we use the GCC wrapper in a user environment, the wrong assembler
will be called. This is not usually a problem, but sometimes it is
(e.g., when using G++ 4.1.1 with binutils 2.16 or so).
svn path=/nixpkgs/trunk/; revision=6862
Glibc. This is useful when building GCC.
* gcc-wrapper: the dynamic linker has a different name on x86_64 and
powerpc.
* gcc-wrapper: "glibc" -> "libc", because someday we might support
different C libraries.
* gcc: don't do a multilib build (e.g., 32-bit support on x86_64),
don't need it.
* gcc: merge in support for static builds.
* gcc: various simplifications in the compiler/linker flags, hope they
work.
svn path=/nixpkgs/trunk/; revision=6823
contains arbitrary information about a package, like this:
meta = {
homepage = "http://gcc.gnu.org/";
license = "GPL/LGPL";
description = "GNU Compiler Collection, 4.0.x";
};
The "meta" attribute is not passed to the actual derivation
operation, so it's not a dependency --- changes to "meta" attributes
don't trigger a recompilation.
Now we have to standardise some useful attributes ;-)
svn path=/nixpkgs/branches/usability/; revision=5024
Even though you could just inherit binutils it is conceptually cleaner (I
think) to make these tools available here.
svn path=/nixpkgs/trunk/; revision=4967
* added an experimental fetchdarcs function, based on fetchsvn
(there are no expressions yet that use this function)
svn path=/nixpkgs/trunk/; revision=4615
in /usr/lib/libGL.so.1 (which will typically load a driver in
/usr/X11R6/lib/modules/dri). This has been tested on a i915
graphics card; it should work with most open source X.org drivers.
For NVidia's proprietary drivers (which we cannot build ourselves
anyway), some more symlinks are necessary; I'll add those later.
So to get hardware-accelerated Quake 3, do:
$ nix-env -p /nix/var/nix/profiles/opengl -i xorg-sys-opengl
$ nix-env -i quake3-demo
$ quake3
svn path=/nixpkgs/trunk/; revision=4613
allow the OpenGL implementation to be overriden through the
OPENGL_DRIVER environment variable. If it is not set, we use the
implementation installed in the profile
/nix/var/nix/profiles/opengl, allowing easy late binding by the
user, e.g.,
$ nix-env -p /nix/var/nix/profiles/opengl -i nvidia-sys-opengl
might install the NVidia OpenGL implementation.
The code that does this is not specific to Quake 3: it has been
factored out into build-support/opengl/mesa-switch.sh. Presumably
any application that requires hardware-accelerated OpenGL needs it.
* Add the Quake 3 demo to the cache.
svn path=/nixpkgs/trunk/; revision=4612
server's certificate. This is perfectly safe: we don't care whether
the server is being spoofed --- only the cryptographic hash of the
output matters.
svn path=/nixpkgs/trunk/; revision=4377
right thing. This is necessary to make libtool detect g++
properly. (Fixes the `libtool: compile: unable to infer tagged
configuration' error on FreeBSD when building Berkeley DB 4.4.)
svn path=/nixpkgs/trunk/; revision=4370
* Use the system Perl on all non-i686-linux platforms.
* Don't build Python support in libxml2 on most platforms.
svn path=/nixpkgs/trunk/; revision=3019
`*.dll.config' files corresponding to CLR assemblies. I.e., the
full path to native libraries is included in the maps. In effect
this allows us to set the equivalent of an RPATH for assemblies.
* gtk-sharp: use the DLL fixer. It's not perfect yet: I still have to
set the LD_LIBRARY_PATH for monodoc to include the gtk-sharp lib
directory itself, so that it can find the `*sharpglue.so' files.
This seems to be gtk-sharp's fault; it doesn't have an entry for
those libraries in its DLL maps.
svn path=/nixpkgs/trunk/; revision=2330
