the recent luarocks can install for a different interpreter than the one
running luarocks.
Due to the way the update is done on nix, it seems more practical to use
this feature than running the script with different luarocks-nix ?
For update script parallelization, we have started calling builtins.toJSON
on updateScripts, which triggers evaluation of paths and therefore their copying
to Nix store. This breaks update scripts that assume that they exist in nixpkgs
like dwarf-fortress.
https://github.com/NixOS/nixpkgs/issues/61935
Let’s stringify the paths before JSONification to prevent the evaluation.
One bugfix patch is included - merged upstream but not released yet.
knot-resolver wrapper would need to add binaryheap explicitly,
so it's migrated to the automatic LUA path discovery instead.
Similarly rename std.normalize to std_normalize.
Having a dot in the name prevents some changes to the lua infrastructure,
for instance passing attribute names { std._debug }: to a function would fail with
`syntax error, unexpected ',', expecting '.' or '=',`
Also made changes to lua package generation system to account for packages like cqueues
that have one version per lua interpreter.
The csv file also accepts comments now.
* neovim-unwrapped: now use lua environments
* mpv: use lua environments
* luaPackages.inspect: init at 3.1.1-0
* luaPackages.lgi: mark as a lua module
* luaPackages.vicious: mark as a lua module
* lua: generate packages from luarocks
* luarocks-nix: update
* removed packages already available in nixpkgs
* adressing reviews
update script can now accept another csv file as input with -c
* Remove obsolete comment
All Python packages now have an updateScript. The script calls
`update-python-libraries` and passes it the position of the derivation
expression obtained using `meta.position`. This works fine in case a Nix
expression represents only a single derivation. If there are more in it,
`update-python-libraries` will fail.
To make updating large attribute sets faster, the update scripts
are now run in parallel.
Please note the following changes in semantics:
- The string passed to updateScript needs to be a path to an executable file.
- The updateScript can also be a list: the tail elements will then be passed
to the head as command line arguments.
* Print Dependency failures as well as direct failures and update to Python3
some package fail due to non-exposed dependencies and would thus not appear in the list, for example gcj
* hydra-eval-failures: simpler hashbang
