[00:00] <nacc> wgrant: can you tell me in 2:8.35-3.2's changelog what the prior version was?
[00:00] <nacc> wgrant: if you have it handy
[00:05] <wgrant> nacc:  https://launchpadlibrarian.net/192002399/pcre3_8.35-3.2.dsc
[00:05] <wgrant>  https://launchpadlibrarian.net/192002400/pcre3_8.35.orig.tar.gz
[00:05] <wgrant>  https://launchpadlibrarian.net/192002401/pcre3_8.35-3.2.debian.tar.gz
[00:05] <nacc> wgrant: thanks!
[00:12] <nacc> wgrant: found a few more examples of this happening, but it only affets ubuntu's publishes because debian's history in lp doesn't go back far enough (i guess) ipvsadm 1:1.24-2 and netty 1:3.1.0.CR1-1
[00:13] <nacc> in those cases, `pull-lp-source` actually ends up failing because it uses the wrong dsc file for checksum verification
[00:13] <wgrant> nacc: I don't see a 1.24-2 in Ubuntu
[00:14] <wgrant> It goes 1.24+1.21-1.4ubuntu then 1:1.24-2
[00:14] <wgrant> Unless the orig differs? Surely not.
[00:14] <nacc> check, you're right
[00:14] <wgrant> Oh hey the orig differs
[00:14] <wgrant> Sweet
[00:14] <nacc> heh
[00:14] <nacc> so a different issue :)
[00:15] <wgrant> Ah, and the old orig was only in Hoary, so pre-Soyuz.
[00:15] <wgrant>  https://launchpadlibrarian.net/1273397/ipvsadm_1.24-1.dsc
[00:15] <wgrant>  https://launchpadlibrarian.net/1273398/ipvsadm_1.24.orig.tar.gz
[00:15] <wgrant>  https://launchpadlibrarian.net/1273399/ipvsadm_1.24-1.diff.gz
[00:16] <wgrant>  https://launchpadlibrarian.net/38886650/netty_3.1.0.CR1.orig.tar.gz
[00:16] <wgrant>  https://launchpadlibrarian.net/38886651/netty_3.1.0.CR1-1.diff.gz
[00:16] <wgrant>  https://launchpadlibrarian.net/38886653/netty_3.1.0.CR1-1.dsc
[00:16] <nacc> thanks!
[00:16] <nacc> and actually, even with my parent overrides possibilities for these two, i can't do anything, as pull() will fail, which we can't really handle -- as i have no contents to commit into the tree, i wonder if we will need to do something like -- we can't import this particular version support :/
[00:17] <nacc> smoser: --^
[00:19] <nacc> smoser: actually, probably the 'fix' for this is that hte importer can't do this on its own, but will fail and an admin has to go in and fix-up the history by hand (basically, with the files that wgrant provided and a git-dsc-commit --tree-only and git commit -p ... and then the importer can maybe pick up again from there
[00:19] <nacc> smoser: will look at it tmrw
[00:19] <wgrant> That might be reasonable. But you will probably also eventually find something that you can't import at all -- life finds a way.
[00:20] <nacc> yeah, we've gotten pretty close (at least for server pacakges in main) to importing everything
[00:20] <nacc> but yes, i'm sure we'll hit more issues as time goes on :)
[01:22] <nacc> wgrant: how stable are those URLs, that is -- if they are usable generally, could I implement a generic override for file shadowing -- so then i could say that particular publish of this srcpkg which I know will fail to import should use these urls instead (and still validate the tarballs against that dsc)?
[01:33] <wgrant> nacc: Those URLs won't ever change, unless we decide to start pruning old sources (which seems unlikely).
[01:33] <wgrant> We do prune old primary archive binaries after several years, but never old sources.
[01:35] <sarnold> just out of curiosity, how much storage is in the librarian? :)
[01:37] <wgrant> sarnold: It's backed by Swift nowadays, but it's only somewhere around 50TB.
[01:37] <wgrant> If we didn't aggressively prune PPAs, it would be a little bit larger...
[01:37] <sarnold> wgrant: that's less than I expected
[01:38] <wgrant> sarnold: We were very space constrained until 18 months ago, so we found some fairly inventive space-saving measures.
[01:42] <nacc> wgrant: ok, good to know! i'll see what i can devise tomorrow. Thanks for your help!
[01:45] <wgrant> nacc: I can obtain the librarian URLs for any primary archive package trivially, so feel free to ask if you need any more. We don't expose them through the API because dget wants them all in one dir (hence the /ubuntu/+archive/primary/+files redirect) and they don't work for private packages.
[01:45] <wgrant> The +files redirect works fine except when there are duplicate filenames :/
[09:39] <juliank> Hmm, the last message I received via email from https://bugs.launchpad.net/ubuntu/+source/apt/+bug/1637801 is #12
[09:39] <ubot5`> Ubuntu bug 1637801 in apt (Ubuntu) "Incorrect Russian translation of "apt list --upgradeable" results" [Undecided,Fix committed]
[09:40] <juliank> Well, actually I only have the emails from Gunnar Hjalmarsson (gunnarhj) in my gmail
[09:41] <juliank> Anyone has an idea why I only get half of the messages? I'm also subscribed to all apt bugs, but I seem to only get half of the bug reports as well
[09:43] <juliank> Apparently LP sends the emails to juliank@ubuntu.com which forwards to jak@jak-linux.org which forwards to julian.klode@gmail.com...
[09:44] <juliank> Now, I'm not sure where in there it gets lost
[09:50] <juliank> Hmm, apparently I'm not supposed to set ubuntu.com as my primary address. Strange that this worked at all then for the past years given bug 5292
[09:50] <ubot5`> bug 5292 in Launchpad itself "People setting preferred contact address to @ubuntu.com" [Low,Triaged] https://launchpad.net/bugs/5292
[09:57] <wgrant> juliank: LP doesn't manage the aliases, so we don't really know how that works -- Canonical IS (#canonical-sysadmin) generates the aliases using Launchpad profiles.
[09:58] <juliank> Right, I think that's what I understood from the bug report
[09:59] <juliank> I still wonder where half of the emails went. I switched LP to use my gmail directly know, but that looks a bit bad in the LP ui
[16:26] <nacc> wgrant: could i get the correct orig tarball for ipvsadm 1:1.24-2 (the one at https://launchpad.net/ubuntu/+source/ipvsadm/1:1.24-2 doesn't match the checksum listed)
[20:26] <nacc> cjwatson: --^ maybe you can provide that so I can close out my importer bug?
[21:34] <cjwatson> nacc: I believe it's https://launchpadlibrarian.net/15550377/ipvsadm_1.24.orig.tar.gz
[21:34] <cjwatson> yeah, that matches
[21:39] <nacc> cjwatson: thanks!