[01:12] hey, this might be a dumb question but how do I make a version higher than 2~svn###? The package is not from SVN but it is newer [01:13] 2.1 would not be accurate, its still 2.0, should it be 2-1~git###? I don't really get LP's versioning [01:15] would 2ppa1 superseed 2~svn###? [01:21] I mean, how can a git source superseed a svn source? [01:22] 2 is > 2~ [01:23] quatzal: http://www.debian.org/doc/debian-policy/ch-controlfields.html#s-f-Version describes the interpretation of Debian package versions [01:23] Launchpad doesn't have its own versioning scheme [01:24] thanks guys, and can a git package superseed a svn package? in this case its because the svn was abandoned [01:25] quatzal: The ~svnXXXX style is just a convention that some people use for svn versions. [01:25] it's not mandated, so you can always adjust the convention such that a git version supersedes an svn version, sure [01:26] thanks wgrant [01:26] :) [10:12] wgrant: in what way? my current patch set doesn't do anything particular about PPAs - I basically made DS lookup handle the alias, and put symlink creation at the end of the publisher [10:15] wgrant: I know one of Rick's goals was to make it easier for developers to stay current by default so that we stop having to separately persuade so many people to move forward, so that does seem to suggest making aliases work for PPAs too [10:30] cjwatson: Right, but we'll want to think about PPAs before we move anybody onto something unmaintainable, I suspect [10:36] Still not totally sure what you mean [10:37] cjwatson: We don't want to convince everyone to move onto some primary archive devel alias, only to discover that it's not supportable for PPAs. [10:38] So, my current patch will create the symlinks for PPAs too, although there's no UI for it or anything [10:38] I'm certainly open to discussion about whether that's a good thing [10:38] Except that we don't republish each PPA every time. [10:40] Hmm. [10:41] Just so I have numbers (not as a proposal), roughly how long does it take to run a no-op publish on all PPAs? [10:41] PPAs don't get the new series cloned [10:41] So that doesn't even necessarily make sense at all [10:41] Um [10:41] Probably a while [10:41] The publisher is pretty terrible [10:42] Maybe an hour or two [10:42] But then most of them will 404 [10:42] And apt will be sad. [10:42] I suppose one possible model would be to remember whether something has been uploaded to "next" rather than "saucy", and cause those to be cloned, or something ... [10:42] But I must go and help with children [10:42] (And I'm not sure that makes much sense) [10:42] Yeah [10:43] But it's almost definitely not symlink-simple for PPAs.