This provides a means to build a PHP package based on a list of
extensions from another.
For example, to generate a package with all default extensions
enabled, except opcache, but with ImageMagick:
php.withExtensions (e:
(lib.filter (e: e != php.extensions.opcache) php.enabledExtensions)
++ [ e.imagick ])
So now we have only packages for human interaction in php.packages and
only extensions in php.extensions. With this php.packages.exts have
been merged into the same attribute set as all the other extensions to
make it flat and nice.
The nextcloud module have been updated to reflect this change as well
as the documentation.
Make mkExtension put headers in the dev output and use them, instead of
a different part of the current source tree, when referring to another
extension by using internalDeps.
This means external extensions can be built against the internal ones.
This means php packages can now refer to other php packages by looking
them up in the php.packages attribute and gets rid of the internal
recursive set previously defined in php-packages.nix. This also means
that in applications where previously both the php package and the
corresponding version of the phpPackages package set had to be
specified, the php package will now suffice.
This also adds the phpWithExtensions parameter to the
php-packages.nix, which can be used by extensions that need a fully
featured PHP executable.
This plugin allows configuring the URLs generated by Catalyst (and
therefore by Hydra) to be relative instead of absolute, which makes it
automatically behave correctly when Hydra is accessed both directly
and behind a reverse proxy.
Upstream dropped python2 support in 1.45, so the `nixpkgs-update` bot has not
been successful in bumping it.
While this still does not run the test suite, it gets us closer.
This reverts commit 09dde57e93.
Apparently, the rust-cbindgen bump wasn't the cause for the firefox
build error reported in https://github.com/NixOS/nixpkgs/pull/83247
(we could reproduce the build error even after
09dde57e93 applied).
For some reason it must have succeeded on hydra, as it's in the cache,
tricking us in believing 76458f89f4 broke
it initially.
So the build seems flaky of some sort - we haven't yet determined
whether it's luck, compiling with the right CPUs or something else. :-/
There's still some investigation to be done
(https://github.com/NixOS/nixpkgs/issues/84283), but no need to keep an
ineffective revert around.
It was using libusb-compat instead of libusb1, and the former no longer propagates the latter.
While at it, also move cmake to nativeBuildInputs and format the expression.
Allow build pass by disabling test. Isolated issue to
test_sockets.py::TestAIOSockets::test_sock_close_add_reader_race.
This test is supposed to be skipped, but it isn't for some reason,
so we disable it instead.
See uvloop#284 (https://github.com/MagicStack/uvloop/pull/284)
for full details. Don't know why this test isn't properly skipped.
