[00:19] <mwhudson> wgrant: general rebuild testing question
[00:19] <mwhudson> i did a rebuild test and millions of builds failed
[00:19] <mwhudson> once i've fixed the problems, is it easiest just to make a new ppa and start again?
[00:19] <wgrant> mwhudson: If the builds failed, you can just retry them.
[00:20] <wgrant> But you can't retry a build that was successful but wrong.
[00:20] <wgrant> You need to either upload a new version of the successful packages, or create a new PPA.
[00:20] <mwhudson> i guess there is a script somewhere to retry all failures?
[00:20] <mwhudson> i don't think i have any successful but wrongs this time
[00:21] <wgrant> Something like "for build in lp.archives.getByReference(reference='~some/ppa/path').getBuildRecords(build_status='Failed to build'): build.retry()"
[00:21] <wgrant> I don't know if there's an existing script for it.
[00:22] <mwhudson> cool thanks
[00:22]  * mwhudson gets back to installing headers somewhere builds can find them...
[02:17] <blr> wgrant: collander schemas are kind of nice, might be worth adding better validation to turnip's api using that at some point.
[02:17] <blr> err colander
[02:20] <wgrant> I've not looked at colander in depth.
[02:21] <blr> it lets you provide a validation schema for an endpoint and handles error messages appropriately
[20:30] <blr> morning
[23:20] <mwhudson> ppas for debian pls