| bluesabre | 3 | 03:38 |
|---|---|---|
| xnox | hm, interesting. | 04:12 |
| xnox | slangasek: without a matching boost-mpi-source1.54 above will result in uninstallables in universe. | 04:13 |
| * xnox goes to prepare it | 04:14 | |
| slangasek | xnox: ok, thanks | 04:14 |
| * xnox mumble mumble archive reorg mumble =) | 04:15 | |
| xnox | thanks! | 04:29 |
| slangasek | no rpob | 04:30 |
| ScottK | xnox: That's not archive reorg's fault. Before we split the package, we just didn't provide the mpi packages at all. | 05:11 |
| xnox | ScottK: I mean if archive reorg plans would be in place, we would not need to split source packages at all. | 05:12 |
| ScottK | How so? | 05:13 |
| xnox | https://wiki.ubuntu.com/ArchiveReorganisation/Components | 05:14 |
| infinity | ScottK: One of the theories behind archive reorg (ie: flattening the components to free/non-free instead of {un,}supported-{non-}free is that we could let supported build-dep on unsupported and just mark binaries supported. | 05:14 |
| infinity | ScottK: I'm less convinced this is actually a sensible way to declare things, since we probably want to support things we use to build things, but in a handwavy way, I get the intent. | 05:15 |
| slangasek | we don't really want to support them; feel free to peruse http://reqorts.qa.ubuntu.com/reports/m-r-package-team-mapping.html for some of the crazier stuff that doesn't actually matter to us except transitively and we shouldn't actually provide security support for | 05:16 |
| infinity | slangasek: Well, we don't want to support everything, but we want to support a lot of things. Means a lot more explicit seeding, I guess, for toolchains and things we actually really should support? | 05:17 |
| infinity | But I can see, for instance, why we don't care about supporting things like aalib, except to make sure they're buildable. | 05:18 |
| slangasek | build-essential is already seeded, and done ;) | 05:18 |
| infinity | Which was what I meant by "handwavy"... We need to support a lot of things that are currently only transitively supported, but not necessarily all of it. | 05:18 |
| infinity | slangasek: There's a lot more than build-essential that matters. :P | 05:18 |
| infinity | slangasek: We don't seed most of X, for instance, cause it's all transitive. | 05:19 |
| infinity | In fact, that's true of most libraries. | 05:19 |
| infinity | And some we might not care about, but others, we really do. | 05:19 |
| slangasek | we would seed the binary deps, not the source deps | 05:19 |
| infinity | Err, right. Sorry, X and libs are poor examples. | 05:20 |
| infinity | But there are other toolchain bits that aren't build-essential. | 05:20 |
| infinity | Anyhow, I get the theory. Some day, it might be less of a theory. | 05:20 |
| xnox | argh, why is lintian hanging in doing select, when executed under sbuild?! | 05:26 |
| xnox | annoying. | 05:26 |
| infinity | xnox: Probably trying to write to the controlling terminal, which sbuild doesn't provide. | 06:05 |
| xnox | hm, but lintian hasn't changed recently, nor e.g. sbuild. i wonder if it's something else like a hook or some such. | 06:06 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!