[08:41] wgrant: sorry, what's a TTB? [08:41] I can follow the rest of that [08:41] tomwardill: TranslationTemplatesBuild [08:42] aha! [08:42] yes, that feels roughly analogous [08:42] thanks [08:42] Takes a branch, runs some predefined stuff on it, uploads translations to LP [08:42] So has no related Ubuntu series, even. [08:42] yeah, that makes sense [10:41] Agreed, I think that's the best approach [10:42] Ideally make distro_arch_series / distro_series / pocket properties of the build work based on that [10:44] yeah, that's what I think I've got [10:44] (ish, it's a bit of a mess atm, currently tidying up) [15:17] No url for because broke the chain. [15:17] well. [15:18] cjwatson: is that part of the work you did with the views? [15:23] tomwardill: one sec [15:23] * cjwatson peers at +activereviews [15:24] yeah, I just had a look through that [15:25] I could have sworn I did something here, but where [15:25] I definitely remember seeing a branch for it [15:25] because we had the discussion about the rebase chain becoming a tree [15:25] Maybe that was something else [15:26] Oh, I did OCI *project* views (4a1852da9d872d815daa0bed1dd36ef9bcc2f699) [15:26] I haven't done any recipe views, and they have no URLs defined at the moment [15:26] Do you need me to work something out there? [15:26] https://code.launchpad.net/~cjwatson/launchpad/+git/launchpad/+merge/376106 [15:26] ah, yes [15:27] cjwatson: ideally if you have time, as I suspect it'd be quicker for you to do it agian than me [15:27] OK [15:27] Let me fix buildbot first [15:28] cool, I'll work around this for now and continue implementing extraBuildArgs [17:50] buildbot issues were exposed by converting away from urllib.urlopen, but I think they're fundamentally a txfixtures bug. I've pushed https://github.com/testing-cabal/txfixtures/pull/13 which should fix that (test failures there are fixed by https://github.com/testing-cabal/txfixtures/pull/12), and I've poked Free about reviewing them [17:51] Ran 1 tests with 0 failures, 0 errors, 0 skipped in 1.038 seconds. [17:51] weeeeeee [17:51] extraBuildArgs for OCIRecipeBuildBehaviour. [17:51] Now to unpick the mess I've made, and ensure that it's actually the values that I wanted [17:55] cjwatson: that doesn't look like it was a whole lot of fun to find [18:39] tomwardill: Not exactly [23:21] lifeless: Thanks for the testing-cabal invite. Do you think I could have PyPI access to txfixtures? [23:21] cjwatson: of course [23:22] I wish pypi had teams [23:22] Yeah, totally [23:25] added you to a few random projects [23:26] Thanks, I like randomness [23:27] speaking of random, you should consider porting launchpad to rust [23:28] A nice small project [23:28] side project for when you're not busy [23:29] :) [23:29] but this is actually a serious suggestion [23:30] I've been watching the way you've been adding things in - more and more microservice style additions [23:30] which is great [23:30] I presume you're ported or nearly so to run live on py3 at this point [23:31] which means you'll have mypy gradual typing [23:32] anyhow, point is that rust productivity has been steadily climbing (as has golangs for fairness), but in the last 6-12 months I think things have gotten to the point where for production services, developer productivity now exceeds that of python [23:33] Released txfixtures 0.4.3 [23:34] so I think if Launchpad wants to be around for another 5-10 years, a sensible question to ask is what is the most developer efficient platform for it to be in, and a dead-end fork of an experimental branch of Zope probably isn't it :) [23:34] It's conceivable that we'd look at Rust for an experimental service at some point, though there's some politics too of course [23:34] the rest of the conclusions pretty much follow :P [23:34] I don't think it's really true that we're on a dead-end fork of Zope any more - I upgraded to newest versions of everything late last year [23:35] Certainly Zope has less mindshare and I want to get off e.g. zope.server [23:35] I retract my statements then [23:35] I had thought that a lot of the stuff we had was experiments that never gained traction [23:35] I don't disagree with everything you say, although there is the minor practical note that I haven't learned Rust yet :) [23:36] I think we are probably in a better situation upstream-wise than you remember [23:36] well done, thats fantastic news [23:37] At the moment I think we have two small zope.* forks [23:37] I think you will very much like rust [23:37] and hey - its Gnome aligned :P [23:38] zope.session to stick with an old session cookie algorithm, and zope.testrunner for a couple of tweaks to test ID handling and some changes to thread leak handling (I got all the subunit stuff back upstream again, which was the big chunk) [23:43] I think I take a more incrementalist attitude (it was ever thus); there's still a lot we can do to move LP onto things developers find familiar, and at the moment (modulo finishing the py3 port) I think we're in a good position to migrate to them gradually [23:44] while possibly looking to entirely different platforms for microservices [23:44] certainly any migration would have to be incremental :P [23:47] * cjwatson sleeps [23:47] night! thanks for the chat