Post-release cleanups
Release v3.35
tesh: display the run command ASAP, not after its timeout
Restaure INCLUDE(CheckCCompilerFlag) in cmake It was actually used, my bad.
Remove some restrictions on enable_model-checking now that it's only for the stateless MC
Remove the stateful model-checking from the archive. It's not working anymore It's a real pity to remove all this great code, as it was doing very audacious things, but I need to do so to move forward. This code is very fragile and badly tested, making it very difficult to build upon. The next time this is implemented (if any), I guess that a compiler-approach (e.g. a clang plugin providing a full introspection over the compiled program) would be more robust than a fully runtime approach tricking with the debug information to get the info. That's too difficult and fragile. Again, I'm deeply sorry I had to remove this code, as it requires me to admit that I failed to maintain this great code over the years. But facts are stubborn.
Add a cmake option to disable McMini tests, and rename some existing ones Old configurations should be updated seamlessly.
The Dart module is getting removed from cmake. Don't use it
Allow to disable EIGEN3 even if it's found on disk
add a make target named 'examples' that is easier to remember than 'tests' vs 'test'
Start a new development cycle
Release 3.34
A few spelling mistakes and many replacements: [Ss]imgrid -> SimGrid. [ci-skip]
Add a target to recompile all ns3 tests (and only them)
Remove a file generated by cmake
Cosmetics in cmake - use ON/OFF instead of 1/0 - display all MC related options together - mark some crufy variables as advanced to hide them
More logs around the configuration of the MC variants
Only compile stateless MC when libevent is found
Rename SIMGRID_HAVE_MC into SIMGRID_HAVE_STATEFUL_MC (so that MC can be optional again)
Compile the safe part of MC in default mode too Safety properties (with no checkpointing) are not requiring anything weird nowadays. Liveness properties, communication determinism and state-equality reductions are not considered safe as they still need some memory introspection black magic.