Fixes for gmsh parser for periodic meshes

[vikaskurapati]

Still draft as I am not sure if this fix goes well with the project, and I used AI to generate this code. @davschneller, do these changes give you a hint on why the dummy periodic mesh (uploaded in this comment) was getting stuck with PUMGen?

Contains AI generated code (GPT-Codex-5.3 as part of copilot CLI)
periodic.zip

[davschneller]

Looking at it—it's exactly what the PR fixes: namely the Affine keyword before the affine transform seems to have tripped the parser. I would classify the token differently, however (as string signifies something in ", i.e. quotation marks). More like a generic keyword or the likes.
Consuming one token for unknown is fine IMO.

Though it seems like there is a chance you'll get only 12 entries, not 16? https://gmsh.info/doc/texinfo/gmsh.html#index-Affine-_007b-expression_002dlist-_007d-_007b-transform_002dlist-_007d (maybe I was wrong there)

Also the parser for GMSH in version 4 should be good before this PR already (at least that's what worked for me back then—but great to see that we can now also handle version 2).

[vikaskurapati]

Looking at it—it's exactly what the PR fixes: namely the Affine keyword seems to have tripped the parser. I would classify the token differently, however (as string signifies something in ", i.e. quotation marks). More like a generic keyword or the likes. Consuming one token for unknown is fine IMO.

OK.

Though it seems like there is a chance you'll get only 12 entries, not 16? https://gmsh.info/doc/texinfo/gmsh.html#index-Affine-_007b-expression_002dlist-_007d-_007b-transform_002dlist-_007d (maybe I was wrong there)

I am unsure about it too.

Also the parser for GMSH in version 4 should be good before this PR already (at least that's what worked for me back then—but great to see that we can now also handle version 2).

Yes, I was using version 2 because most of my meshes are adapted from the older remnants, and I did not want to spend a lot of time modifying them all.

So do you think it is good enough to merge this, should I mark it ready for review? Or is it not worth having in the main branch, and I just use this only when necessary from this branch?

[davschneller]

I'd merge it. We should support it IMO. :)

(same for eventually Netcdf etc.; not sure if Simmodeler supports it)