[00:06:12] ^demon|away: have you been able to configure phpstorm for running the unit tests? [13:43:15] <^demon|away> ori: I don't use phpstorm, so no. [15:13:37] ^d: I tried searching for "label:Code-Review=+2,user=self" in Gerrit (and the same with variations on explicitly specifying me in place of "self"), which the docs seem to say should work, but it doesn't seem to. Any idea? [15:18:42] label:Code-Review+2,self ? [15:19:07] https://gerrit.wikimedia.org/r/#/q/owner:self+label:Code-Review%252B2%252Cself,n,z [15:19:14] Krenair: No luck on that one either. [15:19:18] WFM [15:19:35] ^ that link WFM too [15:19:36] oh, hmm, only with the owner: [15:20:02] anomie: I think the ,self only shows self-merges [15:20:18] anomie: https://gerrit.wikimedia.org/r/#/q/owner:self+label:Code-Review%252B2,n,z ? [15:20:22] anomie, perhaps try with a status:merged filter? [15:20:40] legoktm: With owner:self it would. But without owner:self? [15:20:45] Krenair: Ha, that seems to work! [15:20:57] "label:Code-Review<0" -> HTTP 500 ISE [15:21:08] "label:Code-Review<=-1" -> totally OK [15:21:25] https://gerrit.wikimedia.org/r/#/q/owner:self+label:Code-Review%252B2%252Cself,n,z only shows me self merges. [15:21:53] that still has owner:self legoktm [15:21:58] https://gerrit.wikimedia.org/r/#/q/status:merged+label:Code-Review%252B2%252Cself+-age:1y,n,z [15:22:21] ah [15:22:22] okay [15:22:25] https://gerrit.wikimedia.org/r/#/q/status:merged+label:Code-Review%252B2%252Cself,n,z should work to show you actual merges you've triggered [15:22:37] yep :D [15:23:17] Sadly, https://gerrit.wikimedia.org/r/#/q/status:merged+label:Code-Review%252B2%252Cself+OR+owner:self,n,z gives a 500 [15:32:58] * ^d sighs [15:33:00] <^d> f'ing gerrit [16:17:48] * hasharMeeting kicks to the GitHub greenest pastures [16:23:02] could I get a review on https://gerrit.wikimedia.org/r/#/c/207151/ ? [16:31:14] legoktm: can you remind me why we added the no-op logger in the first place? [16:32:34] bd808: originally we didn't want stuff logged locally, but people asked for it so I made global rename do that a while back, but didn't update user merge for it [16:33:03] k. I'll +2. looks like a simple change [16:34:14] thanks! [16:34:25] legoktm: Mind if we do our meeting in about a half an hour? I won't be available at the usual time [16:35:39] Keegan: I have to drive somewhere in exactly half an hour :/ would 11 PDT work? (1:30 from now) [16:35:57] That's fine, sir. [16:37:27] Thanks! [17:53:22] anomie: is there a phab task about getting sorted results from a generator-based query? [17:55:00] tgr: I'd probably have declined it, the SQL queries involved would make it difficult to make happen. For generator=search and generator=prefixsearch we now have the ability for returning an "index" from the generator that can be used client-side. [17:58:02] do you need to involve SQL? you can just sort the result in memory, I imagine [17:59:41] You'd have to preserve the original ordering of the data coming from the generator somewhere, do queries, merge results from various sources, then match up with the original data and sort based on that original ordering. [18:00:10] And then deal with the problem of resolved redirects, etc. [18:02:16] in my happy ignorance of API internals, that sounds relatively simple to me [18:03:45] generators always deal with pages, right? so just record page ids in the generator output and sort the final result by them [18:03:50] "Preserve" might be a bit of a pain. "Match up" somewhat more so, since the original data might have been Title objects, page ids, and/or rev ids. "Resolved redirects" are really a pain. [18:04:32] You might not have the page ids given to you, and one of those SQL queries is "look up needed page data from what you have, in a batch". [18:05:16] Looking up the page id for each of 5000 titles, one at a time, gets slow. [18:05:23] I see [18:07:02] anyway, if it does not fall into the "declined on sight" category, I'll just open a task and try to get familiar with the code [18:07:30] If you want to try to work on it, feel free. [18:11:43] legoktm: poke. [19:11:07] ok so I'm trying to read a text file with likely invalid unicode characters and running into [19:11:08] UnicodeDecodeError: 'utf-8' codec can't decode byte 0xf6 in position 28557: invalid start byte [19:11:30] and I'm just using a standard: with open('invalidusernames.txt') as f: text = f.read() [19:11:33] (using python3) [19:11:35] any ideas? [19:11:59] full code: http://fpaste.org/218750/14308531/raw/ [19:15:10] legoktm: what does "file -bi invalidusernames.txt" say? [19:15:25] text/plain; charset=unknown-8bit [19:15:45] how was it generated? [19:15:49] the file is also on terbium at /home/legoktm/invalidusernames.txt [19:15:59] kk, looking [19:16:22] it's the output of the checkUsernames.php script, which finds invalid usernames [19:17:36] I could modify the script to not output the invalid username, but it's pretty slow so I'd rather not do another run of it [19:18:10] can you find the user page for one of theseu sers? [19:18:16] so we can cross-reference the encoding in the file and the actual name? [19:18:20] ori: so...the script works under python2 :P [19:18:45] 'works' in that it's treating everything as ascii [19:18:51] we could find it, but you wouldn't be able to access it since they're invalid [19:19:11] well I'm just trying to get the user ids out of the file so it's good enough [19:20:47] legoktm: k. for the future: [19:20:52] with open('invalidusernames.txt', encoding='ascii', errors='ignore') as f: [19:20:54] in python3 [19:21:12] ah, thanks :D [21:50:15] anomie: I found yet another way to get a dump of gerrit activity: `ssh gerrit gerrit query --format=JSON -- owner:self status:merged -age:1y limit:10000`