[01:21:31] Change on 12en_wikipedia a page Wikipedia:Huggle/Feedback was modified, changed by Lowercase sigmabot III link https://en.wikipedia.org/w/index.php?diff=597308114 edit summary: Archiving 1 discussion(s) to [[Wikipedia:Huggle/Feedback/Archive 18]]) (bot [01:21:33] Change on 12en_wikipedia a page Wikipedia:Huggle/Feedback/Archive 18 was modified, changed by Lowercase sigmabot III link https://en.wikipedia.org/w/index.php?diff=597308116 edit summary: Archiving 1 discussion(s) from [[Wikipedia:Huggle/Feedback]]) (bot [10:04:34] [02huggle3-qt-lx] 07benapetr pushed 031 commit to 03master [+2/-0/±1] 13http://git.io/HwkyQg [10:04:35] [02huggle3-qt-lx] 07benapetr 038d4731b - docs new dia [10:30:08] Change on 12mediawiki a page Manual:Huggle was modified, changed by Justus Nussbaum link https://www.mediawiki.org/w/index.php?diff=916438 edit summary: added [[Category:Manual]] using [[w:Wikipedia:HotCat|HotCat]] [11:15:48] [02huggle3-qt-lx] 07benapetr pushed 031 commit to 03master [+0/-0/±2] 13http://git.io/nhqV5A [11:15:49] [02huggle3-qt-lx] 07benapetr 03287bbfb - made it slower so that it doesn't hurt eyes, there is no point in displaying this huge banned for 0.2 sec [11:33:20] [02huggle3-qt-lx] 07benapetr pushed 031 commit to 03master [+0/-0/±3] 13http://git.io/YxrT6A [11:33:21] [02huggle3-qt-lx] 07benapetr 0351d44e8 - lock page if it's not supposed to be reverted [11:57:29] [02huggle3-qt-lx] 07benapetr pushed 031 commit to 03master [+0/-0/±16] 13http://git.io/yDQivQ [11:57:30] [02huggle3-qt-lx] 07benapetr 03a8b40e0 - cleaned up configuration and renamed some of them [13:36:14] [02huggle3-qt-lx] 07benapetr pushed 031 commit to 03master [+0/-0/±12] 13http://git.io/4bDNNQ [13:36:15] [02huggle3-qt-lx] 07benapetr 03c987acf - fixes improved some things and renamed some confusing variables [13:36:41] se4598: if there was anything not clear in dev documentation let me know so I can extend it [13:37:05] it's hard to decide on these things from point of view of person who write almost all the code :P most of thigs look clear to me, even if they aren't [13:37:41] with docs you mean http://tools.wmflabs.org/huggle/docs/head/ ? [13:38:12] or the directory "documentation" in git? [13:38:53] Change on 12en_wikipedia a page Wikipedia:Huggle/Huggle3 Beta was modified, changed by Petrb link https://en.wikipedia.org/w/index.php?diff=597369536 edit summary: /* Known issues */ [13:40:20] se4598: website [13:40:22] [02huggle3-qt-lx] 07se4598 pushed 031 commit to 03master [+0/-0/±1] 13http://git.io/Xe2Jrw [13:40:23] [02huggle3-qt-lx] 07se4598 039621706 - fixed argument check at Core::MonthText(int n) [14:02:37] petan: How do I use line breaks in the (project) config file? [14:04:45] e.g. for dewiki the report user has to insert a new section, so aiv-ip or aiv-user has to contain a heading and also the section content (the report reason) [14:08:54] aiv-ip:* {{IPVandal|$1}} $2$3 ~~~~ [14:08:55] aiv-user:* {{Vandal|$1}} $2$3 ~~~~ [14:09:34] do I have to create a nice template which substitutions a new section? [14:09:42] as helper [14:09:47] Put there something like aiv-ip:* {{subst:HuggleAIV|$1}} $2 $3 ~~~~ [14:09:59] create template HuggleAIV that contains the heading [14:10:04] ok [14:10:39] Huggle will just literally post this string into report page [14:10:47] so if there is any mediawiki magic it will happen [14:11:31] and can you please look at that diff from yesterday: https://de.wikipedia.org/w/index.php?title=Wikipedia:Vandalismusmeldung&diff=prev&oldid=127963969 [14:12:02] that seems to be a not dected edit conflict [14:12:23] indeed [14:12:41] (that the report from Spuk968 is empty is because we don't have the vars set yet) [14:12:47] what version of huggle that user has? [14:13:07] 947... [14:13:09] mhm [14:13:10] interesting [14:14:00] mediawiki says that the section was added 2! seconds before that edit [14:14:30] (the previous edit) [14:15:10] meh,..... my edit adding that section was two seconds earlier that edit from Spuk happened [14:16:27] * se4598 has to fight to write (english) in a way that it's clear and understandable [14:24:57] petan: Do we have a documentation for the config pages (not that I need one)? And do we have somewhere a ToDo/notes pad (maybe like a etherpad.wikimedia.org) [14:35:16] !todo [14:35:17] if you don't know what to work on, here is list: http://tools.wmflabs.org/huggle/docs/head/todo.html [14:35:33] http://tools.wmflabs.org/huggle/docs/head/classHuggle_1_1Configuration.html [14:35:41] se4598: you mean this ^ [14:35:43] ? [14:36:36] Change on 12de_wikipedia a page Wikipedia:Huggle/Config was modified, changed by Se4598 link https://de.wikipedia.org/w/index.php?diff=127998899 edit summary: [+160] /* Reporting */ +aiv-ip +aiv-user (HG3) [14:43:16] not really. That's unflexible. I don't wanna spam the repo or bugzilla e.g. for above bug with undetected edit conflict. Just something easy editable like a wikipage or a etherpad [14:44:43] unflexible for what? [14:45:21] If you want to document the code, then use doxygen format and write the documentation directly into the source code, that is surely best way to document things [14:45:26] or you mean to-do? [14:45:39] for all bug reports there is bugzilla [14:45:59] I don't mind if you use etherpad but I just don't know what you need it for [15:18:56] petan: What e.g. does this assigment when the value is not present on that page?: Configuration::HuggleConfiguration->LocalConfig_ReportPath = Configuration::ConfigurationParse("aiv", config); [15:19:43] Will it do nothing and still have the same value from the initialisation? [15:23:12] se4598: empty string [15:24:08] se4598: static QString ConfigurationParse(QString key, QString content, QString missing = ""); [15:24:20] returns content of variable missing in case that key is missing [15:24:44] so Configuration::ConfigurationParse("blabla", config, "key is missing"); would return "key is missing" in case there would be no such a key [15:25:32] [02huggle3-qt-lx] 07benapetr pushed 031 commit to 03master [+0/-0/±4] 13http://git.io/_jsSew [15:25:33] [02huggle3-qt-lx] 07benapetr 0346d075d - implemented TS in reportuser timer function [15:26:08] then why we initialize that whole stuff (with random default values) at Configuration::Configuration() ? Doesn't make sense [15:29:04] [02huggle3-qt-lx] 07benapetr pushed 031 commit to 03master [+0/-0/±1] 13http://git.io/fyW6aQ [15:29:05] [02huggle3-qt-lx] 07benapetr 03cf874d6 - fix that function needs to return after displaying error message [15:30:40] se4598: that has multiple reasons, primary reason is that uninitialized variables are always problem :-) because they can segfault your application with pretty much cryptic reason, also there were first all these variables and then I started doing these ConfigurationParse functions for all of them [15:31:04] so in past there was no ConfigurationParse for most of these variables [15:31:53] and as I was adding it, I wasn't removing their initialization, because at some point, it doesn't really break anything, also in case someone wanted to access any of these variables before we call respective load function, huggle wouldn't segfault at least [15:33:12] makes sense, but now most/some of the variables are having no reasonable default, making also things difficult for other projects != enwiki [15:33:58] reasonable default can be always inserted, but TBH things like LocalConfig_ReportPath are perfectly OK to be empty strings as default [15:34:37] because huggle will consider such a project as it doesn't have any AIV configured (and will disable AIV reporting) which is what we want for projects where they don't have AIV [15:34:58] you can't just make huggle post random AIV reports on some random page which isn't even watched by anyone :-) [15:36:58] features like AIV must be properly configured on wiki or disabled [15:38:52] fixed it for dewiki by adding https://de.wikipedia.org/w/index.php?diff=127998899 or you would post nothing, possible never saved by mediawiki. That's difficult to debug then, I'll add some sane defaults that doesn't break too much [15:42:29] yes these 2 variables probably should either have some defaults or there could be some function that checks if configuration is sane and if not, it disable AIV have a look at 878 configuration.cpp [15:42:49] // sanitize if (Configuration::HuggleConfiguration->LocalConfig_ReportAIV == "") { Configuration::HuggleConfiguration->LocalConfig_AIV = false; [15:44:05] you know what I will fix this now and some other thigns I need to fix, don't touch that file now [15:44:17] I will soon push a change that completely overwrites configuration.cpp [15:44:19] petan: wait I'm just working on it [15:44:23] ok [15:44:33] give me 2 minutes [15:49:10] meh, fuck, wheres the ui_reportuser.h generated from [15:49:22] qmake does it [16:00:15] [02huggle3-qt-lx] 07se4598 pushed 031 commit to 03master [+0/-0/±4] 13http://git.io/kGmI6Q [16:00:16] [02huggle3-qt-lx] 07se4598 03025e781 - honor vandal-report-reason [16:00:56] petan: your turn, sorry for taking longer [16:01:07] no problem [16:04:58] * se4598 can easily grab something to eat while building/compiling although he has a fast computer [16:16:33] [02huggle3-qt-lx] 07se4598 pushed 031 commit to 03master [+0/-0/±1] 13http://git.io/AfouoQ [16:16:34] [02huggle3-qt-lx] 07se4598 0397576b1 - Show recent edits of reported user in descending order (newest top) [16:20:41] [02huggle3-qt-lx] 07benapetr pushed 032 commits to 03master [+0/-0/±3] 13http://git.io/3WVszw [16:20:42] [02huggle3-qt-lx] 07benapetr 03180d62e - fix prevent conflict in case that talk page of user didn't exist [16:20:43] [02huggle3-qt-lx] 07benapetr 03b81606b - Merge branch 'master' of github.com:huggle/huggle3-qt-lx [16:32:45] petan: re QtWebKit/QWebView -> QtWebKitWidgets/QWebView [16:33:02] that wasn't me but the stupid UI editor [16:34:03] aha and what that actually changes? [16:34:08] does it make any difference at all? [16:35:01] the "Report Reason" which actually get overwriten by LocalConfig_ReportDefaultReason [16:36:13] yes, the change is almost pointless in retrospective [16:47:50] petan: that may broke it for QT4 suggest me googleing it, will change back [16:48:08] "New webkit shipping with QT5 has a new structure. QWebView, QWebpage, etc. are now part of QtWebKitWidgets." [16:48:24] if it broke then CI should be jumping all around [16:50:37] apparently it doesn't, so then I will leave it. [16:53:24] is there a reason why we do both QT4 and QT5? [16:55:24] I have at my Qt Creator only 5.2 installed [16:58:23] se4598: yes, Qt4 is default in ubuntu, debian and most other distros [16:58:30] Qt5 is too new, basically only on windows [17:00:06] [02huggle3-qt-lx] 07benapetr pushed 031 commit to 03master [+0/-0/±3] 13http://git.io/ooHZxQ [17:00:07] [02huggle3-qt-lx] 07benapetr 03053b6bd - reformatted now it's a bit more readable though it's not so clear where these functions come from, there wasn't really any reason for these functions to be static anyway :P [17:17:16] petan: installed 4.8.5 (1.8 GB btw) and tryed to run it with Qt Creator......... segfault... [17:19:38] I interpret the debugger that this happens inside QMutex called from Huggle::Syslog::Syslog line 17 [18:50:23] se4598: I am running qt4 too and it works just fine [18:51:08] only qt4 or with qt5 together? [18:51:15] I have both [19:06:15] If I want to improve the reliability if a user is already reported, can I eventually introduce a RegExp into the project configuration (it may work with a generic one for en+dewiki at least too)? Currently the indicator is if the "username" occurs somewhere on the report page [19:07:20] I want to restrict that it only matches if the username is in a report template [19:08:49] hmm.... yes [19:09:03] but you would need to make this regex in project config [19:09:18] by default this regex could just check for username [19:09:39] regex should have $1 that get replaced with escaped string of username [19:10:01] yeah, exactly that [19:10:22] QRegExp::escape already exists [21:32:45] [02huggle3-qt-lx] 07benapetr pushed 031 commit to 03master [+0/-0/±5] 13http://git.io/Lki4dw [21:32:46] [02huggle3-qt-lx] 07benapetr 034af6034 - formatted code [21:53:24] btw se4598 keep in mind that people don't need to use any template to report user [21:53:41] if they just add a message to that page with name of user it's also valid report [21:53:47] so your regex may not work [21:56:53] imho not that big deal in that case because pure vandalism reports are 99,999% done with that template at dewiki. Without a restrictered check you would'nt able to report a user if he has posted on that page (there are also discussions about reports on that page) [22:00:34] btw: it's great to code here, b/c I'm learning some c++ and I code finally. I can't find great ideas or much motivation to code for myself only :P [22:01:31] but not so great to use/learn c++ now because in the next weeks a have a exams where will have to code in JAVA :D [22:04:59] petan: also look how hg2 has done that, basically regexp on templates: https://github.com/huggle/huggle/blob/master/Requests/ReportRequests.vb#L165 [22:05:15] hardcoded [22:05:27] se4598: I don't care how we did that in hg2 that was completely broken [22:05:44] if hg2 was working ok we wouldn't need to create new version :P [22:06:51] hg2 is totally fine for me except that it throws sometimes exceptions on fetching newly created pages :P [22:07:29] who cares linux or mac users xD