move from astyle to clang-format

[MFraters]

I thought there was already an issue about moving from astyle to clang-format, but I would like to look into this again and I am considering moving to clang-format after the release of version 1.1. Will take another look at what dealii did here with their clang-format setup.

[tjhei]

Interesting! We have talked about this in ASPECT as well, so it would be good to coordinate. I am generally happy with the way we have set it up in deal.II.

[MFraters]

Good to hear you are generally happy with the deal.ii setup. I agree that it would be good to coordinate, so that deal.ii, aspect and gwb use the same formatting. This would I think just mean using the deal.ii formatting :)

I personally don't have a strong preference styling, as long as it is consistent. I do get the impression that clang-format is generally better integrated in various tools.

I am happy for gwb to bite the bullet before aspect to see what issues we encounter, but I guess there is already a lot of experience from the deal.ii side of switching formatting tools, right?

[tjhei]

Well, we spent some effort in making clang-format (the right version!) to be easy to install by hosting binaries. You could just use those for gwb.

Nothing fancy about switching. You just make any open PRs (or local branches people have) difficult to rebase. The switch itself is easy.

[MFraters]

So it sounds like it would be ideal to do it before the gwb hackathon, but after the release. So that would be tomorrow (I need to make the gwb release tomorrow, and the hack starts the evening of the day after). On the other hand, I think that a few weeks after the hackthon, the pull requests that are still open, will mostly stall anyway, so I could also delay it a bit to have more time to discuss it withing aspect.

I feel we actually already had this discussion several times during aspect hackathon's, and I got the feeling no one was really opposed to the change to clang-formt, but I also didn't find a record of it.

[gassmoeller]

Related ASPECT issue is geodynamics/aspect#6803.

Generally, I think ASPECT would consider switching as well.