This introduces VirtualBox version 5.1.6 along with a few refactored
stuff, notably:
* Kernel modules and user space applications are now separate
derivations.
* If config.pulseaudio doesn't exist in nixpkgs config, the default is
now to build with PulseAudio modules.
* A new updater to keep VirtualBox up to date.
All subtests in nixos/tests/virtualbox.nix succeed on my machine and
VirtualBox was reported to be working by @DamienCassou (although with
unrelated audio problems for another fix/branch) and @calbrecht.
If people want to disable support for PulseAudio they can still
explicitly use pulseaudio = false in their nixpkgs config.
But even with enabled PulseAudio support, it's still optional, enabled
at runtime and can be turned off in VirtualBox settings as well.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
Issue: #15005
Putting the kernel modules into the same output path as the main
VirtualBox derivation causes all of VirtualBox to be rebuilt on every
single kernel update.
The build process of VirtualBox already outputs the kernel module source
along with the generated files for the configuration of the main
VirtualBox package. We put this into a different output called "modsrc"
which we re-use from linuxPackages.virtualbox, which is now only
containing the resulting kernel modules without the main user space
implementation.
This not only has the advantage of decluttering the Nix expression for
the user space portions but also gets rid of the need to nuke references
and the need to patch out "depmod -a".
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
The opam package manager relies on external solvers to determine package
management decisions it makes related to upgrades, new installations,
etc.
While, strictly speaking, an external solver is optional, aspcud is
highly recommended in documentation. Furthermore, even having a
relatively small number of packages installed quickly causes the limits
of the interal solver to be reached (before it times out).
Aspcud itself depends on two programs from the same suite: gringo, and
clasp.
On Darwin, Boost 1.55 (and thus Gringo) do not build, so we only support
Aspcud on non-Darwin platforms.
1) add pcre dependency (for some reason builtin_pcre doesn't work)
2) Disable dependencies that are currently not supported by the
expression. Most users should not need those. These are disabled to
prevent cmake from picking them up from system and causing impurities.
Once there is a user who needs these they will have to update the
expression.
3) disable some OSX detection code that relies on /usr/bin/sw_vers
that chooses c++ library, silences warnings and sets macosx-version-min.
macosx-version-min is already set by nix using MACOSX_DEPLOYMENT_TARGET
environment variable.
The Cryptol REPL has a hard dependency on Z3, but the rest of the
library uses SBV to support multiple solvers. Ensure that Z3 is
available for `pkgs.cryptol`, which is likely to be installed via
nix-env for REPL usage, but do not change pkgs.haskellPackages.cryptol,
which is likely to be used as a dependency (in Nix expressions).
It's not clear to me what this is achieving, plus for some reason this
is causing an evaluation error in hyperterm. So let's hope it's not
really needed...
So far we don't yet need the Qt 5 build for qtkeychain because the two
packages that depend on it are still using Qt 4. However, the next
upstream version of Tomahawk for example already uses Qt 5, so let's
prepare for that.
Tested building against Tomahawk Git master with qt5.qtkeychain.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
A few additional changes:
1. We load patches from debian for improving privacy and security
2. Now with QT 5.6 instead of 5.5
3. We strip bundled python code and use proper upstream instead
This moves libsystemd.so and libudev.so into systemd.lib, and gets rid
of libudev (which just contained a copy of libudev.so and the udev
headers). It thus reduces the closure size of all packages that
(indirectly) depend on libsystemd, of which there are quite a few (for
instance, PulseAudio and dbus). For example, it reduces the closure of
Blender from 430.8 to 400.8 MiB.
This builds elisp to setup an emacs buffer with the packages given
available. See shlevy/nix-buffer for more information.
Currently only modifies $PATH.
gnomepanel was part of Gnome 2 and is currently broken.
There seemed to be no runtime dependency to gnomepanel and building also
seems to work fine without it.
It appears that packageOverrides no longer overrides aliases, so
aliases like
dbus_tools = self.dbus.out;
dbus_daemon = self.dbus.daemon;
now use the old, non-overriden version of dbus. That seems like a
pretty serious regression in general, but for this particular problem,
I've fixed it by replacing dbus_daemon by dbus.daemon and dbus_tools
by dbus.
This removes locales, bash completion and crap like that. This cuts
6.5 MiB from the NixOS system closure (which unfortunately contains
two copies of util-linux, because of the need to break a dependency
cycle with systemd).
The -rc kernels are quite likely to break out-of-tree modules and thus
cause unnecessary Hydra failures.
(Note that linux_testing already has `hydraPlatforms = [];` but that
does not prevent the package from being built since it has reverse
dependencies. Arguably that could be considered undesirable and thus
fixing that could be considered the proper fix, but this should do
for now.)
After splitting the DejaVuSans.ttf file into a multiple output in the
dejavu_fonts Nixpkgs expression it is not possible to install in the
user profile due to the collision. The attached patch makes a new
package without the collision for user environment installing.
From fae78903c6ce56eda70a1a9a6914c41d248b15e8 Mon Sep 17 00:00:00 2001
From: Karn Kallio <kkallio@skami.org>
Date: Sat, 3 Sep 2016 14:09:36 -0400
Subject: [PATCH] dejavu-fonts : Prepare an environment package without
collision.
The xapianBindings version is now directly tied to the input xapian
version, to ensure that they're in synch.
Also, as of this version, sphinx is required to build documentation
for the python bindings.
After making multiple outputs in the mesa_glu package the headers are
not included in the mesa attribute. The attached patch puts them in it.
From ced24208a300bea8234e7898ae6fec34fbd67289 Mon Sep 17 00:00:00 2001
From: Karn Kallio <kkallio@skami.org>
Date: Thu, 1 Sep 2016 16:18:23 -0400
Subject: [PATCH] mesa: Add the mesa glu headers to the mesa attribute.
We now have a newer version and the older version didn't work anymore
anyway because it depended on sqlalchemy7 which was itself broken,
because it depended on an older version of sqlite.
- Add support for python bindings
- make neuron respect standard pythonpath prefix
- force exec_prefix == prefix to respect standard nix file hierarchy
- normalize indentation
- propagate dependencies necessary for nrniv_makefile usage
- Add support for darwin
This was one of the ways to build packages, we are trying
hard to minimize different ways so it's easier for newcomers
to learn only one way.
This also:
- removes texLive (old), fixes#14807
- removed upstream-updater, if that code is still used it should be in
separate repo
- changes a few packages like gitit/mit-scheme to use new texlive
With the default kernel and thus with the build I have tested in
74ec94bfa2, we get an error during
modules_install:
make[2]: execvp: /nix/store/.../bin/bash: Argument list too long
I haven't noticed this build until I actually tried booting using this
kernel because make didn't fail here.
The reason this happens within Nix and probably didn't yet surface in
other distros is that programs only have a limited amount of memory
available for storing the environment and the arguments.
Environment variables however are quite common on Nix and thus we
stumble on problems like this way earlier - in this case Linux 4.8 - but
I have noticed this in 4.7-next as well already.
The fix is far from perfect and suffers performance overhead because we
now run grep for every *.mod file instead of passing all *.mod files
into one single invocation of grep.
But comparing the performance overhead (around 1s on my machine) with
the overall build time of the kernel I think the overhead really is
neglicible.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
The darwin stdenv currently provides clang-3.7; however,
a) go-1.7 currently expects a compiler that supports "-fdebug-prefix-map"
arguments (which clang-3.8 supports but clang-3.7 does not) and
b) even with clang-3.8, go-1.7 misinterprets the result of its runtime probes
for support for the "-no-pie" flag, thereby failing to build runtime/cgo.
This patch resolves (a) by building go-1.7 with clang-3.8 and resolves (b) by
teaching go how to correctly probe "-no-pie" error messages from clang.
Fixes#17755.
As noted in the issue tracker, we have been carrying two versions of the
keybase client: the old node.js client and the new Go client. The old
client is unusable, but takes up the valuable `keybase` name. This
revision removes the old client entirely and replaces it with the new one.
