| jcastro | well if the intent is to get the drivers in the distro then probably not? | 13:41 |
|---|---|---|
| tseliot | ? | 14:51 |
| tseliot | jcastro: what is wrong with having a PPA now? | 14:54 |
| jcastro | well, that only solves half the problem? | 14:55 |
| tseliot | we don't have the resources for what you last suggested | 14:56 |
| tseliot | at least as far as I know | 14:56 |
| jcastro | well, it doesn't need to be perfect, because the blingers know about the PPA | 14:56 |
| jcastro | it just needs to not be crushingly slow | 14:57 |
| tseliot | maintaining all that would be a nightmare | 14:57 |
| tseliot | I'm talking about actual kernel patching, not only packaging changes | 14:58 |
| jcastro | maybe we can ask for more help then? | 14:58 |
| tseliot | if you can find what I suggested in my last email to the technical board, then things should be ok | 14:59 |
| jcastro | ok, I'll talk to leanne and see what we can do there | 15:00 |
| tseliot | jcastro: please make it clear that I won't be the one doing the work | 15:01 |
| jcastro | right | 15:02 |
| jcastro | tseliot: out of curiosity, how often does that happen? | 15:02 |
| jcastro | please don't say all the time, heh | 15:02 |
| ricotz | tseliot, kernel-patching? this gpl mistake was kind of a review problem of the kernel team | 15:02 |
| tseliot | jcastro: the last example: LP: #1479913 | 15:02 |
| jcastro | that seems to be for fglrx? | 15:03 |
| tseliot | jcastro: pretty often, I would say | 15:03 |
| tseliot | jcastro: it could have easily affected nvidia | 15:03 |
| ricotz | jcastro, no, nvidia too | 15:03 |
| ricotz | aka flush_workqueue | 15:04 |
| tseliot | I had a similar problem with nvidia in wily | 15:04 |
| tseliot | yes, that one | 15:04 |
| tseliot | imagine having to fix that on even more drivers | 15:05 |
| tseliot | (which I already do for private OEM projects) | 15:08 |
| jcastro | is this usually connected to a new backported kernel? | 15:08 |
| jcastro | so like, if I'm on "normal" 14.04 without backported kernels this shouldn't be a problem right? | 15:09 |
| tseliot | not really, it happened in vivid too | 15:09 |
| tseliot | it was just an abi change | 15:09 |
| tseliot | (same bug report I mentioned) | 15:09 |
| jcastro | ok so the ask from you would be what, test kernels with the proprietary drivers before being published? | 15:09 |
| tseliot | we already have such system in place. Something went wrong (and will go wrong again) | 15:10 |
| tseliot | when kernels break binary drivers, the kernel team ask me to fix them | 15:11 |
| jcastro | and if you don't fix them the kernel team publishes the kernel anyway? | 15:11 |
| tseliot | that's what happened. The thing is I should fix the drivers before the kernels are published but I didn't have enough time to do it (it was a bit late when I was told) | 15:12 |
| tseliot | they usually wait for me to finish | 15:13 |
| jcastro | ok, so it's all process stuff then | 15:13 |
| tseliot | yep | 15:14 |
| jcastro | I'll ask will to talk to leann and figure out what's up | 15:14 |
| jcastro | so you guys can concentrate on the drivers themselves | 15:14 |
| tseliot | yes, you might want to ask if her team has the resources to allocate, to fix any new drivers that may be uploaded (should the drivers really become part of Ubuntu) | 15:15 |
| jcastro | hey so thinking outloud, don't flip out when I say this... | 15:20 |
| jcastro | but would the work be cut down if we concentrated on LTS-only? | 15:20 |
| tseliot | :) | 15:21 |
| tseliot | cut down to an extent, yes, reasonable, not so much | 15:24 |
| ricotz | by reducing the support to 9 months of normal releases, there are not many supported versions anyway ;) | 15:24 |
| tseliot | right | 15:24 |
| ricotz | tseliot, e.g. do you care for precise? | 15:24 |
| jcastro | tseliot: ok, I was just thinking about low hanging fruit | 15:25 |
| tseliot | ricotz: I have to, we still have OEM projects that run precise | 15:25 |
| jcastro | like, if you're behind on the drivers for X,Y, and Z but the LTS is up to date, then it's not so bad. | 15:25 |
| ricotz | I was specifically asked in the past for updated nvidia driver for precise in xedgers though | 15:25 |
| ricotz | ok | 15:25 |
| tseliot | jcastro: again, prove that you have the resources to do it. That's the main problem | 15:26 |
| ricotz | jcastro, currently there is nothing to skip, utopic is gone already | 15:26 |
| jcastro | tseliot: I don't have any, it's just us and mmarley, and maybe jderose | 15:26 |
| ricotz | or hire new resrouces ... | 15:26 |
| tseliot | jcastro: that is obviously not enough | 15:27 |
| jcastro | yeah which is why I was thinking of low hanging fruit | 15:27 |
| jcastro | I recognize that this is difficult | 15:27 |
| tseliot | what ricotz said. Possibly a small team | 15:27 |
| tseliot | adding a PPA is very easy, especially for gamers | 15:28 |
| jcastro | yeah | 15:29 |
| jcastro | ok so what happens when this kernel issue hits the people with the PPA? | 15:29 |
| tseliot | the packages in main/restricted are officially supported | 15:29 |
| ricotz | ( in relation to that: not sure how the ubuntu-x team is doing after maarten is gone ) | 15:30 |
| tseliot | tjaalton and Sarvatt maintain the X stack now. I don't think we are going to replace Maarten | 15:31 |
| tseliot | jcastro: I will try to help but any such task won't have a high priority on my TODO list. After all, some breakage is expected. ppa-purge to the rescue ;) | 15:32 |
| tseliot | a messed up PPA is much better than a messed up archive (and the OEM projects that depend on it) | 15:33 |
| jcastro | ack | 15:33 |
| mamarley | The PPA can still have updates for all supported releases, right? | 15:46 |
| jcastro | yeah | 15:47 |
| tseliot | yep | 15:50 |
| tseliot | ricotz: by the way, the new code in my 355 branch works now. You might want to have a look at the changes | 15:56 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!