GitLab

#177 Handle Valgrind scripts more like the usual build ones (i.e. mkdir and cd done in Yaml rather than in Python directly).

#177 Valgrind shell script replaced by a Python script: the idea is to provide one script per target. A new stage is introduced to do so. The gitlab Yaml file has been commented to speed-up the 'usual' process and see if the valgrind stage works as intended (probably not: not tested locally at all...).

#1490 Replace Python script by shell script to run sonar-scanner. Temporarily deactivate clang static analysis.

#1490 Fix sonarqube script (unfortunately not nicely: tee doesn't seem to work properly, and I therefore use shell to make do). It's an intermediate step: clang analysis should probably be run in previous stage.

#1471 I was wrong assuming I could remove the workaround in the script that filter Doxygen messages (and that is related to Doxygen run under Linux).

#1483 Script for the warning of Doxygen: add one item to ignore (warning about too much nodes) and remove the one we had to introduce for #1471: new version of Doxygen removed the need for that.

#1427 Improve the integration of Libmesh, using for it the IncludeWithoutWarning trick exactly as for others library (previously it was done under the hood elsewhere in the code...). Some types have been modified to comply with new libmesh API (essentially int that have become int64_t).

#1394 Add explicitly in SourceList.cmake files that that are generated files that should not be edited manually.

#1390 Also don't consider warning about markdown file (Doxygen doesn't understand the markdown link which is required in Gitlab for proper navigation).

#1383 Add a script to interpret a compilation log and returns how many warnings were found. The point is to add a phase in gitlab-ci checking the warnings.

#1198 Scripts: the remaining Python scripts have been updated and cleaned-up; some that were not actually used were removed.

#1198 Clean-up of scripts in progress: README has been completed with the ones already processed. Some that were not useful were removed (now we get continuous integration for instance no need to have a script to check there is no case issue: Linux VM will tell it soon enough!).

substitute_env_variable.py was in fact poorly written and has been rewritten more cleanly. count_number_of_lines.py has been extended to provide reports per module; some folder that were unduly taken into account have been removed. cloc has been updated from v1.70 to v1.80.

#1198 Slight reworking of 1D_Ensight_Mesh.py, which has been put in a dedicated directory. A README to describe all scripts has been initiated.