enable meson, update unl0kr, etc

added 2 commits

cc618230 - unl0kr: move to my fork for now
860caf11 - unl0kr: update submodule

Compare with previous version

By Caleb Connolly on 2023-06-07T18:28:13

added 6 commits

b7f3ddf2 - unl0kr: update for fbdev/libinput fixes
5da00877 - switch to meson + temp unl0kr fork
1f23baa7 - unl0kr: move to subprojects/ for meson
5a3b5cd5 - WIP: storages: sdm845 support
1bfab8d4 - gui: use transparency instead of hidden flag for back button
43be5f1c - WIP: minor fixes and sdm845 support

Compare with previous version

By Caleb Connolly on 2023-06-10T20:51:01

added 1 commit

1d3afaf3 - build system fixes for packaging

Compare with previous version

By Caleb Connolly on 2023-06-10T22:56:38

added 2 commits

e47da49f - build system fixes for packaging
dd2e3b2a - TODO: Calebs status report

Compare with previous version

By Caleb Connolly on 2023-06-10T23:18:47

Did some more hacking on this, hit some really weird bug on axolotl where I just can't flash userdata, the rootfs file being 5.1G (when non-sparse) perhaps contributes to this?? I flashed manually using netcat from debug-shell (yeah...) and after rebooting the initramfs mount_subpartitions still failed. So idk what's going on here and if it relates to ondev2 at all even.

Otherwise, the main issue seems to be handling /dev/mapper/blah properly, the initramfs for some reason mounts the boot partition via /dev/dm-X which makes detecting which partition it's loop mounted from hard, if we can make it consistently use the alternative path like /dev/mapper/sda9p1 this gets easier. Then it's just a case of teaching the storages.c code to handle this.

I think that and the image size issues are the only issues left for this to at least be basically functional, there's plenty of other smaller bugs but as you said we should get meson in and get a PoC first.

I'll try and spend some time cleaning up this MR so it's in a better shape to get merged and fix these last few issues soon.

By Caleb Connolly on 2023-06-10T23:23:37

added 1 commit

b0954230 - tests: add a test framework and a test for storages

Compare with previous version

By Caleb Connolly on 2023-06-11T22:16:41

added 1 commit

b7f203a8 - tests: add a test framework and a test for storages

Compare with previous version

By Caleb Connolly on 2023-06-11T22:16:49

added 4 commits

879ee5e1 - WIP: minor fixes and sdm845 support
af414e3b - build system fixes for packaging
300cda6a - TODO: Calebs status report
b7231452 - tests: add a test framework and some unit tests

Compare with previous version

By Caleb Connolly on 2023-06-11T23:15:53

changed the description

By Caleb Connolly on 2023-06-11T23:19:49

changed the description

By Caleb Connolly on 2023-06-11T23:20:10

Did some more hacking on this, hit some really weird bug on axolotl where I just can't flash userdata, the rootfs file being 5.1G (when non-sparse) perhaps contributes to this?? I flashed manually using netcat from debug-shell (yeah...) and after rebooting the initramfs mount_subpartitions still failed. So idk what's going on here and if it relates to ondev2 at all even.

Weird, I don't know either what's going on there :\

Otherwise, the main issue seems to be handling /dev/mapper/blah properly, the initramfs for some reason mounts the boot partition via /dev/dm-X which makes detecting which partition it's loop mounted from hard, if we can make it consistently use the alternative path like /dev/mapper/sda9p1 this gets easier. Then it's just a case of teaching the storages.c code to handle this.

Good approach!

I think that and the image size issues are the only issues left for this to at least be basically functional, there's plenty of other smaller bugs but as you said we should get meson in and get a PoC first.

Yep. On that note, I took some of the commits from this MR, cleaned them up, made CI work, fixed a bug that caused random colors instead of selecting a theme etc, and pushed them to master. Added notes to the commits with what I changed. Now the big buildsystem change for meson is in!

Root cause the failed unmount issue (I added -l to umount args which ought to hide the issue, not great though)

Finally figured it out: b5e1fa24

Ensure that localisation hasn't regressed

Not sure about that, I've tested all commits with test/run-test-host-losetup.sh. This test doesn't support switching the language currently (even without the buildsystem changes). But if it did regress we can fix it up again, the important part is having the big meson change done.

tests: add a test framework and some unit tests

Great work, looking forward to getting that in as well! Once the related storage changes aren't WIP anymore: can you rebase on master, extend CI so it runs your unit tests too and make sure it passes?

Other notes:

when increasing warning_level in meson.build, it shows the same errors as in your IDE regarding wrong enums etc. let's do that at some point and fix the errors.
thoughts on copying & using strscpy from the kernel in ondev2.git? 5cd46b50

By Oliver Smith on 2023-06-30T00:42:12

marked the checklist item Root cause the failed unmount issue (I added -l to umount args which ought to hide the issue, not great though) as completed

By Oliver Smith on 2023-06-30T00:47:07

added status::mr-stale label

Sorry to bother you @calebccff,

but we've detected that this merge request hasn't seen any recent activity. If you need help or want to discuss your approach with developers you can ping @postmarketOS. You can also ask on matrix in #devel:postmarketos.org or #postmarketos-devel on OFTC. If no further activity occurs in this MR, postmarketOS developers may close it in the future.

Thanks for your contribution.

By * postmarketOS Bot on 2024-06-17T16:45:10

closed

enable meson, update unl0kr, etc

Merge request reports

Activity