Conversation
b3e6619 to
c447c0c
Compare
c447c0c to
03f8258
Compare
03f8258 to
57c4449
Compare
|
Do you want to get a Hydra jobset for this? |
|
@vcunat could we get a Hydra jobset? |
49adb92 to
325b485
Compare
|
The GNOME 45 PR now also have 2141f32 (sway) b3575ab (i3lock) 5b579d0 (hyprland) a7c25f4 (i3lock-color) 610aabe (wayfire) e1e8282 (celluloid) b871990 (wxSVG), and I-don't-know-why 9f5f74b (cogl) Update: also initial fixup for weston, string-machine, stone-phaser, mission-center, hyprpicker, hyprpaper, fbida in a3338b0...aa6f1d8 |
|
Jobset created: https://hydra.nixos.org/jobset/nixpkgs/pr-257301-cairo-1.18 I doubt that it will be possible to get significant amounts of builds for this PR done this week, with all-rebuilding rush starting on Wednesday due to security fixes. I guess we'll also wait with gnome jobset until this progresses? (it also contains the update and it doesn't make much sense to me for these jobsets to compete against each other for builders) |
|
I triggered it for x86_64-linux and aarch64-darwin for now. |
|
Builds finished, FYI. |
325b485 to
0db3542
Compare
|
Uh, the today's push... do we really need to rebuild 100k builds now on Hydra? I mean, there are lots of other stuff to build, e.g. other branches asking for a jobset incidentally. |
|
No, go ahead and cancel it. edit: I didn't realise that pushing would automatically trigger a Hydra build. |
|
I had one eval per day timer in there, but I triggered this one myself :-) I thought it was mainly the two extra commits shown by GitHub UI. |
|
So I suppose you'll soon consider this ready to merge, without another Hydra eval? There are only 10 days left to merge breaking changes that should reach NixOS 23.11. |
|
I'm still going through the first Hydra comparison, so we'll definitely need at least one more. All I've pushed so far is a rebase that was a prerequisite to applying @bobby285271's other changes. I won't push again until I've made as much progress as I can based on the current comparison. |
|
OK, I stopped the eval timer. Ping me when you want it. |
Currently, this is propagated from cairo, but that will soon no longer be the case.
Cairo will soon no longer propagate libGL. The DPF submodule needs this:
src/pugl/pugl.h:39:14: fatal error: GL/gl.h: No such file or directory
39 | # include <GL/gl.h>
| ^~~~~~~~~
compilation terminated.
(See also the fix for string-machine, which has the same build issue)
Cairo will soon no longer propagate libGL. The DPF submodule needs this:
src/pugl/pugl.h:39:14: fatal error: GL/gl.h: No such file or directory
39 | # include <GL/gl.h>
| ^~~~~~~~~
compilation terminated.
Currently, this is propagated from cairo, but that will soon no longer be the case.
Currently, this is propagated from cairo, but that will soon no longer be the case.
Cairo will soon no longer propagate libGL. Run-time dependency egl found: NO (tried pkgconfig and cmake)
Currently, this is propagated from cairo, but that will soon no longer be the case.
Currently, this is propagated from cairo, but that will soon no longer be the case.
- libiconv and expat are no longer used. - libintl and xcbutil no longer appear to be used. - The GL backend has been removed. - We can no longer disable PDF support, since something always propagates zlib. - Not sure whether the Requires.private is still needed, so I've left it.
0db3542 to
2c0a4f4
Compare
|
@bobby285271 I've fixed cogl in a different way, just linking libGL and avoiding the need for a patch. |
|
@vcunat I've now addressed all the direct new build failures from the last Hydra run, so I think it's time for another. |
There was a problem hiding this comment.
Been using cairo 1.18.0 as part of GNOME 45 branch: https://github.com/NixOS/nixpkgs/pull/247766/files#diff-c1b6675856a14491b8c610758d37ad4372e04a3f25a6049926c738f21121e5b0
Eval looks great with minimal breakage ( https://hydra.nixos.org/eval/1800745 )
|
After taking various things into consideration, let's finish this during the upcoming |
|
And we probably still need to fix cross there I guess (did not test, but I saw the |
|
Oh, whoops. |
|
I noticed that |
| , x11Support? !stdenv.isDarwin, libXext, libXrender | ||
| { lib, stdenv, fetchurl, fetchpatch, gtk-doc, meson, ninja, pkg-config, python3 | ||
| , docbook_xsl, fontconfig, freetype, libpng, pixman, zlib | ||
| , x11Support? !stdenv.isDarwin || true, libXext, libXrender |
There was a problem hiding this comment.
Good observation, my instinct thinks this is an unintended change and || true should be removed but tagging @NixOS/darwin-core for decision on disabling or enabling on darwin by default.
|
Mhhh, and this is a year old. Probably best to leave enabled then? (It has been part of the last 2 stable releases) |
|
I would prefer not to pull in X11 libraries unnecessarily, but what’s the effect on closure size? |
Hydra comparisons
Description of changes
Fixes #256519.
Things done
sandbox = trueset innix.conf? (See Nix manual)nix-shell -p nixpkgs-review --run "nixpkgs-review rev HEAD". Note: all changes have to be committed, also see nixpkgs-review usage./result/bin/)