| tsimonq2 | I was just about to ask about a potential SDDM merge, but I see RikMills took care of it in a PPA :) https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/3997/+sourcepub/16135555/+listing-archive-extra | 14:45 |
|---|---|---|
| tsimonq2 | Just needs testing | 14:45 |
| tsimonq2 | I also just synced the latest qt6-speech upload | 14:45 |
| ErichEickmeyer[m | Bridge terminated? | 14:46 |
| ErichEickmeyer[m | Nvm, guess not. | 14:46 |
| tsimonq2 | sgmoore, arraybolt3, Eickmeyer, RikMills: I'm about to make a decision that will be somewhat disagreeable, so allow me to explain myself... given a) the speed of things last cycle, b) the time it takes to get things through the NEW queue, and c) the amount of work we still have ahead of us this cycle, I'm going to go ahead and get Oracular in sync with the KF6 progress in Experimental. Not much is | 15:50 |
| tsimonq2 | left Debian-side; I'm making this decision knowing full well that we'll have a good amount of packages caught in proposed, but I'm willing, able, and prepared to help with that effort. I'd rather have 5 months to get everything through to -release than 4. Big picture, I truly believe this is the correct decision, despite the short-term frustration it will cause. | 15:51 |
| Eickmeyer | tsimonq2: I say JFDI. | 15:53 |
| Eickmeyer | Rip off the bandaid. | 15:53 |
| tsimonq2 | All 60 of them that were ready have been accepted. | 16:08 |
| tsimonq2 | Just time for retries now. | 16:08 |
| tsimonq2 | (Dep trees) | 16:08 |
| tsimonq2 | I'm sure I'll get the emails :P | 16:08 |
| AaronRainbolt | <tsimonq2> "sgmoore, arraybolt3, Eickmeyer..." <- "Break the dam! Release the river!" | 16:33 |
| AaronRainbolt | alright I'mma go and see more about why Qt6 isn't migrating. | 16:34 |
| tsimonq2 | Thanks! | 16:38 |
| tsimonq2 | also yes we broke the dam :D | 16:39 |
| AaronRainbolt | https://objectstorage.prodstack5.canonical.com/swift/v1/AUTH_0f9aae918d5b4744bf7b827671c86842/autopkgtest-oracular/oracular/arm64/b/beads/20240513_044156_dbdcb@/log.gz ⁉️ | 16:42 |
| AaronRainbolt | This is that same beads package again, still failing with a symbol lookup error because it's puling a library from Qt 6.4.2 rather than pulling it from 6.6.2. But here's the spooky thing - this was triggered with all-proposed=1 (by me) and libqt6core5compat6 6.6.2 is in -proposed. | 16:43 |
| AaronRainbolt | so... I guess that means the beads package itself is doing something funky. | 16:43 |
| AaronRainbolt | hmm, Build-Depends on qtbase5-dev... | 16:44 |
| tsimonq2 | c++filt _ZN5QUtf818convertFromUnicodeE11QStringViewPN20QStringConverterBase5StateE | 16:46 |
| tsimonq2 | QUtf8::convertFromUnicode(QStringView, QStringConverterBase::State*) | 16:46 |
| tsimonq2 | hmmm | 16:46 |
| ErichEickmeyer[m | AaronRainbolt: That can't be right. | 16:47 |
| AaronRainbolt | well it is | 16:53 |
| AaronRainbolt | could be wrong but that's what's there and it builds right | 16:53 |
| AaronRainbolt | I assume that qt6 is somehow replacing qt5 using this core5compat thing | 16:53 |
| tsimonq2 | Correct | 16:54 |
| TimAndersson[m] | <AaronRainbolt> "This is that same beads..." <- sounds odd. We have some odd apt fallback mechanisms we've noticed in autopkgtest, this may be linked to these issues. I'll take a look at the log tomorrow morning and see if the failure is autopkgtest related or something deeper (i'm one of the autopkgtest admins) | 16:55 |
| tsimonq2 | \o/ well that's convenient :D | 16:55 |
| AaronRainbolt | so my guess is that qtbase5-dev is somehow referencing a very particular version of qt6 and that's where it goes wrong. Why else would it install a mixture of Qt 6.6.2 and Qt 6.4.2 other than something weird like what Tim just mentioned? | 16:55 |
| AaronRainbolt | TimAndersson[m]: Thank you! | 16:55 |
| AaronRainbolt | <AaronRainbolt> "so my guess is that qtbase5-..." <- Actually this can't be right, since when beads is built it pulls the right libqt6core5compat6. | 17:03 |
| AaronRainbolt | So OK, new working theory is "Tim knows best", just going to let that package sit until tomorrow then. | 17:03 |
| AaronRainbolt | Calibre seems to be stuck on Qt WebEngine dying when it attempts to run unit tests, going to see if it's transient, and if not, what caused it. | 17:07 |
| AaronRainbolt | ok, change of plan, dayjob has a critical task for me to finish so I have to drop everything and do that. | 17:22 |
| RikMills | AaronRainbolt: calibre is not transient. I already retried | 17:44 |
| AaronRainbolt | kk, good to know | 17:44 |
| AaronRainbolt | and that's pretty bad news potentially | 17:45 |
| RikMills | AaronRainbolt: not great, but as a leaf package it could be removed fron -release temporarily if really needed | 17:46 |
| RikMills | tbh, calibre is one of the packages that I don't know why we ship a .deb | 17:50 |
| RikMills | it changes so fast uptream that we are always out of date by a mile on stable releases | 17:50 |
| RikMills | I always use the release from thier website | 17:51 |
| ErichEickmeyer[m | calibre ships in Edubuntu and is part of the Ubuntu Studio publishing suite (not shipped by default but supported/part of the ubuntustudio-publishing seed). | 19:12 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!