microg
System network location provider does not report locations
In v0.3.3.240913, the system network location provider would not return network locations even if bound by the system. This breaks network location in apps that only use the system network location provider and not the Google fused location provider, so this mostly affects free software apps.
It may be better to release 0.3.3.1 to include this fix since it breaks most free software.
I have updated before the note about the bug was released and unfortunately reverting back to 0.3.2.240913 doesn't help. Upon reinstalling the previous version, the network location service is still broken, My Location (FOSS from GitHub) reports that Network Location service has been stopped and microG keeps crashing.
Releasing a fix as soon as possible would be highly appreciated. The votes above clearly suggest that I'm not the only one. Thanks for your excellent work!
@mar-v-in Isn't possible to create a branch starting at the commit 52a08ba (the commit of the 0.3.3 release) and then merge only the location fix in that branch? So you can create a tag from that branch and you can avoid the commits added after the release.
The fix in 2149212967ed0038a8b22f60726ecc39476e5c96 is incomplete, it can result in network location requests staying active infinitely (which is not much better than not getting started at all), so further fixing is needed. I'm mostly done with that and I can backport the relevant parts to 0.3.3 to create a 0.3.3.1 version afterwards - and/or provide a nightly based on then-current master.
@mar-v-in Does the latest commit fix "network location requests staying active infinitely"?
@sconim
The commit 511afe8, the commit date is last week but it was pushed today.
If it would be possible/not hard to do a last pull from Weblate right before next release, even if it's a bugfix release, I'd help our work tremendously, I have no easy way of checking easily how things look in-app. Building so far hasn't worked for me and I have no idea how to change the package names. I ended up having to use things like Lucky Patcher on a new release (CalyxOS hasn't launched 0.3.3 because of this bug) which really wasn't ideal. Updates to microG are months apart, so it's complicated to get an overall pic.
I'm mostly done with that and I can backport the relevant parts to 0.3.3 to create a 0.3.3.1 version afterwards - and/or provide a nightly based on then-current master.
@mar-v-in When you do the release, please mark the new release (v0.3.3.1 or v0.3.4) as stable on F-Droid, as there shouldn't no longer be big issues. If it isn't marked we will continue to receive reports for v0.2.27, that it is just a waste of time.
It really is time to get out of 0.2.27, microG has gotten amazing improvements since then and defaulting to 0.2.27 makes microG no justice, when I was using 0.2.27 it was a pretty bad exp.
https://gitlab.com/CalyxOS/calyxos/-/issues/2722 Calyx sometimes bundles newer commits of microG not yet in release, but they've been waiting for this; it'd be nice to know if this is fixed in those newer commits or if there's still some more work to be done
A new release is out which supposedly fixes this: https://github.com/microg/GmsCore/releases/tag/v0.3.4.240913
Thank you to everyone involved :)