Starting with 3.10, #! script handling can be built modularly (or not
at all). By default the nixpkgs builder sets everything modular, but
since our initird init is a #! script this creates a chicken-and-egg
problem on NixOS.
Signed-off-by: Shea Levy <shea@shealevy.com>
This introduces the ability to mount filesystems (only). Also, the description
is now less kickstart specific as in the long term we want to move away from
kickstart syntax to Nix attribute sets.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
Changes since 0.16-1:
- Used Python type instead of variable name (hamzy)
- Fix detection of valid EFI system partition during autopart. (dlehman)
Full changelog can be found in the spec file in the package or at:
https://git.fedorahosted.org/cgit/blivet.git/tree/python-blivet.spec
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
demjson is a Python JSON module that reads, writes and validates
JSON-encoded data; compliant with RFC 4627.
It also includes a lint checker, jsonlint, which can be used to validate
JSON documents for strict conformance to the RFC specification; as well
as to reformat them, either by re-indenting or for minimal/canonical
JSON output.
Homepage: http://deron.meranda.us/python/demjson/
I'm sure that sqlite works on more platforms than linux, but I cannot
verify that myself (I only have linux). So I'll leave it up to the users
of other platforms to expand this as needed.
This version is preliminary because it quite heavily depends on pykickstart
(through blivet) and the roadmap is to have a nice NixOS attrset-based
specification of partitions.
Currently the main purpose for this is in preparation for the Hetzner nixops
backend, but we might want to make this part of the standard NixOS installer.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
Kinda icky to not have archives available here, but I got an error during VM
tests because of an sha256sum mismatch, hence the update. Maybe Hydra has cached
this?
And yes, I checked twice, it wasn't a broken download - there really *is* a new
upstream version available.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
This adds the correct store paths for mount, umount and lsof to blivet as these
commands are still generic enough to _not_ add them as a dependency.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
First of all, the path to wipefs didn't work at all, and though it is
documented, the "-f" flag only works when used as a long option ("--force").
This is probably fixed upstream in util-linux, but using the long-option will
stay compatible.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
by default (as it was before), kde depends on a libcanberra with its
gtk dependency removed so a kde system can be gtk-free.
However, both firefox and chromium depend on gtk+ and libcanberra too.
So this leads to 2 versions being installed. this option combats that
by letting the kde use the other libcanberra
Otherwise the user won't be able to override anything specified in
NixOS' /etc/fonts/conf.d, since ~/.config/fontconfig is included from
upstream fonts/conf.5/50-user.conf.
This should ensure the test casnes are always running in the same order,
regardless of the target machine. We're just using the class name here, which
should be sufficient enough to address the issue.
Now the following build should be fixed:
http://hydra.nixos.org/build/5425811
Big thanks to @rbvermaa for being a *really* great help debugging the problem
quickly, because I couldn't reproduce it here (the run order of the test cases
on my machine were 'accidentally' right).
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
This still doesn't have all possibly required dependencies, but at least we can
do basic partitioning, which is our primary goal. At least when it comes to the
Hetzner target of nixops. But even for partitioning when installing NixOS this
library could be _very_ useful.
Test cases currently don't work because they're filled with syntax errors and
some references to the mocking library are missing.
As you can see in propagatedBuildInputs, here are the promised overrides for
Python support of libselinux and cryptsetup.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
This again is only optional to avoid too much dependencies when bootstrapping
small systems or when constrained to RAM disks of lower size. It is needed for
blivet as well, which will override the option in its dependency list.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
Thought this would be needed for blivet, but it wasn't the case. They seem to
have their own mini-implementation. But it might be useful for other Nixers, who
knows?
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
The reason this is optional is because we might want to use it for bootstrapping
in some constellations. And we really don't want whole lot of dependencies in
those situations.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
Needed for blivet and this is part of Anaconda (Fedora's installation system).
The reason I'm packaging this is because of blivet and because it's quite well
decoupled from Anaconda itself, so it can be used for other purposes.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
Wow, this is one of the most dangerous programs I've seen in a while. It not
only tries to probe for a package manager to install dependencies but also
tries to execute a whole bunch of programs in $PATH. That's why I decided to
override the postFixup phase for now in order to get rid of the current $PATH
and meanwhile getting the basics working.
So, I'm still not sure how to do the best implementation here on NixOS without
allowing winswitch to be too invasive and without restricting it too much so
that it's of no use.
But let's figure that out once we trimmed down the radiation level of this
"living" thing ;-)
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
A recent X update broke VirtualBox guest additions (vboxvideo driver version
mismatch, desktop won't start). This fixes it.
Here is the error log:
(II) "glx" will be loaded by default.
(II) LoadModule: "glx"
(II) Loading /nix/store/kzvmnjlps51q4piqmwr7zbmxcg2z9vgk-xorg-server-1.13.4/lib/xorg/modules/extensions/libglx.so
(II) Module glx: vendor="X.Org Foundation"
compiled for 1.13.4, module version = 1.0.0
ABI class: X.Org Server Extension, version 7.0
(==) AIGLX enabled
Loading extension GLX
(II) LoadModule: "vboxvideo"
(II) Loading /nix/store/4kbxi00h8xsmfgbws2qqh674lcfp03h6-VirtualBox-GuestAdditions-4.2.14-3.2.46/lib/xorg/modules/drivers/vboxvideo_drv.so
(II) Module vboxvideo: vendor="Oracle Corporation"
compiled for 10.12.0, module version = 1.0.1
Module class: X.Org Video Driver
ABI class: X.Org Video Driver, version 12.0
(EE) module ABI major version (12) doesn't match the server's version (13)
(II) UnloadModule: "vboxvideo"
(II) Unloading vboxvideo
(EE) Failed to load module "vboxvideo" (module requirement mismatch, 0)
(II) LoadModule: "vboxmouse"
(WW) Warning, couldn't open module vboxmouse
(II) UnloadModule: "vboxmouse"
(II) Unloading vboxmouse
(EE) Failed to load module "vboxmouse" (module does not exist, 0)
(EE) No drivers available.
Fatal server error:
no screens found
The second failure, and the last one I'm going to try today:
http://hydra.nixos.org/build/5404634
On the bright side there is at least the fact that version 1.4.10 has failed on
Darwin already, so I guess we don't have a lot of Mac users using Synergy.
Latest (failed) build of 1.4.10:
http://hydra.nixos.org/build/5359408
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
Seems that crypto++ in nixpkgs doesn't build on Darwin, so let's use bundled
crypto++ until the version in nixpkgs works well.
This refers to the following build:
http://hydra.nixos.org/build/5404516
Hopefully, this will fix it on Mac OS X, because I don't have a Darwin machine
for testing.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
I'm heavily using synergy for daily work, so I'm most probably going to watch
out for changes/improvements/bugs :-)
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
This uses recurseForDerivations directly after using callPackage magic to ensure
that the input attributes can be overriden *and* nix-env shows the package as in
recurseIntoAttrs.
The reason for making this optional is because there probably is only a minority
of people who want to use XMPP and we don't want to introduce an additional
dependency for the majority, do we?
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
Chromium 28.0.1500.52 finally is stable, so the release channels are now:
stable: 28.0.1500.52 (builds fine, tested)
beta: 28.0.1500.52 (same as stable)
dev: 29.0.1541.2 (patch rebased, builds fine, tested)
The user namespace patch doesn't apply for version 29, so I had to rebase it
against the current trunk (revision 207742).
And as version 27 is outdated, we no longer need to distinguish versions for
patching the hardcoded gcc path in core/core.gypi.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
Version 1.2.0 is way too old in order to build the latest chromium (29) version,
so let's get it up to date (especially because no other package is referencing
ninja, so it should be non-critical).
The dependency on unzip is not needed here, because GitHub also provides
archives in tar.gz format.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
I'm still wondering why noone has reported this, but I found out about this
while trying to introduce someone to NixOS, eventually wondering why it is going
to install version 29 when using "nix-env -i chromium".
So, in hope that everyone out there using the package is using the attribute,
let's make _stable_ the default here.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
This fixes a copy & paste error I made in 486e918, which resulted in PHP being
built with the bundled version of GD instead of the one we have in nixpkgs.
Thanks to @peti for noticing.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
Integration tests don't seem to work right now, so let's see if we can figure
out a way to enable them later. But at least running unit tests is better than
not running any tests :-)
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
Nowadays, multiple monitor setups are quite common, so I suppose we'd want
support for that. Especially because users might get confused if synergy is
unable to pick the right screen resolution and thus cause edges to be cut off
from the available pointing area.
The postPatch hook is to force cmake into thinking that we have XRRNotifyEvent,
which we _do_ have with the xrandr version shipped in nixpkgs. Automatic
detection from CMakeLists.txt fails here because it tries to search for the
symbol within the libX11 store path.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
This brings in support for encryption and thus requires the crypto++ library as
an additional dependency. Unfortunately the upstream integration isn't quite the
way we'd like it to be, so we need to add a small patch to ignore the bundled
version and use the package from nixpkgs.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
