For PRs whose commits end up as HEAD of master like
bcc5c141bf,
there might be workflow runs associated with both PRs and pushes.
Only the push event is the one that will run fully and should be used to compare against.
So far it didn't distinguish between the two events, causing it to
sometimes pick the wrong one and then fail to download non-existent
artifacts: https://github.com/NixOS/nixpkgs/actions/runs/12092053414/job/33721377179
This commit fixes that by ensuring that the push event result is
compared against
A free-form `attrsOf str` option that is merged with the /etc/os-release
builder, allowing downstreams to customise arbitrary os-release fields.
This is separate from the variant option, as using an attribute set
merge means one gets an infinte recursion when making extraOSReleaseArgs
a recursive set, and the variant attribute is useful to define elsewhere
or multiple times.
Ditto for /etc/lsb-release.
Signed-off-by: Fernando Rodrigues <alpha@sigmasquadron.net>
It's almost 2025; we don't need to use 3-bit colour anymore. Let's use
the proper colour code for NixOS' light blue:
ea1384e183/logo/README.md (colours)
Signed-off-by: Fernando Rodrigues <alpha@sigmasquadron.net>
can be used like:
```nix
mpv.override { scripts = [ mpvScripts.mpv-image-viewer.image-positioning ]; }
```
Co-authored-by: Arne Keller <2012gdwu+github@posteo.de>
Commit e14483d6a6 fixed a bug in the `ini`
type with `listsAsDuplicatedKeys = true;`: multiple list declarations
weren't merged, but latter declarations shadowed the former without any
error.
The fix brought another issue to surface however: before, the
`plugin-load-add` declaration in the MySQL test shadowed the
`auth_socket.so` setting in the module. But now the attempt
to merge a list and a single declaration breaks because of
`types.either` seeing a mix of declarations from the left
AND right type.
Turning the `plugin-load-add` in the module into a list triggers the
correct merging behavior and thus fixes the evaluation error (and
merging behavior of `plugin-load-add`)!
This wasn't an issue for mysql itself (empty `plugin-load-add` in the
test) and neither for mariadb (the `auth_socket.so` isn't added for
this).
