[12:08] <cjwatson> Could I have reviews of https://code.launchpad.net/~cjwatson/launchpad/+git/launchpad/+merge/378473 and https://code.launchpad.net/~cjwatson/lazr.restful/py3-no-cmp/+merge/378531 ?
[12:08] <cjwatson> Both short
[12:26] <cjwatson> ilasc: thanks
[12:26] <ilasc> +1
[12:26] <cjwatson> Landing pappacena's audit trail stuff, if I can get buildbot to not be stupid
[12:28] <ilasc> :)
[13:00] <cjwatson> tomwardill: Do you think I can make the index on OCIFile (build, layer_file_digest) unique?
[13:01] <tomwardill> I think so
[13:01] <cjwatson> Can't see why a given build would ever have more than one file with the same layer_file_digest
[13:01] <tomwardill> yeah, same
[13:01] <cjwatson> tomwardill: And in https://code.launchpad.net/~cjwatson/launchpad/+git/launchpad/+merge/374673, William asked if we might need to look up OCIFile by layer_file_digest globally (i.e. without a context build), for index purposes.  What do you think?  It'd be easy enough to add an OCIFile (layer_file_digest) index
[13:03] <tomwardill> I don't think it's required in the normal operation of things, but might be useful for garbo?
[13:10] <cjwatson> tomwardill: Do we not need it for sharing of layer files?  I've forgotten exactly how you were doing that
[13:11] <tomwardill> oh, good point!
[13:11] <tomwardill> yes, it's needed for that
[13:11] <cjwatson> The versions of your draft code that I could find didn't do sharing, I think
[13:12] <cjwatson> But yeah, presumably there are two uses: one is reassembling layers when doing a registry push, and the other is sharing
[13:15] <tomwardill> the end result is here: https://git.launchpad.net/~twom/launchpad/tree/lib/lp/oci/model/ociregistryclient.py?h=docker-registry-upload#n120
[13:15] <tomwardill> but there's a part in the build artifact retrieval that checks if the file already exists in the librarian
[13:16] <cjwatson> I found https://git.launchpad.net/~twom/launchpad/tree/lib/lp/archiveuploader/ociupload.py?h=docker-registry-upload but that seems to add unconditionally
[13:18] <tomwardill> indeed
[13:19] <tomwardill> fairly sure I did actually write that code...
[13:19]  * tomwardill is unsure where though
[13:19] <cjwatson> OK, no big deal, we knew that was draft
[13:45] <cjwatson> tomwardill: Could you have a look over the last four commits in https://code.launchpad.net/~cjwatson/launchpad/+git/launchpad/+merge/374673 and make sure you're OK with them before I land that?  The OCIRecipe.ociproject → OCIRecipe.oci_project renaming will require a bit of work on your end, I'm afraid, but the rest should be transparent
[13:45] <tomwardill> righto, will have a look
[13:53] <tomwardill> cjwatson: they look good to me
[14:02] <cjwatson> Thanks, will land after the standup
[14:46] <cjwatson> I'll sort out the test failures from Thiago's branches
[15:35] <cjwatson> Could I please have reviews of https://code.launchpad.net/~cjwatson/launchpad/+git/launchpad/+merge/378599 and https://code.launchpad.net/~cjwatson/launchpad/+git/launchpad/+merge/378600 ?  The first is a testfix
[15:38] <cjwatson> Urgh, and db-devel is broken now too, faff faff faff
[15:53] <cjwatson> OK, I also need a review of https://code.launchpad.net/~cjwatson/launchpad/+git/launchpad/+merge/378602 to fix db-devel
[16:04] <tomwardill> +1 on the latter
[16:05] <tomwardill> I'll have a look at the others in a mo
[16:05] <cjwatson> cheers
[16:05] <tomwardill> trying to extract myself from the mess I've made in the tests
[16:05] <cjwatson> Tell me about it ...
[16:06] <cjwatson> I should go for a walk though, doing too much at once
[17:48] <cjwatson> tomwardill: Have you had a chance to look at the other testfix at least?
[17:49] <cjwatson> Would like to unblock devel before I finish today
[17:49] <tomwardill> argh, sorry. Looking now
[17:50] <tomwardill> got lost in trying to constract tarfiles in python :)
[17:50] <tomwardill> +1 to both :)
[17:50] <cjwatson> tomwardill: Did you find lp.services.tarfile_helpers?
[17:51] <cjwatson> I realise that's in LP and not in launchpad-buildd, but it might give an idea
[17:51] <cjwatson> And thanks
[17:56] <tomwardill> aha, now I realise why I had a known file checked in
[17:56] <tomwardill> otherwise the digests change every time (due to file access times, etc)
[17:57] <tomwardill> can probably mock that out though
[17:57] <tomwardill> mock/fixture/appropriate
[18:00] <cjwatson> An alternative would be to commit documentation/scripting/whatever for building the test file.  I suspect mocking etc. will be less effort though
[18:00] <tomwardill> yeah
[18:01] <tomwardill> but it's beer, pizza and DnD night, so it'll have to wait till tomorrow :)
[18:03] <cjwatson> Heh.  Enjoy
[20:49] <jelmer> cjwatson: our plan for breezy is to create a 3.1 branch which will support python 2 and 3, and only support python 3 in trunk
[20:50] <jelmer> We'll merge any changes necessary for launchpad into the 3.1 branch
[20:50] <jelmer> (such as changes necessary for SVN support)
[22:00] <cjwatson> jelmer: OK, that sounds reasonable - thanks for letting me know
[22:00] <cjwatson> I started trying to see if I could work out what's up with subvertpy, but it's a slow debugging process and I'm not sure whether I'll be up to it
[22:04] <cjwatson> And this is in parallel with plugging away at ~everything else :-)
[22:23] <roadmr> subvertpy mangles mentally into "pervy" and downhill from there