Quinson [Mon, 30 Apr 2012 23:41:16 +0000 (01:41 +0200)]
use -1 as default kill time to make sure that our process don't get killed on startup
Martin Quinson [Mon, 30 Apr 2012 22:06:22 +0000 (00:06 +0200)]
work around issues by passing the hostname instead of the host to processCreate()
Martin Quinson [Mon, 30 Apr 2012 22:05:39 +0000 (00:05 +0200)]
adapt to the lastest (and probably short-lived) API change in MSG
Martin Quinson [Mon, 30 Apr 2012 22:03:00 +0000 (00:03 +0200)]
revive this file that is very handful to debug the bindings
Martin Quinson [Fri, 27 Apr 2012 21:02:30 +0000 (23:02 +0200)]
adapt to lastest cleanups in the core library
Martin Quinson [Fri, 27 Apr 2012 10:01:24 +0000 (12:01 +0200)]
remove the now useless Process.ifInterruptedStop()
Martin Quinson [Fri, 27 Apr 2012 09:57:36 +0000 (11:57 +0200)]
implement the ability to kill processes without relying on ifInterruptedStop
Samuel Lepetit [Fri, 27 Apr 2012 09:08:36 +0000 (11:08 +0200)]
Merge branch 'master' of git://scm.gforge.inria.fr/simgrid/simgrid-java
Martin Quinson [Thu, 26 Apr 2012 16:22:12 +0000 (18:22 +0200)]
remove a method that is deprecated since a very long time
Martin Quinson [Thu, 26 Apr 2012 16:20:48 +0000 (18:20 +0200)]
Make Process.kill(process) an instance method, not a static one
It feels more OOP this way, AMHA.
Samuel Lepetit [Thu, 26 Apr 2012 14:58:15 +0000 (16:58 +0200)]
cache for pid/ppid/host in Process.java
Martin Quinson [Wed, 25 Apr 2012 23:46:20 +0000 (01:46 +0200)]
Merge branch 'master' of scm.gforge.inria.fr:/gitroot/simgrid/simgrid-java
Martin Quinson [Wed, 25 Apr 2012 22:41:43 +0000 (00:41 +0200)]
check error conditions of task receiption [Samuel Lepetit]
navarro [Wed, 25 Apr 2012 11:15:23 +0000 (13:15 +0200)]
Use of SIMGRID_LIB_PATH and not SIMGRID_LIB
navarro [Wed, 25 Apr 2012 11:09:58 +0000 (13:09 +0200)]
Remove unused tesh files.
navarro [Wed, 25 Apr 2012 09:18:07 +0000 (11:18 +0200)]
Fix compilation problem with library path.
Cosmetics
Martin Quinson [Tue, 24 Apr 2012 08:22:02 +0000 (10:22 +0200)]
add a release date to the previous stable, and document what I know of what were done since then
Martin Quinson [Tue, 24 Apr 2012 08:17:11 +0000 (10:17 +0200)]
Add dsend and simulatedSleep to the binding, and add an example
Martin Quinson [Tue, 24 Apr 2012 07:05:21 +0000 (09:05 +0200)]
Stop using SIMGRID_ROOT at compile time, use SIMGRID_LIB/INCLUDE instead
At configuration time, the SIMGRID_ROOT variable is used as a hint in
the FindSimgrid chunk, when looking for our headers and libraries. If
they cannot be located anyway, the user is given the opportunity to
manually give the mandated pathes in ccmake for example.
Then, at compilation time (ie, once the configuration is done), we
should not use SIMGRID_ROOT anymore, but only SIMGRID_LIB and
SIMGRID_INCLUDE because otherwise, we ignore specific settings that
the user may have provided in ccmake.
Martin Quinson [Mon, 23 Apr 2012 15:32:30 +0000 (17:32 +0200)]
add a pimple to our code to silence some useless warning from eclipse -- sorry
Martin Quinson [Mon, 23 Apr 2012 15:30:53 +0000 (17:30 +0200)]
add some parenthesis around a truth value, as suggested by eclipse
Stephane Genaud [Tue, 17 Apr 2012 15:24:16 +0000 (17:24 +0200)]
reflect real location for simgrid.jar
navarro [Fri, 13 Apr 2012 09:28:47 +0000 (11:28 +0200)]
Fix some link problème.
navarro [Fri, 13 Apr 2012 09:04:16 +0000 (11:04 +0200)]
Print simgrid_lib (dbug)
Arnaud Giersch [Mon, 26 Mar 2012 16:10:42 +0000 (18:10 +0200)]
Build argv properly.
navarro [Mon, 5 Mar 2012 10:07:14 +0000 (11:07 +0100)]
Replace MSG_get_host_number with MSG_hosts_as_dynar.
Laurent Bobelin [Thu, 16 Feb 2012 15:09:42 +0000 (16:09 +0100)]
Process kills now seems to work. May need some additional checking ...
unknown [Thu, 16 Feb 2012 11:39:29 +0000 (12:39 +0100)]
Cosmetics
navarro [Thu, 16 Feb 2012 10:29:38 +0000 (11:29 +0100)]
Load SG_java but rename libSG_java for SG_java on windows.
navarro [Thu, 16 Feb 2012 10:21:42 +0000 (11:21 +0100)]
Protect windows specific needs.
unknown [Thu, 16 Feb 2012 10:09:34 +0000 (11:09 +0100)]
Load libSG_java instead of SG_java
unknown [Thu, 16 Feb 2012 10:08:54 +0000 (11:08 +0100)]
Remove deprecated function getErrCode.
unknown [Thu, 16 Feb 2012 10:06:00 +0000 (11:06 +0100)]
Use and create dll for simgrid java.
unknown [Fri, 13 Jan 2012 15:16:21 +0000 (16:16 +0100)]
Don't try to use git with windows and cmake
navarro [Mon, 13 Feb 2012 12:08:02 +0000 (13:08 +0100)]
Update commands to install simgrid-java
navarro [Mon, 30 Jan 2012 09:56:28 +0000 (10:56 +0100)]
Make a pdf for simgrid-java
Martin Quinson [Tue, 31 Jan 2012 09:28:43 +0000 (10:28 +0100)]
msg.h lives in a subdir, actually
Martin Quinson [Tue, 31 Jan 2012 09:08:04 +0000 (10:08 +0100)]
give more hints to that stupid cmake
Martin Quinson [Fri, 27 Jan 2012 21:02:08 +0000 (22:02 +0100)]
doc cleanups
Martin Quinson [Fri, 27 Jan 2012 20:58:27 +0000 (21:58 +0100)]
also fix the way we look for the simgrid library
Martin Quinson [Fri, 27 Jan 2012 20:53:17 +0000 (21:53 +0100)]
properly look for the SimGrid header files; stop relying on the SIMGRID_ROOT
This environment variable is used as an hint at configure time, the
first time you run ccmake. But the value provided to ccmake is only
what's needed to get the thing running.
Martin Quinson [Fri, 27 Jan 2012 20:43:42 +0000 (21:43 +0100)]
I changed that path recently
Martin Quinson [Fri, 27 Jan 2012 20:43:25 +0000 (21:43 +0100)]
look for msg.h instead of gras.h (cosmetics only: GRAS is soon to be deprecated)
navarro [Thu, 19 Jan 2012 10:32:48 +0000 (11:32 +0100)]
Write doc for how to use examples
Christopho [Fri, 6 Jan 2012 16:49:37 +0000 (17:49 +0100)]
Clean smx_ctx_java_stop, but it does not work yet
Christopho [Fri, 6 Jan 2012 16:48:36 +0000 (17:48 +0100)]
Don't destroy the Java thread object yet, we still have to call it
Navarrop [Mon, 19 Dec 2011 14:03:46 +0000 (15:03 +0100)]
With DMSG_USE_DEPRECATED we can use function MSG_get_errno()
Navarrop [Mon, 19 Dec 2011 14:03:12 +0000 (15:03 +0100)]
Update way to have processes to run from Simgrid 3.7
Navarrop [Mon, 19 Dec 2011 14:02:41 +0000 (15:02 +0100)]
Rename MSG_TASK_CANCELLED by MSG_TASK_CANCELED
Laurent Bobelin [Wed, 16 Nov 2011 09:47:49 +0000 (10:47 +0100)]
Corrected typos as I had 5 minutes to waste before a meeting at 11.
Laurent Bobelin [Mon, 14 Nov 2011 10:32:10 +0000 (11:32 +0100)]
Corrected some typos, clarified the index.
Navarrop [Mon, 14 Nov 2011 10:16:49 +0000 (11:16 +0100)]
Add button for website.
Navarrop [Thu, 10 Nov 2011 16:46:14 +0000 (17:46 +0100)]
Add a remove section from simgrid documentation.
Navarrop [Thu, 10 Nov 2011 16:01:36 +0000 (17:01 +0100)]
Add a sync target.
Rewrite *.html with postprocesser.
Laurent Bobelin [Thu, 10 Nov 2011 14:18:55 +0000 (15:18 +0100)]
Added an index page for javadoc