| *** persia has quit IRC | 04:17 | |
| *** tristan has joined #buildstream | 05:32 | |
| *** ChanServ sets mode: +o tristan | 06:53 | |
| *** tiagogomes has joined #buildstream | 08:25 | |
| *** tiagogomes has quit IRC | 08:29 | |
| *** tiagogomes has joined #buildstream | 08:48 | |
| tristan | So, fwiw, I have *almost* a perfectly replicated GNOME system, and I think there are a couple of regressions | 09:18 |
|---|---|---|
| tristan | would have to try a ybd build to verify if they are indeed regressions | 09:18 |
| ironfoot | kewl | 09:19 |
| tristan | Looks like we can boot to gnome-initial-setup, but then something breaks in the handoff to regular session | 09:19 |
| tristan | I suspect this was not tested after usr merge | 09:19 |
| ironfoot | :) | 09:19 |
| tristan | and maybe systemd was upgraded again, not sure | 09:19 |
| ironfoot | i think that we haven't been testing almost anything for a long time | 09:20 |
| * tristan shudders | 09:20 | |
| ironfoot | we didn't have enough time for that | 09:20 |
| ironfoot | if it was my fault, apologies | 09:20 |
| tristan | I dont think I have time to find out who's fault it is | 09:21 |
| tristan | heh | 09:21 |
| tristan | anyway, this is a throw-away system | 09:21 |
| tristan | just a proof of concept that we've converted | 09:21 |
| tristan | and that it still works | 09:21 |
| ironfoot | I still dream with a CI that would run some tests in a deployed system after building it.. | 09:21 |
| tristan | Next will be to build GNOME release modulesets and flatpak runtimes with BuildStream, and boot *that* | 09:22 |
| tristan | that will be worth debugging, not this | 09:22 |
| tristan | Although, I do want to get the handoff working again | 09:22 |
| * tristan also finds a problem with the damn 'compose' element and system integration | 09:23 | |
| tristan | looks like I'm carrying over the files which system integration commands *create* | 09:23 |
| tristan | but I'm failing to carry over the files which system integration commands *modify* | 09:23 |
| *** ssam2 has joined #buildstream | 09:31 | |
| tristan | ugh, debugging gnome session, gdm, accountservice and gnome-initial-setup has to be the most annoying thing | 11:01 |
| tristan | /usr/libexec/gdm-x-session[677]: (gnome-initial-setup:774): AccountsService-CRITICAL **: act_user_get_uid: assertion 'ACT_IS_USER (user)' failed | 11:01 |
| tristan | gdm[666]: GdmSession: Setting user: '[Invalid UTF-8]' | 11:02 |
| tristan | AccountsService: ActUserManager: failed to load user [Invalid UTF-8] | 11:02 |
| tristan | etc | 11:02 |
| ironfoot | ugh | 11:02 |
| * tristan really annoyed | 11:03 | |
| tristan | that took me like 2 weeks to "get right" | 11:03 |
| tristan | it's very delicate | 11:03 |
| tristan | and then, just go ahead and upgrade systemd, churn the system, yeah; sure it'll just work | 11:03 |
| ssam2 | unloved code | 11:03 |
| tristan | no: this is actually very, very delicate | 11:04 |
| tristan | you cant just upgrade stuff, not boot the system and go through login/initial setup, and call it a day | 11:04 |
| tristan | right about now I could use the workspaces feature | 11:05 |
| tristan | now it's going to be a hellish ride of: A.) checkout these modules separately B.) Stage commits with added printf statements C.) Rebuild, rinse, repeat | 11:06 |
| tristan | Ok, enough torture for today | 12:26 |
| tristan | Have to figure my way around this obstacle | 12:26 |
| tristan | Theres no way I can afford to spend this week swimming in the regressed integration hell of gdm | 12:27 |
| tristan | There's also no way I can really blog about this if I cant say "just run bst build gnome-system-image.bst and you will have a bootable gnome system" | 12:28 |
| tristan | So, consider me thoroughly peeved about baserock moving ahead and committing stuff to master without verifying that things actually boot and login properly. | 12:28 |
| * tristan will try to run a ybd build of baserock master overnight and see what kind of image it gives me anyway | 12:29 | |
| * paulsher1ood wonders if there is an old gnome system definition which was known to boot | 12:34 | |
| tristan | there certainly is, I can try building a conversion of *that* | 12:34 |
| tristan | but then, it's not a conversion I can really propose to baserock | 12:35 |
| tristan | "Hey, so I have conversions working ! Just please roll back all of your commits since last year" | 12:35 |
| paulsher1ood | "Hey, so I have conversions working ! and in the process it seems you guys broke the GNOME definitions at some point" | 12:36 |
| tristan | Yeah, sorry I'm just a bit frustrated | 12:36 |
| paulsher1ood | ack | 12:36 |
| tristan | there's probably some middle ground I can walk | 12:37 |
| paulsher1ood | :) | 12:37 |
| *** tristan has quit IRC | 12:45 | |
| *** ssam2 has quit IRC | 16:59 | |
| *** ssam2 has joined #buildstream | 17:01 | |
| *** ssam2 has quit IRC | 17:07 | |
| *** ssam2 has joined #buildstream | 17:24 | |
| *** tristan has joined #buildstream | 17:40 | |
| *** ChanServ sets mode: +o tristan | 17:40 | |
| *** ssam2 has quit IRC | 18:11 | |
| *** ssam2 has joined #buildstream | 18:27 | |
| *** ssam2 has quit IRC | 18:30 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!