search

ExtremeFLOW / neko

/ᐠ. ｡.ᐟ\ᵐᵉᵒʷˎˊ˗
https://neko.cfd/
Other
159 stars 27 forks source link

Harmonize factory module filenames #1207

Closed timofeymukha closed 3 months ago

timofeymukha commented 3 months ago

Most modules were named _fctry but some rogue ones (mostly coded be my) were _factory. Better to have the same naming, so I rounded off everything to _fctry.

Edit: not module names but rather the file names for the modules.

timofeymukha commented 3 months ago

To save space in the declaration, I'm in favour of fctry

Yeah, everything is fctry now. But my message was confusing, this only changes filenames to _fctry.f90, the module names were already correct.

timfelle commented 3 months ago

I guess we should be consistent on the contract or not across the different pars of the code. But yes the fctry is probably fine. In any case it is generally good to have module xyz in file xyz.f90. I have to open up files relatively often to figure out where the module i am looking for is ;P

timofeymukha commented 3 months ago

In any case it is generally good to have module xyz in file xyz.f90

Yeah, this is exactly what I address here, 3 file names changed to be consistent with module name.

timofeymukha commented 3 months ago

Nice intel bonk.

timofeymukha commented 3 months ago

@timfelle Just to add, I think the consistency between file and module name is more or less policy, so one can issue PRs to address inconsistencies, like this one. One biggie is coef.f90, with the module name coefs. But things are a bit more difficult now with the name shadowing taken into account. The coef_t variables are called coef in a lot of routines. So maybe we wait for _m module names for that one, I don't know.

timfelle commented 3 months ago

@timfelle Just to add, I think the consistency between file and module name is more or less policy, so one can issue PRs to address inconsistencies, like this one. One biggie is coef.f90, with the module name coefs. But things are a bit more difficult now with the name shadowing taken into account. The coef_t variables are called coef in a lot of routines. So maybe we wait for _m module names for that one, I don't know.

Yea, question is wether we should use a module xyz_m for file xyz.f90. Even though i can see someone like pFUnit does not work if the module and file names are identical.

timofeymukha commented 3 months ago

Even though i can see someone like pFUnit does not work if the module and file names are identical.

If that is the case, I suppose it forces our hand.

timfelle commented 3 months ago

Even though i can see someone like pFUnit does not work if the module and file names are identical.

If that is the case, I suppose it forces our hand.

Well it's for the .pf files, where it is struct, but thats also for their preprocessor thing.

timofeymukha commented 3 months ago

The CI somehow really got cursed here.

timofeymukha commented 3 months ago

CI finally blessed the PR, I think we can merge.