julia_15: run test suite and mark as broken

[samuela]

Fixes #121101.

Motivation for this change

#121101

Things done

• Tested using sandboxing (nix.useSandbox on NixOS, or option sandbox in nix.conf on non-NixOS linux)
• Built on platform(s)
  • NixOS
  • macOS
  • other Linux distributions
• Tested via one or more NixOS test(s) if existing and applicable for the change (look inside nixos/tests)
• Tested compilation of all pkgs that depend on this change using nix-shell -p nixpkgs-review --run "nixpkgs-review wip"
• Tested execution of all binary files (usually in ./result/bin/)
• Determined the impact on package closure size (by running nix path-info -S before and after)
• Ensured that relevant documentation is up to date
• Fits CONTRIBUTING.md.

[7c6f434c]

Somewhat against that as someone who uses Julia for non-numeric stuff. On the other hand, if 1.6 works strictly better everywhere and 21.05 branch-off is close, maybe we should simply drop 1.5

[samuela]

Somewhat against that as someone who uses Julia for non-numeric stuff. On the other hand, if 1.6 works strictly better everywhere and 21.05 branch-off is close, maybe we should simply drop 1.5

I hear that, and marking as broken does inconvenience some users. Although I think it is worth noting that julia is primarily designed to be a numerical computing language, the likes of matlab, python/numpy, etc. So a considerable chunk of the user base will be using it for those purposes, myself included. Finding pernicious floating point bugs in your code is a really really bad time. For me personally it could mean having to retract academic papers.

I'm all for killing off 1.5 though as you suggest though. I don't think there will be too much reason to keep it around once 1.6 lands.

[7c6f434c]

I'm all for killing off 1.5 though as you suggest though. I don't think there will be too much reason to keep it around once 1.6 lands.

… which I think has landed in the branch, if not in the channel

[samuela]

Not quite yet, iiuc @ninjin will be making that happen in a forthcoming PR.

[ghost]

On Thu 29 Apr 2021, Samuel Ainsworth wrote: Not quite yet, iiuc @ninjin will be making that happen in a forthcoming PR.

Yes, needs some mild polishing: regular expressions into patches and removal of cusses.

[ghost]

On Thu 29 Apr 2021, ninjin wrote: On Thu 29 Apr 2021, Samuel Ainsworth wrote: > > Not quite yet, iiuc @ninjin will be making that happen in a forthcoming PR. Yes, needs some mild polishing: regular expressions into patches and removal of cusses.

All right, binary packages for v1.0.5 and v1.6.1 are now in a PR [1]. [1]: #123188 The 21.05 feature freeze is now five days away [2], so I guess we need a decision here. Personally, barring a working v1.6.x expression that builds from source (which I deem highly unlikely any time soon) I do think it makes sense to keep `julia_15` around. But, with tests disabled as it otherwise will not build and marked as broken so that the user can explicitly enable it if desired. Thoughts? [2]: #121972

[samuela]

@ninjin Yeah, I think that all makes sense. IMHO either we should remove julia_15 or mark it as broken.

[ghost]

On Sun 16 May 2021, Samuel Ainsworth wrote: @ninjin Yeah, I think that all makes sense. IMHO either we should remove julia_15 or mark it as broken.

Agreed. Could you add a comment above `broken = true;` stating that it fails to pass its test suite? I guess also throw in a rebase while you are at it. Other than that this is good to merge in my book and should really be so before #121972 hits is three days.

[samuela]

Agreed. Could you add a comment above broken = true; stating that it fails to pass its test suite? I guess also throw in a rebase while you are at it. Other than that this is good to merge in my book and should really be so before #121972 hits is three days.

Roger that, done in 1e7f3bf.

[SuperSandro2000]

The 21.05 feature freeze is now five days away [2]

We can always backport stuff.

[ghost]

On Tue 18 May 2021, Samuel Ainsworth wrote: > Agreed. Could you add a comment above `broken = true;` stating that it fails to pass its test suite? I guess also throw in a rebase while you are at it. Other than that this is good to merge in my book and should really be so before #121972 hits is three days. Roger that, done in 1e7f3bf.

Super! I would squash them into a single commit, but frankly that could be a bit pedantic. Good to merge from my perspective.