[09:57:25] huggle 3 is not removing from queue the edits that where reverted, pages with several revisions are shown, known bug? [09:57:33] hg3 on linux [10:48:15] [bugzilla] [Bug 61347] HG 3 crashed https://bugzilla.wikimedia.org/show_bug.cgi?id=61347 [10:49:22] Hello Everyone [11:04:37] nitinagarwal3006: hi [11:04:58] Alchimista: more details pls [11:05:05] Alchimista: reverted by huggle? or others? [11:05:12] Alchimista: did you configure huggle to remove such edits? [11:05:16] hi [11:05:36] addshore: we really need to fix travis :D [11:05:38] Are u Petr Bena [11:05:43] nitinagarwal3006: yes [11:05:46] ok [11:05:58] I saw ur mail on mailing list [11:06:04] cool :) [11:06:17] for C++ and Python developers [11:06:31] I have replied to ur mail with more details about me [11:07:11] ah yes I see cool, do you want to join our dev team? it's rather simple [11:07:31] basically all you need to do is add your name on list at https://meta.wikimedia.org/wiki/Huggle/Members [11:08:23] I will give you direct push access as soon as you commit at least 3 patches to any of your github.com/huggle repositories [11:08:53] ok [11:08:59] From where shall I start off [11:09:52] Which of the tasks shall I pick off [11:10:43] !todo [11:10:44] if you don't know what to work on, here is list: http://tools.wmflabs.org/huggle/docs/head/todo.html [11:10:50] we have a lot of things there :P [11:11:13] I think that first, rather easy task is to get source code and set up your build environment [11:11:32] ok [11:11:34] Thanks [11:11:43] petan: YES [11:11:49] once your build environment is ready you can start with some easy tasks, for example you can add yourself to list of devs in about form :) [11:12:27] addshore: because some of my commits really do break builds, and I don't even notice, for example latest HEAD doesn't seem to be buildable at all :P [11:12:40] petan: I would like to its open source, will it be participating in Gsoc 2014. [11:14:04] [02huggle3-qt-lx] 07benapetr pushed 031 commit to 03master [+0/-0/±5] 13http://git.io/CVLrZA [11:14:05] [02huggle3-qt-lx] 07benapetr 035492542 - fixed some bugs cleanup up the config a bit too, no idea why pieces of config init were in core's init [11:14:48] nitinagarwal3006: I don't think so, but we could consider it.. however huggle is rather small project, maybe it could be part of wikimedia's Gsoc participation [11:16:04] petan: the option is marked on preferences -> feed, and at least other reverts are shown, i'm not sure if huggle ones are. [11:16:49] Alchimista: there are several options that can affect this, however keep in mind that huggle sees differently reverts made by anons (IP) and trusted (whitelisted) users [11:16:49] petan: ok, After sometime, is there a way to get fund to work on the project. [11:17:16] Alchimista: if you change your queue settings and check that you want to filter out reverts, you shouldn't see them in queue (edits that are reverts) [11:17:48] Alchimista: if you don't want to see also edits that were reverted, you need to also check remove all edits that were reverted in feed settings. this doesn't work 100% though [11:18:06] petan: at least it should show only the last edit of an user on an article,even if they're not reverted [11:18:15] example: https://pt.wikipedia.org/w/index.php?title=Circunfer%C3%AAncia&action=history [11:18:40] Change on 12meta_wiki a page Huggle/Members was modified, changed by Nitinagarwal3006 link https://meta.wikimedia.org/w/index.php?diff=7580787 edit summary: /* Core developers */ [11:18:45] nitinagarwal3006: this is open source project, we are all volunteers and basically work for free. There is no fund I know of [11:18:53] i had to revert all those edits manually, because that page had a lot of calls on queue [11:19:00] petan: ok Thanks [11:19:55] Alchimista: it doesn't do that by default, it can be configured to do so, but it's not always 100% because of heavy caching [11:20:18] sometimes the page is edited immediately after it's cached, in that case huggle doesn't even know there is newer version [11:20:42] huggle 3 is using different kind of caching than 2, that's why it's so much faster [11:21:10] I could eventually make an option that would automatically switch to latest revision in case there is some [11:21:56] [02huggle3-qt-lx] 07benapetr pushed 031 commit to 03master [+0/-0/±1] 13http://git.io/vekZeA [11:21:57] [02huggle3-qt-lx] 07benapetr 039b86481 - fixed CI [11:22:00] i think that option on default would be better, if there is a newer revision with same title and same editor, remove the older from queue and replace ti by the newer [11:22:15] because older reviews can be seen on history [11:22:19] Change on 12meta_wiki a page Huggle/Members was modified, changed by Nitinagarwal3006 link https://meta.wikimedia.org/w/index.php?diff=7580838 edit summary: /* Core developers */ [11:23:11] petan: you fixed the CI? [11:23:17] [02huggle3-qt-lx] 07benapetr pushed 03421 commits to 03travis [+53/-7/±1502] 13http://git.io/gvJPjQ [11:23:18] [02huggle3-qt-lx] 07benapetr 039eb482b - removed -qq option to see what is wrong in travis [11:23:23] :P [11:23:35] addshore: lol, that's a confusing message. I fixed the unit test (broken code) I didn't fix the travis though [11:23:39] lol, pushed 421 commits xD [11:23:44] heh [11:24:07] Change on 12meta_wiki a page Huggle/Members was modified, changed by Nitinagarwal3006 link https://meta.wikimedia.org/w/index.php?diff=7580862 edit summary: /* Core developers */ [11:24:34] Alchimista: yes but history is not cached, unlike the edit [11:24:52] Alchimista: so huggle first load the edit from operating memory, and then it query mediawiki for history [11:26:59] petan: the cached edit has what information? The page title and user? [11:27:12] no [11:27:14] or just the page title? [11:27:21] Alchimista: let me check the code [11:28:20] post processed cached edit has: Title, User, All page flags, Timestamp, Content of page, Content of diff, Content of user's talk page, and several other information [11:28:42] it basically contains everything you need to identify whether it's vandalism or not [11:28:59] if you unplugged your etherned cable while using huggle 3 you could still load next edits from queue [11:29:04] * ethernet [11:31:18] so when there is a new edit, can't it be checked against those in queue to see if some has the same title, same user and older timestamp? sorry if i'm making some confusion, i'm tryng to understand hg3 architecture [11:32:01] [02huggle3-qt-lx] 07benapetr pushed 031 commit to 03travis [+0/-0/±1] 13http://git.io/IgoqcQ [11:32:02] [02huggle3-qt-lx] 07benapetr 032ff45db - more dependencies, let's see :o [11:32:18] Alchimista: yes it can and it does [11:33:00] even if they're not reverted? I haven't seen that behaviour [11:33:19] Alchimista: when there is a new edit to same edit and you have checked "remove old edits to same edit if newer edit is parsed" then yes [11:33:34] you need to check this option in preferences [11:33:56] or whatever it's called [11:34:26] Remove old edits if newer edit to same page is parsed [11:34:31] that's it [11:34:50] let me restart hg and try again [11:35:14] it should work without restart, but it start working only on edits that are parsed after you check this [11:35:30] this rule is not applied recursively on existing edits [11:35:44] it could be implemented though :P [11:36:31] addshore: current labels in preferences are confusing, if you could think of better wording maybe update them :P [11:36:49] i had that option checked, but provably then the problem was with older edits, prior to the option marking [11:36:51] short and clear explanation of each option would be cool [11:37:02] that's possible [11:37:19] if you start huggle with parameter -v you will also see when edit is removed becaues of this option [11:37:32] it will tell you "removing old edit to bla bla..." [11:38:47] hg is docked at ubuntu launcher, so there is no back screen [11:39:07] but i'll try that too, to check more information [11:43:30] [02huggle3-qt-lx] 07benapetr pushed 031 commit to 03travis [+0/-0/±1] 13http://git.io/4qhNuw [11:43:31] [02huggle3-qt-lx] 07benapetr 0326a9103 - more deps [11:43:52] Alchimista: you can always type "huggle -v" in terminal [11:43:59] its binary is in /usr/bin [11:45:49] [02huggle3-qt-lx] 07benapetr pushed 031 commit to 03travis [+0/-0/±1] 13http://git.io/_rp_jw [11:45:50] [02huggle3-qt-lx] 07benapetr 03a6eac43 - yay [11:45:53] addshore: almost there :D [11:46:00] xD [11:46:02] we had some dep mixup / whatever [11:46:14] i know it petan, i've added to terminal just for convenience. by the way, is it possible to *install* a second hg with a diferent name? I'm thinking to have a stable huggle to usual usage, and a huggle-dev separate where i could try py without damage the stable one [11:46:18] qtchooser : Breaks: qt4-dev-tools (<= 4:4.8.4+dfsg-0ubuntu3~) but 4:4.8.1-0ubuntu5~precise1~test1 is to be installed [11:46:19] Breaks: qt4-linguist-tools (<= 4:4.8.4+dfsg-0ubuntu3~) but 4:4.8.1-0ubuntu5~precise1~test1 is to be installed [11:46:44] Alchimista: yes, just call it different or move it to different file name? :P [11:47:01] it's 1 binary file, it can run standalone [11:48:35] addshore: I will make a different repository becaue I don't want to fuck up history [11:48:41] I will need zillions of commits to do this [11:49:05] when making sudo make install isn't possible to give a diferent name? Or does it have to be on ./configure --prefix=newdir? [11:49:22] addshore: or maybe what about dropping whole "travis" branch after we fix it? [11:49:25] without merge [11:49:28] addshore: would that work? [11:50:31] Alchimista: if you want to do that using apt then not, but why don't you just copy /usr/bin/huggle to ~/bin/huggle1 [11:50:43] then install new version of huggle, you will have old version in ~/bin/huggle1 [11:50:51] it's just a 1 file [11:53:05] hum, i'll try that, huggle1 for stable, and huggle for dev [11:53:08] thanks! [11:53:09] [02huggle3-qt-lx] 07benapetr pushed 031 commit to 03travis [+0/-0/±1] 13http://git.io/QSkvqA [11:53:10] [02huggle3-qt-lx] 07benapetr 035b89ab4 - let's start from clean tree [11:57:21] [02huggle3-qt-lx] 07benapetr pushed 031 commit to 03travis [+0/-0/±1] 13http://git.io/hnKyGQ [11:57:22] [02huggle3-qt-lx] 07benapetr 037035d48 - maybe we could just follow the documentation of this ppa [11:57:39] Alchimista: you realize there is no stable version? :P [11:57:44] that one from ppa is beta version [12:31:43] petan: yap, when i say stable means the one i hadn't breack yet :P [12:39:35] addshore: I figure out what is wrong [12:39:58] addshore: for some reason it's not possible to have both qt4 and qt5 on same system now. So we need to create 2 paralel environments [12:40:03] addshore: do you know how to do that? :P [12:49:46] [02huggle3-qt-lx] 07benapetr pushed 031 commit to 03travis [+0/-0/±1] 13http://git.io/gj-aIQ [12:49:47] [02huggle3-qt-lx] 07benapetr 03a565182 - rm qt4 [12:53:09] hehe [12:53:11] petan: yes [12:53:41] are you working on a branch? where shouold I poke? [12:53:52] or should I start from a fresh branch? [12:54:39] [02huggle3-qt-lx] 07addshore created branch 03travisQtSplit 13http://git.io/nbHzDw [12:54:51] addshore: in travis branch - DO NOT MERGE IT [12:54:58] that branch needs to be nuked later [12:55:05] I was going to do it on my own branch :P [12:55:09] it contains hundreds of commits I don't want to have in master [12:55:19] aha [12:55:36] you can do that anywhere, but try to avoid merging anything useless [12:56:15] which of these is needed for qt5 which for qt4 and which for both? [12:56:15] "sudo apt-get install -y g++ libqt4-dbus libqt4-dev libqt4-network libqt4-opengl libqt4-webkit libqtwebkit-dev libqtgui4 libqtcore4 libqt4-xml qt4-dev-tools qt4-qmake libqt5webkit5-dev qt5-default qtquick1-5-dev qtlocation5-dev qtsensors5-dev qtdeclarative5-dev libsqlite3-dev" [12:56:27] that is what I am now trying to figure out [12:56:32] haha xd [12:57:04] [02huggle3-qt-lx] 07benapetr pushed 031 commit to 03travis [+0/-0/±1] 13http://git.io/zcjipQ [12:57:05] [02huggle3-qt-lx] 07benapetr 031f276a4 - qtquick1-5-dev [12:57:09] addshore: I will send you working conf for qt5 soon [13:11:15] [02huggle3-qt-lx] 07benapetr pushed 031 commit to 03travis [+0/-0/±1] 13http://git.io/R7NhBA [13:11:16] [02huggle3-qt-lx] 07benapetr 033aac92f - fix [13:11:21] addshore: this is working qt5 config ^ [13:11:26] now I will make qt4 one [13:14:26] !travis.yml@qt5 is http://tools.wmflabs.org/paste/view/02aa113e [13:14:26] Key was added [13:14:32] addshore ^ [13:14:59] !paste [13:15:00] http://toolserver.org/~mmovchin/paste/ [13:15:04] !paste del [13:15:04] Successfully removed paste [13:15:14] !paste is http://tools.wmflabs.org/paste [13:15:14] Key was added [13:17:44] petan: okay :) now make me the Qt4 one ;p [13:17:57] can pushh the change to the brancha s soon as you tell me :P [13:19:46] [02huggle3-qt-lx] 07benapetr pushed 031 commit to 03travis [+0/-0/±1] 13http://git.io/I3AWRw [13:19:47] [02huggle3-qt-lx] 07benapetr 038e40aab - qt4 [13:20:06] that works for qt4? [13:21:19] addshore: dunno [13:21:23] addshore: testing now [13:21:24] xD [13:21:26] addshore: why this doesn't stop https://travis-ci.org/huggle/huggle3-qt-lx/builds/19331700 [13:21:37] Done. Your build exited with 0. [13:21:46] but it's still in running [13:22:04] refresh :P [13:22:08] your qt4 one failed [13:23:02] [02huggle3-qt-lx] 07benapetr pushed 031 commit to 03travis [+0/-0/±1] 13http://git.io/SGyFtA [13:23:03] [02huggle3-qt-lx] 07benapetr 033f9d8a9 - maybe using --qt4 is better for qt4 :D [13:23:08] lol [13:24:26] [02huggle3-qt-lx] 07benapetr pushed 031 commit to 03travis [+0/-0/±1] 13http://git.io/ap1_RA [13:24:26] [02huggle3-qt-lx] 07benapetr 0369a5d20 - qt4-dev-tools [13:26:10] looks like it working [13:27:24] yeah [13:27:49] will wait for it to pass then push mine to a branch :) [13:27:56] ok [13:28:27] addshore: if you didn't know what to do you can try build huggle with python on windows :DD [13:28:31] it's pretty much not easy [13:28:50] python support for MS windows suck [13:30:50] xD [13:31:31] right it passed :) [13:31:47] [02huggle3-qt-lx] 07addshore pushed 031 commit to 03travisQtSplit [+4/-0/±1] 13http://git.io/7Rb9aQ [13:31:48] [02huggle3-qt-lx] 07addshore 03f5bf708 - Use different travis environments for QT4 vs QT5 This also splits each part of the travis execution into its own bash script for easier reading and use! [13:31:49] ok can I nuke travis now [13:31:54] petan: ^^ [13:31:57] lets see what this does first ;p [13:32:02] ok [13:32:13] :o [13:32:48] addshore: you prefer directories starting with low case? [13:32:55] we could rename Documentation then [13:33:12] it depends what the DIR is for, but yes this is better lowercase :P [13:33:20] ok [13:33:20] .travis.yml matches /travis/ [13:33:42] petan: this should also mean the CI tetss go slightly faster now as the QT4 and QT% will be done in parallel [13:33:58] [02huggle3-qt-lx] 07benapetr pushed 031 commit to 03master [+7/-7/±0] 13http://git.io/85jKUA [13:33:59] [02huggle3-qt-lx] 07benapetr 032554ca5 - rename docs [13:34:19] the last successfull build before everything broke took 12 mins 42 [13:34:33] yes I know [13:34:35] although before they could take up to 20 mins [13:34:42] but it also takes a lot of time to install all packages [13:34:55] it would be better if these were preinstalled on image [13:35:10] indeed, but meh! [13:35:23] im racing QT4 vs QT5 [13:36:16] [02huggle3-qt-lx] 07benapetr pushed 031 commit to 03master [+0/-0/±2] 13http://git.io/Dug2fw [13:36:17] [02huggle3-qt-lx] 07benapetr 034a61a68 - clean up fixed comments and stuff [13:36:28] racing? [13:36:37] qt4 will be faster for sure [13:36:41] because qt5 needs a lot of packages [13:37:20] sent you a pic of the race in hangout xD [13:37:42] qt4 has echoed 490 lines but qt5 has done 1456 [13:39:19] petan: also in theroy we could make travis push these builds somewhere so that they can be downloaded [13:39:48] QT4 dones in 6 mins 35 [13:40:01] addshore: no point in that, linux builds are in ppa [13:40:07] true [13:40:20] I wanted to set up nightly builds there [13:40:26] but I can't find out how [13:40:38] they say we would need to host source in bazaar instead of git :( [13:40:47] I prefer git [13:41:06] QT5 done in 8 mins 36 [13:41:30] in theory we could push these binaries somewhere to labs and tar / deb them but these would only run on x64 ubuntu [13:41:31] not only did we fix the CI but we sped it up ;p [13:41:48] ok merge this [13:42:05] [02huggle3-qt-lx] 07addshore pushed 032 commits to 03master [+8/-0/±2] 13http://git.io/qXcp6A [13:42:06] [02huggle3-qt-lx] 07addshore 033ddefce - Merge pull request #34 from huggle/travisQtSplit Use different travis environments for QT4 vs QT5 [13:42:07] [02huggle3-qt-lx] 07addshore deleted branch 03travisQtSplit [13:42:10] done [13:43:36] [02huggle3-qt-lx] 07benapetr deleted branch 03travis [13:43:44] ok [13:44:30] petan: you should go and close that bug you opened with travis now ;p [13:44:40] yes [13:44:43] @search travis [13:44:43] Results (Found 2): travis, travibug, [13:44:48] !travisbug [13:44:51] !travibug [13:44:51] https://github.com/travis-ci/travis-ci/issues/1982 [13:46:04] :) lovbely [13:49:12] btw did IRC bot stop working or they just started ignoring this channel lol [13:49:20] because I didn't see it for some time [13:51:15] it probably just broke bit :P [16:34:03] addshore: here? [18:46:11] okay, there's a couple weird things going on here in unstable-land [18:46:29] i don't see an undo button anymore. could be pebcak. [18:47:09] and I'd like to see warnings being canceled if reverts are canceled first [18:48:15] and at least once, doing an AGF revert actually linked the wrong user in the edit summary [18:48:31] that's a little weird [18:54:12] just FYIs, getting it on the radar. [18:55:17] nfd9001: there ever was undo button in hg3? [18:55:24] I am sure that was in development now [18:56:49] nfd9001: how do you cancel reverts? when it sent a warning on canceled revert? that never happened to me [18:57:16] can you provide a screenshot or logs where a warning was sent and rollback not? [18:57:25] because I just can't believe it :D [19:04:15] petan: err, lemme see [19:07:10] petan: looks like I was just confused with that one. "sending message to ___" seems to show up whether or not the prior edit was canceled [19:07:44] also, i'd just started using hg3, so i wasn't sure if there was supposed to be one yet [19:08:48] a "canceled" revert being one that you stop from an "are you sure?" popup [19:09:03] and here's the diff with the AGF gaffe: https://en.wikipedia.org/w/index.php?title=Zander_Diamond&diff=prev&oldid=596457535 [19:48:12] [bugzilla] [Bug 61764] Can't request for deletion using the "Request deletion of a page!"-window https://bugzilla.wikimedia.org/show_bug.cgi?id=61764 [19:48:52] nfd9001: it's "Sending message" not "sent message" :P [19:48:56] but yes it's confusing :D [19:58:25] [02huggle3-qt-lx] 07benapetr pushed 031 commit to 03master [+0/-0/±1] 13http://git.io/-PcTfw [19:58:27] [02huggle3-qt-lx] 07benapetr 033d160b6 - moved "sending message to user" was confusing in normal log [20:28:56] petan you there? [20:58:44] @notify josve05a [20:58:44] I'll let you know when I see josve05a around here [21:42:38] addshore: why you didn't add u in about box? [21:55:55] [02huggle3-qt-lx] 07benapetr pushed 031 commit to 03master [+0/-0/±3] 13http://git.io/YXHt8w [21:55:56] [02huggle3-qt-lx] 07benapetr 037a26319 - gui * updated team * inserted some items to history so that we can implement undo operation [22:08:45] [02huggle3-qt-lx] 07benapetr pushed 031 commit to 03master [+6/-5/±1] 13http://git.io/qs95ww [22:08:45] [02huggle3-qt-lx] 07benapetr 038d8a6ce - reorganized html [22:26:49] neat! [22:28:13] hey, i noticed that H3 is sometimes inserting level 4 warnings where they don't completely make sense? [22:28:40] i thought they were based on the other warnings recently added; is it badness score or something? [22:45:59] petan: cause I havnt touched this one much yet ;p [23:33:40] [02huggle3-qt-lx] 07benapetr pushed 031 commit to 03master [+1/-0/±9] 13http://git.io/WbGAgA [23:33:41] [02huggle3-qt-lx] 07benapetr 0301f4d80 - option user can now force the page to load latest revision even if they open different revision of page from queue, this makes it behave like older huggle [23:54:28] [02huggle3-qt-lx] 07benapetr pushed 031 commit to 03master [+0/-0/±1] 13http://git.io/6aoq2g [23:54:30] [02huggle3-qt-lx] 07benapetr 0345aa440 - make selected edit more visible