Martin Quinson [Thu, 8 Jun 2017 05:08:24 +0000 (07:08 +0200)]
eradicate all remaining manual refcounting on ActivityImpl
Martin Quinson [Thu, 8 Jun 2017 04:56:24 +0000 (06:56 +0200)]
refcount should be automatic now, and initialized to 0
Martin Quinson [Thu, 8 Jun 2017 00:22:52 +0000 (02:22 +0200)]
AH! there was a remaining manual refcount stain
Martin Quinson [Thu, 8 Jun 2017 00:06:54 +0000 (02:06 +0200)]
cosmetics (+ sleep at the end of test to avoid problems)
Martin Quinson [Thu, 8 Jun 2017 00:00:38 +0000 (02:00 +0200)]
improve the debug messages of smx_network
Martin Quinson [Wed, 7 Jun 2017 23:59:46 +0000 (01:59 +0200)]
Correctly deal with simcall returning a intrusive_ptr of nullptr
Martin Quinson [Wed, 7 Jun 2017 23:54:44 +0000 (01:54 +0200)]
Make sure that the user don't try a detached receive
Martin Quinson [Wed, 7 Jun 2017 23:54:16 +0000 (01:54 +0200)]
correct refcounting of ActorPtr
Martin Quinson [Wed, 7 Jun 2017 22:29:34 +0000 (00:29 +0200)]
Add an exhaustive test for point to point communications
Martin Quinson [Wed, 7 Jun 2017 09:17:05 +0000 (11:17 +0200)]
further WIP on the ActivityImplPtr feature. Now it compiles (but fails)
ActivityImplPtr are all over the place, and the manual refcounting is
dead. Most tests still fail, for some reason that I don't understand
yet. That's a WIP...
Notable point:
MC wants to retrieve the pointers to the ActivityImpl in a remote
process, even if their type is ActivityImplPtr (that is,
intrusive_ptr<ActivityImpl>).
The good thing is that in this case, the raw pointer is stored in the
simcall parameters, and the *Ptr type is reconstructed by the marshal
function.
So I just have to generate the <simcall>__getraw__<param/res>()
functions that return the raw pointer without building a *Ptr type. It
compiles, but the MC is broken ATM so maybe it's not sufficient.
I'm committing anyway because I want to add simpler tests for the
communications that would make the debug of this feature easier to
achieve.
Martin Quinson [Wed, 7 Jun 2017 09:13:26 +0000 (11:13 +0200)]
Declare and use ActorImplPtr
Martin Quinson [Wed, 7 Jun 2017 07:20:13 +0000 (09:20 +0200)]
Reduce the displayed backtrace
Martin Quinson [Tue, 6 Jun 2017 06:16:34 +0000 (08:16 +0200)]
first attempt (ongoing WIP)
Martin Quinson [Tue, 13 Jun 2017 20:15:05 +0000 (22:15 +0200)]
kill all trailling whitespaces
This commit could have been called "VI VI VI, the editor of the beast"
:)
Martin Quinson [Tue, 13 Jun 2017 20:06:10 +0000 (22:06 +0200)]
please codacy
Martin Quinson [Tue, 13 Jun 2017 19:58:34 +0000 (21:58 +0200)]
please vera++: don't pass vectors by copy
degomme [Tue, 13 Jun 2017 16:24:35 +0000 (18:24 +0200)]
Update changelog for SMPI before release, this time.
Martin Quinson [Tue, 13 Jun 2017 16:03:57 +0000 (18:03 +0200)]
please sonar, damnit
Frederic Suter [Tue, 13 Jun 2017 15:50:54 +0000 (17:50 +0200)]
Merge branch 'master' of git+ssh://scm.gforge.inria.fr//gitroot/simgrid/simgrid
Frederic Suter [Tue, 13 Jun 2017 15:50:44 +0000 (17:50 +0200)]
plug leak
Martin Quinson [Tue, 13 Jun 2017 15:20:49 +0000 (17:20 +0200)]
Energy: New function to update the consumption of all hosts at once
Takishipp [Tue, 13 Jun 2017 15:00:50 +0000 (17:00 +0200)]
update requested changes
Takishipp [Tue, 13 Jun 2017 10:51:00 +0000 (12:51 +0200)]
remove all the events that are called and destroyed one time
Takishipp [Tue, 13 Jun 2017 09:54:30 +0000 (11:54 +0200)]
remove DefineVariableEvent
Takishipp [Tue, 13 Jun 2017 09:50:05 +0000 (11:50 +0200)]
remove DefineContainerEvent
Frederic Suter [Tue, 13 Jun 2017 07:07:40 +0000 (09:07 +0200)]
Merge branch 'master' of git+ssh://scm.gforge.inria.fr//gitroot/simgrid/simgrid
Martin Quinson [Mon, 12 Jun 2017 20:33:05 +0000 (22:33 +0200)]
try to please sonar in a new and innovative way
it's not quite happy about all our callbacks that don't use all their
parameters. Will see if commenting their names better fits its taste.
It may well find this disgusting too. Will see.
Martin Quinson [Mon, 12 Jun 2017 20:24:40 +0000 (22:24 +0200)]
use the current syntax in this doc snippet
Martin Quinson [Mon, 12 Jun 2017 20:23:57 +0000 (22:23 +0200)]
Merge branch 'master' of scm.gforge.inria.fr:/gitroot/simgrid/simgrid
Frederic Suter [Mon, 12 Jun 2017 12:35:36 +0000 (14:35 +0200)]
a few smells less
Takishipp [Mon, 12 Jun 2017 11:57:04 +0000 (13:57 +0200)]
remove some useless structures
Frederic Suter [Mon, 12 Jun 2017 11:27:13 +0000 (13:27 +0200)]
try to reduce the ugliness of file management
still a long way to go
Frederic Suter [Mon, 12 Jun 2017 08:22:00 +0000 (10:22 +0200)]
this is now useless
Martin Quinson [Sun, 11 Jun 2017 22:23:53 +0000 (00:23 +0200)]
I'm at ns3.26 now
Martin Quinson [Sun, 11 Jun 2017 19:50:18 +0000 (21:50 +0200)]
allow to declare an actor with parameter without going for a full class object
Martin Quinson [Sun, 11 Jun 2017 09:45:42 +0000 (11:45 +0200)]
another insider trick
Martin Quinson [Sat, 10 Jun 2017 19:33:58 +0000 (21:33 +0200)]
NS3 probably don't like bandwidths in Bytes, so use bits
Martin Quinson [Fri, 9 Jun 2017 20:54:55 +0000 (22:54 +0200)]
cosmetics
Martin Quinson [Fri, 9 Jun 2017 13:17:08 +0000 (15:17 +0200)]
more doxygen fixes
Martin Quinson [Fri, 9 Jun 2017 13:07:53 +0000 (15:07 +0200)]
fix a kind of parse error from doxygen
Frederic Suter [Fri, 9 Jun 2017 04:31:32 +0000 (06:31 +0200)]
kill dead code and plug leak
Frederic Suter [Thu, 8 Jun 2017 21:55:04 +0000 (23:55 +0200)]
why doesn't distcheck told me that !?!?
Frederic Suter [Thu, 8 Jun 2017 21:42:51 +0000 (23:42 +0200)]
oops, another module was dropped
Frederic Suter [Thu, 8 Jun 2017 21:33:33 +0000 (23:33 +0200)]
remove useless includes
Frederic Suter [Thu, 8 Jun 2017 20:45:07 +0000 (22:45 +0200)]
fix by not using lib at all!
Frederic Suter [Thu, 8 Jun 2017 20:00:13 +0000 (22:00 +0200)]
Merge branch 'master' of git+ssh://scm.gforge.inria.fr//gitroot/simgrid/simgrid
Frederic Suter [Thu, 8 Jun 2017 19:59:52 +0000 (21:59 +0200)]
try to fix build
still need this include
Martin Quinson [Thu, 8 Jun 2017 19:27:18 +0000 (21:27 +0200)]
Merge branch 'master' of github.com:simgrid/simgrid
Frederic Suter [Thu, 8 Jun 2017 19:10:11 +0000 (21:10 +0200)]
distcheck cannot be trusted
Frederic Suter [Thu, 8 Jun 2017 19:00:52 +0000 (21:00 +0200)]
getting closer: no more MSG_STORAGE_LEVEL
next step, stop using storage_lib in surf at all, now that it's a mere
map
Frederic Suter [Thu, 8 Jun 2017 16:49:43 +0000 (18:49 +0200)]
redefine sg_storage_t
Martin Quinson [Thu, 8 Jun 2017 15:29:00 +0000 (17:29 +0200)]
Merge pull request #184 from Takishipp/signals
getting sg_instr_AS_end through a signal structure
Takishipp [Thu, 8 Jun 2017 14:41:29 +0000 (16:41 +0200)]
getting sg_instr_AS_end through a signal structure
Martin Quinson [Thu, 8 Jun 2017 12:58:25 +0000 (14:58 +0200)]
Merge branch 'master' of github.com:simgrid/simgrid
Martin Quinson [Thu, 8 Jun 2017 12:56:05 +0000 (14:56 +0200)]
Merge pull request #183 from Takishipp/leak_memory
Correct the leak of memory that correspond to EndLinEvent
Takishipp [Thu, 8 Jun 2017 12:22:35 +0000 (14:22 +0200)]
correct the leak of memory that correspond to EndLinEvent
Frederic Suter [Thu, 8 Jun 2017 06:45:03 +0000 (08:45 +0200)]
use ordered map when order is needed
Frederic Suter [Thu, 8 Jun 2017 05:34:16 +0000 (07:34 +0200)]
use pimpl/piface combo for s4u storage
not perfect but one more towards killing storage_lib (and xbt_lib)
Frederic Suter [Wed, 7 Jun 2017 23:06:08 +0000 (01:06 +0200)]
reindent
Frederic Suter [Wed, 7 Jun 2017 22:07:25 +0000 (00:07 +0200)]
kill dicts and unused code
Martin Quinson [Wed, 7 Jun 2017 15:52:26 +0000 (17:52 +0200)]
fix a race condition spotted by TSan
I was hopping that it will help the msg-dht-parallel test that is
broken since a few days, but it's not sufficient. Will see what TSan
has to say next night.
Martin Quinson [Wed, 7 Jun 2017 15:52:04 +0000 (17:52 +0200)]
doc: move the option index to the top of page to make it visible
Martin Quinson [Wed, 7 Jun 2017 15:39:30 +0000 (17:39 +0200)]
small doc improvement
Martin Quinson [Wed, 7 Jun 2017 15:30:36 +0000 (17:30 +0200)]
create a NetZone::onCreation signal and use it from instr
Martin Quinson [Wed, 7 Jun 2017 14:51:39 +0000 (16:51 +0200)]
plug a memleak around instr
bcamus [Wed, 7 Jun 2017 13:15:06 +0000 (15:15 +0200)]
fixing codacy warnings on multi-core VM tests + using a dedicated platform file for these tests (instead of cluster.xml)
bcamus [Wed, 7 Jun 2017 09:56:35 +0000 (11:56 +0200)]
add test for multicore vm
Martin Quinson [Tue, 6 Jun 2017 06:34:33 +0000 (08:34 +0200)]
Merge branch 'master' of scm.gforge.inria.fr:/gitroot/simgrid/simgrid
Frederic Suter [Tue, 6 Jun 2017 06:30:31 +0000 (08:30 +0200)]
try to catch some leaks
Frederic Suter [Tue, 6 Jun 2017 01:46:51 +0000 (03:46 +0200)]
fix S4U chord example \o/
Frederic Suter [Tue, 6 Jun 2017 01:02:02 +0000 (03:02 +0200)]
tiny stuff
Martin Quinson [Mon, 5 Jun 2017 20:30:55 +0000 (22:30 +0200)]
reindent
Martin Quinson [Sun, 4 Jun 2017 07:10:57 +0000 (09:10 +0200)]
make it possible to return intrusive_ptr in simcalls
Martin Quinson [Sat, 3 Jun 2017 20:09:37 +0000 (22:09 +0200)]
please codacy a bit
Martin Quinson [Fri, 2 Jun 2017 15:04:57 +0000 (17:04 +0200)]
Merge pull request #179 from Takishipp/signals
call sg_instr_new_host via signal call
Martin Quinson [Fri, 2 Jun 2017 14:51:49 +0000 (16:51 +0200)]
rename a method
Takishipp [Fri, 2 Jun 2017 12:32:04 +0000 (14:32 +0200)]
remove the useless prototype of sg_instr_new_host
Martin Quinson [Fri, 2 Jun 2017 09:32:39 +0000 (11:32 +0200)]
ns3 routes are always symmetrical, don't dupplicate them
Martin Quinson [Fri, 2 Jun 2017 09:00:33 +0000 (11:00 +0200)]
cleanups after PR #177
Martin Quinson [Fri, 2 Jun 2017 08:47:05 +0000 (10:47 +0200)]
cosmetics and reindent
Martin Quinson [Fri, 2 Jun 2017 08:16:56 +0000 (10:16 +0200)]
further cut the sorted diff to not overflow travis
Martin Quinson [Fri, 2 Jun 2017 08:15:25 +0000 (10:15 +0200)]
please sonar a tiny bit
Martin Quinson [Fri, 2 Jun 2017 05:40:39 +0000 (07:40 +0200)]
git-format: don't fail if there is nothing to do
Martin Quinson [Fri, 2 Jun 2017 05:38:49 +0000 (07:38 +0200)]
Don't try to assist the eclipse parser here: it hurts the feeling of clang
Martin Quinson [Fri, 2 Jun 2017 08:56:33 +0000 (10:56 +0200)]
Merge pull request #177 from Takishipp/sd_exit
Use automated call to sd_exit
Takishipp [Fri, 2 Jun 2017 08:51:55 +0000 (10:51 +0200)]
add nochecking for sd initialization
Takishipp [Fri, 2 Jun 2017 08:41:53 +0000 (10:41 +0200)]
add clean-atexit test
Takishipp [Fri, 2 Jun 2017 08:13:46 +0000 (10:13 +0200)]
call sg_instr_new_host via signal call
Takishipp [Fri, 2 Jun 2017 07:54:40 +0000 (09:54 +0200)]
add checking version sd_init
Martin Quinson [Thu, 1 Jun 2017 18:03:19 +0000 (20:03 +0200)]
The previous change actually (fix #178). Thanks @bcamus!!
Martin Quinson [Thu, 1 Jun 2017 15:11:24 +0000 (17:11 +0200)]
Another attempt at getting the simgrid.jar working on Mac OSX
Fix the @rpath of the copy that will land into the jarfile, not the
original one.
(this is about #178)
Martin Quinson [Thu, 1 Jun 2017 14:38:22 +0000 (16:38 +0200)]
when fixing the rpath on Apple, search simgrid-java where it is
Martin Quinson [Thu, 1 Jun 2017 14:04:17 +0000 (16:04 +0200)]
split the java library packing thing, and try to do the right thing for apple
This may help for #178
Martin Quinson [Thu, 1 Jun 2017 04:02:25 +0000 (06:02 +0200)]
please codacy: make a private field useful
Martin Quinson [Thu, 1 Jun 2017 03:27:02 +0000 (05:27 +0200)]
wanna be exhaustive test for the VM sharing
Martin Quinson [Thu, 1 Jun 2017 02:54:07 +0000 (04:54 +0200)]
Allow multicore VMs, and take this into account in the sharing
This does not break any test, because it's not well tested yet. Yup,
the ChangeLog is a bit overselling so far. A new test torturing this
to make sure we get the right result in many sharing situation is
still to be written. examples/msg/cloud-simple will be taken as a
basis for that.
Martin Quinson [Wed, 31 May 2017 15:52:37 +0000 (17:52 +0200)]
destroying a VM that was not shutdown is OK
Even more because sometimes in Java, the VM is already half destroyed
when the garbage collector wants to finish it.
Takshipp [Wed, 31 May 2017 16:09:01 +0000 (18:09 +0200)]
add the missing atexit call
Martin Quinson [Wed, 31 May 2017 15:39:48 +0000 (17:39 +0200)]
make it possible to destroy a VM that is already half-destroyed
Martin Quinson [Wed, 31 May 2017 15:30:57 +0000 (17:30 +0200)]
reindent and cosmetics