| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
| |
The problem here was that the attributes were added to the resulting
derivation of mkDerivation instead of the input attribute set.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
| |
Well, it won't work without it, duh!
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
|
|
|
|
| |
Right now, this is done by replacing all NIX_PATH references with
<nixpkgs...> by the path from Hydra input and silently hope that we
won't replace the wrong reference ;-)
Of course we're only replacing it in files that end with a .nix suffix,
so we shouldn't accidentally replace documentation.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
| |
Should make it easier to patch the channel expressions.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
| |
It was redundant anyway, see previous commit (e08f399).
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
|
| |
No longer necessary since NixOS/nixpkgs@a2a3508. Now overriding with
`systemd = null` should be enough to disable support for systemd (which
we don't want anyway).
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
|
| |
Yes, I know that it's only two more commits we don't really care about,
but I needed my Hydra to evaluatio the jobset again in order to test
channels.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
| |
In the latest not-yet-public implementation of Hydra channel
improvements, we now have a meta attribute to designate a channel.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
| |
This is because I've changed the the WIP implementation to use a product
type instead of a subtype.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
| |
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
|
|
|
|
| |
Those channels only depend on whether the build of the particular
machine succeed and currently don't include any VM tests.
Also, this is quite repetitious, because we're recursing through the
whole set of machines again instead of re-using the corresponding Hydra
jobs.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
| |
The third argument in hydra-build-products is actually the base
directory and not a name, Alzheimer's near!
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
| |
Regression introduced by a610560bb4b5a05eb480a4dbab3e24c553727047.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
|
| |
The channel now serves more or less as a generic channel, which has no
constituents and thus should always be the most up-to-date channel for
vuizvui.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
|
|
| |
In the end we want to have something like (import <vuizvui> {}), which
exports all stuff we need for building machines, packages, providing
custom helper functions and much more. And all that by using a channel
instead of setting NIX_PATH directly :-)
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
|
|
| |
We're going to create several channels and we don't want to code
duplicates across vuizvui. This essentially not only creates a channel
but also ties it to constituents, which make sure that channels are only
updated whenever all constituent builds are successful.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
|
| |
Mainly to make it a bit easier to distinguish, but also because we want
to make the release attribute set recursive, so we don't run into weird
evaluation errors.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
| |
Forgot to remove the slash from the end of the replacement of the
transformation expression in tar.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
| |
DUH! Finger fart, as I was writing a tar.bz2 in the first place and
later changed it to tar.xz.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
|
|
|
| |
With my previous implementation (NixOS/hydra#203) the tarball was
created by Hydra itself on the fly. Our new implementation actually
generates the tarball via Nix and Hydra only picks up the build product.
And exactly that's what we're doing here.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
| |
Fast-forward for 23 commits, which include fixes for sound, pci,
pm-sleep and nios2.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
| |
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
| |
Plus 30 commits more (akpm, drm-fixes, media-fixes).
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
| |
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|\
| |
| |
| |
| |
| |
| |
| |
| | |
Adds the openlab user to the dialout group to allow access to 3D
printers and other USB serial devices.
This might be just temporary and in the long term we might want to use a
udev rule after all to allow access to *all* USB devices to the openlab
user.
|
| |
| |
| | |
by default usb <-> rs232 interfaces which are used to connect to 3d printers are owned by the group `dialout`.
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
By default, there are 10 workspaces defined and since the module system
changes, we need to assign default priorities to all values instead of
the full attribute set, because whenever there is an override of the
default, the whole attrset is replaced by the override instead of just
the values.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|/
|
|
|
|
|
|
| |
It's mainly to test whether the workspace assignment is done correctly
and if not, the screenshot on the test will be showing one of the
default workspaces instead.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
| |
Patch is from pull request NixOS/nixops#280 and can hopefully be dropped
in favour of the latest master version.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
|
| |
We now no longer can (or even should) refer to pkgs.lib within modules
since the merge of the module system improvements in
NixOS/nixpkgs@889f72b.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
I wanted to use it for my countless XTerms, so I patched xterm+zsh to
use madvise() with MADV_MERGEABLE on every allocation and the results
weren't *that* bright:
pages_shared: 83
pages_sharing: 1498
pages_unshared: 942
pages_volatile: 1
With 20 terminals, so roughly 5.85 MB "saved" and ~9 for 120 instances.
That really isn't worth the CPU overhead, thus I'm turning it off again.
So, I need a better strategy to support my habit of running hundreds of
terminals (or better: shells, because that's really what's actually
eating up my RAM).
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
| |
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
| |
|
| |
|
| |
|
|
|
|
|
|
| |
Also, since NixOS/nixpkgs@6b467ad we no longer need to use allowBroken.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
| |
So far only emulateWheel has been set, but it doesn't work if you don't
also enable the module.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
| |
Its upstream package is called "printrun" and includes Pronterface.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
| |
Contains a bunch of fixes especially some related to the Hetzner
backend.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
|
| |
This reverts commit fc53a72f48ae6b567e7bf9901a2bd3657853633c.
Applied upstream at NixOS/nixpkgs@224ed7e.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
|
|
|
| |
They're just a few commits from 7fc377e..6c310bc and 24 changed files,
with 156 insertions and 93 deletions, so this really isn't necessary,
but I want to have that -rc6 instead of -rc5.
Actually, this isn't really -rc6 anyway, but -rc6 plus 9 more commits.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
|
| |
I can't stand it anymore when my excessive use of "lots of xterms" is
getting my machines into the need to use swapspace. The last time one of
my machines got to a crawl was with >300 open xterms.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
|
| |
This is very useful if you're running a lot of the same instances of
particular applications. For example virtual machines or even lots of
terminals.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
|
|
|
| |
This reverts the split up of the config from 55ec3a1 and puts all the
stuff from the hardware configuration file into the main config. The
reason for this is that vuizvui is structured so that the machine
definition *IS* the actual hardware configuration which deviates from a
common profile.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
| |
|
|
|
|
|
|
|
| |
Installing the GRUB bootloader to non-existing devices of mmrnmhrm to
for example dnyarri is not going to help anyone, right? ;-)
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
|
| |
Useful to call the machine from configuration.nix like this:
(import <vuizvui/machines> {}).aszlig.mmrnmhrm.use
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
|
| |
This is actually a store path consisting of the vuizvui sources
themselves and currently doesn't depend on anything else, so it usually
won't fail when mandatory tests or builds fail.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
|
|
|
|
|
| |
The generalization is done by symlinking include dirs and libraries to
the corresponding suffixed versions *only* if there isn't a ...5 suffix
already.
We now no longer need to pass QUAZIP_INCLUDE_DIR and QUAZIP_LIBRARIES as
cmakeFlags, because the Tomahawk cmake build is able to find it by
itself.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
| |
Same as with 9ea16e0 and d3bfedd.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|