The current type for the busId options are too relaxed, a stricter
constraint should be imposed to guard against typos which result
in Xorg unable to start.
This commit restricts the type to adhere to the B/D/F notation[1] for
addressing devices as expected by the module option.
[1] - https://wiki.osdev.org/PCI#Configuration_Space_Access_Mechanism_.231
Apparently this was never properly tested and never worked. When the IPFS repo needs upgrading, the first call to ipfs, which is run before running the migration, fails with the error message "Error: ipfs repo needs migration".
To fix this, simply run the migration before any `ipfs config` calls but don't run it when `dataDir` is empty and we need to call `ipfs init`.
Writing a NixOS test for this would require keeping at least two versions of IPFS in Nixpkgs, which we don't currently do.
* python3Packages.fenics: fix build, pin to older boost
Looking at upstream, there are various issues with newer boost.
(At least some of them have been since fixed)
For now, fix the build by using a version of boost that works
with the current version.
Error here was complaining about `std::min_element`,
which is no longer available, apparently, due to newer boost
no longer (transitively) including <algorithm>.
This was added in C++17, so I'm not sure the cmake flag
specifying dolfin built with C++11 makes sense or is used.
Leaving for now :).
* nixos/tests/fenics: fix name of machine/node in script
Still fails for now.
* python3Packages.fenics: fix accidentally changed strings in subst
Looks like in migration to pkg-config this was erroneously
changed from `pkgconfig` (python package, and source string)
to `pkg-config` (nix package name, tool name).
(see 9bb3fccb5b)
Fixes the NixOS test.
1. Update the default values of several addresses-related settings
that have been changed by upstream.
2. Make `dns.address` take multiple addresses. This is needed
for dual stack, now working by default.
Until now, this script has used the version of pandoc from unstable.
This means that running the script on the same version of Nixpkgs
could produce different results, and meant that when Pandoc's output
was changed, random PRs were changing the whole manual when they ran
the script to regenerate docs[1][2].
Here I've changed the manual to use a consistent version of pandoc —
the one from the latest release tag, which will avoid this problem in
future. This will avoid this problem in future. The only time we'll
need to worry about pandoc output changes is when we bump the version
used in this script.
I also considered using the version from the current Nixpkgs branch,
but decided against it as it's unlikely that e.g. the person bumping
Pandoc will remember to regenerate the manual.
[1]: https://github.com/NixOS/nixpkgs/pull/162550
[2]: https://github.com/NixOS/nixpkgs/pull/168535
GNOME 42 needs two wallpaper pictures – for the default (light)
colour scheme and for the dark one. Because we are clearing out
the paths in `gsettings-desktop-schemas` to prevent closure
from bloating, we need to set them in the NixOS module.
Since the wallpaper for the default colour scheme is dark,
will relegate it to the dark colour scheme and switch
to a light blue variant for the default colour scheme.
That one has inverted roundel for the NixOS logo but
it is the only light-ish background that has the logo
of the same size and placing as the dark wallpaper.
This module exposes a config.system.build.kexecBoot attribute,
which returns a directory with kernel, initrd and a shell script
running the necessary kexec commands.
It's meant to be scp'ed to a machine with working ssh and kexec binary
installed.
This is useful for (cloud) providers where you can't boot a custom image, but
get some Debian or Ubuntu installation.
With version 17 of Keycloak, the Wildfly based distribution was
deprecated in favor of the one based on Quarkus. The difference in
configuration is massive and to accommodate it, both the package and
module had to be rewritten.
This describes how to build your own installer medium with some custom
firmware/drivers, using an Intel MacBook as an example - on which WiFi
doesn't work out of the box, due to it being nonfree.
Fixes#15162.
Co-authored-by: Lucas Hoffmann <lucc@users.noreply.github.com>
Co-authored-by: Robert Hensing <roberth@users.noreply.github.com>