| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This was one of the places where types.loaOf was still in place and it
got removed a while[1] ago and this in turn causes evaluation to fail
for quite a few machines:
The option value `boot.initrd.luks.devices' in `...' is not of type
`attribute set of submodules'.
I've not only changed all the machines to use attribute sets but also
fixed the check in core/tests.nix, because comparing against a list when
the actual type is an attribute set will result in all the LUKS tests to
be part of *all* channels, no matter whether you're actually using LUKS.
[1]: https://github.com/NixOS/nixpkgs/commit/20d491a317d9956ddca80913f07
Signed-off-by: aszlig <aszlig@nix.build>
Cc: @Profpatsch
Cc: @sternenseemann
|
|
|
|
|
|
|
| |
This has also been removed [1] upstream and therefore needs to be
removed here.
[1]: https://github.com/NixOS/nixpkgs/commit/b2f3bbd3fb958601a7357e39d66f226e065d76c1
|
|
|
|
|
|
|
| |
This package, the module and its test have been removed [1], so to
unbreak Hydra we need to remove it as well.
[1]: https://github.com/NixOS/nixpkgs/commit/aebf9a4709215c230e5841d60e2
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
I even mentioned the full names of the test attributes in
bc4203d221c87a0622d089aaa61884f86d40fd44:
tests.wireguard.wireguard-basic-linux-5_4
tests.wireguard.wireguard-basic-linux-latest
tests.wireguard.wireguard-generated-linux-5_4
tests.wireguard.wireguard-generated-linux-latest
tests.wireguard.wireguard-namespaces-linux-5_4
tests.wireguard.wireguard-namespaces-linux-latest
tests.wireguard.wireguard-wg-quick-linux-5_4
tests.wireguard.wireguard-wg-quick-linux-latest
So I forgot about the additional "wireguard-" prefix, so Hydra fails to
evaluate again:
in job 'tests.nixos.wireguard.basic-linux-5_4':
evaluation aborted with the following error message: 'cannot find attribute `nixos.wireguard.basic-linux-5_4''
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Commit 89c3cda819da5dc32e4a29ff8f2aad118bde64c7 changed the wireguard
attributes to use nested attributes because it was changed upstream.
However what the commit has missed is that the attributes are not just
plain attributes but come with a kernel version suffix, for example like
this:
tests.wireguard.wireguard-basic-linux-5_4
tests.wireguard.wireguard-basic-linux-latest
tests.wireguard.wireguard-generated-linux-5_4
tests.wireguard.wireguard-generated-linux-latest
tests.wireguard.wireguard-namespaces-linux-5_4
tests.wireguard.wireguard-namespaces-linux-latest
tests.wireguard.wireguard-wg-quick-linux-5_4
tests.wireguard.wireguard-wg-quick-linux-latest
Adding the "-linux-${version}" suffix should fix evaluation.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
A few days ago, the wireguard tests were consolidated[1] into a single
attribute with various subattributes, similar to how the virtualbox
tests and many other tests are structured.
I fixed the attributes and also added the wg-quick attribute that was
also introduced in the meantime[2].
This should fix the evaluation errors occuring whenever a channel
contains a machine configuration with wireguard enabled.
[1]: https://github.com/NixOS/nixpkgs/commit/41bd6d2614749d12ce5ded3e991555b56ea6b2dc
[2]: https://github.com/NixOS/nixpkgs/commit/abf60791e2bd274d39e0f18def46795798f9aefd
Signed-off-by: aszlig <aszlig@nix.build>
Cc: @Profpatsch
|
|
|
|
|
|
|
|
|
|
| |
The option has been renamed[1] from just "docker-containers" to
"virtualisation.oci-containers.containers", so our checks no longer
evaluate.
[1]: https://github.com/NixOS/nixpkgs/commit/2f7747526cc80844a506c4aa147
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
| |
With the merge of [1], the `.enable` check for
`config.services.dokuwiki` has been removed.
[1]: https://github.com/NixOS/nixpkgs/pull/83769
|
|
|
|
|
|
|
|
|
|
|
| |
The nesting.clone and nesting.children options were renamed[1] upstream
and they're now contained in one single submodule option, which means
that for our tests we only need to check for config.specialisation
instead of nesting.clone and nesting.children separately.
[1]: https://github.com/NixOS/nixpkgs/commit/ec2d28e3233c4c04fe6afe13e82
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
|
|
|
|
| |
All of the NixOS options in services.graphite.pager have been removed in
37ffa6ea51f3a3d7e43736b9a5ce3c149dd9d1ee, but since we're using the
"enable" option to check whether we need to run tests, we'll get an eval
error.
Removing the check fixes the evaluation error.
Signed-off-by: aszlig <aszlig@nix.build>
Reported-by: @devhell
|
|
|
|
|
|
|
|
|
| |
Another thing I missed in ffa7a45ba0aa27c926472e23587423bc07a6db53 is
that not only the module got renamed, but its interface has changed as
well and instead of an "enable" option, it's now just an attribute set
of submodules and the service(s) are enabled whenever it's non-empty.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
|
|
|
|
| |
The buildkite-agent module, NixOS test and option got renamed[1] to its
plural form, also supporting multiple agents.
While none of our machines are using the module, we nevertheless need to
fix the module option attribute so that our channels will evaluate.
[1]: https://github.com/NixOS/nixpkgs/commit/f003810989c58746db9ea52f623
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
| |
This test is only really useful in test environments and is imported
only within the QEMU test driver. The NixOS module option is not
available otherwise, which in turn will cause the evaluation to fail.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
|
| |
Eval error introduced by 0f0e3cef6297d1334a3b1c27bb125490850be84f.
The virtualisation.dockerPreloader attribute doesn't include an "enable"
option, but instead is configured by specifying a list of images.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
It has been a while, since we did a full update of all the tests, which
we last did in 3ea85a699a65dbdd4f633d21e70be6859b0f3f8f (April 2018).
Since then, a whole lot of tests were added and a few of them got
removed as well. Since we only did incremential updates whenever
something broke, we didn't have any newer tests.
We should now have conditions of all the tests we have in nixpkgs
master[1] minus a few special ones, which are only related to upstream
image generation (eg. the ec2 test).
[1]: https://github.com/NixOS/nixpkgs/commit/637bb9fa984238f2f7e6235ca11
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This fixes the following evaluation error:
The option `services.xserver.displayManager.slim' can no longer be
used since it's been removed. The SLIM project is abandoned and their
last release was in 2013.
Because of this it poses a security risk to your system.
Other issues include it not fully supporting systemd and logind
sessions.
Please use a different display manager such as LightDM, SDDM, or GDM.
You can also use the startx module which uses Xinitrc.
Here is the nixpkgs upstream pull request removing SLiM:
https://github.com/NixOS/nixpkgs/pull/73251
Since I was using a custom theme for SLiM and actually liked the
minimalism, it's probably time to start patching LightDM soon. For now
however, I'll stay with a default LightDM configuration and wait until
I'm getting annoyed :-)
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The test has been removed upstream[1] for the following reason[2]:
> BeeGFS is a fantastic high performance file system. However, after
> running it for a few months last year the meta data daemon kept
> crashing randomly in production use (support from the manufacturer
> requires a paid support contract). The current version that is nixpkgs
> is outdated and I do not want to maintain it any longer. The build
> system is purely based on make file and is hard to maintain.
[1]: https://github.com/NixOS/nixpkgs/commit/c3e17efaab0b53b4f45454aaaea7690506a8fcde
[2]: https://github.com/NixOS/nixpkgs/pull/66352
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
| |
`runCommandLocal` was added to nixpkgs in
https://github.com/NixOS/nixpkgs/pull/74642
to speed up trivial `runCommand` derivations by always building them
locally. We have a few places where that’s good to use.
|
|
|
|
|
|
|
|
|
| |
This test has been recently removed upstream [1].
[1]: https://github.com/NixOS/nixpkgs/commit/50ea99cbc18d3f480a773de5250b4ef9c7f6d514
Signed-off-by: devhell <devhell@mailfresser.de>
Cc: @aszlig
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Besides using the default value from options, I could have just used
mkOptionDefault back then when I was writing this, but fortunately using
the default attribute from the option itself now resulted in an eval
error.
Had I used mkOptionDefault, the value just would have been silently
changed to the default value, since in
https://github.com/NixOS/nixpkgs/pull/63810 the default value now has
priority 1000 instead of 1500 (mkOptionDefault).
So since there is no longer an option default, we now can just set our
binary cache in a singleton list (with priority 100) and be done.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
|
| |
The test got removed in NixOS/nixpkgs@552c2236254271f12342d3ad7a0f7c0bb
with the following comment:
The package/service is broken. Upstream is dead
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
| |
The test has been removed in nixpkgs because it is unmaintained:
https://github.com/NixOS/nixpkgs/commit/5fccac2b8d77a660a968809519710a5b
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
| |
The attribute linuxPackages_copperhead_stable has been removed in
NixOS/nixpkgs@ddce094ddf456434e15a7af315be1f6de4136fd7 because it's
unmaintained.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
|
|
|
| |
This has been removed in NixOS/nixpkgs@c3eefe801a55586445125d49f159c1c0.
@xeji writes in NixOS/nixpkgs#39991:
> fleet is deprecated and unmaintained upstream. panamax is broken and
> depends on fleet. Remove packages, modules and tests.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
| |
Those also have subtests and they also test different variations of
options, so let's actually create a check for every single variations.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
| |
The Kubernetes tests actually contain subtests, so let's specify them so
we don't run into errors during Hydra evaluation because the result
would be a generic attribute set instead of a derivation.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
|
| |
While hasPackage is not yet very accurate when it comes to detecting
whether a package is in systemPackages (it only matches the name
attribute), let's make at least two of the most frequently used browsers
part of our machine specific tests.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
| |
We want to make sure nix-instantiate --parse will catch errors very
early, because evaluation of the test definitions takes a ton of time.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
New tests as of NixOS/nixpkgs@9df35787f16636628a2c7392cc4bfeff0e815242:
buildbot, ceph, deluge, docker-edge, docker-tools,
docker-tools-overlay, fwupd, gjs, iftop, kubernetes.dns,
kubernetes.rbac, matrix-synapse, nix-ssh-serve, novacomd, openldap,
plotinus, powerdns, predictable-interface-names, prosody, quagga,
rabbitmq, rspamd, rxe, strongswan-swanctl, systemd, transmission,
xautolock, xrdp, yabar
The main kubernetes test no longer exists and has been split into
kubernetes.dns and kubernetes.rbac in
NixOS/nixpkgs@709b6f664e472ae2075bbee83bfb1dd1ba95e0e7.
I also added a small hasPackage helper for gjs, which we're going to use
for packages in the system path very soon.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
|
| |
This was changed in NixOS/nixpkgs@844c08ec0a5b79abb16d9de5bd39ed492f4.
Instead of linuxPackages_copperhead_hardened we now have
linuxPackages_copperhead_lts and linuxPackages_copperhead_stable.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
|
| |
Since NixOS/nixpkgs@c6f7d4367894047592cc412740f0c1f5b2ca2b59 the system
argument is now localSystem to be more consistent with crossSystem.
This fixes evaluation of all machine configurations.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
|
| |
This test has been removed in NixOS/nixpkgs@21dcead5f0c3d548ad4642a4e86,
so we no longer need to check for it.
Fixes an error during channel evaluation.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
|
|
|
| |
Since NixOS/nixpkgs@c1bed05e341756f7fdfa73035c7f1e078e49be3d it is now
called ipv4.addresses.
The reason why the evaluation on our behalf has failed despite the
option being aliased is that we try to access the option while checking
for eligible tests for a particular machine.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
|
|
| |
Since NixOS/nixpkgs@3a2ee6e5c3bfae64c604859c7cc1ee6b069ea5b5 the
attribute name for the kernel packages is now
linuxPackages_copperhead_hardened instead of
linuxPackages_hardened_copperhead, so we need to change it in our
conditions as well.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
New tests as of NixOS/nixpkgs@27a4a5511dc2cf9756943458d7a752617e4ad6e3:
acme, atd, beegfs, couchdb, dovecot, env, gitolite, grafana, graphite,
home-assistant, initrd-network-ssh, kernel-copperhead, kernel-latest,
kernel-lts, mesos, mutableUsers, mysqlBackup, netdata, nghttpx,
owncloud, prometheus, radicale, statsd, sudo, switchTest, vault,
xmonad, zookeeper
I've removed the pgjwt test, because the attribute is not reachable
anymore within nixos/release.nix as it has been commented out in
NixOS/nixpkgs@40ed22650737a444fee7883e69e852fd4b808133.
Also I didn't add the postgis test, because checking for the
configuration attribute is a bit tedious and chances are very low that
we'd have a machine in Vuizvui that uses it.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
| |
Openstack has been removed from nixpkgs (see PR #32637). Hydra's eval is
failing because of this, so we should probably remove the tests here
too.
|
|
|
|
|
|
|
|
|
|
|
|
| |
For example if you have something like cabal2nix as a lazy package,
using "cabal2nix --shell . > default.nix" will result in the store path
of cabal2nix being present in the resulting default.nix as the first
line.
We obviously do not want that, so let's redirect stdout of nix-store -r
to /dev/null.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
It's a bit weird to need to write something like this:
nix-build pkgs -A vuizvui.aszlig.vim
We're already within vuizvui, and the filesystem paths are already
structured in a way so that the above is in pkgs/aszlig/vim, so the need
to add the vuizvui attribute to that seems quite pointless and
unnatural.
I have tested this by evaluating a few machines, tests and packages.
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
| |
The configuration option boot.kernel.sysctl is of type attrsOf, so we
can't assume that the attribute "user.max_user_namespaces" is actually
there.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
New tests as of NixOS/nixpkgs@77dfa05bd5a296476143f02f2cfc0e47bdded01a:
* cloud-init
* hardened
* ldap
* snapper
Removed tests:
* grsecurity
Renamed tests:
* installer.simpleUefiGummiboot -> installer.simpleUefiSystemdBoot
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
There are two other occasions where packageOverrides are still in use,
one of them is @sternenseemann's fliewatuet machine and another one is
@devhell's package profile.
I've replaced every other occurence of packageOverrides and replaced it
with overlays and checked the store path hash after evaluation.
The reason why I left @sternenseemann's fliewatuet alone was that for
I wasn't quite sure whether "bluez = pkgs.bluez5" is still needed or
intentional to pin it to version 5. Because if it's not the case the
packageOverrides can just be dropped because bluez is already bluez5 in
upstream <nixpkgs>.
For @devhell's package profile, I did the conversion, but the store path
turned out to be a different one than what it was prior to the change.
I did take a quick look at the requisites of the drv and found that sox
was one of the different paths that led to the change in the final store
path.
This obviously needs to be tested and/or investigated first.
Other than that, the main reason why I'm moving everything to overlays
is that it's the replacement for packageOverrides and also has a better
way to pass through chains of overrides than packageOverrides had.
My guess is that after NixOS 17.09, the old packageOverrides function
will be removed, so let's make sure we're ready for that.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
Cc: @devhell, @sternenseemann
|
|
|
|
|
|
|
| |
The option value was missing a .host, which caused the evaluation to
fail because of a missing attribute.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This makes sense to put into the core module here, because we already
have the public key in there as well and if we want to use the binary
cache outside of a channel it's a good idea to provide it here by
default.
Being a default of course also means that users can easily override this
using:
nix.binaryCaches = mkForce [ "something else" ];
I've also removed the redundant setting of that option in all of the
machine configurations.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
Cc: @sternenseemann
|
|
|
|
|
|
| |
Added in NixOS/nixpkgs@dc87b5587e3fb39e96e6b965a26f324daf7508f8.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
New tests as of NixOS/nixpkgs@b0179351036a870da8c1a89c991a359a73aab0e2:
* containers-restart_networking
* nginx
* pgjwt
* postgresql (now with subtests per package)
* sddm (now with subtest "autoLogin" and "default")
* wordpress
The following tests were renamed:
* kde5 -> plasma5
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
| |
Fixes evaluation.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This fixes the attributes accordingly to the following two upstream
commits:
* NixOS/nixpkgs@f21d4d00152726c02f870cb20faba4aca18ef49a
* NixOS/nixpkgs@8eb4d2afbc1d61d316073f25f5885a46ccb37ea2
The first renames desktopManager.kde5 to desktopManager.plasma5 and the
second one removes the kde5 attribute from the top-level packages set
and puts applications in kdeApplications instead.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
|
|
|
|
|
| |
Haha, so that's what I got from introducing callSubTests in
NixOS/nixpkgs@b3337edd0b77f04d0abe4a29642a5e4fdb4ad692.
While in general this is an improvement, it makes it harder for vuizvui
to gather tests that were added or removed.
Maybe we should write an updater for this soon.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
New tests as of NixOS/nixpkgs@b207bf523b8c49b037a4a87456d61f63043cd2c6:
* containers-extra_veth
* containers-hosts
* containers-macvlans
* containers-physical_interfaces
* containers-tmpfs
* glance
* hound
* keystone
* leaps
* mongodb
* nat.firewall-conntrack
* pam-oath-login
* samba
* sddm
The following tests were removed:
* kde4
* panamax
I've also added a few checks on whether containers are actually enabled,
because some tests like containers-bridge would run if containers with a
bridge were defined but enableContainers was false.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This allows to add packages to vuizvui.lazyPackages which then aren't
directly installed onto the system but instead built by the Hydra and
only fetched from it as soon as a binary of one of these packages is
executed.
Doing this only within a NixOS module however isn't enough, because by
default gc-keep-outputs is false, so a garbage collect on the Hydra
instance would remove the packages we wrap in vuizvui.lazyPackages.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|