Since commit e44038bccab0cae, cairo-1.0.typelib contains an absolute
path to cairo in the nix store so that no $LD_LIBRARY_PATH hacks are
needed. However, this did not yet work for lgi, because lgi does
dlopen("libcairo.so.2") without a full path, too.
To make this work, this commit ensures that lgi first uses
gobject-introspection to load libcairo. This uses the full path provided
by the typelib. Afterwards, dlopen("libcairo.so.2") does not hit the
filesystem anymore since the library is already loaded.
This commit adds a patch that reorders some code in lgi's cairo
initialisation. Previously, this started with core.module('cairo', 2),
which is where the dlopen happens. Now, this code is moved down and
instead core.gi.cairo.resolve is used to load the definitions of some
enums first. This part of the code goes through gobject-introspection
and causes libcairo to be loaded.
Signed-off-by: Uli Schlachter <psychon@znc.in>
Nixpkgs is a collection of packages for the Nix package manager. It is periodically built and tested by the Hydra build daemon as so-called channels. To get channel information via git, add nixpkgs-channels as a remote:
% git remote add channels git://github.com/NixOS/nixpkgs-channels.git
For stability and maximum binary package support, it is recommended to maintain
custom changes on top of one of the channels, e.g. nixos-17.09 for the latest
release and nixos-unstable for the latest successful build of master:
% git remote update channels
% git rebase channels/nixos-17.09
For pull-requests, please rebase onto nixpkgs master.
NixOS Linux distribution source code is located inside
nixos/ folder.
- NixOS installation instructions
- Documentation (Nix Expression Language chapter)
- Manual (How to write packages for Nix)
- Manual (NixOS)
- Community maintained wiki
- Continuous package builds for unstable/master
- Continuous package builds for 17.09 release
- Tests for unstable/master
- Tests for 17.09 release
Communication:
Description
Languages
Nix
96.3%
Shell
1.8%
Python
0.7%
Perl
0.4%
C
0.3%
Other
0.1%
