| StevenK | wgrant: https://code.launchpad.net/~stevenk/launchpad/garbo-abort-info/+merge/154592 is updated | 00:22 |
|---|---|---|
| wgrant | StevenK: Much better, thanks | 00:25 |
| wgrant | r=me | 00:25 |
| === wedgwood is now known as wedgwood_away | ||
| === jamesh_ is now known as jamesh | ||
| === almaisan-away is now known as al-maisan | ||
| === al-maisan is now known as almaisan-away | ||
| === wedgwood_away is now known as wedgwood | ||
| === benji___ is now known as benji | ||
| xnox | mgz: i'd like to quickly chat to you about jubany when you have a moment or two =) | 14:36 |
| === matsubara_ is now known as matsubara | ||
| mgz | xnox: lets do it | 15:03 |
| xnox | mgz: i've read all the README and README_DEPLOYMENTS info. Roughly, my understanding is sudo into pkg_import user, schroot into quantal, source / update path to include the magic scripts directory & run those. | 15:08 |
| xnox | mgz: but for example, i'm not quite sure what requeue_package --full will do and whether it is "safe" to run on out of date packages. | 15:08 |
| xnox | mgz: also I'm curious about what will need to be done at s-series opening. | 15:09 |
| mgz | --full is not safe as such | 15:09 |
| xnox | that was my thinking. | 15:09 |
| mgz | as it discards all existing history, and attempts to recreate it from what's published | 15:09 |
| xnox | that's not what I want =) | 15:10 |
| mgz | generally packages are out of date for one of two reasons: | 15:10 |
| mgz | * a bug or limitation of some kind in constructing the history from a certain point, often packaging error, has also been stuff like lack of support of latest .xz format changes | 15:11 |
| mgz | * some history got rewritten after an import ran, so the importer can't reconcile what it has with what now exists | 15:12 |
| mgz | really #2 is just a subcase of #1. but #2 is the only kind where --full might actually fix things | 15:13 |
| xnox | I think I have a latter case, where --full might actually help. All revisions were committed by pkg-importer only, ye | 15:13 |
| xnox | yet the latest upload by me, is not being imported. | 15:14 |
| xnox | mgz: has anything been done after raring switched to uploads to -proposed with a copy to -release? as I bet the importer is going crazy with packages appearing & disappearing in -proposed. | 15:14 |
| xnox | and later reappearing in -releases (as this is the course of the events usually) | 15:14 |
| mgz | nothing has been done, and that does sound like it might break things | 15:15 |
| cjwatson | well, *usually* they disappear in -proposed and appear in release in the same publisher cycle | 15:15 |
| xnox | cjwatson: sure, unless is haskell and it's in -proposed for a while before it migrates, or if the package is huge and takes time to build on armhf, by which time the source package in -proposed is published already. | 15:16 |
| cjwatson | if it's in -proposed for a while before it migrates then it hasn't disappeared from -proposed | 15:16 |
| cjwatson | and how long it takes to build has no bearing on the relative timing of disappearance from -proposed vs. appearance in release | 15:17 |
| xnox | ok. | 15:19 |
| cjwatson | if that *is* the problem then it's plain bad luck on when proposed-migration finishes | 15:20 |
| cjwatson | if it happens to finish very close to the start of the publisher then it might race in that particular way. but it's not clear whether that is in fact the problem anyway :) | 15:20 |
| xnox | right. I just wonder how come we have 417 packages failing with bug 714622, with some of them fairly straight forward (e.g. lp:ubuntu/libsemanage where nobody but package-importer has touched the branches) | 15:21 |
| _mup_ | Bug #714622: import fails when lp branch has been push --overwrite'n <Ubuntu Distributed Development:Confirmed for canonical-bazaar> < https://launchpad.net/bugs/714622 > | 15:21 |
| xnox | and there was no --overwrite. | 15:21 |
| mgz | that does sound like something else is tickling that symptom | 15:22 |
| xnox | mgz: let me run pkg-import here locally for libsemanage to see if there are any other bugs with it, and if there aren't I'll just requeue full it. | 15:25 |
| mgz | xnox: on the updates needed for S front, I got myself in a muddle last time, there's a thread on the udd list from October you can refer to | 15:41 |
| mgz | in short, there are just some hardcoded values in lp:udd to change | 15:41 |
| mgz | part of the release process should be stopping the importer before release, then enabling again after the copy-over script has finished | 15:41 |
| xnox | mgz: https://wiki.ubuntu.com/NewReleaseCycleProcess only has "notify to restart package-import" Can you expand that check-list with other steps that are needed to be done? | 15:49 |
| mgz | wiki/sso is being slow over letting me log in... | 15:54 |
| mgz | xnox: done. also updated ReleaseProcess page | 16:07 |
| xnox | mgz: cjwatson: let's see if I can do those bzr importer tasks this release. | 16:37 |
| * xnox has access to jubany ;-) | 16:37 | |
| mgz | sounds good to me | 16:41 |
| === deryck is now known as deryck[lunch] | ||
| === yofel_ is now known as yofel | ||
| === matsubara is now known as matsubara-lunch | ||
| === deryck[lunch] is now known as deryck | ||
| === matsubara-lunch is now known as matsubara | ||
| === wedgwood is now known as wedgwood_away | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!