| Message ID | CAPfwWagUzJd_v+n2JR=P-yhutJ3VudHeJ+mLeXgKJGaCv3prcw@mail.gmail.com |
|---|---|
| State | New |
| Headers | show |
| Series | [bug#73135] Enable parallel tests for GTK | expand |
Hi, The Man <squishypinkelephant@gmail.com> skribis: > In the middle of testing a set of patches that caused a large number of > rebuilds, I noticed the gtk package was taking an obscenely long time to > run checks. Upon inspection I found out it was not running tests as > parallel. > For example, https://ci.guix.gnu.org/build/5623913/details took >3000! > seconds. > The fix is obvious and will save time on future rebuilds. > diff --git a/gnu/packages/gtk.scm b/gnu/packages/gtk.scm > index 3d4766713d..f0f6a0cd9a 100644 > --- a/gnu/packages/gtk.scm > +++ b/gnu/packages/gtk.scm > @@ -1195,6 +1195,7 @@ (define-public gtk > (string=? (%current-system) "aarch64-linux")) > "--no-suite=wayland_gles2_failing" > "--no-suite=wayland_gles_failing")) > + #:parallel-tests? #t I wonder why #:parallel-tests? defaults to #f in ‘meson-build-system’, contrary to what other build systems do. I think we should create a branch where we change it to #t, and merge it once we’re done rebuilding all the packages and fixing any relevant failure. Would you like to give that a try? Thanks, Ludo’.
Sounds like a good idea to me. On Fri, Oct 4, 2024, 10:14 AM Ludovic Courtès <ludo@gnu.org> wrote: > Hi, > > The Man <squishypinkelephant@gmail.com> skribis: > > > In the middle of testing a set of patches that caused a large number of > > rebuilds, I noticed the gtk package was taking an obscenely long time to > > run checks. Upon inspection I found out it was not running tests as > > parallel. > > For example, https://ci.guix.gnu.org/build/5623913/details took >3000! > > seconds. > > The fix is obvious and will save time on future rebuilds. > > diff --git a/gnu/packages/gtk.scm b/gnu/packages/gtk.scm > > index 3d4766713d..f0f6a0cd9a 100644 > > --- a/gnu/packages/gtk.scm > > +++ b/gnu/packages/gtk.scm > > @@ -1195,6 +1195,7 @@ (define-public gtk > > (string=? (%current-system) > "aarch64-linux")) > > "--no-suite=wayland_gles2_failing" > > "--no-suite=wayland_gles_failing")) > > + #:parallel-tests? #t > > I wonder why #:parallel-tests? defaults to #f in ‘meson-build-system’, > contrary to what other build systems do. > > I think we should create a branch where we change it to #t, and merge it > once we’re done rebuilding all the packages and fixing any relevant > failure. > > Would you like to give that a try? > > Thanks, > Ludo’. >
Hi,
The Man <squishypinkelephant@gmail.com> writes:
> Sounds like a good idea to me.
Implemented! See: https://ci.guix.gnu.org/jobset/meson-parallel-tests
It hasn't started building anything yet (it's buildin guix at that
revision for now).
There's an RSS feed to keep track of it:
https://ci.guix.gnu.org/events/rss/?specification=meson-parallel-tests
Hello, Maxim Cournoyer <maxim.cournoyer@gmail.com> writes: > Hi, > > The Man <squishypinkelephant@gmail.com> writes: > >> Sounds like a good idea to me. > > Implemented! See: https://ci.guix.gnu.org/jobset/meson-parallel-tests > > It hasn't started building anything yet (it's buildin guix at that > revision for now). > > There's an RSS feed to keep track of it: > https://ci.guix.gnu.org/events/rss/?specification=meson-parallel-tests No issues whatsoever found: https://ci.guix.gnu.org/jobset/meson-parallel-tests Will prepare the branch merge request and close this one.
diff --git a/gnu/packages/gtk.scm b/gnu/packages/gtk.scm index 3d4766713d..f0f6a0cd9a 100644 --- a/gnu/packages/gtk.scm +++ b/gnu/packages/gtk.scm @@ -1195,6 +1195,7 @@ (define-public gtk (string=? (%current-system) "aarch64-linux")) "--no-suite=wayland_gles2_failing" "--no-suite=wayland_gles_failing")) + #:parallel-tests? #t #:phases #~(modify-phases %standard-phases (add-after 'unpack 'generate-gdk-pixbuf-loaders-cache-file