rubiks: init at 20070917 #38803

timokau · 2018-04-11T19:02:03Z

Motivation for this change

Package rubiks.

Things done

Tested using sandboxing (nix.useSandbox on NixOS, or option build-use-sandbox in nix.conf on non-NixOS)
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 nox --run "nox-review wip"
Tested execution of all binary files (usually in ./result/bin/)
Fits CONTRIBUTING.md.

dotlambda · 2018-04-12T17:17:34Z

pkgs/development/libraries/science/math/rubiks/default.nix

+  ];
+
+  meta = with stdenv.lib; {
+    homepage = https://github.com/graph-algorithms/edge-addition-planarity-suite;


That doesn't seem correct

Ma27 · 2018-04-13T19:30:03Z

do we really want to add a package from 2007 into our package set?

timokau · 2018-04-13T20:39:56Z

Why not? I don't think rubiks cubes have changed significantly since then.

Ma27 · 2018-04-13T21:01:02Z

that’s true, it might’ve reached some kind of EOL phase though. Before each release we fix tons of packages because of independencies (such as am inconpatible GCC versions) and I think that packages from 2007 will be quite painful to keep building..

timokau · 2018-04-13T21:13:35Z

It works for now without any special attention. I'm packaging this as a sage dependency, which means theres at least one other group interested in keeping it buildable.

dotlambda · 2018-04-14T08:15:38Z

pkgs/development/libraries/science/math/rubiks/default.nix

+
+  installFlags = [
+    "DESTDIR=${placeholder "out"}"
+    "PREFIX="


Normally, setting the prefix is preferred.

I'm sure I had a reason for that, I'll check what happens if I leave it out.

Okay it works if I just set

"PREFIX=${placeholder "out"}"

But shouldn't that be done automatically anyways?

No, it's --prefix that is set by default: https://nixos.org/nixpkgs/manual/#ssec-configure-phase

Does it work if you do just "PREFIX=$out"?

dotlambda · 2018-04-14T08:32:30Z

pkgs/development/libraries/science/math/rubiks/default.nix

+      mit # Dik T. Winter's software
+    ];
+    maintainers = with maintainers; [ timokau ];
+    platforms = platforms.all;


dotlambda · 2018-04-14T17:23:49Z

Considering the number of patches, this is maybe better suited for a sage overlay.

Ma27 · 2018-04-14T17:38:14Z

agreed, that's what I'm worried about - we need several patches ATM to keep it working and I fear that the number will grow with the time it lives in nixpkgs which will cause additional (and IMO avoidable) overload especially in the last weeks before a release.

It's absolutely not that I want to discourage @timokau's work, I'm just not sure if we actually want to add as many packages as possible and rather say no to some new packages (e.g. when they're way to old).
I don't know about the further context and status of this package, but are there any people reachable that could keep the project alive? I (personally) believe that such cases upstream needs to get their stuff fixed before we can make this available in our package set.

timokau · 2018-04-15T10:26:54Z

agreed, that's what I'm worried about - we need several patches ATM to keep it working and I fear that the number will grow with the time it lives in nixpkgs which will cause additional (and IMO avoidable) overload especially in the last weeks before a release.

The patches are effectively the sage developers way to maintain it. I'm not sure why they prefer it to altering the source tarball in this case, since the tarball is generated by them anyways, but maybe they just want to keep it reproducible from the upstream packages. Debian, Gentoo and Arch all package this (and all the other packages I've created PRs for).

When the package breaks somebody may fix it, probably me. If nobody does, it gets marked broken and someone might still fix it in the future. I am pretty sure there are multiple people here interested in sage, and nix might attract more once its packaged in a proper way (I think nix and sage more or less attrackt similar groups of people).

If I create an overlay, all the maintenance burden is on me alone. I don't know if I can or want to maintain that.

It's absolutely not that I want to discourage @timokau's work, I'm just not sure if we actually want to add as many packages as possible and rather say no to some new packages (e.g. when they're way to old).
I don't know about the further context and status of this package, but are there any people reachable that could keep the project alive? I (personally) believe that such cases upstream needs to get their stuff fixed before we can make this available in our package set.

The sagemath upstream effectively maintains this. They have a big community and are pretty active. For reference:
https://trac.sagemath.org/search?q=rubiks

Apparently somebody already suggested a proper fork (instead of the "pach fork" they currently maintain): https://trac.sagemath.org/ticket/21493

matthewbauer · 2018-04-19T17:29:58Z

pkgs/development/libraries/science/math/rubiks/default.nix

+  };
+
+  preConfigure = ''
+    export INSTALL="${coreutils}/bin/install"


You can actually do this as an attribute instead of configure:

INSTALL="${coreutils}/bin/install";

But is that better? It would clutter the attributes of the package and basically depend on a side-effect of how attributes are handled (and passed as environment variables) in nix, right?

this the standard way to set environment variables.

matthewbauer · 2018-04-19T17:32:35Z

I'm inclined to merge this. 4 patches is a lot but not ridiculous. Eventually we will have "inclusion" rules for what can go in Nixpkgs but right now there are none AFAIK (except maybe against nonfree stuff).

@GrahamcOfBorg build rubiks

GrahamcOfBorg · 2018-04-19T17:33:12Z

No attempt on x86_64-darwin (full log)

The following builds were skipped because they don't evaluate on x86_64-darwin: rubiks

Partial log (click to expand)


a) For `nixos-rebuild` you can set
  { nixpkgs.config.allowUnsupportedSystem = true; }
in configuration.nix to override this.

b) For `nix-env`, `nix-build`, `nix-shell` or any other Nix command you can add
  { allowUnsupportedSystem = true; }
to ~/.config/nixpkgs/config.nix.

GrahamcOfBorg · 2018-04-19T17:33:31Z

Success on x86_64-linux (full log)

Attempted: rubiks

Partial log (click to expand)

/nix/store/jy9knxp7nmw80jkf932axrs1b4p9k4hi-coreutils-8.29/bin/install dietz/mcube/mcube ut/bin
/nix/store/jy9knxp7nmw80jkf932axrs1b4p9k4hi-coreutils-8.29/bin/install dietz/cu2/cu2 ut/bin
/nix/store/jy9knxp7nmw80jkf932axrs1b4p9k4hi-coreutils-8.29/bin/install dik/dikcube ut/bin
/nix/store/jy9knxp7nmw80jkf932axrs1b4p9k4hi-coreutils-8.29/bin/install dik/size222 ut/bin
post-installation fixup
shrinking RPATHs of ELF executables and libraries in /nix/store/4clr9jnh8rila1pblsifr5mdz9pra9k6-rubiks-20070912
strip is /nix/store/j75dgadrff2d1fyc4fczmcgqkid2imdx-binutils-2.30/bin/strip
patching script interpreter paths in /nix/store/4clr9jnh8rila1pblsifr5mdz9pra9k6-rubiks-20070912
checking for references to /build in /nix/store/4clr9jnh8rila1pblsifr5mdz9pra9k6-rubiks-20070912...
/nix/store/4clr9jnh8rila1pblsifr5mdz9pra9k6-rubiks-20070912

GrahamcOfBorg · 2018-04-19T17:33:49Z

Success on aarch64-linux (full log)

Attempted: rubiks

Partial log (click to expand)

/nix/store/0svz5all13h1lxkfbx5y9vvy6r4cy422-coreutils-8.29/bin/install dietz/mcube/mcube ut/bin
/nix/store/0svz5all13h1lxkfbx5y9vvy6r4cy422-coreutils-8.29/bin/install dietz/cu2/cu2 ut/bin
/nix/store/0svz5all13h1lxkfbx5y9vvy6r4cy422-coreutils-8.29/bin/install dik/dikcube ut/bin
/nix/store/0svz5all13h1lxkfbx5y9vvy6r4cy422-coreutils-8.29/bin/install dik/size222 ut/bin
post-installation fixup
shrinking RPATHs of ELF executables and libraries in /nix/store/7h489mh054fxrpjlqvcja1zwd0lz42av-rubiks-20070912
strip is /nix/store/j7d4mr0ikv974ig7yzhknpsq288js4bs-binutils-2.30/bin/strip
patching script interpreter paths in /nix/store/7h489mh054fxrpjlqvcja1zwd0lz42av-rubiks-20070912
checking for references to /build in /nix/store/7h489mh054fxrpjlqvcja1zwd0lz42av-rubiks-20070912...
/nix/store/7h489mh054fxrpjlqvcja1zwd0lz42av-rubiks-20070912

GrahamcOfBorg · 2018-04-19T17:34:01Z

Success on x86_64-linux (full log)

Attempted: rubiks

Partial log (click to expand)

/nix/store/4clr9jnh8rila1pblsifr5mdz9pra9k6-rubiks-20070912

GrahamcOfBorg · 2018-04-19T17:34:14Z

Success on aarch64-linux (full log)

Attempted: rubiks

Partial log (click to expand)

/nix/store/7h489mh054fxrpjlqvcja1zwd0lz42av-rubiks-20070912

GrahamcOfBorg · 2018-04-19T17:35:14Z

No attempt on x86_64-darwin (full log)

The following builds were skipped because they don't evaluate on x86_64-darwin: rubiks

Partial log (click to expand)


a) For `nixos-rebuild` you can set
  { nixpkgs.config.allowUnsupportedSystem = true; }
in configuration.nix to override this.

b) For `nix-env`, `nix-build`, `nix-shell` or any other Nix command you can add
  { allowUnsupportedSystem = true; }
to ~/.config/nixpkgs/config.nix.

matthewbauer · 2018-04-19T17:38:31Z

One thing that might be neat and lessen the maintenance burden would be an automatic importer of these Sage packages. It's common to have things like this for instance "cabal2nix" or "emacs2nix". Perhaps someone wants to try something like "spkg2nix".

A good start for the above would be creating a shared file like pkgs/science/math/sage/default.nix that holds all of the sage stuff. Not sure if that makes sense it just might help with clutter in all-packages.nix at least.

GrahamcOfBorg · 2018-04-19T17:39:16Z

Success on aarch64-linux (full log)

Attempted: rubiks

Partial log (click to expand)

/nix/store/7h489mh054fxrpjlqvcja1zwd0lz42av-rubiks-20070912

timokau · 2018-04-20T08:20:19Z

One thing that might be neat and lessen the maintenance burden would be an automatic importer of these Sage packages. It's common to have things like this for instance "cabal2nix" or "emacs2nix". Perhaps someone wants to try something like "spkg2nix".

I actually do have such an exporter, which is what I started out with. But then I migrated all the packages to system packages one-by-one. The generated packages are not very pretty (because spkgs simply don't have as much information as a nix package, sometimes transitive dependencies are assumed, no difference between build inputs and native build inputs etc.). Also most sage dependencies are already packaged, and re-generating duplicates of those defeats the point of packaging it with system packages.

Edit: For comparison, here is the auto-generated rubiks.nix (could be improved a bit of course, but not to the level of hand-written): https://gist.github.com/timokau/dd42ccaf273afdbcbd85de913db8d6f7

GrahamcOfBorg added 8.has: package (new) 10.rebuild-darwin: 1-10 10.rebuild-linux: 1-10 labels Apr 11, 2018

dotlambda reviewed Apr 12, 2018

View reviewed changes

timokau force-pushed the rubiks-init branch from b7f8acd to 7e67c51 Compare April 13, 2018 20:39

dotlambda reviewed Apr 14, 2018

View reviewed changes

timokau force-pushed the rubiks-init branch from 7e67c51 to 70c2048 Compare April 14, 2018 08:29

dotlambda reviewed Apr 14, 2018

View reviewed changes

timokau force-pushed the rubiks-init branch from 70c2048 to 5929251 Compare April 14, 2018 08:36

rubiks: init at 20070917

1403a61

timokau force-pushed the rubiks-init branch from 5929251 to 1403a61 Compare April 14, 2018 08:37

GrahamcOfBorg added 10.rebuild-darwin: 0 and removed 10.rebuild-darwin: 1-10 labels Apr 14, 2018

dotlambda approved these changes Apr 14, 2018

View reviewed changes

matthewbauer reviewed Apr 19, 2018

View reviewed changes

NixOS deleted a comment from GrahamcOfBorg Apr 19, 2018

matthewbauer merged commit d0d7894 into NixOS:master Apr 22, 2018

timokau deleted the rubiks-init branch April 22, 2018 17:27

timokau mentioned this pull request Apr 27, 2018

rubiks: fix build #39608

Merged

8 tasks

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

rubiks: init at 20070917 #38803

rubiks: init at 20070917 #38803

timokau commented Apr 11, 2018 •

edited

dotlambda Apr 12, 2018

Ma27 commented Apr 13, 2018

timokau commented Apr 13, 2018

Ma27 commented Apr 13, 2018

timokau commented Apr 13, 2018

dotlambda Apr 14, 2018

timokau Apr 14, 2018

timokau Apr 14, 2018

dotlambda Apr 14, 2018

timokau Apr 14, 2018

dotlambda Apr 14, 2018

dotlambda commented Apr 14, 2018

Ma27 commented Apr 14, 2018

timokau commented Apr 15, 2018

matthewbauer Apr 19, 2018

timokau Apr 20, 2018

Mic92 Apr 21, 2018

matthewbauer commented Apr 19, 2018 •

edited

GrahamcOfBorg commented Apr 19, 2018

GrahamcOfBorg commented Apr 19, 2018

GrahamcOfBorg commented Apr 19, 2018

GrahamcOfBorg commented Apr 19, 2018

GrahamcOfBorg commented Apr 19, 2018

GrahamcOfBorg commented Apr 19, 2018

matthewbauer commented Apr 19, 2018

GrahamcOfBorg commented Apr 19, 2018

timokau commented Apr 20, 2018 •

edited

rubiks: init at 20070917 #38803

rubiks: init at 20070917 #38803

Conversation

timokau commented Apr 11, 2018 • edited

Motivation for this change

Things done

Choose a reason for hiding this comment

Ma27 commented Apr 13, 2018

timokau commented Apr 13, 2018

Ma27 commented Apr 13, 2018

timokau commented Apr 13, 2018

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

dotlambda commented Apr 14, 2018

Ma27 commented Apr 14, 2018

timokau commented Apr 15, 2018

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

matthewbauer commented Apr 19, 2018 • edited

GrahamcOfBorg commented Apr 19, 2018

GrahamcOfBorg commented Apr 19, 2018

GrahamcOfBorg commented Apr 19, 2018

GrahamcOfBorg commented Apr 19, 2018

GrahamcOfBorg commented Apr 19, 2018

GrahamcOfBorg commented Apr 19, 2018

matthewbauer commented Apr 19, 2018

GrahamcOfBorg commented Apr 19, 2018

timokau commented Apr 20, 2018 • edited

timokau commented Apr 11, 2018 •

edited

matthewbauer commented Apr 19, 2018 •

edited

timokau commented Apr 20, 2018 •

edited