What really puzzles me is the fact, that I didn't have the issue for my
intel build and even with adding back inherit gsettings to the tracker
recipe I'm unable to reproduce the issue for raspberrypi now.
Whatever, tracker doesn't install glib schemas so the commit should be
correct in any case
On Mon, Dec 11 2023 at 11:41:50 AM +00:00:00, Ross Burton
<[email protected]> wrote:
On 11 Dec 2023, at 11:37, [email protected]
<mailto:[email protected]> wrote:
On Mon, Dec 11 2023 at 11:21:48 AM +00:00:00, Ross Burton
<[email protected] <mailto:[email protected]>> wrote:
So the class is redundant, yes. Can you also add a check to the
gsettings.bbclass so its harmless instead of crashing?
something like this?
gsettings_postinstrm () {
[ -d $D${datadir}/glib-2.0/schemas ] && glib-compile-schemas
$D${datadir}/glib-2.0/schemas
}
glib-compile-schemas hasn't an option to not fail if the directory
was not found
But maybe crashing isn't that bad. Helps to keep the recipes more
clean I think?
Crashing is bad: failing with an error message is better. And
recipes may have an option for gsettings (unlikely, I admit, but
possible) and then they’ll need to mess about with conditional
inherits.
Ross
-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#107328):
https://lists.openembedded.org/g/openembedded-devel/message/107328
Mute This Topic: https://lists.openembedded.org/mt/103076498/21656
Group Owner: [email protected]
Unsubscribe: https://lists.openembedded.org/g/openembedded-devel/unsub
[[email protected]]
-=-=-=-=-=-=-=-=-=-=-=-
