#instantbird log on 09 06 2013

All times are UTC.

00:00:20 <-- nhnt11 has quit (Quit: Instantbird 1.5a1pre -- http://www.instantbird.com)
00:00:21 <clokep> http://lxr.instantbird.org/instantbird/source/chat/protocols/irc/ircHandlers.jsm#102
00:06:04 --> mconley has joined #instantbird
00:08:54 <-- mconley has quit (Input/output error)
00:09:39 <flo-retina> clokep: is bug 2089 checkin-needed?
00:09:43 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=2089 nor, --, ---, qheaden, ASSI, Debug logs are binary only
00:10:18 <flo-retina> the indent is messed up in sendBinaryData :(
00:11:01 <clokep> flo-retina: If something is messed up then it isn't, qheaden will fix it and I can check it in or you can tomorrow.
00:11:22 <flo-retina> it's just 2 spaces to add ;)
00:12:16 <flo-retina> clokep: if you get qheaden to fix it, ask him to also include a more descriptive commit message ;)
00:12:47 <clokep> OK. :)
00:13:45 <-- florian has quit (Quit: Instantbird 1.5a1pre -- http://www.instantbird.com)
00:14:32 <instantbot> Check-in: http://hg.instantbird.org/instantbird/rev/c52c3dd7677f - Nihanth Subramanya - Bug 2066 - New conversation tab should suggest chat rooms - Implement parsing of LIST data in js-irc, r=clokep,aleth.
00:14:33 <instantbot> Check-in: http://hg.instantbird.org/instantbird/rev/0bc4f51cc78b - Nihanth Subramanya - Bug 2066 - New conversation tab should display chat rooms, r=fqueze,aleth
00:14:34 <flo-retina> Good night
00:14:35 <instantbot> Check-in: http://hg.instantbird.org/instantbird/rev/419735388432 - aleth - Bug 2141 - Don't include ib-sender span in magic copy, r=fqueze.
00:14:36 <instantbot> Check-in: http://hg.instantbird.org/instantbird/rev/9db515149f68 - Quentin Headen - Bug 2132 - Sender name not set in Yahoo conversations, r=aleth,clokep.
00:14:37 <instantbot> Check-in: http://hg.instantbird.org/instantbird/rev/a17743c6cca1 - aleth - Bug 2142 - Allow tab completion of nicks which have left the room, r=fqueze.
00:16:40 <instantbot> clokep@gmail.com set the Resolution field on bug 2066 to FIXED.
00:16:49 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=2066 enh, --, 1.5, nhnt11, RESO FIXED, New conversation tab should suggest chat rooms
00:17:11 <instantbot> clokep@gmail.com set the Resolution field on bug 2141 to FIXED.
00:17:18 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=2141 nor, --, 1.5, aleth, RESO FIXED, Don't include ib-sender span in magic copy
00:18:12 <instantbot> clokep@gmail.com set the Resolution field on bug 2132 to FIXED.
00:18:18 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=2132 nor, --, 1.5, qheaden, RESO FIXED, Sender name not set in Yahoo conversations
00:18:43 <instantbot> clokep@gmail.com set the Resolution field on bug 2142 to FIXED.
00:18:45 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=2142 enh, --, 1.5, aleth, RESO FIXED, Allow tab completion of nicks which have left the room
00:29:41 <instant-buildbot> build #454 of win32-onCommit is complete: Failure [failed compile]  Build details are at http://buildbot.instantbird.org/builders/win32-onCommit/builds/454  blamelist: aleth <aleth@instantbird.org>, Nihanth Subramanya <nhnt11@gmail.com>, Quentin Headen <qheaden@phaseshiftsoftware.com>
00:47:10 <instant-buildbot> build #461 of macosx-onCommit is complete: Success [build successful]  Build details are at http://buildbot.instantbird.org/builders/macosx-onCommit/builds/461
00:55:28 <-- Mook_as has quit (Quit: Mook_as)
00:58:09 <instant-buildbot> build #464 of linux-onCommit is complete: Success [build successful]  Build details are at http://buildbot.instantbird.org/builders/linux-onCommit/builds/464
00:59:26 <-- dionisos has quit (Ping timeout)
00:59:37 --> dionisos has joined #instantbird
01:47:25 <-- clokep has quit (Quit: Instantbird 1.5a1pre -- http://www.instantbird.com)
01:55:17 <-- dionisos has quit (Ping timeout)
01:55:28 --> Mook has joined #instantbird
01:55:45 --> dionisos has joined #instantbird
02:12:26 <-- dionisos has quit (Ping timeout)
02:12:35 --> dionisos has joined #instantbird
03:37:18 --> nhnt11_phone has joined #instantbird
03:42:40 <-- nhnt11_phone has quit (Quit: AndroIRC - Android IRC Client ( http://www.androirc.com ))
03:56:31 <instant-buildbot> build #962 of linux-nightly-default is complete: Success [build successful]  Build details are at http://buildbot.instantbird.org/builders/linux-nightly-default/builds/962
03:57:34 --> mconley has joined #instantbird
04:04:31 <-- mconley has quit (Input/output error)
04:11:15 <-- wnayes has quit (Quit: wnayes)
04:20:07 <-- chrisccoulson has quit (Client exited)
04:56:25 --> YH has joined #instantbird
05:27:00 <-- YH has quit (Quit: Instantbird 1.3 -- http://www.instantbird.com)
05:32:59 <-- Mook has quit (Quit: Mook)
05:49:06 <instantbot> qheaden@phaseshiftsoftware.com requested review from florian@instantbird .org for attachment 2835 on bug 2089.
05:49:11 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=2089 nor, --, ---, qheaden, ASSI, Debug logs are binary only
05:50:45 <instant-buildbot> build #983 of macosx-nightly-default is complete: Success [build successful]  Build details are at http://buildbot.instantbird.org/builders/macosx-nightly-default/builds/983
05:55:10 <-- dionisos has quit (Ping timeout)
06:01:35 <instantbot> qheaden@phaseshiftsoftware.com requested review from clokep@gmail.com for attachment 2836 on bug 2111.
06:01:37 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=2111 nor, --, ---, qheaden, ASSI, Yahoo Web Client Causes Endless Typing Notification
06:07:04 <-- EionRobb has quit (Quit: Leaving.)
06:23:49 --> jb has joined #instantbird
06:33:21 <-- jb has quit (Ping timeout)
06:33:35 --> FireFly_TB has joined #instantbird
06:57:58 <instant-buildbot> build #1063 of win32-nightly-default is complete: Failure [failed compile]  Build details are at http://buildbot.instantbird.org/builders/win32-nightly-default/builds/1063
07:17:37 <-- qheaden has quit (Quit: ZNC - http://znc.in)
07:18:58 --> jb has joined #instantbird
07:23:59 --> chrisccoulson has joined #instantbird
07:56:54 <-- FireFly_TB has quit (Ping timeout)
08:06:35 <-- Suiseiseki has quit (Ping timeout)
08:08:25 --> Suiseiseki has joined #instantbird
08:10:25 <-- jb has quit (Ping timeout)
08:33:59 --> chrisccoulson_ has joined #instantbird
08:34:20 <-- chrisccoulson has quit (Ping timeout)
08:40:50 <-- chrisccoulson_ has quit (Quit: OSError: [Errno 130] Owner died)
08:40:57 --> chrisccoulson has joined #instantbird
09:16:47 <-- flo-retina has quit (Quit: Instantbird 1.5a1pre -- http://www.instantbird.com)
09:33:05 --> aleth has joined #instantbird
09:33:05 * ChanServ sets mode +h aleth 
09:40:37 --> dionisos has joined #instantbird
09:44:54 <-- aleth has quit (Quit: Ciao)
09:44:58 --> aleth has joined #instantbird
09:44:58 * ChanServ sets mode +h aleth 
09:47:34 <-- dionisos has quit (Ping timeout)
09:47:43 --> dionisos has joined #instantbird
09:49:55 <-- gerard-majax has quit (Ping timeout)
09:51:43 <aleth> Wow, looks like lots happened last night :)
09:53:22 <aleth> Lots of fixing followups before landing :D Not bad...
09:54:22 <-- dionisos has quit (Ping timeout)
09:54:46 --> dionisos has joined #instantbird
09:55:32 <aleth> "I would also like us to stop pointlessly calling getChatRoomDefaultFieldValues all the time" heh, so did I, and the suggestion got shot down at the time ;)
10:01:07 --> flo-retina has joined #instantbird
10:01:08 * ChanServ sets mode +qo flo-retina flo-retina 
10:02:17 <aleth> I was expecting the newtab with IRC LIST enabled to be slower :)
10:03:19 <flo-retina> aleth: have you preff'ed it on?
10:03:23 <aleth> Sure.
10:03:39 <flo-retina> are you connected to freenode?
10:03:49 <-- dionisos has quit (Ping timeout)
10:03:54 <aleth> That's the next thing I'll try.
10:03:54 <aleth> Obviously there is a /lot/ of optimization to be done
10:04:01 --> dionisos has joined #instantbird
10:04:09 <aleth> And it's not a super useful feature without better search
10:04:47 --> clokep has joined #instantbird
10:04:47 * ChanServ sets mode +o clokep 
10:05:14 <flo-retina> aleth: hmm, how will we offer channels likes #developers that aren't in the list result? Do we need a new API for the stats/logger service to be able to generate the chat room fields from a log file? :-S
10:05:19 <flo-retina> clokep: good morning :)
10:05:40 <-- dionisos has quit (Ping timeout)
10:05:50 <aleth> flo-retina: I don't see how that's different to using Join Chat apart from the case where we may want to store passwords
10:05:50 --> dionisos has joined #instantbird
10:06:02 <flo-retina> aleth: "the suggestion got shot down at the time" really? I hope that wasn't by me. :-S
10:06:11 <aleth> I think it was clokep actually ;)
10:06:35 <clokep> Hmm?
10:06:55 <aleth> We really need to drop all those UTF-8 encoding warnings when doing LIST.
10:07:00 <flo-retina> aleth: if you have IRC in mind, there's likely no problem because (when there's no password), the only relevant field is the channel name
10:07:29 <flo-retina> aleth: do you think we should drop them only when doing list, or more generally?
10:08:14 <aleth> flo-retina: I'm not sure. For LIST it seems to make sense as it's not an incoming message in a conversation.
10:11:32 --> novabyte has joined #instantbird
10:14:24 <-- novabyte has quit (Quit: bye bye)
10:19:43 <-- dionisos has quit (Ping timeout)
10:20:15 --> dionisos has joined #instantbird
10:21:55 <-- dionisos has quit (Ping timeout)
10:21:59 --> dionisos has joined #instantbird
10:22:37 --> novabyte has joined #instantbird
10:23:31 <-- novabyte has quit (Quit: bye bye)
10:26:43 <-- dionisos has quit (Ping timeout)
10:28:11 <instantbot> clokep@gmail.com granted review for attachment 2835 on bug 2089.
10:28:16 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=2089 nor, --, ---, qheaden, ASSI, Debug logs are binary only
10:28:49 <clokep> Anyone have ideas on bug 2152?
10:28:52 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=2152 maj, --, ---, nobody, UNCO, Messages appearing twice
10:29:03 --> dionisos has joined #instantbird
10:29:23 <instantbot> clokep@gmail.com granted review for attachment 2836 on bug 2111.
10:29:25 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=2111 nor, --, ---, qheaden, ASSI, Yahoo Web Client Causes Endless Typing Notification
10:30:34 <aleth> clokep: That's very strange
10:47:24 <-- dionisos has quit (Ping timeout)
10:47:53 --> dionisos has joined #instantbird
10:48:41 * flo-retina is now known as flo-retinal
10:49:31 * flo-retinal is now known as flo-retina
10:49:34 <-- dionisos has quit (Ping timeout)
10:49:38 --> dionisos has joined #instantbird
10:52:09 <flo-retina> clokep: likely a broken bot.
10:52:26 <flo-retina> I don't think it could be us displaying the same message twice, as the timestamps of the duplicated messages are different
10:52:49 <aleth> The Bubbles screenshot is puzzling
10:52:52 <clokep> I agree.
10:54:26 <-- clokep has quit (Quit: Instantbird 1.5a1pre -- http://www.instantbird.com)
11:02:00 --> gerard-majax has joined #instantbird
11:03:22 <-- gerard-majax has quit (Connection reset by peer)
11:03:25 <-- dionisos has quit (Ping timeout)
11:03:47 --> gerard-majax_ has joined #instantbird
11:03:59 --> dionisos has joined #instantbird
11:08:04 <-- dionisos has quit (Ping timeout)
11:08:07 --> dionisos has joined #instantbird
11:34:30 <-- dionisos has quit (Ping timeout)
11:34:42 --> dionisos has joined #instantbird
11:42:56 --> clokep_ has joined #instantbird
11:42:58 <instantbot> New Core - IRC bug 2153 filed by aleth@instantbird.org.
11:43:00 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=2153 nor, --, ---, nobody, NEW, Don't log UTF-8 conversion warnings during LIST
11:44:39 <instantbot> New Instantbird (UI) bug 2154 filed by aleth@instantbird.org.
11:44:44 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=2154 enh, --, ---, nobody, NEW, Improve styling of participant count in awesometab
11:46:25 <-- dionisos has quit (Ping timeout)
11:46:39 --> dionisos has joined #instantbird
11:47:32 <instantbot> New Instantbird (UI) bug 2155 filed by aleth@instantbird.org.
11:47:34 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=2155 nor, --, ---, nobody, NEW, Existing conversations should be ranked to the top
11:50:43 --> nhnt11_phone has joined #instantbird
11:50:49 <nhnt11_phone> Hi
11:51:06 <aleth> nhnt11_phone: Congrats on landing that patch :)
11:51:34 <nhnt11_phone> aleth: are you sure we care about pushing existing convs to the top of the list? It will most likely be useless after we have stats ranking
11:51:47 <nhnt11_phone> aleth:Heh, thanks. Lots of optimization to do...
11:52:03 <nhnt11_phone> I'll be online later to get started on it.
11:52:11 <aleth> nhnt11_phone: Existing convs have an unbeatable recency value ;)
11:52:35 <aleth> It's quite possible the ranking patch will automatically fix it of course.
11:53:13 <nhnt11_phone> There's also the point that only convs on hold are the ones that need to go to the top
11:53:36 <aleth> Hmm, that's possible.
11:53:40 <nhnt11_phone> I dont think "switch to tab" falls in the awesometab's target use case
11:53:49 <clokep_> I strongly disagreee.
11:54:06 <nhnt11_phone> Unless there are many windows i guess
11:54:27 <nhnt11_phone> I have to go
11:54:28 <aleth> For now I don't think you should make the distinction between on hold or not.
11:54:29 <nhnt11_phone> Bbl
11:54:41 <-- nhnt11_phone has quit (Quit: AndroIRC - Android IRC Client ( http://www.androirc.com ))
11:55:29 <flo-retina> aleth, nhnt11_phone: I also think that we shouldn't worry too much about the ordering of conversations until the ranking lands.
11:56:00 <aleth> flo-retina: Definitely, I only filed it so I wouldn't forget by the time ranking lands...
11:56:46 --> mconley has joined #instantbird
12:00:08 <instantbot> New Core - IRC bug 2156 filed by aleth@instantbird.org.
12:00:11 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=2156 nor, --, ---, nobody, NEW, Figure out how to handle ISUPPORT info for LIST
12:03:17 <flo-retina> any obvious reason why the participant count isn't displayed for channels on hold?
12:03:31 <aleth> In the awesometab?
12:04:00 <flo-retina> yes
12:04:08 <aleth> I think it's not displayed because there is no place yet to display it
12:04:24 <aleth> (Cramming it into the topic is just a TODO really)
12:04:44 <-- dionisos has quit (Ping timeout)
12:05:03 --> dionisos has joined #instantbird
12:06:08 <flo-retina> ok ^^
12:10:13 <-- mconley has quit (Input/output error)
12:11:45 <-- dionisos has quit (Ping timeout)
12:12:00 --> dionisos has joined #instantbird
12:42:08 <-- dionisos has quit (Ping timeout)
12:55:40 <instantbot> New Instantbird (UI) bug 2157 filed by aleth@instantbird.org.
12:55:42 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=2157 nor, --, ---, nhnt11, NEW, Call getChatRoomDefaultFieldValues lazily for awesometab roomInfos
12:57:06 <flo-retina> :)
13:03:17 <flo-retina> compiling purplexpcom gives me plenty of clang warnings that can probably disappear if we add a few MOZ_FINAL
13:09:51 <instantbot> New Instantbird (UI) bug 2158 filed by aleth@instantbird.org.
13:09:53 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=2158 nor, --, ---, nobody, NEW, [a11y] Ensure the newtab is accessible
13:15:41 --> dionisos has joined #instantbird
13:18:10 --> Mic has joined #instantbird
13:18:11 * ChanServ sets mode +h Mic 
13:39:01 --> dew1 has joined #instantbird
13:39:56 <-- dew has quit (Ping timeout)
13:43:54 <clokep_> aleth: I have no idea what "No, this bug is about when to log the conversion problem *with a separate warning*, not about changing the conversion behaviour." means. :-S
13:44:25 <aleth> It means "let's make this less noisy"?
13:44:52 <flo-retina> (and less CPU intensive ;))
13:47:38 * clokep_ is still confused what your goal in that bug is. :-S
13:47:47 <clokep_> I understand the "make this less noisy"
13:47:53 <clokep_> I don't understand how you propose to do it at all.
13:48:45 <aleth> I don't have a good proposal.
13:49:25 <clokep_> :P
13:52:31 <flo-retina> clokep_: I think aleth's proposal is to find a gsoc student to poke at the code and propose something, then have you r- a couple times, and me r- yet another time, and finally check it in ;).
13:54:01 <clokep_> I can get on board w/ that proposal.
13:55:35 <Mic> "Hey, you! You're getting paid for it, aren't you? Could you please have a look at..." ;)
13:56:17 <clokep_> You'd be a good manager.
13:56:47 <aleth> s/good/ ;)
13:56:52 <Mic> :D
13:56:58 <Mic> I need to go, have a nice day!
13:57:02 <-- Mic has quit (Quit: Instantbird -- http://www.instantbird.com)
13:57:51 <clokep_> aleth: That'll error, good try though. ;)
13:58:34 <aleth> I was lucky we haven't added sed support yet ;)
13:58:42 * clokep_ still wants to do that. :)
13:58:42 <flo-retina> "I need to go, have a nice day!" typical manager behavior too. "Hey you, fix this. Ok, now I've got to go do some stuff. Enjoy!" :-D
13:59:52 <flo-retina> sed? Or being a manager? :-P
14:00:21 <clokep_> sed. :P
14:00:30 <clokep_> We could use Mic's  cool diffing code.
14:11:36 --> jb has joined #instantbird
14:16:49 <-- dionisos has quit (Ping timeout)
14:17:03 --> mconley has joined #instantbird
14:18:57 --> dionisos has joined #instantbird
14:21:27 <-- dionisos has quit (Ping timeout)
14:25:12 --> dionisos has joined #instantbird
14:31:08 <instantbot> New Core - Twitter bug 2159 filed by clokep@gmail.com.
14:31:09 <instantbot> clokep@gmail.com requested review from florian@instantbird .org for attachment 2839 on bug 2159.
14:31:10 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=2159 nor, --, ---, nobody, ASSI, Port bug 884319
14:32:10 * clokep_ thinks flo- retina will enjoy that bug. :P
14:34:55 <flo-retina> bah...
14:35:57 <clokep_> ;)
14:36:02 <clokep_> I mean you already reviewed the code once.
14:36:06 <clokep_> Soooooo should be easy?
14:41:34 --> nhnt11 has joined #instantbird
14:41:38 <nhnt11> Hello
14:49:38 <clokep_> Hi
14:53:08 <aleth> nhnt11: So, some first feedback from using the awesometab today: 1) It works really well and is already snappy 2) as long as we're not receiving LIST data, which causes trouble even with no awesometab open. 3) The LIST results are not very useful atm unless one already knows the name of what one is looking for, this confirms we really do need better search for that. 4) You're so right about the tab closing/opening feeling weird.
14:53:41 <nhnt11> aleth: Thanks for the feedback, glad you're enjoying it ;)
14:53:45 <aleth> i.e. nothing unexpected I guess ;)
14:53:53 <nhnt11> But yeah receiving the LIST results is bad :(
14:54:00 <nhnt11> a bad experience*
14:54:25 <nhnt11> The enumerator was a great improvement
14:54:33 * nhnt11 updates his nightly
14:54:37 <aleth> You should feel good about everything working well once we have the results though ;)
14:54:50 <nhnt11> :)
14:54:59 <-- nhnt11 has quit (Quit: Instantbird 1.5a1pre -- http://www.instantbird.com)
14:55:02 --> nhnt11 has joined #instantbird
14:55:54 <nhnt11> aleth, flo-retina: So would you rather I look at more optimization tonight or started something else?
14:56:15 <aleth> nhnt11: I think optimization will be easier once ranking is present.
14:56:48 <aleth> But if you can see things to improve in discrete patches, go ahead of course ;)
14:56:55 --> wnayes has joined #instantbird
14:56:56 * clokep_ thinks the getDefault... optimization should be done easily / quickly.
14:56:59 <clokep_> And ranking is more important.
14:57:26 <aleth> That one and bug 2153 are fairly easy wins.
14:57:31 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=2153 nor, --, ---, nobody, NEW, Don't log UTF-8 conversion warnings during LIST
14:57:38 <aleth> Though the latter does lack a proposal :D
14:57:41 <nhnt11> aleth, clokep_: I think I can finish off bug 2155, bug 2157 quickly right now
14:57:45 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=2155 nor, --, ---, nobody, NEW, Existing conversations should be ranked to the top
14:57:46 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=2157 nor, --, ---, nhnt11, NEW, Call getChatRoomDefaultFieldValues lazily for awesometab roomInfos
14:58:02 <nhnt11> Until ranking lands, I really would like existing convs to be at the top :P
14:58:04 <aleth> What happened to the opacity bug btw?
14:58:10 <nhnt11> Yeah I remembered that yesterday
14:58:13 <nhnt11> I should finish it up
14:58:25 <aleth> iirc there was some feedback from flo in it on how to make it work smoothly...
14:58:37 <nhnt11> Yeah
14:59:28 <nhnt11> aleth: What exactly is your experience while LIST results are being received?
14:59:40 <nhnt11> I really don't see any lag at all for moznet.
14:59:52 * nhnt11 tries freenode
14:59:57 <aleth> The UI lags for freenode (my guess is that is mainly due to log output)
15:00:15 <nhnt11> Ah yes. It's lagging
15:00:23 <nhnt11> But it's done already so..
15:01:05 <nhnt11> Hmm, filtering is pretty fast too
15:01:38 <aleth> Filtering is snappy enough for me, but not very useful as I can't filter for what I want ;)
15:01:49 <nhnt11> Fair enough
15:02:01 <nhnt11> I was just worried about filtering so many items being slow
15:02:10 <nhnt11> But I see no issues :)
15:02:35 <nhnt11> I just thought of a potential problem, if someone has 2 accounts on the same network there will be duplicates of all the channels :-/
15:02:54 <aleth> That was on your etherpad a long time ago iirc ;)
15:03:03 <nhnt11> yeah..
15:03:16 <nhnt11> i think it was stated more generally though
15:03:20 * nhnt11 opens the etherpad
15:04:43 * nhnt11 can remove the patches for 2066 from his mq :D
15:05:11 <aleth> Right, no issues, no errors in the console, ... "_
15:05:14 <aleth> :)
15:06:35 <aleth> Time for a blogpost? :D
15:07:09 <aleth> (or maybe when we pref it on by default)
15:08:49 <nhnt11> A blogpost is long overdue :(
15:08:54 * nhnt11 will do that now
15:11:47 <aleth> nhnt11: I was referring to the instantbird.org blog, but sure ;)
15:12:04 <nhnt11> Ah
15:16:58 <flo-retina> aleth: "The UI lags for freenode (my guess is that is mainly due to log output)" guessing is often wrong for performance issues ;)
15:17:40 <aleth> flo-retina: Sure, but it's definitely an easy win ;)
15:18:16 <aleth> My guess is based on watching the log messages scroll by in the terminal... that's not free.
15:18:21 <flo-retina> but it's only 10% of the time in yesterday's profiles
15:18:38 <aleth> I never expected it to solve the whole problem.
15:19:12 <aleth> But I do hope it may solve most of the responsiveness issue.
15:19:17 <flo-retina> nhnt11: We identified yesterdays that we have 2 different performance problems: 1. Excessive CPU usage. 2. UI freeze.
15:19:40 <flo-retina> I would like you to fix 2. so that we can pref on, before we move all our reviewing/polishing efforts to ranking.
15:20:02 <nhnt11> ok
15:20:24 <flo-retina> and if you can fix some stuff contributing to 1. quickly, that's nice.
15:20:29 * nhnt11 isn't sure where to start to fix the UI freeze...
15:20:34 <nhnt11> possibly make some stuff async
15:20:41 <flo-retina> but I expect we will keep working on that for a while, with iterative improvements.
15:20:48 <aleth> Right.
15:21:06 <flo-retina> nhnt11: yeah, stuff needs to be async.
15:21:20 <flo-retina> nhnt11: look at how convbrowser.xml queues messages that need to be displayed
15:21:22 <aleth> A lot of it already is async.
15:21:51 <nhnt11> flo-retina: Do you think an executeSoon on _addChat would help?
15:22:08 <flo-retina> that's part of the solution
15:22:17 <flo-retina> I'm in a meeting right now
15:22:19 <nhnt11> I thought of another thing
15:22:23 <nhnt11> ok sure
15:22:38 <aleth> nhnt11: The thing to find out is whether you have to queue at the protocol level or the stats service level or both
15:22:38 <nhnt11> anyway, when a newtab gets an updated notification it refreshes all the items it's already displaying
15:22:48 <nhnt11> but when stuff is /added/, other stuff doesn't need to be refreshed
15:22:51 <flo-retina> aleth: stats level
15:23:16 <nhnt11> From the profiling yesterday it seems _addChat is the most expensive thing happening
15:23:37 <aleth> Oh cool, then that's already discovered :)
15:28:02 <aleth> flo-retina: Can you link to the profiler output of _addChat so nhnt11 can see it?
15:28:12 <nhnt11> It was linked yesterday I think
15:42:04 <aleth> That profile has lots of interesting stuff in it...
15:48:18 --> nhnt11-testing has joined #instantbird
15:48:29 <nhnt11-testing> Hmm, I'm running a LIST on freenode right now
15:48:32 <nhnt11-testing> No lag :)
15:48:53 <nhnt11-testing> Looks like executeSoon(this._addChat) works
15:49:02 <aleth> Removing the encoding warnings on their own doesn't improve the UI freeze much.
15:49:24 * nhnt11-testing goes to try this on a debug build
15:49:27 <-- nhnt11-testing has quit (Quit: Instantbird 1.5a1pre -- http://www.instantbird.com)
15:52:04 <aleth> Ah, I see why I thought it did before :D I had a command line flag set that always opened the error console.
15:52:24 <aleth> But the window manager remembered it was last minimized so I didn't see it.
15:52:50 <nhnt11> executeSoon isn't helping on a debug build :(
15:53:10 <nhnt11> totally frozen...
15:53:54 <flo-retina> nhnt11: not really surprising
15:53:55 <aleth> nhnt11: Executesoon isn't really queueing. Take a look at the convbrowser code flo mentioned
15:54:00 <flo-retina> nhnt11: can you pastebin your executesoon stuff?
15:54:02 <nhnt11> ok
15:54:23 <aleth> nhnt11: http://lxr.instantbird.org/instantbird/source/chat/content/convbrowser.xml#311
15:54:36 <nhnt11> flo-retina: I just put an executeSoon around addChat in the forEach, it got rid of the lag completely on a normal build
15:54:38 <nhnt11> aleth: thanks
15:55:21 <nhnt11> http://pastebin.instantbird.com/321897
15:55:23 <nhnt11> flo-retina: ^
15:55:29 <flo-retina> the code you need is: 1. Insert in the queue and return really if the boolean was true. 2. Set a boolean to true. 3. Execute soon a function that will take some stuff out of the queue and process it for a few ms, then call execute soon until the queue is empty. If the queue is empty, set the boolean to false.
15:55:48 <flo-retina> s/return really/return early/
15:56:12 <aleth> Basically you can copypaste from the code I linked ;)
15:56:16 <nhnt11> I understand. :)
15:56:26 <nhnt11> I've seen the convbrowser queue code before
15:56:31 <flo-retina> aleth: I dislike that code btw
15:56:50 <flo-retina> aleth: it sucks because if more than one thing is playing with similar code at once, we still freeze the UI because the timers add up :(
15:57:12 <flo-retina> would be nice if we had a way to check how much time was spent since the last time we returned to the event loop
15:57:19 <aleth> flo-retina: True, for multiple tabs being filled at the same time it fails.
15:58:44 --> qheaden has joined #instantbird
15:58:54 <qheaden> Hello everyone.
16:00:18 <nhnt11> aleth: Is there a reason that code stores the next pending message index rather than just shifting the pendingMessages array?
16:00:19 <nhnt11> Hi qheaden
16:00:26 <qheaden> Hey nhnt11. :)
16:00:35 <nhnt11> (I'm guessing it's because shifting is expensive? :P )
16:00:49 <flo-retina> nhnt11: it's needed for the progressbar
16:00:55 <nhnt11> Ah right
16:01:06 <nhnt11> Cool
16:01:16 <flo-retina> aleth: btw, the progress bar is really annoying to deal with if I switch that code to using an enumerator instead of an array :(
16:01:17 <qheaden> flo-retina, clokep_: So it looks like the soft pencils down date is in 10 days. How are we going to submit code samples to Google? Should I just send them the Yahoo prpl files?
16:01:27 <flo-retina> aleth: I would basically need to send the total message count separately :(
16:01:55 <flo-retina> qheaden: it doesn't really matter. If I were you I would just send them the content of chat/protocols/yahoo/
16:02:06 <aleth> flo-retina: I would also like to obsolete the progress bar
16:02:06 <qheaden> Okay
16:02:15 <flo-retina> aleth: what would you put instead of it?
16:02:29 * nhnt11 wonders what he should submit
16:02:37 <nhnt11> Maybe I should just tarball all the patches I've written? :P
16:02:38 <aleth> flo-retina: Only display a few messages to start with... obviously that's a big project.
16:02:38 <flo-retina> nhnt11: it's not as easy for you.
16:03:00 <flo-retina> nhnt11: but I think hg export'ing all the changesets for which you are the author would be reasonable.
16:03:15 <aleth> flo-retina: My point is that ideally we don't want to be in the situation where we need a progress bar.
16:03:27 <flo-retina> aleth: hmm
16:04:08 <aleth> But I haven't yet got an idea for how to do it that I am convinced will work :-/
16:04:13 <flo-retina> aleth: makes sense, but sounds like a larger project than my relatively-short-term of getting rid of the UI freeze when displaying a large log.
16:04:26 <qheaden> aleth: Could you briefly help me test out Yahoo conferences again?
16:04:33 <aleth> flo-retina: Yeah...
16:04:34 * nhnt11 should write a script to export those patches
16:04:37 <qheaden> Now that I have packet logging, I can see what's going on.
16:04:43 <aleth> qheaden: OK
16:05:19 <aleth> flo-retina: It may not be so bad to send the total message count separately.
16:06:30 <flo-retina> nhnt11: Don't write a script, just ask hg:
16:06:30 <flo-retina> hg export `hg log |grep -B 1 nhnt11@|grep changeset|cut -d ':' -f 3`
16:07:12 <nhnt11> flo-retina: By write a script I meant work out what I need to grep :P Thanks :D
16:07:34 * flo-retina doesn't want you to waste time poking at shell scripting _now_ :)
16:07:55 <nhnt11> flo-retina: I wasn't going to do it _now_ :P
16:10:05 <aleth> nhnt11: Btw shift is probably more expensive as well ;)
16:11:59 <-- dionisos has quit (Ping timeout)
16:23:31 <nhnt11> So my GSoC diff is 6187 lines long, with 1644 added lines and 842 removed
16:23:36 <nhnt11> It felt like more than that :P
16:24:40 <flo-retina> nhnt11: strange, I get " 71 files changed, 2535 insertions(+), 800 deletions(-)"
16:24:54 <flo-retina> (when I put |diffstat at the end of the command I gave you)
16:25:32 <nhnt11> flo-retina: I just did |cat gsoc.diff |grep ^+|wc -l| because I'm a noob
16:26:04 <flo-retina> nhnt11: ah, so you couldn't resist the urge of writing shell scripts _now_, and you wrote one that doesn't work? :-P
16:26:17 <nhnt11> Ok, when I do diffstat I get 1571 insertions, 769 deletions
16:26:20 <nhnt11> flo-retina: I was curious :P
16:26:28 <nhnt11> Oh and 57 files changed
16:26:50 <aleth> nhnt11: Ranking will give you another couple of extra lines ;) Not that it matters. Quality counts
16:27:00 <nhnt11> Sure, I was just curious like I said
16:27:34 * aleth finds 71 files changed impressive considering it was only a couple of months :D
16:27:53 <flo-retina> aleth: gerv can to much better ;)
16:28:07 <aleth> flo-retina: Yeah, it didn't take him long at all ;)
16:28:24 * nhnt11 wonders why he's getting 57 files changed
16:29:17 <nhnt11> What did gerv do?
16:30:06 <flo-retina> nhnt11: change all the license headers :-P
16:30:17 <nhnt11> oh :D
16:33:59 * clokep_ goes home. :P
16:35:49 <-- clokep_ has quit (Quit: http://www.mibbit.com ajax IRC Client)
16:35:57 * flo-retina is jealous
16:49:50 <qheaden> My determined goal by the end of next week is to have conferences fully working.
16:50:13 <qheaden> Some time next week, we can possibly move chatdev to a Yahoo conference to test things.
16:50:48 * qheaden begins to wonder why chatdev isn't named devchat.
16:54:52 <-- flo-retina has quit (Quit: Instantbird 1.5a1pre -- http://www.instantbird.com)
16:56:42 <nhnt11> looks like the queue didn't help :(
16:56:50 <nhnt11> debug build still freezes..
16:56:59 <nhnt11> oh wait, there are errors
16:57:00 <nhnt11> oops
16:58:45 <nhnt11> Well i fixed it, and it still freezes
17:00:58 --> Mook_as has joined #instantbird
17:00:59 <nhnt11> Interesting
17:01:12 <nhnt11> LIST seems to be over, but it's still at 100% CPU usage
17:02:37 <-- wnayes has quit (Ping timeout)
17:02:43 --> qlum has joined #instantbird
17:04:05 --> nhnt11-testing has joined #instantbird
17:04:21 <nhnt11-testing> Well on my normal build it's snappy...
17:05:23 <-- nhnt11-testing has quit (Quit: Instantbird 1.5a1pre -- http://www.instantbird.com)
17:07:06 --> wnayes has joined #instantbird
17:10:41 --> clokep has joined #instantbird
17:10:41 * ChanServ sets mode +o clokep 
17:13:42 <clokep> qheaden: Because it's named after #maildev.
17:13:55 <clokep> So it's "chat developers", why would devchat make more sense?
17:14:19 <qheaden> Ahh, that makes sense. I was thinking of it saying "developer chat" or something.
17:15:46 --> nhnt11-testing has joined #instantbird
17:15:59 <qheaden> aleth: Are you friends with clokep on Yahoo?
17:16:18 <nhnt11-testing> Hmm, my CPU usage is down to 30%  while doing LIST...
17:16:25 <nhnt11-testing> I guess that's an improvement at least
17:16:38 <nhnt11-testing> This is on my normal build of course, debug build is still at 100
17:17:59 <nhnt11-testing> Also, 25 chats are added to the queue at a time, and 25 chats are added in each call to addPendingChats, so it appears the queuing isn't helping much
17:18:12 <nhnt11-testing> (Aside from delaying the addition of these 25 chats)
17:18:25 <nhnt11-testing> (but that can be done with a simple exectueSoon anyway)
17:18:49 <-- gerard-majax_ has quit (Ping timeout)
17:19:08 <nhnt11-testing> Changing it to 250 channels per batch results in the queue being of some use, but that's like creating a problem for the solution
17:19:34 <aleth> qheaden: no
17:20:13 <qheaden> aleth: That's fine. In fact, don't friend him. I would like to soon test inviting people who aren't on the contacts list again.
17:21:06 <aleth> nhnt11: Why did you change it to 25 in the first place?
17:21:15 <clokep> Florian asked for it.
17:21:24 <aleth> Yes, but why?
17:21:42 <clokep> I don't read minds.
17:21:50 <aleth> I was asking nhnt11 ;)
17:22:15 <aleth> nhnt11: If the CPU load is down that already sounds like a plus. Not everyone has a new macbook ;)
17:23:26 <aleth> nhnt11: Do you have a diff of your current changes?
17:25:57 <nhnt11> aleth: It's faster to transfer 25 wrapped objects at a time
17:26:14 <aleth> But is it 10 times faster?
17:26:19 <nhnt11> nah
17:26:27 <nhnt11> It did help a bit though
17:26:35 <aleth> Sounds like a workaround for a problem you may have fixed by the queueing
17:26:39 <nhnt11> aleth: I'll get you a diff.
17:26:40 <nhnt11> yeah
17:26:42 <nhnt11> it kinda was.
17:26:54 <-- nhnt11-testing has quit (Client exited)
17:28:20 <nhnt11> aleth: http://pastebin.instantbird.com/322078
17:29:48 <aleth> nhnt11: Have you considered batching 250 and providing the data via an enumerator
17:30:02 <nhnt11> aleth: yes
17:30:11 <nhnt11> I'll try it next
17:31:44 <qheaden> It looks like the fix for bug 2150 is simply undoing the patch for bug 2094. :-S
17:31:48 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=2150 maj, --, ---, nobody, NEW, Conference invites don't work
17:31:49 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=2094 nor, --, 1.5, qheaden, RESO FIXED, Required key 52 wasn't found.
17:32:21 <qheaden> And when I undo the work for 2094, I still can't reproduce the error in that bug. Something I must have changed somewhere else that fixed that issue.
17:33:34 <aleth> Strange. Required key now found? :P
17:35:03 <qheaden> Seems to be that way.
17:35:12 <qheaden> Maybe the key was on vacation. :P
17:36:04 * nhnt11 is trying the enumerator
17:36:27 --> nhnt11-testing has joined #instantbird
17:36:54 <-- clokep has quit (Quit: Instantbird 1.5a1pre -- http://www.instantbird.com)
17:37:18 <nhnt11-testing> Hmm, no difference on my normal build as expected
17:37:23 * nhnt11-testing tries the debug build
17:37:24 <-- nhnt11-testing has quit (Quit: Instantbird 1.5a1pre -- http://www.instantbird.com)
17:38:01 <nhnt11> My debug build is using 100% cpu even before LIST :/
17:38:08 <nhnt11> Ok it stopped, I think it was receiving vcards
17:38:30 <nhnt11> Bah, still completely unusable
17:38:46 <aleth> nhnt11: Are you dump()ing timings to see if your changes are improving things?
17:39:14 <nhnt11> No, I'm just seeing if the UI response is improving
17:39:17 <aleth> Debug builds also freeze when opening big conversations etc...
17:39:28 <aleth> not that that is a good thing of course.
17:41:16 <nhnt11> aleth: Is there any use dumping timings? I'm not improving the performance of any operations, I'm just delaying operations so that the UI doesn't freeze
17:41:54 <nhnt11> hmm, but on the plus side, it should no longer freeze so badly that the user is timed out
17:41:58 <nhnt11> let me try
17:41:59 <aleth> Really it would help if you could use the profiler.
17:42:13 <nhnt11> Yeah, I'll learn how to use it tonight
17:42:31 <aleth> I'm not sure it's straightforward at all. Afaik flo had to hack some things.
17:42:35 --> nhnt11-testing has joined #instantbird
17:42:41 <nhnt11> Well I can try :)
17:42:41 <aleth> We're hoping some work going on for gsoc will change that.
17:42:55 <nhnt11> Let's see if nhnt11-testing gets timed out...
17:44:23 <aleth> Another idea might be to replace your executeSoons with executeIn50ms or whatever and see if the debug build /still/ freezes.
17:44:40 <nhnt11> Hmm
17:45:02 <nhnt11-testing> I'm still here \o/
17:45:04 <aleth> It seems to me you don't know yet why it is freezing...
17:45:11 <aleth> nhnt11: :)
17:47:12 * nhnt11 should research more on how the event queue works
17:49:24 * nhnt11 googled and didn't learn much that he didn't already know from 2 minutes of reading
17:53:13 <nhnt11> Well, I'll go ahead and submit a patch
17:54:00 <aleth> If it's improved the situation it's already worthwhile...
17:54:19 <aleth> What functions are taking so long (in the debug build) before returning control to the UI that you end up freezing?
17:55:29 <nhnt11> aleth: It appears when the function completes, IRC already has more channels ready, so another batch is added before control is returned to the UI
17:55:49 <nhnt11> So I was thinking of increasing the number of channels returned per batch further and seeing if that helps
17:56:01 <nhnt11> So that the UI has a chance to respond before another batch of channels is returned
17:56:04 <nhnt11> I may be completely wrong
17:56:24 <-- nhnt11-testing has quit (Client exited)
17:56:27 <instantbot> qheaden@phaseshiftsoftware.com requested review from clokep@gmail.com for attachment 2840 on bug 2150.
17:56:29 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=2150 maj, --, ---, qheaden, ASSI, Conference invites don't work
17:56:40 * qheaden still isn't too satisfied with that patch. :-(
17:58:15 <aleth> "I was thinking of increasing the number of channels returned per batch further and seeing if that helps" or maybe the service should signal that it is ready for the next batch rather than a fixed batch size?
17:58:23 <Mook_as> hrm, normally you'd want smaller batches in that case, to give more chances of running actual UI
17:59:59 --> nhnt11-testing has joined #instantbird
18:00:21 <instantbot> qheaden@phaseshiftsoftware.com requested review from clokep@gmail.com for attachment 2841 on bug 2093.
18:00:24 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=2093 nor, --, ---, qheaden, ASSI, "Adding chat buddy twice" error when inviting someone to a chatroom
18:00:42 <nhnt11> aleth: I'm thinking I'm trying to fix this problem in the wrong place
18:00:49 <nhnt11> Doing a /list completely freezes the UI too
18:01:03 <nhnt11> (Without ever opening an awesometab)
18:01:16 <nhnt11> nhnt11-testing is completely unresponsive...
18:01:39 --> gerard-majax_ has joined #instantbird
18:01:39 <nhnt11-testing> Ah, I'm back
18:02:45 <aleth> nhnt11: So for debug builds, just handling the messages is too much?
18:03:12 <nhnt11> aleth: Yeah, I'm thinking I should queue LIST messages on the protocol side
18:03:28 <-- nhnt11-testing has quit (Quit: Instantbird 1.5a1pre -- http://www.instantbird.com)
18:03:33 <aleth> nhnt11: You would probably have to queue them at the socket.
18:04:31 <nhnt11> aleth: Why not queue them at the handler?
18:04:46 <aleth> If this is really a problem, why fix it only for IRC?
18:04:52 <nhnt11> Although yeah it's better practice to queue it at the socket level
18:04:52 <nhnt11> yeah
18:05:45 <nhnt11> Yeah it should be done at socket level
18:05:53 <nhnt11> I made the handler for LIST return without doing anything
18:05:55 <nhnt11> still freeze
18:06:19 * nhnt11 gets to play with yet another part of Ib
18:07:36 * nhnt11 doesn't want to do this now
18:07:48 <nhnt11> I think I'll go get something to eat/drink and come back to hack at this
18:08:14 <aleth> iirc we actually discussed that we might have to do something like this a long time ago, when you first started looking at LIST
18:14:04 <nhnt11> I'll be back later
18:14:24 * nhnt11 is now known as nhnt11_away
18:20:15 <-- nhnt11_away has quit (Ping timeout)
18:21:24 * qheaden has homework to do.
18:21:29 <qheaden> Bye all.
18:38:22 --> nhnt11 has joined #instantbird
18:39:30 <-- Mook_as has quit (Quit: Mook_as)
18:40:52 <-- nhnt11 has quit (Ping timeout)
18:42:03 --> nhnt11 has joined #instantbird
18:42:22 --> Mook_as has joined #instantbird
19:18:11 <-- chrisccoulson has quit (Quit: OSError: [Errno 130] Owner died)
19:21:57 * aleth is looking at bug 2153
19:22:00 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=2153 nor, --, ---, nobody, NEW, Don't log UTF-8 conversion warnings during LIST
19:22:20 <-- gerard-majax_ has quit (Quit: Ex-Chat)
19:22:25 --> gerard-majax__ has joined #instantbird
19:31:01 <-- jb has quit (Ping timeout)
19:32:32 <-- gerard-majax__ has quit (No route to host)
19:32:38 --> gerard-majax_ has joined #instantbird
19:33:14 <-- gerard-majax_ has quit (Client exited)
19:33:20 --> gerard-majax__ has joined #instantbird
19:39:10 <-- gerard-majax__ has quit (Quit: Ex-Chat)
19:39:13 --> gerard-majax_ has joined #instantbird
19:45:40 <instantbot> aleth@instantbird.org requested review from clokep@gmail.com for attachment 2842 on bug 2153.
19:45:45 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=2153 nor, --, ---, aleth, ASSI, Don't log UTF-8 conversion warnings during LIST
19:47:19 <-- gerard-majax_ has quit (Quit: Ex-Chat)
19:47:26 --> gerard-majax__ has joined #instantbird
19:48:23 <instantbot> aleth@instantbird.org cancelled review?(clokep@gmail.com) for attachment 2842 on bug 2153.
19:48:24 <instantbot> aleth@instantbird.org requested review from clokep@gmail.com for attachment 2843 on bug 2153.
19:48:25 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=2153 nor, --, ---, aleth, ASSI, Don't log UTF-8 conversion warnings during LIST
19:50:38 <-- gerard-majax__ has quit (Ping timeout)
19:51:15 --> gerard-majax__ has joined #instantbird
19:53:22 <-- gerard-majax__ has quit (Quit: Ex-Chat)
19:53:25 --> gerard-majax_ has joined #instantbird
19:57:03 <-- aleth has quit (Quit: Ciao)
20:11:32 --> jb has joined #instantbird
20:15:53 <-- jb has quit (Ping timeout)
20:21:01 --> jb has joined #instantbird
20:38:16 <-- Mook_as has quit (Ping timeout)
20:40:09 --> Mook_as has joined #instantbird
20:40:32 --> flo-retina has joined #instantbird
20:40:32 * ChanServ sets mode +qo flo-retina flo-retina 
20:41:18 <-- nhnt11 has quit (Input/output error)
20:41:23 --> nhnt11 has joined #instantbird
20:50:03 * flo-retina is trying to read the log, and is confused by a lot of things that have been said that seem just wrong
20:50:25 <flo-retina> aleth: the gsoc project is to get the browser debugger working; the profiler already works
20:53:03 <flo-retina> nhnt11: is it possible to have a status update/summary of where you are/how I can help?
20:54:27 <nhnt11> flo-retina: I made the stats service use a queue to add chats, and it helped, but not fully. I then did more debugging (that I should have done before) and decided there was a bigger problem in that even if list results weren't processed at all, just receiving them was causing a UI lag
20:55:05 <nhnt11> So aleth suggested introducing queuing at the socket level, which I haven't started working on yet
20:56:12 <-- jb has quit (Ping timeout)
20:56:32 <flo-retina> nhnt11: we should profile that before starting any work on it
20:56:53 <nhnt11> Right, honestly I didn't want to start on it myself so I waited for you to come back online :P
21:03:14 --> jb has joined #instantbird
21:05:08 * nhnt11 deduces after looking at socket.jsm that onDataReceived is the call that needs to be queued
21:05:34 <nhnt11> flo-retina: So how do I use the profiler?
21:11:18 <flo-retina> nhnt11: so what's the code you tried?
21:11:40 <nhnt11> flo-retina: To queue adding chats?
21:11:45 <flo-retina> yes
21:12:12 <nhnt11> flo-retina: http://pastebin.instantbird.com/322078
21:12:48 <-- mconley has quit (Input/output error)
21:19:56 <-- jb has quit (Ping timeout)
21:21:32 --> florian has joined #instantbird
21:26:04 <flo-retina> nhnt11: so I'm profiling right now without this patch on an optimized build. I don't get any UI freeze :-S
21:26:16 <-- qlum has quit (Quit: Getting the <censored> out.)
21:26:46 <nhnt11> Yeah it's not bad. Like I said, I just noticed slightly choppy scrolling.
21:26:56 <nhnt11> And teeny delays while typing if I looked for them.
21:27:02 <nhnt11> The patch prevents me timing out on a debug build though
21:27:25 <flo-retina> so many what we would need to do is test on a slower machine
21:27:42 <-- Mook_as has quit (Quit: Mook_as)
21:27:45 * nhnt11 will take out his XPS
21:28:42 <flo-retina> or we can optimize on the debug build... with the possibility that we are doing pointless work (if the slow down caused by the debug build isn't similar at all to the slow down we would have on a slower machine)
21:31:25 <-- florian has quit (Quit: Instantbird 1.5a1pre -- http://www.instantbird.com)
21:37:45 --> florian has joined #instantbird
21:41:49 <nhnt11> flo-retina: Er, so how do I use the profiler?
21:42:02 <flo-retina> you read your emails
21:42:18 <nhnt11> A notification popped up just as you said that
21:42:19 <nhnt11> thanks
21:42:32 <flo-retina> well, you asked that just when I was clicking the "send" button :-P
21:43:30 <flo-retina> so, the UI isn't frozen, but it's still a bit janky
21:44:07 <flo-retina> http://people.mozilla.com/~bgirard/cleopatra/#report=7df4fecb1a6ec0e4c86563a504e23a5f557c2916
21:45:26 <-- florian has quit (Quit: Instantbird 1.5a1pre -- http://www.instantbird.com)
21:45:35 <flo-retina> I'll try again without scrolling at the same time
21:45:50 --> florian has joined #instantbird
21:48:42 <nhnt11> This is what I got: https://people.mozilla.com/~bgirard/cleopatra/#report=22ca009399b7c3b52c2b3c2f973671f4897e910b
21:49:31 <nhnt11> That's with the list command returning without doing anything
21:49:49 <nhnt11> handler*
21:50:01 <flo-retina> ah
21:50:12 <flo-retina> yeah, I was surprised that scriptError was taking so much!
21:51:17 <flo-retina> nhnt11: well, that profile mostly shows we should spend less time doing debug logs, and less time calling Socket.resetPingTimer
21:52:20 <flo-retina> nhnt11: wait, you had the error console open. That changes completely the results.
21:52:40 * nhnt11 doesn't remember if he did
21:52:45 <nhnt11> I'll try it again
21:53:02 <flo-retina> you don't need to remember, given the time spent in consoleBindings.xml ;)
21:54:03 <flo-retina> http://people.mozilla.com/~bgirard/cleopatra/#report=3de11e768335477014c33c9bf047cd285343a626 without scrolling this time
21:58:04 <-- florian has quit (Quit: Instantbird 1.5a1pre -- http://www.instantbird.com)
21:58:28 <nhnt11> Hmm
21:59:10 --> chrisccoulson has joined #instantbird
22:00:20 --> florian has joined #instantbird
22:00:56 <nhnt11> flo-retina: You're on a normal build, correct?
22:01:28 <flo-retina> built with   ac_add_options --disable-debug --enable-optimize --enable-profiling
22:01:38 <nhnt11> ok
22:01:53 <nhnt11> flo-retina: And these profiles are without the queuing patch?
22:02:08 <nhnt11> http://people.mozilla.com/~bgirard/cleopatra/#report=037d2bb9c8549c226cb465bdf07e02e9b73fddf8
22:02:10 <flo-retina> yes
22:02:15 <nhnt11> That's with the queuing patch
22:02:21 <nhnt11> And with list enabled
22:06:06 <flo-retina> I wonder if we should try dropping nsIClassInfo on RoomInfo objects
22:08:00 <-- florian has quit (Quit: Instantbird 1.5a1pre -- http://www.instantbird.com)
22:08:39 --> florian has joined #instantbird
22:10:29 <flo-retina> nhnt11: something I was also wondering: would we get more significant profiler results if instead of using a debug build to slow things done, we put all the CPU cores at 100% first, and then launched the stuff we want to profile
22:11:44 <nhnt11> hmm
22:13:53 <nhnt11> I was thinking of trying on a VM with limited resources
22:14:09 <nhnt11> But I don't really want to spend time setting up a VM...
22:15:05 <nhnt11> I guess that's not the same as what you suggested anyway
22:21:30 <flo-retina> I was thinking writing a C program that just does an infinite loop, and launching it in the background as many times as my machine has cores.
22:22:08 <nhnt11> That would work I guess
22:23:28 <flo-retina> I think I'm too tired to write anything useful tonight and should rather go to bed
22:24:02 <nhnt11> flo-retina: I wrote it
22:24:08 <nhnt11> All my coes are at 98% ;)
22:24:13 <nhnt11> cores*
22:24:22 <nhnt11> Not that there was anything to write.. meh
22:27:17 <flo-retina> I wasn't talking about the C code :-P
22:27:27 <nhnt11> obviously :P
22:27:39 <flo-retina> I would like to optimize the QueryInterface/getInterfaces stuff
22:28:01 <flo-retina> but I'm too tired to test correctly the results of my changes, so I could produce a patch, but that would be a bit 'random'
22:28:16 <nhnt11> Entropy of fatigue :P
22:28:25 <nhnt11> Hey, that shortens to EOF :P
22:29:05 <nhnt11> So here's my profile with all cores at 98-99%: http://people.mozilla.com/~bgirard/cleopatra/#report=b72eff431c3abefad11c749a7674450d369611c5
22:30:17 <flo-retina> that's with the queing patch?
22:30:21 <nhnt11> Yeah.
22:30:38 <flo-retina> maybe it really helps?
22:30:49 <flo-retina> if it does we should file a bug/clean it up/review it
22:30:58 <nhnt11> It does seem to help yes.
22:31:06 <nhnt11> I'll upload a patch.
22:33:28 <nhnt11> flo-retina: All of these profiles are while doing a LIST on freenode btw
22:33:45 <flo-retina> I was doing it with a moznet + a freenode account
22:33:50 <flo-retina> not sure if that makes a significant difference
22:34:02 <nhnt11> Yeah moznet was on too
22:34:52 <flo-retina> nhnt11: in http://pastebin.instantbird.com/322078 I was wondering what the point of the _addChatToQueue method was
22:35:03 <flo-retina> couldn't you just inline it?
22:35:31 <flo-retina> and avoid the if (this._addingPendingChats) test for each room
22:35:48 <nhnt11> flo-retina: I figured the same code may be needed later if we add chats from cache
22:36:05 <flo-retina> so it's a premature optimization? :-P
22:36:14 <nhnt11> Lol. Ok.
22:36:57 <nhnt11> flo-retina: Oh, I seem to have given you a slightly older version of the patch
22:37:14 <nhnt11> I made the onRoomInfoAvailable api return an enumerator instead of an array of RoomInfo
22:37:59 <nhnt11> (and changed channels per batch to 250)
22:38:31 <nhnt11> I'll do some testing and see if it's making enough of a difference to include in the patch
22:39:45 <nhnt11> Actually I like the enumerator more even if it doesn't make much of a difference... but I don't really have a reason for liking it :P
22:51:09 <flo-retina> Good night
22:51:15 <nhnt11> night
23:08:47 --> mconley has joined #instantbird
23:18:43 <instantbot> New Instantbird (UI) bug 2160 filed by nhnt11@gmail.com.
23:18:44 <instantbot> nhnt11@gmail.com requested review from florian@instantbird .org for attachment 2844 on bug 2160.
23:18:46 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=2160 nor, --, ---, nobody, NEW, UI lags while stats service is receiving chat rooms
23:21:22 <-- nhnt11 has quit (Quit: Instantbird 1.5a1pre -- http://www.instantbird.com)
23:47:21 <-- mconley has quit (Input/output error)