| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
| |
Silly me for not noticing that I was using the pkgs version rather than
the module version. Thanks to @aszlig for pointing out my idiocy. :)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Version 5.0.0 of Radare2 introduced a few changes in symbol naming,
which causes our patcher to fail because it's unable to find the
"/usr/share/factorio" string and the getSystemWriteData() method.
The upstream change[1] in question for example now uses underscores for
certain characters in the symbol/comment name, so for us to be
backwards-compatible I added a check against the major version of
Radare2 to determine whether we should add underscores when needed.
[1]: https://github.com/radareorg/radare2/commit/aaa930ab261a31e58b1257df06db02481cd3bd55
Signed-off-by: aszlig <aszlig@nix.build>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The following nix functions allow easily creating derivations for
building a signed releases directory for project(s) to be served via
e. g. HTTP.
* buildGitTarball: builds a reproducible .tar.gz for a given git
revision or tag (similar to git archive, but we don't actually
reuse it in favor of fetchgit).
* bundleSignedReleases: symlinks tarballs generated using
buildGitTarball and accompanying (manually provided) signatures into a
directory and verifies the signatures to ensure buildGitTarball is
donig what it's supposed to.
|
|
|
|
|
|
|
|
|
| |
Only does very simple dispatching to pkgs.lowdown and pkgs.chroma, but
is at least significantly faster than the default source and about
filters bundled with cgit (which is not really a challenge as they use
python and pygments).
Added to enable my cgit setup until we can have TVL's //tools/cheddar.
|
| |
|
|
|
|
|
|
|
| |
Mistyped with my evaluation fix and added lib.optional instead of
lib.optionals, so a list got into systemPackages which breaks evaluation
in a hard to debug way (and also made me discover a hillarious segfault
in nix).
|
|
|
|
|
|
|
|
| |
Starting with Kernel 5.6 adding this package is no longer necessary.
Since the kernelPackages.wireguard attribute returns `null` for that
version, evaluation fails.
cc @Profpatsch
|
|
|
|
| |
Ugh.
|
|
|
|
| |
Group somewhat semantically instead of ordering alphabetically.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The kernel-lts and kernel-latest tests seem to have been removed (not
sure when exactly, didn't blame). Instead we can use
nixosTests.kernel-generic."linux_${major}_${minor}". This allows us to
easily check for tests for arbitrary kernel versions as well.
However there is currently a bug that these are returned as functions
instead of derivations at the moment which will hopefully be fixed by my
upstream PR soon. Until that happens we check if the test is a
derivation or not with lib.isDerivation.
cc @aszlig for comments / further improvements. Main intention right now
is to fix eval in a future proof way.
|
|
|
|
|
| |
The normal scribus attribute doesn't evaluate on nixos-unstable
currently (2.7 pillow is marked as insecure).
|
|
|
|
|
| |
This is currently mostly to add the dot character used as separator in
dot time (<https://dotti.me>).
|
|
|
|
|
|
|
| |
This seems very neat, and after a brief test I'm going to see if I can
switch to it permanently.
This is also the beginning of my journey to move to sway eventually.
|
| |
|
|
|
|
|
| |
This patch version implements Ingo's suggestion of READ_ALLOWED_PATH:
https://inbox.vuxu.org/mandoc-tech/20210331173434.GA57338@athene.usta.de/
|
|
|
|
|
|
|
| |
whitby.tvl.fyi has virtualhosts configured with the names "b" and "cl"
which redirect to b.tvl.fyi and cl.tvl.fyi respectively. For this to
work in e. g. your browser as well, we need to add entries to
networking.hosts. See also http://cl/2908.
|
|
|
|
|
| |
This should make headcounter.org/hydra and cache.nixos.org take
precedence over cache.tvl.su which is needed less often.
|
|
|
|
|
|
|
|
| |
Finally, after a lot of hussle with the IT department they've come to
the conclusion that, in fact, their eduroam cert was not generated
properly thereby preventing me from using iwd. It looks like this only
affected a minority of devices (presumably those with proper
implementation).
|
|
|
|
|
|
|
|
|
|
|
|
| |
This doesn't do much other than enabling cache.tvl.su as a binary cache
currently, but we can additional settings in the future which are neat
for working with tvl's depot.
I contemplated adding an option to add <depot> to the nixPath, but I
don't want it too desparately right now and it is kind of annoying to
implement for vuizvui, as the default core/common.nix module overrides
any values you set and it is hard to merge the nixPath values, so we'll
probably need an vuizvui.extraNixPath option etc. pp.
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
| |
i3-auto-layout is not yet in nixos-unstable unfortunately, so it
triggers an eval error on hydra. Temporarily hack around it by only
including it in extraPackages if pkgs contains the package in question
(i. e. if rebuilding from master locally).
cc @devhell
|
|
|
|
|
| |
Add i3 extraPackages `i3-auto-layout` and `rofi`. This is much saner
than shoving everything into systemPackages.
|
|
|
|
| |
According to NixOS' wiki this should be good for something.
|
| |
|
|
|
|
| |
It was unused and required a broken hnix pin.
|
|
|
|
|
| |
Luckily most people seem to be using Signal nowadays, so I can get rid
of this crap.
|
| |
|
|
|
|
| |
In case pipewire wants it or I want to use the jack shim.
|
|
|
|
|
| |
blueman is annoying me a bit atm and bluetoothctl plus pactl seem to be
enough mostly.
|
| |
|
| |
|
|
|
|
|
| |
Once again PVP causes a configure failure which doesn't translate into a
build failure. Fixed by bumping upstream repository.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
We had a bunch of instances of
https://github.com/NixOS/nix/issues/2176,
where nix would exit with a “killed by signal 9” error.
According to Eelco in that issue, this is perfectly normal behaviour
of course, and appears if the last command in a loop closes `stdout`
or `stdin`, then the builder will SIGKILL it immediately. This is of
course also a perfectly fine error message for that case.
It turns out that mainly GNU coreutils exhibit this behaviour …
Let’s see if using a more sane tool suite fixes that.
|
|
|
|
|
|
|
|
|
|
| |
Port of <https://cl.tvl.fyi/c/depot/+/2657>. Original message:
`forstdin` iterates over the tests in the test directory, and by
default it does *not* fail if an inner loop returns an error, unless
`-o okcodes` is given, a list of exit codes that indicate success.
Now it fails if a loop returns ≠ 0.
|
|
|
|
|
| |
Since we pass the config, the home directory config isn’t picked up
anymore, so I need to pass it via `extraConfig`.
|
|
|
|
| |
Let's keep it simple, everything else seems to just end in tears.
|
|
|
|
|
|
| |
Copied whatever Profpatsch is using. 5% seems a little extreme, but with
treshhold 2% tail /dev/zero still froze my laptop for a solid minute and
a half.
|
|
|
|
| |
I actually use this now.
|
| |
|
|
|
|
| |
Hearing a lot of good things lately, time to see what it is about.
|
|
|
|
| |
Sigh.
|
|
|
|
|
| |
So the previous idea didn't work as expected, instead I'm going to use
the directory name Syncthing would use anyway.
|
|
|
|
| |
This reverts commit d46aa4ec882dcb9ac0e09da6e1aecb361d90d287.
|
|
|
|
|
| |
This doesn't seem to work properly, Syncthing will still add its default
`Sync` directory.
|
|
|
|
| |
https://youtu.be/LB8bhHyPLgk?t=418
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
nix-env-diff is a tiny utility which parses and compares the output of
`nix-env -qaP --out-path`, printing all changed and added out paths (or
attributes if desired). This facilitates a simple way to determine
rebuilds or changed reverse dependencies when working on nixpkgs in a
similar way as nixpkgs-review and ofborg do it.
Both the new evaluation and the base evaluation to compare against have
to be created manually using nix-env, which in turn also allows
considering attribute sets that are normally not evaluated do to missing
`lib.recurseIntoAttrs`. As an example, here is an example building all
changed attributes in `ocaml-ng.ocamlPackages_4_12`:
```
nix-env -qaP -A ocaml-ng.ocamlPackages_4_12 --out-path -f . \
| nix-env-diff --attrs ./base-ocamlPackages_4_12 \
| xargs -n 1 -P 4 nix-instantiate --quiet -A \
| xargs nix-store --realise --builders 'ssh://edwin'
```
`./base-ocamlPackages_4_12` contains the result of the `nix-env`
invocation executed on the master branch.
|
|
|
|
|
|
|
|
|
| |
Shared by adisbladis in #nixos-chat this is a very stupid fish function
which bonks a nixpkgs package into the PATH in the current shell. This
is often useful if you need a package ad-hoc and opening a nix-shell is
too much hassle or you find yourself perpetually opening and closing a
nix-shell. Extremely racy with the garbage collector, but this is yolo
ops territory anyways.
|
| |
|