| ScottK | slangasek: (since you're the furthest west release team member) - It looks like the latest gcc4.4 upload is breaking many/all (not sure) powerpc builds. | 02:41 |
|---|---|---|
| ScottK | (or anyone else). | 02:41 |
| ScottK | Given where we are in the release cycle, not sure if that's a push the panic button item or not. | 02:42 |
| ScottK | I have access to powerpc hardware and I'm trying to verify reverting fixes it. | 02:42 |
| * holstein has *old* ppc hardware | 02:48 | |
| holstein | but if you need testing | 02:48 |
| holstein | let me know | 02:48 |
| slangasek | ScottK: do you have an example broken build? | 02:52 |
| ScottK | slangasek: kipi-plugins | 02:52 |
| ScottK | qt4-x11 | 02:53 |
| ScottK | kdeedu | 02:53 |
| slangasek | ok | 02:53 |
| ScottK | There may be more. | 02:53 |
| ScottK | So far I've just found C++ examples. | 02:53 |
| slangasek | this won't be a quick package to revert, so I'll wait for you to verify that reverting fixes it and in the meantime will poke at the diff (in between having dinner) | 02:54 |
| slangasek | the failure mode is just a FTBFS, right? | 02:54 |
| ScottK | Yes. | 02:54 |
| ScottK | Not a misbuild at least. | 02:54 |
| slangasek | ok; so once fixed we can do a mass giveback | 02:54 |
| ScottK | Yep | 02:57 |
| lamont | slangasek: do you want the ppc buildds on manual while you wait? | 02:58 |
| ScottK | lamont: I vote yes if that counts. | 03:00 |
| ScottK | slangasek: My build with the previous gcc4.4 already got past where the current one failed. | 03:01 |
| ScottK | I consider it confirmed. | 03:01 |
| lamont | done | 03:01 |
| ScottK | slangasek: That's about all I can do I think. Over to you. | 03:02 |
| ScottK | I'll check back in and let you know if the build finishes, but I'm confident it will. | 03:02 |
| slangasek | "cannot find -lgcc_s" - ah, that definitely lines up with what I'm seeing in this diff | 03:12 |
| slangasek | though I guess it'll take a bit to track down the faulty bit | 03:12 |
| ScottK | That's good news. At least it's not magically broken. | 03:14 |
| ScottK | I'll leave it with you then. | 03:15 |
| ScottK | slangasek: Don't forget to tell lamont when to turn the builders back on .... | 03:15 |
| ScottK | Good luck. | 03:15 |
| ScottK | slangasek: My test build with the old gcc-4.4 of kipi-plugins succeeded. | 04:16 |
| lamont | so if someone wanted to fix gutenprint to not run -j16 at whereever it does that, arm might like it better (gourd may have fallen over completely, or it may just be very angry and slow) | 04:42 |
| lamont | after checking, it looks to be just very angry and slow, not down-and-out. | 04:50 |
| lamont | but I suspect it would build faster if it were less agressive | 04:50 |
| * lamont sleeps | 04:50 | |
| slangasek | Files in first .deb but not in second | 05:24 |
| slangasek | ------------------------------------- | 05:24 |
| slangasek | -rw-r--r-- root/root /usr/lib/gcc/powerpc-linux-gnu/4.4/libgcc_s.so | 05:24 |
| slangasek | well that probably explains it, don't it | 05:24 |
| slangasek | heh; so all the bits are there in debian/rules*, but a missing make dependency means they get executed out of order | 06:16 |
| slangasek | oh, no, simpler than that; they're just plain listed out of order | 06:17 |
| slangasek | lamont: well, I have a candidate fix now; so turning the powerpc builders back on so we can build it would help :) | 07:21 |
| lamont | so... are we expecting these gcc builds now running to be what we need? | 14:05 |
| lamont | cjwatson: around? | 15:11 |
| ScottK | slangasek, lamont, cjwatson: Looks like powerpc is fixed. What's the best way to give back the failures? All or is it easy enough to just do that ones that failed yesterday/today? | 16:08 |
| * ScottK queued up enough that there's no rush to decide. | 16:10 | |
| lamont | ScottK: there's a script to do it, I was just waiting for 4.5 and arm and my sanity to all be happy | 16:11 |
| lamont | weekend tasks are kind of interfering with my playtime | 16:11 |
| ScottK | lamont: OK. I think there's enough queued on powerpc that waiting for armel won't slow things down. | 16:17 |
| slangasek | lamont: powerpc> hurray! and armel is in progress then, I guess? | 18:40 |
| ScottK | Looks like we'll know in another half a day or so. | 19:05 |
| * slangasek nods | 19:48 | |
| ScottK | Actually gcc-4.4 finished much faster this time. Looks like we are in business. | 21:34 |
| Laney | published? | 21:35 |
| ScottK | Yep. | 21:41 |
| * Laney tries a give-back | 21:41 | |
| ScottK | Laney: lamont has a script he'll run to do a mass give back. | 21:41 |
| ScottK | lamont: I think you can fire when ready. | 21:41 |
| Laney | alright then | 21:42 |
| ScottK | slangasek: ^^^ | 21:49 |
| ScottK | lamont: please tell me you took adare down for a good reason and it didn't just die randomly? | 22:18 |
| * ScottK looks around for an FFe to support http://dylanmccall.blogspot.com/2010/09/new-installer-slideshow-for-ubuntu.html | 23:33 | |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!