SUTER Frederic [Wed, 2 Feb 2022 18:32:29 +0000 (19:32 +0100)]
move the code of isend and irecv out of the handler and into CommImpl
Arnaud Giersch [Thu, 3 Feb 2022 14:33:29 +0000 (15:33 +0100)]
Update references to the mailing list.
Arnaud Giersch [Thu, 3 Feb 2022 14:32:45 +0000 (15:32 +0100)]
Merge branch 'stable'
Augustin Degomme [Thu, 3 Feb 2022 10:02:55 +0000 (11:02 +0100)]
add a word in the doc about deploying multiple processes per node in a single line of the hostfile
Augustin Degomme [Thu, 3 Feb 2022 10:00:48 +0000 (11:00 +0100)]
update changelog
Martin Quinson [Wed, 2 Feb 2022 22:51:54 +0000 (23:51 +0100)]
Exclude an oversized python notebook file from the distributed archives
SUTER Frederic [Wed, 2 Feb 2022 18:10:47 +0000 (19:10 +0100)]
add missing override keywords
SUTER Frederic [Wed, 2 Feb 2022 17:59:17 +0000 (18:59 +0100)]
more progress in simcalls modernization and MC cleanup
SUTER Frederic [Wed, 2 Feb 2022 13:33:26 +0000 (14:33 +0100)]
mark some old simcalls as deprecated (test, test_any, and wait_any)
SUTER Frederic [Wed, 2 Feb 2022 11:16:04 +0000 (12:16 +0100)]
take two on prepare() for Waitany and Testany observers
SUTER Frederic [Wed, 2 Feb 2022 10:03:13 +0000 (11:03 +0100)]
better handling of ActivityImpl::wait_for with dying actors
SUTER Frederic [Wed, 2 Feb 2022 09:28:14 +0000 (10:28 +0100)]
add prepare() to some observers
Martin Quinson [Wed, 2 Feb 2022 08:33:26 +0000 (09:33 +0100)]
activity-lifecycle: Factorize some code
Martin Quinson [Wed, 2 Feb 2022 07:59:38 +0000 (08:59 +0100)]
Make it clear in output that the direct comms are sendto [no-ci]
Martin Quinson [Tue, 1 Feb 2022 21:07:35 +0000 (22:07 +0100)]
this file is too deep to allow XBT_INFO, so keep with printfs
Martin Quinson [Tue, 1 Feb 2022 20:40:46 +0000 (21:40 +0100)]
More verbose on an error experienced on aarch64-darwin [no-ci]
Martin Quinson [Mon, 31 Jan 2022 22:14:39 +0000 (23:14 +0100)]
doc: explain how to install stuff in a docker that runs as a user
Augustin Degomme [Tue, 1 Feb 2022 15:23:24 +0000 (16:23 +0100)]
There is no timestamp in TI events, so they were never dumped during execution, only at the end.
This should save a lot of memory, at the expanse of more file accesses.
Arnaud Giersch [Tue, 1 Feb 2022 13:55:50 +0000 (14:55 +0100)]
Update ChangeLog.
Arnaud Giersch [Tue, 1 Feb 2022 13:55:31 +0000 (14:55 +0100)]
Reduce depth of nested 'if' statements (sonar).
Arnaud Giersch [Tue, 1 Feb 2022 12:22:05 +0000 (13:22 +0100)]
Avoid const cast (sonar).
There is apparently no need for a fake argv[0].
Arnaud Giersch [Tue, 1 Feb 2022 12:21:39 +0000 (13:21 +0100)]
Const reference for large parameter.
Arnaud Giersch [Tue, 1 Feb 2022 12:19:20 +0000 (13:19 +0100)]
Cosmetics: whitespaces and empty statement.
Arnaud Giersch [Tue, 1 Feb 2022 11:07:52 +0000 (12:07 +0100)]
Use declared namespace alias py = pybind11.
Also, use plain py::object which is a sort of smart pointer.
SUTER Frederic [Tue, 1 Feb 2022 11:14:12 +0000 (12:14 +0100)]
please clang
SUTER Frederic [Tue, 1 Feb 2022 10:32:05 +0000 (11:32 +0100)]
Start to modernize the remaining old simcalls related to comms
+ refactor wait, test, waitany, and testany at the activity level
+ use observers for all the simcalls related to these operations
+ add 2 new examples that allows to test or wait for the completion of
any kind of activities stored in a single vector
+ simplify another example thanks to that
+ implement the to_string and dot_label functions in the observers
instead of in mc::Api
This (big) commit breaks the MC has the dependency check still mixes
old and modern simcalls and is thus borken. This is a work in progress
for the greater good, and towards SG4.
SUTER Frederic [Tue, 1 Feb 2022 08:43:47 +0000 (09:43 +0100)]
fix deprecation of Engine.get_clock()
Augustin Degomme [Tue, 1 Feb 2022 08:41:21 +0000 (09:41 +0100)]
fix waitall handling in SMPI replay.
Pending requests were not stored properly since last code overhaul.
Some were not waited on, leading to potentially massive leaks and performance issue.
Martin Quinson [Mon, 31 Jan 2022 20:43:28 +0000 (21:43 +0100)]
CMake does not want versions < 2.8.12 of itself
Arnaud Giersch [Mon, 31 Jan 2022 16:15:57 +0000 (17:15 +0100)]
Kill now unused xbt_dynar_shrink().
Arnaud Giersch [Mon, 31 Jan 2022 07:47:18 +0000 (08:47 +0100)]
Define macro XBT_ATTRIB_DEPRECATED_v335.
Arnaud Giersch [Mon, 31 Jan 2022 07:47:18 +0000 (08:47 +0100)]
Remove deprecated features for next release (3.31).
Arnaud Giersch [Mon, 31 Jan 2022 07:50:57 +0000 (08:50 +0100)]
Merge branch 'stable'
Martin Quinson [Mon, 31 Jan 2022 00:06:56 +0000 (01:06 +0100)]
docker: push all tags
Martin Quinson [Mon, 31 Jan 2022 00:04:24 +0000 (01:04 +0100)]
Merge branch 'master' of framagit.org:simgrid/simgrid
Martin Quinson [Sun, 30 Jan 2022 23:24:52 +0000 (00:24 +0100)]
Start a new dev cycle for v3.31
Martin Quinson [Sun, 30 Jan 2022 23:51:23 +0000 (00:51 +0100)]
bummer
Martin Quinson [Sun, 30 Jan 2022 23:48:38 +0000 (00:48 +0100)]
Build the S4U tuto on top of the stable image again
Martin Quinson [Sun, 30 Jan 2022 23:27:57 +0000 (00:27 +0100)]
Rework the gh action to build the stable dockers
Augustin Degomme [Sun, 30 Jan 2022 20:55:21 +0000 (21:55 +0100)]
SMPI: enforce MPI message ordering. Fix #100 (hopefully)
This adds an id for each message, representing the amount of messages sent from one process to another with the same tag.
On receiver side, the number of received messages from each source/tag must be kept the same way and is compared to the message id.
If the count is not identical, the message is early and overtook another message, so don't match with it yet.
Hopefully this does not cause too much memory overhead, but it can be quite a lot in some cases.
todo: cleanup the way probes are handled
Augustin Degomme [Tue, 11 Jan 2022 20:04:03 +0000 (21:04 +0100)]
fix test action ti-tracing, it needs to behave like a wait
Martin Quinson [Sun, 30 Jan 2022 20:09:49 +0000 (21:09 +0100)]
Release v3.30 (Sunday Bloody Sunday)
Martin Quinson [Sun, 30 Jan 2022 17:04:35 +0000 (18:04 +0100)]
Make sure that min<max in MC_random
Martin Quinson [Sun, 30 Jan 2022 17:01:39 +0000 (18:01 +0100)]
Document the next release. I think we're good to go
Martin Quinson [Sun, 30 Jan 2022 14:10:50 +0000 (15:10 +0100)]
Reorg the cmake in teshuite/mc to factorize things a bit
Martin Quinson [Fri, 28 Jan 2022 23:42:38 +0000 (00:42 +0100)]
Make the Catch2-SimGrid integration available to more tests
Martin Quinson [Fri, 28 Jan 2022 22:12:48 +0000 (23:12 +0100)]
tiny improvement to the doc
Martin Quinson [Fri, 28 Jan 2022 22:03:45 +0000 (23:03 +0100)]
Add a destructor to the NetworkNS3Model to cleanup the ns-3 side
Even if in my tests, calling Simulator::Destroy as indicated by the
ns-3 doc does not actually fix our leakage.
Arnaud Giersch [Fri, 28 Jan 2022 16:29:18 +0000 (17:29 +0100)]
Spell check.
SUTER Frederic [Fri, 28 Jan 2022 14:25:33 +0000 (15:25 +0100)]
[no-ci] use :py:attr: when it is the case
SUTER Frederic [Fri, 28 Jan 2022 14:07:28 +0000 (15:07 +0100)]
fix borken refs
SUTER Frederic [Fri, 28 Jan 2022 13:44:15 +0000 (14:44 +0100)]
more attributes in python bindings
Martin Quinson [Fri, 28 Jan 2022 13:30:14 +0000 (14:30 +0100)]
doc: mark python attributes as attributes
Martin Quinson [Fri, 28 Jan 2022 13:28:57 +0000 (14:28 +0100)]
Don't segfault when Engine::get_instance() is called before the initialization of the Engine
SUTER Frederic [Fri, 28 Jan 2022 11:49:54 +0000 (12:49 +0100)]
use property_readonly_static and update deprecation info
SUTER Frederic [Fri, 28 Jan 2022 10:58:55 +0000 (11:58 +0100)]
this is now forbidden, remove right away
SUTER Frederic [Fri, 28 Jan 2022 10:58:36 +0000 (11:58 +0100)]
update
SUTER Frederic [Fri, 28 Jan 2022 10:43:36 +0000 (11:43 +0100)]
[python] cleanup Engine getters/setters
SUTER Frederic [Fri, 28 Jan 2022 10:07:42 +0000 (11:07 +0100)]
Users should not set the netzone_root themselves. Make it protected
Martin Quinson [Thu, 27 Jan 2022 09:52:35 +0000 (10:52 +0100)]
fix some doc issues
SUTER Frederic [Thu, 27 Jan 2022 15:17:57 +0000 (16:17 +0100)]
[no-ci] more entries
Martin Quinson [Wed, 26 Jan 2022 15:54:49 +0000 (16:54 +0100)]
py: indicate which properties are R/W and which are R-O
Arnaud Giersch [Wed, 26 Jan 2022 13:07:34 +0000 (14:07 +0100)]
Update "fixed bugs" in ChangeLog.
[ci-skip]
Martin Quinson [Tue, 25 Jan 2022 14:45:01 +0000 (15:45 +0100)]
cmake: Maybe we have a nested if issue?
If not, I'm not sure why NixOS is ignoring my manual value of
SIMGRID_PYTHON_LIBDIR
Martin Quinson [Tue, 25 Jan 2022 09:25:09 +0000 (10:25 +0100)]
jenkins: actually pass SIMGRID_PYTHON_LIBDIR to cmake, bummer
Arnaud Giersch [Tue, 25 Jan 2022 14:33:09 +0000 (15:33 +0100)]
Fix git-version in smpi scripts.
Thanks Tom Cornebize for reporting.
Arnaud Giersch [Tue, 25 Jan 2022 14:27:12 +0000 (15:27 +0100)]
Enable ns3 for jenkins builds with sanitizers or valgrind.
[ci-skip]
Arnaud Giersch [Tue, 25 Jan 2022 09:00:18 +0000 (10:00 +0100)]
Update .mailmap (we want you?!)
[ci-skip]
Martin Quinson [Tue, 25 Jan 2022 08:18:46 +0000 (09:18 +0100)]
jenkins: force SIMGRID_PYTHON_LIBDIR on NixOS
Martin Quinson [Tue, 25 Jan 2022 08:13:54 +0000 (09:13 +0100)]
Allow to override the python install path (and document it)
Martin Quinson [Tue, 25 Jan 2022 00:45:04 +0000 (01:45 +0100)]
try to honor CMAKE_INSTALL_PREFIX when installing the python package
Martin Quinson [Tue, 25 Jan 2022 00:04:07 +0000 (01:04 +0100)]
adjust the dependencies between the docker images [no-ci]
Martin Quinson [Mon, 24 Jan 2022 23:43:04 +0000 (00:43 +0100)]
Install python module to Python3_SITEARCH as it should [no-ci]
Arnaud Giersch [Mon, 24 Jan 2022 16:45:02 +0000 (17:45 +0100)]
Deprecate Activity::{get,set}_user_data().
Use Extendable::{get,set}_data instead.
Arnaud Giersch [Mon, 24 Jan 2022 15:58:05 +0000 (16:58 +0100)]
Typed template for Extendable::get_data.
Arnaud Giersch [Mon, 24 Jan 2022 16:28:18 +0000 (17:28 +0100)]
Pointer-to-const again.
Arnaud Giersch [Mon, 24 Jan 2022 14:32:55 +0000 (15:32 +0100)]
Specify SSL/TLS protocol version.
Arnaud Giersch [Mon, 24 Jan 2022 14:07:53 +0000 (15:07 +0100)]
Catch more specific exceptions.
Arnaud Giersch [Mon, 24 Jan 2022 13:18:27 +0000 (14:18 +0100)]
Useless assignments.
Arnaud Giersch [Mon, 24 Jan 2022 13:14:47 +0000 (14:14 +0100)]
Unused python imports.
Martin Quinson [Mon, 24 Jan 2022 14:54:22 +0000 (15:54 +0100)]
bummer
Martin Quinson [Mon, 24 Jan 2022 14:27:26 +0000 (15:27 +0100)]
Merge the docker-tuto-s4u action into docker, to ease synchronization issues
Introduce 3 jobs, one per docker target
Augustin Degomme [Mon, 24 Jan 2022 14:32:02 +0000 (14:32 +0000)]
avoid building shared lib for this one.
SUTER Frederic [Mon, 24 Jan 2022 14:20:46 +0000 (15:20 +0100)]
please sonar
Martin Quinson [Mon, 24 Jan 2022 14:14:32 +0000 (15:14 +0100)]
Install the python bindings
Martin Quinson [Mon, 24 Jan 2022 13:55:14 +0000 (14:55 +0100)]
docs: add the copybutton extension to sphinx
adegomme [Mon, 24 Jan 2022 13:58:20 +0000 (13:58 +0000)]
new attempt
adegomme [Mon, 24 Jan 2022 10:20:04 +0000 (10:20 +0000)]
bigdft - fix fortran compiler for ntpoly
Martin Quinson [Mon, 24 Jan 2022 10:20:16 +0000 (11:20 +0100)]
Enable python in the stable and unstable docker images
Martin Quinson [Mon, 24 Jan 2022 09:55:24 +0000 (10:55 +0100)]
This docker is now based on simgrid/unstable so it should be rebuilt each week-end
[no-ci]
Your Name [Mon, 24 Jan 2022 09:19:38 +0000 (09:19 +0000)]
add fortran bindings for smpi_file_get/set_size
Your Name [Mon, 24 Jan 2022 09:19:05 +0000 (09:19 +0000)]
update bigdft rcfile for ntpoly
Martin Quinson [Sun, 23 Jan 2022 09:38:18 +0000 (10:38 +0100)]
Python: Add platform-failures example, and extend Py to make it work
Martin Quinson [Sun, 23 Jan 2022 01:19:43 +0000 (02:19 +0100)]
Implement the last (?) Python functions needed for the tutorial
Martin Quinson [Sat, 22 Jan 2022 23:46:39 +0000 (00:46 +0100)]
Extend the python bindings and cosmetics
Martin Quinson [Sat, 22 Jan 2022 23:44:53 +0000 (00:44 +0100)]
Really tiny optimization when the communication is failed already
Martin Quinson [Sat, 22 Jan 2022 22:37:37 +0000 (23:37 +0100)]
Extend a test to use Host::route_to
Martin Quinson [Sat, 22 Jan 2022 15:39:29 +0000 (16:39 +0100)]
Merge branch 'master' of framagit.org:simgrid/simgrid
Martin Quinson [Sat, 22 Jan 2022 15:33:23 +0000 (16:33 +0100)]
Further fixups and cosmetics in the doc
Martin Quinson [Sat, 22 Jan 2022 14:32:31 +0000 (15:32 +0100)]
Make Activity::on_completion take a const&, just like Comm::on_completion
I think it's dangerous to have a static function of the same name and
differing prototype in a subclass, even more if the only difference is
the parameter const-ness.