[00:00:15] <ToAruShiroiNeko>	 its 1 AM so feel free to email that to me or the link
[00:03:55] <halfak>	 Hmm... Can't find it. 
[00:04:11] <halfak>	 Nothing special.  I'll have another look. 
[00:04:17] <halfak>	 But don't count on it. 
[00:07:25] <bmansurov>	 halfak: o/, blocked on time ;) I think i'll have a little time on Friday to do more.
[00:07:43] <halfak>	 Great!  :) 
[00:08:02] <halfak>	 Sorry to hit you with a brain dump.  
[00:08:26] <bmansurov>	 no problem
[00:28:47] <halfak>	 I'm got it!
[00:28:59] <halfak>	 I don't know how to fix it, but I have it in a corner. 
[01:40:19] <halfak>	 OK.  Here: https://github.com/halfak/python-sphinx-break-stuff
[01:40:25] <halfak>	 A demonstration of the problem. 
[01:40:28] <halfak>	 Now to file a bug. 
[14:41:43] <Amir1>	 o/ halfak
[14:41:59] <halfak>	 o/ Amir1 
[14:42:14] <halfak>	 Wanna see a bug that came straight from hell and ate 7 of my hours?
[14:42:22] <halfak>	 https://github.com/sphinx-doc/sphinx/issues/2233
[14:45:06] <halfak>	 This is the first time I have found a "compiler bug" in a long time. 
[14:45:18] <Amir1>	 :(
[14:45:21] <halfak>	 Obviously sphinx isn't really a compiler, but it's used by a lot of people. 
[14:45:28] <halfak>	 So similar rules should follow. 
[14:45:44] <halfak>	 I boldly assumed that I was doing something wrong for the first 4 hours. 
[14:46:40] <halfak>	 Then for the next two I whittled the problem down until I could figure out the simplest change that differentiated between expected and non-expected behavior. 
[14:46:58] <halfak>	 Last hour was documentation and trying to figure out how to even *report* the bug in an understandable way. 
[14:47:18] * halfak had a little fun with "from do_things import break_stuff"
[14:47:41] <Amir1>	 OK, so we have to wait for a while until it's fixed?
[14:47:47] <halfak>	 So, we have a fun decision to make now. 
[14:47:55] <halfak>	 1. Fix it ourself or wait
[14:48:12] <halfak>	 2. Merge 1.0.0 because it's otherwise good or wait until the fix is ready
[14:48:36] <halfak>	 So it turns out that this bug makes it so that *you can't read about a single feature in the documentation!*
[14:49:20] <halfak>	 All features fall into this buggy pattern. 
[14:53:25] <Amir1>	 halfak: Is there any kind of workaround/fast fix?
[14:53:47] <halfak>	 Amir1, I haven't found one.  I submitted that bug report right before calling it quits yesterday. 
[14:53:56] <halfak>	 Just signed on 30 minutes ago today. 
[14:54:10] <halfak>	 So, it's worth having a look at sphinx internals. 
[14:54:30] <halfak>	 I'm not sure where to get started though. 
[14:54:40] <halfak>	 I wish they had an IRC channel so I could ask for some clues. 
[14:55:30] <halfak>	 It's a pretty weird bug, so I'm skeptical we'll be able to work it out without a massive time investment.
[14:56:00] * Amir1 thins
[14:56:02] <halfak>	 Then again.  Autodoc is one file and it is only 1600 lines (for being a whole extension)
[14:56:03] <Amir1>	 *thinks
[14:56:07] <halfak>	 https://github.com/sphinx-doc/sphinx/blob/master/sphinx/ext/autodoc.py
[14:56:19] <Amir1>	 I'm reading the bug report so maybe I can find a fast fix
[14:56:28] <halfak>	 :) 
[14:56:45] <halfak>	 BTW, we're on schedule to kick off the edit types campaign today. 
[14:56:57] <Amir1>	 yay!
[14:57:01] <Amir1>	 \o/
[14:57:04] <halfak>	 We finished the pilot.  I'm just finishing notes on how to change the taxonomy. 
[14:58:56] <Amir1>	 awesome
[14:59:02] <Amir1>	 I don't know how to help
[14:59:11] <Amir1>	 I should work on the KDD paper too
[14:59:35] <halfak>	 +1  back to KDD paper. 
[14:59:46] <halfak>	 Let me consider our options re. documentation for a bit. 
[15:00:01] <halfak>	 We can hard-code these docs in the short-term.  
[15:00:18] <halfak>	 It'll be a mess and we won't want to maintain it, but maybe a bug fix will come soon. 
[15:01:36] <Amir1>	 maybe someone find a solution and tell us in the bug report
[15:01:57] <Amir1>	 (making a question in stackoverflow wouldn't be bad)
[15:02:10] <halfak>	 ^ That's a good point. 
[15:02:27] <halfak>	 It's a hard question to ask about, but maybe
[15:03:24] <Amir1>	 btw tomorrow is Wikipedia's 15 birthday and I have a presentation (or two): this is slides for one of them https://docs.google.com/presentation/d/1jKyVWnvOA41RAPyaVHoWUarcWeL_6Gug50DTe_62UnQ
[15:03:47] <Amir1>	 I will talk about ORES a little bit
[15:04:23] <halfak>	 \o/
[15:04:36] <halfak>	 Technical III comes out of nowhere :P 
[15:05:33] <Amir1>	 :))))
[15:07:04] <Amir1>	 I fixed it
[15:07:12] <Amir1>	 thanks :)
[15:07:54] <halfak>	 :D
[15:10:09] <Amir1>	 halfak: one of Arab friends came back to me and said it's a lot of work but he will do it ASAP
[15:10:55] <halfak>	 \o/  I really appreciate it.  Any way we can give credit would be awesome.  
[15:11:32] <Amir1>	 sure but I'm not sure they are technical enough
[15:11:44] <Amir1>	 I'll mention his name in the code
[15:11:46] <Amir1>	 :)
[15:12:24] <Amir1>	 legoktm: hey, do you have time to check this? https://gerrit.wikimedia.org/r/#/c/263184/
[15:12:27] <Amir1>	 please
[15:12:49] <Amir1>	 halfak: can you check PR #4 in edit quality?
[15:13:01] <halfak>	 Amir1, it might be cool at some point to go through our history and build up a list of people who have contributed to the project.  I think that we'd probably have ~50 people. 
[15:13:21] <Amir1>	 +1
[15:13:43] <Amir1>	 Let's do after finishing the IEG
[15:13:49] <Amir1>	 or in the final reports
[15:14:35] <halfak>	 +1
[15:14:46] <halfak>	 Oh yeah!  I forgot about our labelers.  We might push 100 :) 
[15:15:25] <Amir1>	 we should classify them
[15:15:38] <halfak>	 +1
[15:16:00] <halfak>	 Man.  Reviewing this code, I am imagining a great big mess when trying to do both pre-labeling *and* probably-damaging revert in the same script. 
[15:16:28] <halfak>	 Looks like this script is good for the probably-damaging-revert case and that's how it should stay :) 
[15:17:14] * Amir1 is reading https://meta.wikimedia.org/wiki/Grants:IEG/Revision_scoring_as_a_service/Renewal/Final
[15:17:57] <Amir1>	 halfak: what do you suggest? two separate script?
[15:18:02] <halfak>	 Looks like that's a copy-paste of our old report
[22:11:06] <Amir1>	 awight: hey, are you around?
[22:11:23] <Amir1>	 halfak: I just changed the patch based on your notes
[22:11:36] <halfak>	 \o/  I have some changes now too
[22:14:08] <awight>	 Amir1: hello!  Thank you for the great work on the extension
[22:14:19] <awight>	 Sorry I haven't had time to review, yet...
[22:14:26] <Amir1>	 it's okay
[22:15:38] <Amir1>	 After that I want to work on letting people choose the threshold and also disappear "r" when an edit is marked as patrolled
[22:15:49] <Amir1>	 it's probably straightforward
[22:16:11] <Amir1>	 awight: Can you tell someone review it?
[22:21:57] <awight>	 Amir1: I don't have anyone to ask... you could ping legoktm perhaps
[22:22:20] <awight>	 fwiw, I think the interaction with patrolling is much higher priority than an adjustable threshold.
[22:23:38] <Amir1>	 okay :)
[22:23:46] <Amir1>	 I do the patrolling first
[22:24:02] <Amir1>	 thank you for working on this extension
[23:17:38] <halfak>	 Amir1, doing a test run of this script (extract_managing) and something is definitely broken.
[23:18:13] <halfak>	 We're not detecting any reverts at all. 
[23:18:19] <halfak>	 I'm checking our constraints now. 
[23:21:01] <halfak>	 Make sure you get pull before any changes. 
[23:21:07] <halfak>	 I just pushed some fixes. 
[23:58:39] <halfak>	 IT WORKS
[23:58:49] <halfak>	 Amir1, I'll have an update to your PR shortly
[23:58:52] <halfak>	 Seems to work great