Release v0.4 or wait to finish 1.0?

[KnutAM]

Just to have the discussion...

There are some bugs that are fixed on master + some issues with docs (e.g. #908).

It seems like there are some more things that would be good to get into 1.0, so it is still a bit away.

Would an alternative be to release v0.4 now, and then accept that there will be a second round of breaking changes before v1.0 which we could aim to release at/before FerriteCon2024?

[termi-official]

xref https://github.com/Ferrite-FEM/Ferrite.jl/pulls?q=is%3Aopen+is%3Apr+milestone%3Av1.0.0

The GPU PR is likely non-breaking for users so we can delay it.

[koehlerson]

should #780 be part of 1.0 milestone?

[fredrikekre]

I think at this point only breaking things should be holding up 1.0. I believe #780 only adds new features right?

[koehlerson]

https://github.com/Ferrite-FEM/Ferrite.jl/pull/780/files#diff-8db5ed06971ce1004169862f3e499f55666dd630868843be8a53d30ea385e505R110-R119

so this would be okay after 1.0 ?

[KnutAM]

I'm soon posting an extension of @lijas' #789 which will be pretty breaking.
So an argument against releasing an intermediate step (if this should be included) is that this would introduce FacetValues replacing FaceValues, but FaceValues currently replaces FaceVectorValues and FaceScalarValues on master, so that gives two breaking changes which would only be one otherwise...

[KnutAM]

https://github.com/Ferrite-FEM/Ferrite.jl/pull/780/files#diff-8db5ed06971ce1004169862f3e499f55666dd630868843be8a53d30ea385e505R110-R119

so this would be okay after 1.0 ?

Only adding fields isn't breaking AFAIU: code (using public API) that used to work with 1.0 would then still work with 1.1

[termi-official]

Also #631 is technically breaking for users.