| === jbicha_ is now known as jbicha | ||
| LocutusOfBorg | any archive admin to remove fcl/armhf pretty please? this will make kido and tinyxml2 transitions end | 13:50 |
|---|---|---|
| LocutusOfBorg | already removed in debian | 13:51 |
| === kirkland` is now known as kirkland | ||
| === jcastro_ is now known as jcastro | ||
| bdmurray | tjaalton: In your blog post at https://tjaalton.wordpress.com/2016/03/11/no-catalystfglrx-video-driver-in-ubuntu-16-04/ you mention HWE stack and 14.04. Was it discussed with AMD? | 17:39 |
| tjaalton | bdmurray: yep | 17:40 |
| bdmurray | tjaalton: What was / is the plan? | 17:41 |
| tjaalton | plan of what? there is amdgpu-pro -driver (beta) which folks are free to try. it's not in the archive because the packaging is not open source yet aiui | 17:42 |
| bdmurray | tjaalton: Currently people with fglrx installed are encountering failures when trying to upgrade to lts-xenial e.g. https://paste.ubuntu.com/23095154/ | 17:42 |
| tjaalton | well it certainly should just remove fglrx | 17:43 |
| bdmurray | tjaalton: what about "users relying on pro/workstation class features"? | 17:43 |
| tjaalton | they are either screwed or have to download the "pro" driver manually.. | 17:44 |
| tjaalton | this is all in the release notes | 17:44 |
| tjaalton | written by alberto I think | 17:44 |
| bdmurray | slangasek: Do you have an opinion on just removing fglrx for 14.04 users? | 17:47 |
| slangasek | bdmurray: if they're on the HWE stack, I think that's the right answer | 17:47 |
| slangasek | maybe with an added warning in the hwe-support output? | 17:48 |
| bdmurray | slangasek: I was looking through forums or askubuntu and one user was surprised they were on the HWE stack - they had downloaded and installed 14.04.2 or so and didn't know the implications of that. | 17:48 |
| tjaalton | yeah one option is to fall back on 14.04.1 | 17:49 |
| bdmurray | slangasek: So they'd made no concious choice to be using the HWE stack. | 17:49 |
| slangasek | sure | 17:49 |
| slangasek | but I don't think we want update-manager to give them options "upgrade stack or downgrade?" | 17:49 |
| slangasek | we can add some messaging about how to get onto the trusty stack if that's what they need to be using | 17:50 |
| slangasek | (maybe just in the wiki page) | 17:50 |
| bdmurray | slangasek: Okay, so add a warning to hwe-support-status re fglrx tell them to really read the wiki, update the wiki, and upload xserver-xorg-core-lts-xenial w/ breaks on fglrx, fglrx-core, fglrx-updates, and fglrx-updates-core. | 17:52 |
| slangasek | bdmurray: yes, +1 from me for that | 17:54 |
| tjaalton | additional breaks? | 17:54 |
| slangasek | tjaalton: the breaks are needed in order for the resolver to reach the correct solution of removing fglrx, instead of bailing | 17:55 |
| bdmurray | tjaalton: additional? I don't see any for -lts-xenial and fglrx | 17:55 |
| tjaalton | okay then | 17:55 |
| tjaalton | bdmurray: it doesn't provide the driver abi which xserver needs | 17:55 |
| tjaalton | that alone should make it uninstall | 17:55 |
| tjaalton | which has worked fine so far for other drivers at least.. | 17:56 |
| slangasek | that's a nice theory ;) | 17:56 |
| bdmurray | tjaalton: well, it didn't - https://paste.ubuntu.com/23095154/ | 17:56 |
| tjaalton | I saw that | 17:56 |
| slangasek | it works for the other drivers because there's a newer version that can replace them | 17:56 |
| tjaalton | no I mean if a driver wasn't rebuilt | 17:57 |
| slangasek | ok | 17:57 |
| tjaalton | it'd remove the driver and -all | 17:57 |
| slangasek | but the other piece is that fglrx has reverse-dependencies | 17:57 |
| tjaalton | been hit by that a few times.. | 17:57 |
| slangasek | that permutes the weights apt assigns for removal | 17:57 |
| tjaalton | yeah that's probably it then | 17:57 |
| === acheronuk_ is now known as acheronuk | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!