[00:04:11] 3MediaWiki extensions / 3WikidataRepo: It should be possible to add a label in any eligible language - 10https://bugzilla.wikimedia.org/59905#c9 (10Marius Hoch) Adding $wgExtraLanguageNames = array( 'rwr' => 'मारवाड़ी' ); to my local settings made rwr usable and it worked like a charm (couldn't identify anyt... [00:05:32] (03PS1) 10Aaron Schulz: Use slave DB connections for LocalFile cache misses [core] - 10https://gerrit.wikimedia.org/r/129079 [00:06:03] ori: that's been overdue ;) [00:06:08] * AaronSchulz heads to the office [00:06:19] to the office? [00:06:40] stuff happens after 5 sometimes ;) [00:16:58] 3MobileFrontend / 3beta: Talk broken on pages which don't have talk pages - 10https://bugzilla.wikimedia.org/64268 (10Jon) 3NEW p:3Unprio s:3normal a:3None Visit http://en.m.wikipedia.beta.wmflabs.org/wiki/Forty-seven_Roninsasasassa?mobileaction=beta Click talk See infinite ajax loader. Uncaught Ty... [00:18:10] (03CR) 10Ori.livneh: "I submitted to make compileValue public and poked upstream to accept it. Let's give it a coupl" [core] - 10https://gerrit.wikimedia.org/r/129023 (owner: 10EBernhardson) [00:19:26] 3VisualEditor / 3Editing Tools: VisualEditor: [Regression pre-wmf1] The first letter of the reference text disappears after typing in the reference dialog box - 10https://bugzilla.wikimedia.org/63909 (10Gerrit Notification Bot) 5REO>3PAT [00:19:26] 3VisualEditor / 3Editing Tools: VisualEditor: [Regression pre-wmf1] The first letter of the reference text disappears after typing in the reference dialog box - 10https://bugzilla.wikimedia.org/63909#c13 (10Gerrit Notification Bot) Change 129081 had a related patch set uploaded by Esanders: Prevent SurfaceWi... [00:20:26] 3MobileFrontend / 3beta: Talk broken on pages which don't have talk pages - 10https://bugzilla.wikimedia.org/64268#c1 (10Bingle) Prioritization and scheduling of this bug is tracked on Mingle card https://wikimedia.mingle.thoughtworks.com/projects/mobile/cards/1954 [00:22:33] (03CR) 10Legoktm: "Nope, that should be it!" [extensions/TitleBlacklist] - 10https://gerrit.wikimedia.org/r/123128 (owner: 10Gerrit Patch Uploader) [00:25:43] 3MobileFrontend / 3Hygiene: Rewrite ImageOverlay to use OverlayNew - 10https://bugzilla.wikimedia.org/64269 (10Jon) 3NEW p:3Unprio s:3normal a:3Jon Rewrite existing overlays to use OverlayNew [00:29:56] 3MobileFrontend / 3Hygiene: Rewrite ImageOverlay to use OverlayNew - 10https://bugzilla.wikimedia.org/64269#c1 (10Jon) This is the only overlay to use the old code given https://gerrit.wikimedia.org/r/129083 is merged. [00:30:34] (03CR) 10EBernhardson: [C: 032] Update to localCache should sync with bufferCache [extensions/Flow] - 10https://gerrit.wikimedia.org/r/127864 (owner: 10Bsitu) [00:31:17] (03Merged) 10jenkins-bot: Update to localCache should sync with bufferCache [extensions/Flow] - 10https://gerrit.wikimedia.org/r/127864 (owner: 10Bsitu) [00:31:41] 3MediaWiki extensions / 3Flow: Flow: after I hide/unhide, the inverse action isn't in the action menu until I reload the page - 10https://bugzilla.wikimedia.org/64200#c2 (10Gerrit Notification Bot) Change 127864 merged by jenkins-bot: Update to localCache should sync with bufferCache https://gerrit.wikimedi... [00:32:12] 3Wikimedia / 3wikibugs IRC bot: Wikibugs does not always report the real name instead of e-mail prefix when reporting on IRC - 10https://bugzilla.wikimedia.org/18831#c21 (10MZMcBride) \o/ [00:32:57] (03PS5) 10EBernhardson: Resolve complex arguments to LESS helper functions [core] - 10https://gerrit.wikimedia.org/r/129023 [00:41:28] 3VisualEditor / 3Editing Tools: VisualEditor Mobile: "Success! Your edit was saved" message stays in the VE mode unless user scrolls down the page - 10https://bugzilla.wikimedia.org/64270 (10ryasmeen ) 3NEW p:3Unprio s:3normal a:3None Created attachment 15174 --> https://bugzilla.wikimedia.org/atta... [00:45:27] 3MediaWiki extensions / 3Flow: Flow: renderHeader fatal error visiting Flow page without a header - 10https://bugzilla.wikimedia.org/62159 (10bsitu ) 5REO>3RES/FIX [00:46:27] 3VisualEditor / 3Mobile: VisualEditor Mobile: "Success! Your edit was saved" message stays in the VE mode unless user scrolls down the page - 10https://bugzilla.wikimedia.org/64270 (10ryasmeen ) [00:48:28] 3MobileFrontend / 3stable: MediaViewer doesn't use LoadingOverlay - 10https://bugzilla.wikimedia.org/64271 (10Jon) 3NEW p:3Unprio s:3normal a:3None Clicking on an image doesn't instantly show an overlay or a spinner so it is not clear that anything has happened. [00:50:26] 3MobileFrontend / 3stable: MediaViewer doesn't use LoadingOverlay - 10https://bugzilla.wikimedia.org/64271#c1 (10Bingle) Prioritization and scheduling of this bug is tracked on Mingle card https://wikimedia.mingle.thoughtworks.com/projects/mobile/cards/1955 [00:52:35] TimStarling: ping [00:58:37] (03CR) 10Ori.livneh: "My pull request was accepted, so you can amend your patch to update lessc in libs/ rather than have to hack around the visibility issue wi" [core] - 10https://gerrit.wikimedia.org/r/129023 (owner: 10EBernhardson) [01:05:42] 3MediaWiki extensions / 3MassMessage: Use ContentHandler based spamlist for MassMessage instead of parserfunction - 10https://bugzilla.wikimedia.org/61255#c1 (10wctaiwan ) *** Bug 62601 has been marked as a duplicate of this bug. *** [01:05:42] 3MediaWiki extensions / 3MassMessage: Use a ContentHandler-based backend for MassMessage - 10https://bugzilla.wikimedia.org/62601#c3 (10wctaiwan ) 5NEW>3RES/DUP *** This bug has been marked as a duplicate of bug 61255 *** [01:13:05] (03PS1) 10MaxSem: Unclusterfuck per-char and -sentence trimming from ApiQueryExtracts [extensions/TextExtracts] - 10https://gerrit.wikimedia.org/r/129088 [01:13:19] (03CR) 10jenkins-bot: [V: 04-1] Unclusterfuck per-char and -sentence trimming from ApiQueryExtracts [extensions/TextExtracts] - 10https://gerrit.wikimedia.org/r/129088 (owner: 10MaxSem) [01:13:56] 3MobileFrontend / 3stable: Do not redirect undo edits to mobile editor - 10https://bugzilla.wikimedia.org/64256#c2 (10Jon) 5NEW>3RES/WON There is no editor on the mobile site that supports undo. Redirecting to desktop would be an experience we should avoid. Undo for mobile is planned. Let's do that inst... [01:14:43] (03PS2) 10MaxSem: Unclusterfuck per-char and -sentence trimming from ApiQueryExtracts [extensions/TextExtracts] - 10https://gerrit.wikimedia.org/r/129088 [01:14:47] gwicke: you want to talk about language choice? [01:15:01] I've just been talking through it a bit with robla [01:15:33] I don't think we really want to do python in the main request flow [01:16:07] and we have to consider our resources: we have lots of JS devs, lots of PHP devs and limited resources in Java and C++ [01:17:03] I would like scala to be seriously considered, as a fully featured language with an easy learning curve from JS [01:17:31] TimStarling, yes [01:17:39] I think the small feature set of JS makes it less appropriate for projects above 10k lines [01:18:09] for small services like rashomon, JS should be ok [01:18:09] I do like some bits of scala too [01:18:27] the main issue I see with it is that the IO library is fairly fragmented [01:18:39] the standard library is all blocking [01:19:09] that's fine for number crunching, but not so nice for highly concurrent IO [01:19:17] similar to Python really [01:19:30] the performance difference is also not huge, and keeps shrinking [01:19:34] well, blocking IO is fairly easy for devs to understand [01:19:54] I agree, but none of the languages we discussed implements it efficiently [01:20:22] (03PS6) 10EBernhardson: Resolve complex arguments to LESS helper functions [core] - 10https://gerrit.wikimedia.org/r/129023 [01:20:34] Erlang, Go and Haskell otoh all use an event loop below synchronous-looking code [01:20:47] it comes down to resources again [01:21:13] and use case [01:21:21] yes [01:21:39] for number crunching I'd pick scala or C++11 [01:22:02] with scala being about 2x faster than JS, and Cpp again 2x faster than scala [01:22:13] but I mean that a more difficult programming model demands a higher skill level from developers [01:22:34] and highly skilled developers are a resource [01:22:55] <^d> Scala is one of the languages in the "7 languages in 7 weeks" book I picked up. [01:23:00] for better or worse developers these days tend to be fairly familiar with event-based programming and CPS [01:23:43] in a language where all IO is async that actually works fairly well [01:24:06] there are also combinators that help to keep the code readable [01:24:18] https://github.com/caolan/async for example [01:24:24] or promises [01:26:03] the main reason for JS is code sharing and devs though [01:26:46] and performance as far as compact dynamic languages go [01:27:32] maybe we should be thinking about pilot projects for scala [01:27:42] I'm not familiar with it at all [01:28:05] I'm very familiar with C++, and I am pretty sure we don't have the capacity to do large projects in it [01:28:11] I have seen some talks and read some more about it on the weekend [01:28:20] but have never written anything in it [01:28:37] the different concurrency primitives and the blocking issue scare me off a bit [01:29:09] actors for example have special message passing primitives that do the right thing, but normal IO calls will just block the worker thread [01:29:24] <^d> "In this book you’ll get a hands-on tour of Clojure, Haskell, Io, Prolog, Scala, Erlang, and Ruby." [01:29:36] <^d> You could also retitle this book "How to become a hipster programmer in 7 weeks" [01:29:45] ;) [01:29:50] the problem with C++ is that when things go wrong, they go badly wrong [01:30:02] Surely "hipster programmer" isn't a hipster-y enough term. [01:30:04] it's really easy to create bugs that are virtually impossible to fix [01:30:05] yeah, especially in template metaprogramming [01:30:11] 3MediaWiki extensions / 3MultimediaViewer: MultimediaViewer does not display the caption used in the article when used in - 10https://bugzilla.wikimedia.org/62518#c3 (10Gerrit Notification Bot) Change 129089 had a related patch set uploaded by Gergő Tisza: Show caption on thumbnails https://gerrit... [01:30:11] 3MediaWiki extensions / 3MultimediaViewer: MultimediaViewer does not display the caption used in the article when used in - 10https://bugzilla.wikimedia.org/62518 (10Gerrit Notification Bot) 5NEW>3PAT [01:30:15] but C++11 really has some nice stuff [01:30:19] I use thunderbird as my main email client [01:30:27] it crashes about once every day or two [01:30:51] afaik the issue there is more in xpcom [01:30:51] I use Outlook 2011 for Mac. It's quite stable. [01:31:09] circular references from C++ to JS and back etc [01:31:23] core dumps are completely useless, something corrupted memory at some point, then maybe hours or days later, something crashes because of it [01:31:48] it does work for me, but is not as fast as mutt [01:31:52] the thunderbird devs had a crash sprint, they contacted me, asked if I had anything for them (since I had reported a backtrace) [01:32:16] I said sorry, no reproduction procedure, I'll let you know if I find anything useful [01:32:34] they closed the bug, maybe they have an "impossible" status [01:33:06] one data point re scala vs. js, single core: http://benchmarksgame.alioth.debian.org/u64/benchmark.php?test=all&lang=v8&lang2=scala&data=u64 [01:34:11] I don't really want scala because it's faster, I want it because it has more features, so I think it'll scale up to larger code sizes better [01:34:19] and vs. Hack: http://benchmarksgame.alioth.debian.org/u64/benchmark.php?test=all&lang=v8&lang2=hack&data=u64 [01:34:26] (03PS1) 10BryanDavis: Add git submodule update to hhvm build [integration/jenkins-job-builder-config] - 10https://gerrit.wikimedia.org/r/129090 [01:35:35] Parsoid is around 25k lines now [01:36:27] how many devs? [01:36:40] 4 1/3 currently [01:36:46] <^d> Mmmm, hack. [01:36:52] it's not many, compared to say MW core [01:36:56] plus some OPW and GSoC students [01:37:01] and the Wikia folks [01:37:28] the goal of services is to limit the complexity of each individual service [01:37:41] 3MediaWiki extensions / 3MultimediaViewer: MultimediaViewer does not display the caption used in the article when used in - 10https://bugzilla.wikimedia.org/62518#c4 (10Gerrit Notification Bot) Change 129089 merged by jenkins-bot: Show caption on thumbnails https://gerrit.wikimedia.org/r/129089 [01:37:44] also, at that size you probably want to move things out into modules [01:37:58] (03CR) 10BryanDavis: "Edit applied manually in Jenkins starting with build #997." [integration/jenkins-job-builder-config] - 10https://gerrit.wikimedia.org/r/129090 (owner: 10BryanDavis) [01:38:22] the libraries we are using (and contribute to) are another 137k lines [01:38:29] features like type safety start to give benefits when you have a lot of developers with limited familiarity with the code [01:39:27] if they are enforces at compile time, then yes [01:39:32] *enforced [01:39:42] heh, you are comparing with PHP now? [01:40:05] PHP, Dart, .. [01:40:31] do you know about the Hack language? [01:40:34] JS is similar to Python re type enforcement [01:41:16] I heard that Hack is lead by a Haskeller at FB, on HHVM [01:41:18] wrt dart, you can just enforce the static type checker in jenkins. [01:41:23] <^d> http://hacklang.org/ - quick link for anyone who needs a primer on Hack. [01:41:39] Hack is not much of a language right now [01:41:49] yeah, moving target etc [01:41:59] FB added some extra type hinting to PHP, they did it a long time before they declared it to be a new language [01:42:17] it's touted as slightly faster than PHP on HHVM [01:42:29] the new thing that they did recently was they wrote a type checker in (wait for it) ocaml [01:43:00] I guess that's the sense in which it is a hack [01:43:00] I personally see ocaml as somewhat dirty in the functional language camp [01:43:07] side effect etc ;) [01:43:37] instead of building compile-time typing into the actual compiler, they wrote a whole new compiler which you are meant to use as a linter [01:44:02] then if the glorified linter fails you and lets something slip through, there is the usual runtime error layer inherited from PHP [01:44:37] a pragmatic choice for a research-y language I guess [01:44:37] and for extra fun, their ocaml program accepts invalid code with its simplified AST [01:45:06] gwicke: why do you think nonblocking IO is a requirement? [01:45:19] so back to services, I think that JS is a reasonable starting point [01:45:27] anyway, if you are comparing PHP and Scala, I don't think PHP has much going for it apart from inertia [01:45:53] although I think inertia is a serious force, it represents a huge amount of money invested in hiring and training [01:45:57] we can always move individual services to something else [01:46:19] gwicke: lightweight processes as in erlang provides a better programming model anyway for distributed systems [01:46:20] if you are comparing PHP with JS, then PHP has a lot going for it -- it has more features [01:46:34] ori, blocking IO implemented on lightweight threads or an event loop is sadly rare [01:46:38] * ^d is enjoying writing Java again [01:46:48] gwicke: not really [01:46:55] erlang, Haskell and (afaik) Go are the exceptions [01:47:17] scala has actors, there's http://akka.io/ too [01:47:44] PHP is really quite feature-rich now, as of 5.3-5.5 [01:47:45] sure, but the IO libs are JVM (blocking) by default [01:48:01] still not a great choice if you like threads, of course [01:48:07] there are some async libs for http, but you are limited in which third-party libs you can use [01:49:21] yes, but if you construct your application from building blocks that don't have side effects and that communicate by copying data then you don't have to worry about blocking io [01:49:39] there is not much support for writing high-concurrency network services in PHP [01:50:31] ori, before I started hacking JS I spent a lot of time with Twisted [01:50:46] you end up reimplementing anything that involves IO to be non-blocking [01:50:57] twisted is the wrong approach, imo [01:51:06] even low-level stuff like ssl bindings or DNS resolution [01:51:10] so in short, it's a pain [01:51:13] and Python is slow [01:51:29] http://benchmarksgame.alioth.debian.org/u64/benchmark.php?test=all&lang=v8&lang2=python3&data=u64 [01:51:52] yes, and merge sort is elegant when implemented in haskell [01:52:04] that's not a good measure of the suitability of a language to real-world problems [01:52:25] <^d> TimStarling: Have you used Traits yet? [01:52:31] ^d: no [01:52:31] <^d> I haven't found cause to. [01:52:41] I actually like Haskell, just think that we don't have enough developers familiar with it [01:52:59] "eventlogging" currently consists of 11 separate processes on vanadium that communicate using zeromq as a messaging layer. there is no asyncronous io. or rather, there is plenty of asyncronous io, except it's all managed by the os or zeromq [01:53:11] the application code is naive and blocking [01:53:16] gwicke: I can understand ops using python, out of the request flow [01:53:29] at uni I even wrote the odd Haskell package: http://hackage.haskell.org/package/bits-atomic [01:54:04] <^d> I don't mind python. I can hack around in other people's python. Not my first choice when looking to start something new. [01:54:09] <^d> Maybe I just need more time with it. [01:54:12] I like Python as a language for readability etc [01:57:05] ori, it becomes interesting when the number of concurrent connections rises beyond a few hundred [01:57:50] and when you are using other services dynamically and in parallel [01:59:51] erlang does pretty well in c10k-type comparisons [02:00:27] my position is that the way out of foolang vs barlang impasse is to say that the extent to which we have to care about language choices represents a failure to specify requirements for services [02:00:35] similar to node.js afaik [02:01:20] well, REST services obviously have defined interfaces [02:01:32] if we articulate clear requirements that each service ought to be reliable, horizontally scalable, speaks {thrift,avro,datapack,protobuf,json} (choose one), and -- i know this last part won't be popular with you -- implements a management interface that integrates with salt [02:01:34] the implementation language does not matter [02:01:48] then go ahead and implement in whatever language rocks your boat, yeah [02:02:09] you can't implement every service in a different language [02:02:19] you could, but it has a cost [02:02:31] right [02:02:45] TimStarling: you need people to review your code and you need to integrate with existing services and leverage useful libraries [02:02:48] I should say, we are not going to implement every service in a different language, I am not supporting this idea [02:02:59] yeah, me neither [02:03:02] that has the effect of constraining the set of viable languages naturally, without having to have a hard requirement in place [02:03:26] 3VisualEditor / 3Editing Tools: 'Insert Media' in VE fails to search local repo, returns blank. - 10https://bugzilla.wikimedia.org/63989#c3 (10Patrick King) Upgraded mediawiki to 1.22.5. Installed parsoid and VisualEditor as new. Merged basic LocalSettings.php. Uploaded a new file. File and description... [02:03:35] if you consider only technical requirements, you could well end up with a different answer each time [02:03:38] for now I don't see a good reason to add more languages to the mix [02:04:02] (03PS7) 10EBernhardson: Resolve complex arguments to LESS helper functions [core] - 10https://gerrit.wikimedia.org/r/129023 [02:04:23] but resourcing is as important as technical considerations [02:04:37] yes, i agree with that [02:04:42] and the jvm does have that going for it [02:05:10] as far as reviewers in the foundation go, I doubt it [02:05:20] you'd be surprised [02:05:30] for scala? [02:06:00] my guess would be that we have more JS reviewers than Scala reviewers [02:06:34] well, code review is one dimension [02:07:48] a software component requires maintainers, packagers, testers, deployers, instrumentation, etc. and the runtime is more significant than the language for most of these [02:08:50] tho if we do choose to introduce scala and to groom an ecosystem around it it would probably make sense to have some organized training [02:10:06] recent services like parsoid, mathoid, the translation service, PDF rendering and Rashomon have all been done in JS [02:10:27] inertia [02:10:36] yeah, and devs [02:11:02] none of these are a full replacement for the components they set out to replace, and thus none are (as yet) fully successful [02:11:30] hehe [02:11:42] as far as Parsoid goes, I'm not aware of any bidirectional PHP parser [02:11:47] that's not a criticism of parsoid or any of the other items on your list, mind you -- i'm just not prepared to take our very short track record there as evidence that we've made all the right choices [02:12:23] SOA does at least remove some of the technical inertia, it just leaves a formidable amount of human inertia [02:12:48] it does make the decision per-service less critical [02:13:21] as each service is relatively small and can be ported to another language in a limited timeframe [02:13:50] yeah [02:13:59] I know that there is inertia behind JS [02:14:22] I just don't really like JS [02:14:45] I am looking for gentle ways to get out of it [02:14:51] TimStarling: there's always Coco ;) [02:15:16] or coffeescript [02:16:02] i was joking; coco is a mess, not even reaching coffeescript's standard of coherence. the one thing we have written in coco has proven to be an unmaintainable mess [02:16:52] I don't really like PHP either [02:18:13] but it's not so horrible that I want to stand in front of the train [02:18:15] ;) [02:18:15] you know, the train that metaphorically represents its terrible inertia [02:18:19] it's important to me that people can do creative work without having to get sign-offs on every technical choice, which is why i think the right balance entails having very high standards for interoperability / reliability / maintainability, suggesting to people that they are more likely to meet them if they leverage the expertise of their peers rather than go solo [02:19:20] but if they feel very strongly about implementing a service in SNOBOL, and if they are prepared to be held up to the standards i mentioned, then it's their choice [02:20:09] well, there are some issues around packaging etc to be addressed as a community [02:20:29] don't mention packaging, i break out in hives [02:20:30] but yeah, language choice is one of the big selling points [02:20:51] but yes, i agree [02:21:06] packaging is actually not so bad once you get into the habit [02:21:24] it's fairly repetitive for a set of similar services [02:21:28] i am with tim in that i am a bit amazed that you'd stand up for javascript as a backend language [02:21:34] it really is pretty awful [02:21:59] are you talking about the entire package, or just JS as a language? [02:22:11] 3Parsoid / 3serializer: Quotes and whitespace normalized in unedited tags - 10https://bugzilla.wikimedia.org/55461#c18 (10WhatamIdoing) Here's another diff that is probably the same bug: https://en.wikipedia.org/w/index.php?title=Rights&curid=51490&diff=603988349&oldid=601154438 In this one, an oddly... [02:22:21] and mind you, my technical trajectory has been front -> back, so i am a practiced hand at rebutting accusations that javascript is a toy language [02:22:23] I never said that JS was the best *language* on earth [02:22:41] and i can produce snippets of functional javascript that look elegant when considered from afar [02:22:57] but the lack of anything like a solid standard library is just so painful [02:23:22] we had relatively few issues with node [02:24:15] and probably wrote most backend JS in the foundation so far [02:25:00] but again, your use case might be different [02:25:52] http://stackoverflow.com/questions/4443597/node-js-execute-system-command-synchronously [02:25:56] just an example [02:26:12] I need in node.js function result = execSync('node -v'); that will synchronously execute the given command line and return all stdout'ed by that command text. [02:26:16] simple request, no? [02:26:35] answer: "it's fairly easy to do with node-ffi" [02:27:05] of course if you want to do synchronous io you're an idiot a priori as far as node is concerned [02:27:07] http://nodejs.org/api/child_process.html [02:27:40] yes, but node has readFileSync, right? because if your application can't do anything until it has read its config file, for example, it's annoying to use callbacks [02:28:01] but there's nothing equivalent for subprocesses afaik [02:28:07] I'm going for lunch [02:28:49] ori, I can see that this might be annoying for a script, but for a network server you normally want to do everything async [02:29:25] in that use case, the only reasonable sync bit I can think of is reading a config file during startup [02:29:58] it's not *that* hard to hook up a callback though [02:30:18] and you avoid typical deadlock issues with popen [02:30:27] heh [02:30:34] i'm not against people writing things in javascript [02:30:44] i'm just somewhat baffled as to why they'd want to inflict that on themselves [02:30:48] but to each his own :) [02:31:29] i gotta run too, fun conversation tho [02:31:52] if you know something that's actually better for highly concurrent network services & with a good developer mindshare, then let us know [02:32:15] (03CR) 10Spage: [C: 04-1] "use FlowActions, not the global" (031 comment) [extensions/Flow] - 10https://gerrit.wikimedia.org/r/125947 (owner: 10Legoktm) [02:37:11] 3MediaWiki / 3Uploading: The column us_props is missing from the mediawiki.uploadstash table with PostgreSQL - 10https://bugzilla.wikimedia.org/63893#c3 (10Jeff Janes) I only just now realized that this is already fixed in 1.23 in Change-Id: I057b1dd845a5ab7ae05f5597454981a6e2c12698 Do we want to backport... [02:53:40] (03PS1) 10EBernhardson: Update lessphp to 2cc77e3c7b [core] - 10https://gerrit.wikimedia.org/r/129093 [02:53:42] (03PS1) 10EBernhardson: Resolve complex arguments to LESS helper functions [core] - 10https://gerrit.wikimedia.org/r/129094 [02:54:07] (03Abandoned) 10EBernhardson: Resolve complex arguments to LESS helper functions [core] - 10https://gerrit.wikimedia.org/r/129094 (owner: 10EBernhardson) [02:54:41] (03PS8) 10EBernhardson: Resolve complex arguments to LESS helper functions [core] - 10https://gerrit.wikimedia.org/r/129023 [03:02:10] (03CR) 10Mattflaschen: [C: 031] "Reviewed and ready for merge, except that we have to coordinate deployment (so the config matches the code)." [extensions/GettingStarted] - 10https://gerrit.wikimedia.org/r/122856 (owner: 10Phuedx) [03:03:53] a [03:03:59] s/a// [03:05:05] (03PS7) 10Gerrit Patch Uploader: Add log for TitleBlacklist hits [extensions/TitleBlacklist] - 10https://gerrit.wikimedia.org/r/123128 [03:05:07] (03CR) 10Gerrit Patch Uploader: "This commit was uploaded using the Gerrit Patch Uploader [1]." [extensions/TitleBlacklist] - 10https://gerrit.wikimedia.org/r/123128 (owner: 10Gerrit Patch Uploader) [03:15:42] 3MediaWiki extensions / 3WikidataRepo: Time data-type inconsistently zero-pads year value (dates earlier than year 1000?) - 10https://bugzilla.wikimedia.org/64084#c2 (10gnosygnu ) This is a sensible suggestion, but the script was not written by me. I only happened to come across this issue on Catalan Wikiped... [03:30:04] (03PS2) 10Ori.livneh: Update lessphp to 2cc77e3c7b [core] - 10https://gerrit.wikimedia.org/r/129093 (owner: 10EBernhardson) [03:30:08] (03CR) 10Ori.livneh: [C: 032] Update lessphp to 2cc77e3c7b [core] - 10https://gerrit.wikimedia.org/r/129093 (owner: 10EBernhardson) [03:30:26] (03PS9) 10Ori.livneh: Resolve complex arguments to LESS helper functions [core] - 10https://gerrit.wikimedia.org/r/129023 (owner: 10EBernhardson) [03:31:18] (03CR) 10Ori.livneh: [C: 04-1] Resolve complex arguments to LESS helper functions (031 comment) [core] - 10https://gerrit.wikimedia.org/r/129023 (owner: 10EBernhardson) [03:35:11] 3MediaWiki / 3Parser: Local interwiki links with additional non-local interwiki prefixes are not interpreted as links - 10https://bugzilla.wikimedia.org/61357#c6 (10Gabriel Wicke) (In reply to This, that and the other from comment #5) > One of the main ues cases for this is cross-wiki code reuse/transclusion... [03:35:39] (03Merged) 10jenkins-bot: Update lessphp to 2cc77e3c7b [core] - 10https://gerrit.wikimedia.org/r/129093 (owner: 10EBernhardson) [03:37:16] (03PS10) 10EBernhardson: Resolve complex arguments to LESS helper functions [core] - 10https://gerrit.wikimedia.org/r/129023 [03:38:32] (03PS1) 10Gerrit Patch Uploader: Document Special:Diff and Special:PermanentLink [core] - 10https://gerrit.wikimedia.org/r/129095 [03:38:39] (03CR) 10Gerrit Patch Uploader: "This commit was uploaded using the Gerrit Patch Uploader [1]." [core] - 10https://gerrit.wikimedia.org/r/129095 (owner: 10Gerrit Patch Uploader) [03:38:41] 3MediaWiki / 3Special pages: Document Special:PermanentLink and Special:Diff in the interface - 10https://bugzilla.wikimedia.org/45221 (10Gerrit Notification Bot) 5NEW>3PAT [03:38:41] 3MediaWiki / 3Special pages: Document Special:PermanentLink and Special:Diff in the interface - 10https://bugzilla.wikimedia.org/45221#c6 (10Gerrit Notification Bot) Change 129095 had a related patch set uploaded by Gerrit Patch Uploader: Document Special:Diff and Special:PermanentLink https://gerrit.wikime... [03:39:47] (03CR) 10PiRSquared17: [C: 04-1] Document Special:Diff and Special:PermanentLink (031 comment) [core] - 10https://gerrit.wikimedia.org/r/129095 (owner: 10Gerrit Patch Uploader) [03:42:05] (03PS2) 10PiRSquared17: Document Special:Diff and Special:PermanentLink [core] - 10https://gerrit.wikimedia.org/r/129095 (owner: 10Gerrit Patch Uploader) [03:42:41] (03CR) 10PiRSquared17: Document Special:Diff and Special:PermanentLink [core] - 10https://gerrit.wikimedia.org/r/129095 (owner: 10Gerrit Patch Uploader) [03:48:11] 3MediaWiki extensions / 3Popups: Hovercards: animated GIFs cause bottom border of horizontal hovercards to render incorrectly - 10https://bugzilla.wikimedia.org/63215 (10Prateek Saxena) 5PAT>3RES/FIX [03:52:27] 3MediaWiki extensions / 3SemanticForms: "The specified target page _ is invalid" after completing captcha with SemanticForms - 10https://bugzilla.wikimedia.org/61794#c4 (10Al Johnson) This may be related. I also got this as a value to a property field that took a default value from {{{PAGECREATOR}}}: pageI... [04:06:22] (03CR) 10Mattflaschen: "> The bug you are referring to regarding