Gmsh reader update
The proposed merge request includes the following updates:
-
A simplification of the XML interface file for the GmshReader plugin. The XML file has also become more flexible in the sense that the path to the msh files can now contain an arbitrary number of occurrences of the time step and part id in any order, whereas the previous format implicitly expected only one occurrence of each in a predefined order. Instead of providing the specifier character %d (or %06d for the old padded msh format), the user now uses keywords in the path pointing to the msh files which are [step], [partID] and [zeroPadPartID].
-
The minimum version of Gmsh is now forced to be 4.1.0.
-
A major API to Gmsh has changed in Gmsh 4.1.0 so that another variant of the msh 2 format named SeparateViews which is better suited for HPC can be handled by the plugin.
-
Finally, the documentation in the README file has been updated accordingly.
Note that I marked this merge request as WIP because the regression tests currently in place for GmshReader still rely on the old format of the XML interface mshi file and on the old name convention for the msh files. In order to pass the regression tests with the proposed updates, two changes must still be applied but I can't apparently make this correction easily since it involves sha keys. Concretely:
-
Rename all existing ".msh_00000#" files in the regression tests by "_#.msh", where # is an integer ranging from 1 to 4 (the mesh is partitioned in 4 parts). For instance, naca_4part.msh_000001 should become naca_4part_1.msh, domainVelocity_t15.msh_000001 should become domainVelocity_t15_1.msh and domainPressure_t15.msh_000001 should become domainPressure_t15_1.msh
-
Replace the XML interface mshi files with the attached ones.
Please let me know if I can provide any additional information to complete this merge request.
Thank you for your help!
Michel