search

NixOS / nixpkgs

Nix Packages collection & NixOS
MIT License
18.39k stars 14.34k forks source link

gfortran: many versions #61130

Open vcunat opened 5 years ago

vcunat commented 5 years ago

This is a thread to gather feedback. I see we have gfortran in five versions ATM, all being built on Hydra (for the two Linux platforms at least): 4.8, 4.9, 5, 6 and the default 7. I see 8 is missing; I tried 9 would now build by the trivial copy&paste, so I expect we'll add that, but first I suggest revisiting the usefulness of the older versions – perhaps we should prune at least the set being built on Hydra?

/cc random people that might be interested: @markuskowa, @ttuegel, @dtzWill.

markuskowa commented 5 years ago

I do not know how it came along that we have all the gfortran versions in the standard package set. Looking at pkgs/top-level/all-packages.nix it seems that they are set up in lock step with the gcc versions. Unless someone provides a reason that they are really needed to compile legacy code I am in favor of removing old versions. All packages that are in nixpkgs use the default version of gfortran (and seem to work fine with it). I would recommend removing at least versions 4.8, 4.9, 5 from the build set.

veprbl commented 5 years ago

What is the problem with building few versions of gfortran on hydra? I suppose we want to have gfortran 4.8, 4.9 etc. for the same reason we want to have gcc 4.8, 4.9 etc. in general.

edit: this message is supposed to follow the first message from markuskowa, but it received a wrong timestamp edit2: the timestamp got fixed somehow, this message now appears in the right place

vcunat commented 5 years ago

Not really a problem AFAIK, but assuming the builds are very unlikely to be useful, it seems quite a waste of resources. For gcc I guess we might also remove some, but C(++) are far more widespread than fortran nowadays, so It seems harder to determine they won't be useful (e.g. the relatively fresh RHEL/CentOS 7 uses gcc 4.8 by default).

EDIT: I reacted to the post just below – no idea why GitHub switched the two posts.

veprbl commented 5 years ago

@vcunat FWIW I am using gcc 4.8 from nixpkgs to mimic the RHEL7-like environment. I also did use older versions of gfortran when I was suspecting a compiler regression, I did not find any, but I did appreciate the ability to do the check.

veprbl commented 5 years ago

Wow, the message ordering is still messed up. The first message from me appears to be written from the future...

vcunat commented 5 years ago

Time machine discovered by accident? :thinking: I'm sure I've heard of such a story already.

stale[bot] commented 4 years ago

Thank you for your contributions.

This has been automatically marked as stale because it has had no activity for 180 days.

If this is still important to you, we ask that you leave a comment below. Your comment can be as simple as "still important to me". This lets people see that at least one person still cares about this. Someone will have to do this at most twice a year if there is no other activity.

Here are suggestions that might help resolve this more quickly:

  1. Search for maintainers and people that previously touched the related code and @ mention them in a comment.
  2. Ask on the NixOS Discourse.
  3. Ask on the #nixos channel on irc.freenode.net.
philiptaron commented 4 months ago

Five years later:

  1. gfortran (last updated 338ed797fe4a2)
  2. gfortran48 (last updated 48f63c2f2e375)
  3. gfortran49 (last updated 784025c747561)
  4. gfortran6 (last updated 58e0c4bcd9d6c)
  5. gfortran7 (last updated f03ee2475406a)
  6. gfortran8 (last updated 2bac9a5a01bd1)
  7. gfortran9 (last updated 7dc03548d1d61)
  8. gfortran10 (last updated da59984d367f4)
  9. gfortran11 (last updated 6ba5ac89d255c)
  10. gfortran12 (last updated c34a8024bf5f8)
  11. gfortran13 (last updated 4b8b00f56c2c2)
  12. gfortran14 (last updated 75104490c0f17)