Jordi Manyer Fuertes

For a bit of context, I am currently updating `GridapPETSc` to work with v0.3, and find myself having to translate [the following](https://github.com/gridap/GridapPETSc.jl/blob/61de9dca0f484de6833b76e4e83160ce036113d2/test/PartitionedArraysTests.jl#L114): ```julia A = PSparseMatrix(I,J,V,ids,ids,ids=:global) ``` for which I...

> Related with this, I think a new constructor like this would be handy in some situations Yes, I've been doing something similar when building mock distributed models for `GridapDistributed`....

> My preferred option would be to add a new kw-argument in psparse! (discover_cols=true) that skips the generation of the col ids (assuming the the user already provided the final...

@fverdugo Ok, nice to know. Then I guess the PR is ready for review, with the new kwarg changes as agreed.

That would indeed be nice, specially for FEM-related assemblies. I might have time to implement it at some point, same for `partiton_from_colors(ranks,global_to_color)`. I'll make a note and hopefully get back...

@fverdugo I am no expert, but I after some reading I think I might have an idea of what could be happening: The CI files for `PartitionedArrays.jl` and the `gridap`...

@AlexandreMagueresse

Great! Ill try to make a list and sumit a PR (hopefully next week).

@oriolcg I don't have time atm, I'll try to push this further after the summer. I am not focusing complicated problems yet, but obviously anything that assembles repeatedly (like NL...

Hey @DanielBoigk , I'm afraid I do not know enough about FerriteGmsh to understand what is going on. I would however have a look at where both packages are trying...