| freemangordon | Wizzup: hmm, I think we made a mistaker | 07:21 |
|---|---|---|
| freemangordon | we shouldn't have changed gconf, but debhelper (or some other package) instead | 07:21 |
| freemangordon | /usr/share/perl5/Debian/Debhelper/Sequence/root_sequence.pm:[14, 'dh_gconf'], | 07:22 |
| freemangordon | this is on daedalus | 07:22 |
| freemangordon | so, we just need to find a way to install dh_gconf | 07:23 |
| freemangordon | as it not also moves files, but also adds gconf2 dependency to the package | 07:28 |
| freemangordon | maybe we shall just fork debhelper and do partial revert of b4cbfe7d2fd4f1e428500f6de1b2a65d21673ae1 | 07:34 |
| freemangordon | instead of forking gconf | 07:34 |
| Wizzup | freemangordon: I was hoping to avoid forking debhelper | 09:03 |
| freemangordon | then maybe we can add dh_gconf to maemo-system-services-dev | 09:17 |
| freemangordon | but, who will invoke it? | 09:17 |
| freemangordon | we will have to edit all the packages | 09:17 |
| freemangordon | again | 09:17 |
| Wizzup | why can't we go with our fix from yesterday | 09:18 |
| Wizzup | if all it is a depend on gconf, that should be easy | 09:18 |
| Wizzup | or does this interfere with our build process as well | 09:18 |
| freemangordon | I think it *might* interfere with image building, like, if gconf is not installed)or configure, not sure) before a $package is installed, then gconf trigger will not be executed | 09:21 |
| freemangordon | well, maybe you can tweak image building to install gconf first | 09:22 |
| freemangordon | or, add gconf as pre-depends on some of the meta packages | 09:22 |
| Wizzup | all of it kinda sucks | 09:23 |
| freemangordon | I know | 09:23 |
| Wizzup | so the line you shared, doesn't that already invoke dh_gconf? | 09:23 |
| freemangordon | lemme check what it exactly does | 09:23 |
| freemangordon | no, it will not | 09:25 |
| Wizzup | ok | 09:25 |
| Wizzup | sorry, I have to go for the next hour or so, back later | 09:26 |
| Wizzup | I'm also fine with forking dh | 09:26 |
| freemangordon | ok | 09:26 |
| Wizzup | whatever makes this less painful | 09:26 |
| freemangordon | I think that will be easiest | 09:26 |
| Wizzup | btw, if we do that we might as well go straight to trixie ande just build gconf for it, no? | 10:26 |
| Wizzup | I guess maybe that is a bit too ambitious currently | 10:40 |
| Wizzup | freemangordon: so I should hold off on fixing all our gconf pkgs? | 11:02 |
| Wizzup | I don't know quite what know debhelper to build | 11:05 |
| sicelo | starts to look like the gconf->gsettings transition needs to start pretty soon | 12:56 |
| Wizzup | if you have a few spare months :) | 12:57 |
| Wizzup | I did a quick search and it looks like about 90 pkgs uses gconf | 12:58 |
| Wizzup | and we also use gconf in a way that gsettings doesn't like - writing to keys outside of schemas, we'd have to mark some keys differently I tihnk | 12:58 |
| Wizzup | freemangordon: with your recent wifi change, does the wifi actually ever disconnect? | 13:22 |
| uvos__ | the schemas are also all in reverse | 13:26 |
| uvos__ | or at least many | 13:26 |
| uvos__ | reading or writeing a gsettings key for witch no schema exits is an error, but in maemo the settings program's applet owns all the schemas | 13:27 |
| uvos__ | so mce will crash with an assert if not every settings applet is installed | 13:27 |
| uvos__ | same thing for all maemo services | 13:28 |
| uvos__ | this creates a bunch of dependancy loops, since the services require the applets that require the services | 13:28 |
| uvos__ | really the schemas need to be owned by the services | 13:28 |
| Wizzup | yes | 13:29 |
| Wizzup | which means it'll be a significant undertaking :D | 13:29 |
| uvos__ | yeah | 13:30 |
| inky | when i run install.sh from bionic-clown-boot repo (root404.sh is successful) i get permission denied messages | 14:25 |
| inky | adb: error: failed to copy 'scripts/kexebootstart.sh' to '/sdcard/kexebootstart.sh': remote Permission denied | 14:25 |
| inky | like this | 14:25 |
| inky | when i enter recovery, the manual says to enter recovery and | 14:27 |
| inky | >select wipe data/factory reset | 14:27 |
| inky | i just see that laying android photo with ! and opened belly. | 14:28 |
| inky | i cannot wipe data/reset. but i did it from android, factory reset. | 14:28 |
| inky | i also restored everything by flashing the stock system.img | 14:28 |
| uvos__ | inky: you have to press both vol up and vol down at the same time for the menu to pop up over the dead android | 14:33 |
| uvos__ | you may have to have a (fat32 formated, visible in android) sdcard inserted for it to work i dont remember that detail | 14:34 |
| uvos__ | you also have to reboot between rooting and installin | 14:35 |
| uvos__ | g | 14:35 |
| uvos__ | if you reflashed system you also have to reroot | 14:36 |
| inky | will write, have to deviate... | 14:41 |
| utilisateur | cc cv | 15:22 |
| Wizzup | someone is adding Tamil translation to our pkgs :) | 15:32 |
| Wizzup | freemangordon: lmk what you want reverted in debhelper and I can do it | 15:32 |
| Wizzup | https://salsa.debian.org/debian/debhelper | 15:33 |
| sicelo | Wizzup: maybe create an issue on GH listing those gconf pkgs. whenever someone has a spare moment, they could have a look | 16:41 |
| sicelo | perhaps it's possible to work on one (or a few) packages at a time. no idea | 16:41 |
| Wizzup | I do not think it will be too easy | 16:45 |
| Wizzup | many pkgs read/expect the same key | 16:45 |
| Wizzup | freemangordon: genuinely wondering if we should not try for trixie if we will fork gconf and dh anyway | 16:46 |
| Wizzup | we'd be ahead of the curve then | 16:46 |
| sicelo | +1 to that ... | 16:47 |
| * sicelo hides | 16:47 | |
| sicelo | although trixie isn't released yet, but clearly won't be changing much until then | 16:48 |
| Wizzup | I think we can finish daedalus/bookworm since we-re 95% there and almost all the work we do for it helps trixie too | 16:51 |
| Wizzup | but let's then also work on trixie after | 16:51 |
| sicelo | heh, bookworm is EOL mid-next year. Only just realized | 17:00 |
| sicelo | anyway LTS will take it to 2028 | 17:01 |
| Wizzup | yeah | 17:02 |
| Wizzup | doing daedalus/bookworm also allows for a dist upgrade path | 17:07 |
| Wizzup | more easier one hopefully | 17:07 |
| sicelo | btw migrating to trixie will also have the issue of usrmerge ... but i guess shouldn't bite us too hard | 17:11 |
| freemangordon | do we want to spend time now migrating to gsettings? | 17:11 |
| freemangordon | I don;t htink it will be that easy | 17:11 |
| Wizzup | freemangordon: no, we'd build gconf for trixie | 17:12 |
| Wizzup | :D | 17:12 |
| freemangordon | oh | 17:14 |
| freemangordon | no, please | 17:14 |
| freemangordon | what is the issue of bringing dh_gconf back in debhelper? | 17:15 |
| Wizzup | freemangordon: why 'no, please'? | 17:18 |
| Wizzup | freemangordon: there is no issue with doing that | 17:18 |
| Wizzup | but if we do that now, why not do it for trixie too | 17:18 |
| Wizzup | otherwise, the moment we're ready, 2-3 months later, debian will stop supporting our base | 17:18 |
| freemangordon | I am not sure what'd we do with mesa | 17:19 |
| Wizzup | because of dri2? is that a problem already in trixie? | 17:19 |
| freemangordon | don;t know | 17:19 |
| Wizzup | btw, I did say let's finish daedalus first and then do excalibus/trixie | 17:19 |
| freemangordon | ah, okm | 17:19 |
| Wizzup | but I think we shouldn't wait, if we agree we can just build gconf | 17:19 |
| Wizzup | and dh | 17:19 |
| freemangordon | ok, that sounds more like a plan | 17:20 |
| freemangordon | but, lets do daedalus first | 17:20 |
| Wizzup | ok | 17:20 |
| Wizzup | so do you want me to find the commit in dh that removed gconf and revert it | 17:20 |
| freemangordon | so, what we do now? | 17:20 |
| freemangordon | the commit is ^^^ | 17:20 |
| Wizzup | I think we keep our gconf as well (why not), but then also build dh | 17:20 |
| freemangordon | b4cbfe7d2fd4f1e428500f6de1b2a65d21673ae1 | 17:20 |
| freemangordon | no need of gconf (for daedalus) if we build dh | 17:21 |
| Wizzup | true, but we have it already | 17:22 |
| Wizzup | and we'd have to build it for trixie anyway | 17:22 |
| Wizzup | so I won't remove our upstream fork repo | 17:22 |
| Wizzup | but we can remove it from daedalus if you like | 17:22 |
| freemangordon | sure, but we don;t need the change | 17:22 |
| Wizzup | ok | 17:22 |
| freemangordon | yeah, remove it from daedalus and revert that change | 17:22 |
| Wizzup | gconf is gone from daedalus | 17:23 |
| Wizzup | revert that change = build dh with that commit reverted? | 17:23 |
| freemangordon | you can't just revert it, there is conflict because of the translations | 17:24 |
| freemangordon | anyway yes, build dh with that commit revertied and conflicts fixed | 17:24 |
| Wizzup | ok | 17:28 |
| Wizzup | freemangordon: he already removed a tag for regular bookworm | 17:46 |
| Wizzup | shall we build bookworm backports, or how do I find 13.11.4 | 17:46 |
| Wizzup | oh no | 17:47 |
| Wizzup | it is there | 17:47 |
| Wizzup | gitlab just lies | 17:47 |
| freemangordon | the tag is there | 17:52 |
| Wizzup | freemangordon: https://github.com/maemo-leste-upstream-forks/debhelper | 17:56 |
| Wizzup | ok? | 17:56 |
| freemangordon | looks like, yes | 17:58 |
| sicelo | freemangordon: btw, regarding upower & n900/d4 fuel gauges, i once did this - https://gitlab.postmarketos.org/postmarketOS/pmaports/-/blob/4b54fc7f83f9e08e2a80bbeb6a97f871748bf7c6/device/community/linux-postmarketos-omap/0012-bq27xxx_battery-add-voltage-based-capacity-estimatio.patch | 20:43 |
| sicelo | worked pretty good, and perhaps could be used in Leste too. it's a kernel patch which is fine since we roll our own kernel anyway. benefit is - mce & upower need no further intelligence. | 20:44 |
| Wizzup | >dh: error: The addon 14 claimed that dh_gconf was obsolete, but it is not!? | 20:49 |
| Wizzup | lol | 20:49 |
| Wizzup | ah I see | 20:51 |
| Wizzup | freemangordon: the debhelper stuff should be in place | 22:50 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!