Several drivers are built for Darwin on Hydra even though they never
worked and probably never will work there. Mark them as broken to avoid
wasted cycles.
upstream startx unsets DBUS_SESSION_BUS_ADDRESS, which in turn breaks
applications run under the startx display manager. arch has shipped this
patch for years as well.
(with review changes from @Mic92)
OpenGL applications should not depend on the Mesa drivers, since these
are supposed to be loaded at runtime from /run/opengl-driver using the
glvnd library. Furthermore, the drivers output has a large closure
since it depends on LLVM.
The only dependency from dev to drivers came from the xatracker
pkg-config file. This commit moves this file into a new
output called driversdev.
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.
This partially reverts bf3d3dd19b.
I don't know why we weren't getting a default logfile back then but Xorg
definitely provides one now ($XDG_DATA_HOME for regular users and /var/log for
root, see `man Xorg`)
Xorg creates the log-dir in its output path because X crashes if it can't write
to its logfile. On a regular distro, this dir would be installed into the root
to prevent that from happening but with Nix, it sits in the read-only Nix store.
Ironically, when Xorg tries to write here, it fails and crashes.
To make Xorg log to /var/log, we have to stop the build script from trying to
create the log-dir as the sandbox doesn't (and shouldn't) have access to /var.
This creates a runtime dependency on /var when running as root but that should
exist on any Linux system (on NixOS, journald always creates /var/log).
Previously, the startx displayManager required some workarounds for logfiles
which are obsolete now.
patchPhase -> postPatch because overriding the patchPhase prevents patches from
being applied
This reverts commit c778945806.
I believe this is exactly what brings the staging branch into
the right shape after the last merge from master (through staging-next);
otherwise part of staging changes would be lost
(due to being already reachable from master but reverted).
