| desrt | Laney: i agree, more or less | 02:59 |
|---|---|---|
| desrt | Laney: the question is what we should configure as the libdir for g-c-c | 02:59 |
| desrt | i guess it should be the multiarched libdir | 02:59 |
| desrt | and have a hardcoded fallback to /usr/lib like we did with the gio patch | 02:59 |
| desrt | yawnorama | 03:00 |
| Laney | Actually I just found out that g-c-c tells you where to put stuff in its pcfile | 16:55 |
| Laney | so we shouldn't be ignoring that | 16:55 |
| larsu | Laney: this is what we were using before, but isn't is a bit stupid to install into /usr/lib from a package that puts all its other stuff into /usr/lib/x86_64? | 17:02 |
| desrt | Laney: ya.... the directory-in-pkgconfig file thing is stupid | 17:03 |
| Laney | It doesn't really matter in this case as it's not multiarch compatible | 17:03 |
| desrt | it results in 'make install' putting files outside of the instructed --libdir | 17:03 |
| desrt | which means that even though the package is nominally multiarch, it won't actually install properly | 17:03 |
| desrt | since there are different arches installing different files into the same /usr/lib path | 17:03 |
| Laney | well this package just uses the multiarch directory without being enabled for it | 17:04 |
| Laney | I guess we could/should make g-c-c install into those directories so it doesn't preclude any future conversions | 17:04 |
| jbicha | Laney: are you sure that dropping the .pc file for gstreamer-bad is right? bug 1170923 | 17:06 |
| ubot2` | Launchpad bug 1170923 in gst-plugins-bad1.0 (Ubuntu) ".pc file is missing from libgstreamer-plugins-bad1.0-dev" [Undecided,New] https://launchpad.net/bugs/1170923 | 17:06 |
| desrt | Laney: i think the correct fix here is to do the following: | 17:07 |
| desrt | 1) multiarch g-c-c so that it has the same --libdir as the indicator | 17:07 |
| desrt | then either | 17:07 |
| desrt | 2a) multiarch all the other indicators at the same time | 17:08 |
| desrt | or | 17:08 |
| desrt | 2b) have a patch to fallback to /usr/lib for the non-multiarched ones until we get them moved over as well | 17:08 |
| Laney | jbicha: you could add a dummy one which depends on -good if you want | 17:08 |
| larsu | desrt: I don't think there's another indicator that comes with a settings panel... | 17:08 |
| desrt | larsu: doesn't dejadup do a panel as well? | 17:08 |
| larsu | desrt: ya, but no indicator :) | 17:09 |
| * desrt is sure that there are at least a couple of others... | 17:09 | |
| Laney | desrt: why do you need 2 if things respect the directory in the pcfile? | 17:09 |
| desrt | Laney: respecting the directory in the .pc file is a _bug_ | 17:09 |
| desrt | from an upstream standpoint | 17:09 |
| desrt | we have a situation where the debian packaging (due to multiarch stuff) is explicitly instructing the package --libdir=(multiarch) | 17:11 |
| desrt | to have the package ignore that and install into /usr/lib anyway is a clear bug | 17:11 |
| desrt | _and_ it also causes the multiarching to produce separate arch packages that are not co-installable | 17:11 |
| desrt | which is against the entire point of multiarching | 17:11 |
| jbicha | Laney: you mean edit the existing one to add gstreamer-good-@GST_API_VERSION@ ? | 17:12 |
| Laney | FWIW this package has multiarch libdir (because of debhelper 9) but aren't actually Multi-Arch: same | 17:12 |
| Laney | isn't* | 17:12 |
| Laney | so it's not coinstallable by MA anyway | 17:12 |
| desrt | would still be a bug to ignore --libdir | 17:12 |
| larsu | Laney: is an acceptable fix to downgrade to an older debhlper then? | 17:12 |
| Laney | you could pass the lib{,exec}dir you want | 17:13 |
| desrt | well.... | 17:13 |
| desrt | yes | 17:13 |
| desrt | but that sort of misses the point, i feel | 17:13 |
| desrt | debhelper defaults to the new dir for a reason, right? | 17:14 |
| desrt | it seems that debian's intention here is to transition to separating libraries by arch | 17:14 |
| Laney | it imagines that you'll want to do that, yes | 17:14 |
| desrt | all libraries, even if not actively multiarching | 17:14 |
| desrt | so instead of trying to fight against that, why not help push it a long a bit further by doing the same in g-c-c? | 17:14 |
| Laney | I do think it's better to fix g-c-c | 17:14 |
| desrt | so let's do that | 17:15 |
| Laney | but it would also be acceptable to me to install in non-multiarch directories | 17:15 |
| desrt | let's do the right thing now that won't require more work to unwind later | 17:15 |
| * desrt is a big fan of emptying out /usr/lib (except for the multiarch subdirs) | 17:16 | |
| jbicha | is g-s-d multiarchable? if it isn't then it's not clear that multiarching g-c-c helps | 17:16 |
| Laney | there's using those directories and being M-A: same | 17:16 |
| desrt | is there a g-c-c/g-s-d cross-dependency that would be impacted by that? | 17:16 |
| larsu | desrt: I agree, but we have some serious breakage because of this right now. How long will fixing g-c-c take? | 17:16 |
| desrt | larsu: i guess someone just needs to flip the switch there | 17:17 |
| desrt | that's step 1, right? | 17:17 |
| desrt | then step 2..... | 17:17 |
| desrt | the patch would probably be a few lines | 17:17 |
| desrt | or we could fix all the other things installing panels there at the same time | 17:17 |
| larsu | actually, the packages that install into the .pc var should be fine... | 17:17 |
| desrt | larsu: imagining that working in the other direction is hurting my brain | 17:18 |
| desrt | but ya.... | 17:18 |
| larsu | :) | 17:18 |
| Laney | jbicha: RE the pcfile - I think you could add one into -bad which requires the new one in -good; should keep stuff working | 17:18 |
| Laney | got to go now, ttyl | 17:18 |
| jbicha | besides datetime, activity-log-manager, deja-dup, g-c-c-signon, and g-c-c-unity also install panels | 17:21 |
| * desrt gets annoyed by airfrance | 17:24 | |
| desrt | their baggage policy is insanely obnoxious | 17:25 |
| desrt | does anyone know how this works? | 17:25 |
| larsu | desrt: ask in #ubuntu-fr? :P | 17:25 |
| desrt | sigh | 17:25 |
| * desrt finally figures out how to get the information he needs | 17:27 | |
| desrt | i have to go to the german site, then change the language to english | 17:27 |
| desrt | of course | 17:27 |
| === Quintasan_ is now known as Quintasan | ||
| desrt | cool | 17:39 |
| desrt | if i book through airfrance's site, they charge me $10 for the privilege of paying with visa | 17:39 |
| desrt | and if i book through any other normal travel site i have to wonder if i'm going to be allowed to bring a bag or not due to the insanely complicated and non-transparent nature of their fare rules | 17:40 |
| desrt | "..... get ripped off for $10 or worry that i don't have the right kind of ticket? hmm..." | 17:40 |
| * desrt gets ripped off | 17:40 | |
| * desrt finds it highly questionable that there should be two fares that are very difficult to find the difference between, separated by €20 in price, where the higher fare includes one free bag and the fee for having that bag, paid at the airport, is €70 | 17:48 | |
| desrt | it's almost as if they want you to make a mistake...... | 17:48 |
| Smaug | hey all, I am having difficulty connecting to irc on my ubuntu desktop (am on windows right now). I've tried both empathy and kvirc, so I'm thinking it's related to ubuntu and not to a particular software not working. Should I check some firewall settings or something? Any thoughts? | 19:38 |
| czajkowski | Smaug: you might fine help in #ubuntu this isn't really a support channel | 20:08 |
| Smaug | thanks | 20:13 |
| Smaug | ah, should have read the topic | 20:13 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!