[10:50] <Riddell> I've never run new-source before but I'm doing so now, is there anything I should look out for?
[10:51] <Riddell> if I feed it into sync-source.py it seems to run for a few packages before running into errors, it'll take quite a few runs (removing packages already done/errored) to complete
[11:16] <ajmitch> Riddell: oh, about sync-source.py - I've got a local change to LP that I wanted to run past some archive admins, which is special-casing -XfakesyncY revisions where the upstream versions in ubuntu & debian are the same
[11:16]  * ajmitch isn't sure how useful it'll be for archive admins, but if it is, I'll submit it to LP
[11:19] <Riddell> ajmitch: you mean packages that have "fakesync" in their version numbers?
[11:21] <ajmitch> yes
[11:21] <ajmitch> a few have been uploaded recently
[11:22] <ajmitch> it'd just be for convenience, bailing out with a decent error message insted of choking on the orig.tar.gz
[11:23] <ajmitch> this came up because we were checking whether packages with 'build' in the version number were a special case, or just those with 'ubuntu'
[11:34] <Riddell> ajmitch: seems interesting
[11:38] <ajmitch> perhaps, I'll pass it by one of the others before I put a merge proposal in :)
[11:42] <cjwatson> Riddell: new-source> could you stop pleasE?
[11:43] <cjwatson> Riddell: you need to first filter it for things that have been removed from Ubuntu but not blacklisted
[11:43] <cjwatson> Riddell: those all need to be checked by hand
[11:43] <cjwatson> ajmitch: as I said on ubuntu-devel, why not just use 'ubuntu' there?
[11:44] <cjwatson> well, I suppose it does have definitely distinct semantics, but in any case, it'd be nice if people who disagree followed up to my mail
[11:44] <cjwatson> Riddell: if you've already flushed a pile of stuff through, then please go back and check the list to see which ones shouldn't have been reintroduced
[11:45] <Riddell> cjwatson: I have, they're in New queue now
[11:45] <Riddell> how do I check to see which ones shouldn't have been reintroduced?
[11:47] <ajmitch> cjwatson: it's a fairly minor difference wrt ubuntu vs fakesync, it'd let those package be automatically synced without having to be checked through when a new upstream version hits
[11:47]  * ajmitch had missed which mail you were referring to, though
[12:08] <cjwatson> Riddell: you have to trawl through publication histories in LP, and then look through the reasons why they were removed
[12:08] <cjwatson> this takes a while and it's why I'd only done part of them; most of the ones that remain are either relatively new, or difficult
[12:09] <Riddell> cjwatson: and the file new-source-formerly-in-ubuntu is part of that?
[12:09] <cjwatson> ajmitch: if you're making sure the autosync only happens when the upstream versions are different and that otherwise we silently skip those packages without breaking the autosync run, then it sounds like a worthwhile improvement
[12:10] <ajmitch> cjwatson: that was the intent
[12:10] <cjwatson> Riddell: that was my list of packages from new-source that had a prior publication history as of whenever that file is dated, yes
[12:10] <cjwatson> Riddell: of course it won't take account of any packages that might have been reintroduced to Debian having previously been in Ubuntu before that
[12:10] <ajmitch> I don't know how silent you need it, it'll currently print a line about not updating
[12:10] <cjwatson> ... since then
[12:10] <cjwatson> doesn't need to be silent, just needs to not fall over
[12:10] <ajmitch> either way, it's only 10 lines of diff
[12:16]  * ajmitch will follow up with it tomorrow, it's not a part of LP that has tests at the moment
[12:18] <wgrant> ajmitch: You know to ignore the SyncSource tests which don't actually test the current implementation?
[12:19] <ajmitch> wgrant: afaik there aren't tests for scripts/ftpmaster-tools, which is where the change was made
[12:19] <ajmitch> though I'm not surprised that there are tests to be ignored
[12:20] <wgrant> ajmitch: Right. But there's another SyncSource implementation lurking around somewhere from years ago. it was mostly written and tested, but never completed.
[12:20] <ajmitch> wonderful
[14:22] <Riddell> cjwatson: so if the publication history is empty e.g. https://edge.launchpad.net/ubuntu/+source/beid/+publishinghistory that means it should be safe to let into the archive?
[14:27] <cjwatson> yes