[00:01] <cody-somerville> Although it's a provincial holiday so most of the people here in Ottawa who work for Federal government won't have it off, lol.
[00:01] <cody-somerville> oops. so much for family day.
[00:03] <cody-somerville> Does setting authorized_size to None disable the quota checking
[00:03] <cody-somerville> ?
[00:07] <cody-somerville> ah. It's not possible.
[00:07] <cody-somerville> However, weirdly distribution archives still have the authorized_size attribute but it is None.
[00:09]  * cody-somerville wishes running just a single doctest didn't take so long.
[00:46] <wgrant> Heh
[00:46] <cody-somerville> Actually, I guess it is possible.
[00:46] <cody-somerville> but it'll probably cause the upload processor to crash
[00:46] <wgrant> lpnet DB CPU usage has dropped by around 70% of a core in the last week, it appears.
[00:47] <wgrant> cody-somerville: I expect that it's possible.
[00:47] <wgrant> But the form might not allow it.
[00:47]  * wgrant tries.
[00:47] <cody-somerville> wgrant, it allows it
[00:47] <wgrant> Ah
[00:47] <wgrant> Maybe if a PPA has it set it will just crash
[00:47] <wgrant> You're probably right.
[00:47]  * wgrant checks the code.
[00:47] <cody-somerville>         MEGA = 2 ** 20
[00:47] <cody-somerville>         limit_size = self.archive.authorized_size * MEGA
[00:47] <cody-somerville> ^^ that'll crash
[00:47] <cody-somerville> You can't multiple None
[00:47] <wgrant> Indeed. Is that under an if archive.is_ppa?
[00:48] <wgrant>         if upload.is_ppa:
[00:48] <wgrant>             self.checkArchiveSizeQuota(upload)
[00:48] <wgrant> So yeah
[00:49] <cody-somerville> It probably shouldn't do that, it should probably check
[00:49] <cody-somerville> and the code fixed to see if authorized_size is None
[00:49] <cody-somerville> holy crap
[00:49] <wgrant> Yes, but that would make sense.
[00:49] <wgrant> And you know our policy about Soyuz.
[00:49] <cody-somerville> why does it take so long to do a commit?
[00:50] <wgrant> If a test commits we have to tear down the DB.
[00:50] <wgrant> Which takes a couple of hundred ms
[00:50] <cody-somerville> No. bzr commit
[00:50] <wgrant> If it doesn't commit, we can just abort the txn.
[00:50] <wgrant> Ah
[00:50] <cody-somerville> it's been going at it for like 5 minutes
[00:50] <cody-somerville> repacking stuff
[00:50] <wgrant> Ah, your repo hasn't been touched in a while?
[00:50] <cody-somerville> I just branched it like yesterday
[00:50] <wgrant> bzr repacks every order of magnitude of revisions, I believe.
[00:51] <wgrant> Hm, it's possible we've hit a threshold, but I wouldn't think so.
[00:51] <cody-somerville> It says 'Saving data locally - Stage: repacking chk:chk node n/n'
[00:51] <cody-somerville> it's already repacked the inventory
[00:51] <cody-somerville> and other stuff
[00:52] <cody-somerville> I've got another 14k chk nodes to go, lol
[00:52] <wgrant> Hm
[00:52] <wgrant> There's about 118000 revs in devel
[00:52] <wgrant> So I wouldn't expect a repack to happen around now.
[00:52] <cody-somerville> ugh. now it's repacking texts and there is 400k of them, lol.
[00:52] <wgrant> But maybe I misunderstand.
[00:53]  * cody-somerville has a party to go to and really wants to get this branch up for review before he goes :(
[01:19] <lifeless> wgrant: its OOM(base 10) size of pack, and every multiple of each OOM, of that size
[01:19] <lifeless> so 118000 will be doing a 1K rev pack, probably, but that should be near-instant
[01:20] <wgrant> Ah
[01:20] <lifeless> also a brand spanking new repo should be under all the thresholds and not pack at all
[01:20] <lifeless> however, 'branched yesterday' does not imply brand new repo
[21:58] <lcanas> hi guys, I'm having a look at the launchpadlib to get data from the platform. I would like to get the bugs for a given project and I can't manage to find out how .. any clues?