Build logs show:
> configure: WARNING: non-linux system; not building mount
> configure: WARNING: non-linux system; not building swapon
So skip these on non-Linux
Using getOutput prevents eval failures on other platforms.
Things should stay eval'able with NIXPKGS_ALLOW_UNSUPPORTED_SYSTEM=1
Co-authored-by: Artturin <Artturin@artturin.com>
In util-linux 2.39, blkid was improved to verify the checksums for
several types of block devices. In the bcache case, this calculation was
incorrect, leading to a regression where valid bcache cache devices
would not be detected any more. In some cases, this can even break the
boot because a required bcache device does not come up any more.
In https://github.com/util-linux/util-linux/pull/2293, a fix was
introduced upstream. Include these patches to make bcache detection work
properly again.
Many packages have some kind of flag indicating whether or not to build with
systemd support. Most of these default to `stdenv.isLinux`, but systemd does
not build on (and is marked `broken` for) `isStatic`. Only a few packages have
the needed `&& !isStatic` in the default value for their parameter.
This commit moves the logic for the default value of these flags into
`systemd.meta.{platforms,badPlatforms}` and evaluates those conditions using
`lib.meta.availableOn`.
This provides three benefits:
1. The default values are set correctly (i.e. including `&& isStatic`)
2. The default values are set consistently
3. The way is paved for any future non-Linux systemd platforms (FreeBSD is
reported to have experimental systemd support)
Util-linux does not fail when crypt() isn't available. It just doesn't
build sulogin, which in turn breaks the initrd-builder for the
systemd-stage-1 tests.
Introduce new boolean arguments "pamSupport" and "capabilitiesSupport" that
control whether "pam" and "libcap" are compiled in. These flags are true by
default, so this commit does not cause any rebuilds.
Prior to this commit, util-linux supported compilation without systemd
if systemd==null. This commit preserves that behavior, and
additionally triggers it when the global systemdSupport attr is set to
false. The systemdSupport argument is understood by many other
nixpkgs expressions and can be set globally in
~/.config/nixpkgs/config.nix.
Co-authored-by: Sandro <sandro.jaeckel@gmail.com>
This reverts commit 1f0ef842ca. The
underlying issue was fixed in 2ebeb02a99 ("stdenv/setup: tell libtool
about library paths"), so we don't need a workaround in util-linux any
more.
The "minimal" argument just removed some files to reduce package
size. Now that we have multiple output derivations in Nixpkgs (we
didn't when "minimal" was added), we can just throw more outputs at
this problem and get the same effect, with the advantage that those
files are actually there, available for download without a rebuild, if
they're ever needed.
I've tested that with this change, there's no difference in the
closure size of util-linuxMinimal.bin.
We have to set SYSCONFSTATICDIR to stop $bin/bin/agetty looking in
$out for issue or issue.d. SYSCONFSTATICDIR is supposed to be for
configuration that comes with a package, but it's currently only used
for finding an issue, and util-linux doesn't come with any issue
files, so I'm not really sure what the point is...
Removing the locales after they've been built and installed is a big
hack when util-linux gives us an option to disable them directly.
This way, the code in util-linux that deals with locales can be
disabled as well, leading to a (very small) size reduction.
Quoting uuidd(8), as generated prior to this commit:
> By default, the pid file is written to
> /nix/store/6ig1awrlxd‐nadna38kxgfvds9kfr6ny2-util-linux-2.36.2/var/run/uuidd/uuidd.pid.
That's not right...
This was probably needed for scriptrelay(1) before it was rewritten
from Perl to C in util-linux 2.14. I've checked with Diffoscope that
after removing the build input, the result is bit-for-bit identical
except for the different prefix.
continuation of #109595
pkgconfig was aliased in 2018, however, it remained in
all-packages.nix due to its wide usage. This cleans
up the remaining references to pkgs.pkgsconfig and
moves the entry to aliases.nix.
python3Packages.pkgconfig remained unchanged because
it's the canonical name of the upstream package
on pypi.
