lib, treewide: introduce repoRevToName and use it to cleanup most fetch* functions

[oxij]

This a first half of #49862. Basically, this is a noop cleanup that keeps all the fetch* names as they are now, while introducing the new config.fetchedSourceNameDefault option.

Changes

This patch adds lib.repoRevToName function that generalizes away most of the code used for derivation name generation by fetch* functions (fetchzip, fetchFromGitHub, etc, except those which are delayed until latter commits for mass-rebuild reasons).

It's first argument controls how the resulting name will look (see below).

Since lib has no equivalent of Nixpkgs' config, this patch adds config.fetchedSourceNameDefault option to Nixpkgs and then re-exposes lib.repoRevToName config.fetchedSourceNameDefault expression as pkgs.repoRevToNameMaybe which is then used in fetch* derivations.

The result is that different values of config.fetchedSourceNameDefault now control how the src derivations produced by fetch* functions are to be named, e.g.:

• fetchedSourceNameDefault = "source" (the default):

    $ nix-instantiate -A fuse.src
    /nix/store/<hash>-source.drv
  

• fetchedSourceNameDefault = "versioned":

    $ nix-instantiate -A fuse.src
    /nix/store/<hash>-libfuse-2.9.9-source.drv
  

• fetchedSourceNameDefault = "full":

    $ nix-instantiate -A fuse.src
    /nix/store/<hash>-libfuse-2.9.9-github-source.drv
  

See documentation of config.fetchedSourceNameDefault for more info.

[philiptaron]

A few curious questions as I read through this PR.

[oxij]

Okay, so I edited this to use fetchedSourceNameDefault instead of nameSourcesPrettily, and changed the type to be a simple enum.

`nixfmt` check fails, but I have no idea on which file. All files I think are ok to be touched pass `nixfmt` without changes locally. I don't think re-formatting of files under `pkgs/top-level` should be done in this PR. /cc @infinisil

A simple way to see which files fail the checks, except to run those checks yourself, would be nice, though.

[oxij]

ping @AndersonTorres specifically, since your review requested changes that were since resolved, IMHO.

Also, somewhat relatedly, #391322 is a follow-up to 6620964 which changed the version without changing the hash, which would not have happened if @mweinelt used this PR with fetchedSourceNameDefault = "versioned" set. :]

Such errors, how frequent are they? 🤔

Just saying.

[AndersonTorres]

I would complain about nixfmt not having its own commit. But nothing relevant is being lost.

So, LGTM.

[mweinelt]

Such errors, how frequent are they? 🤔

The cause is limitations in the python bulk updater. These commits are under my name, but update-python-libraries created them.

[oxij]

@AndersonTorres Thanks! Also, mass-reformat now ate those nixfmt changes anyway.

@mweinelt Well, okay, but it would have been caught by this infra.

Also, rebased onto master to fix conflicts.

[oxij]

Ping. What's holding this up?

[arianvp]

Took the liberty with @Lassulus to rebase this to solve the merge conflict

[oxij]

Thanks. Rebase LGTM.

[oxij]

Rebased again to fix conflicts. And it is now stuck at Darwin check again. And the previous time it was stuck right there for a month. So, I don't think it's likely this will get any better anytime soon.

[oxij]

Thanks @Lassulus!