#instantbird log on 07 30 2014

All times are UTC.

00:06:01 <-- rosonline has quit (Ping timeout)
00:07:56 --> rosonline has joined #instantbird
00:10:37 --> qheaden has joined #instantbird
00:10:44 <qheaden> Hello all.
00:11:45 <-- qheaden has quit (Input/output error)
00:12:03 --> qheaden has joined #instantbird
00:12:23 <clokep> Hello.
00:13:23 * qheaden is happy to see 64-bit Linux nightlies
00:13:32 <clokep> qheaden: Are you using them? :)
00:13:37 <qheaden> Yep.
00:13:48 <qheaden> Before I would use the one I compile myself.
00:15:01 --> clokep_work has joined #instantbird
00:15:01 * ChanServ sets mode +o clokep_work 
00:32:37 <qheaden> Argh. I checked out the latest comm-central without seeing that the tree is closed due to bustage. :(
00:44:40 <clokep> Yes. :-\
00:44:55 <qheaden> clokep: How long has it been busted?
00:45:07 * qheaden is trying to see if he can hg revert somewhere.
00:45:44 <qheaden> Nevermind. I'm looking on tbpl.
00:46:25 <clokep> qheaden: A few days.
00:46:33 <qheaden> Okay.
01:04:12 <qheaden> I kept getting configure errors about how DIRS should be used in a few moz.build files when PARALLEL_DIRS is used instead. I just replaced PARALLEL_DIRS with DIRS using sed. Is this something that any of you have encountered?
01:04:18 <qheaden> I'm wondering if a patch is needed somewhere.
01:05:02 <qheaden> Cool. Looks like it is a thing: Bug 1045762
01:05:04 <instantbot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=1045762 nor, --, ---, nobody, ASSI, Switch PARALLEL_DIRS to DIRS in .build files.
01:06:02 <clokep> qheaden: Your m-c is probably still too new.
01:06:17 <qheaden> Probably. I just did a checkout with client.py
01:06:28 <-- Mnyromyr has quit (Quit: ChatZilla 0.9.86 [SeaMonkey 1.1.19/2010030105])
01:06:40 <-- clokep_work has quit (Ping timeout)
01:06:50 <qheaden> clokep: I want to add to the replacements this guy made in this bug. Should I perhaps make a separate bug for Ib code?
01:07:20 <clokep> qheaden: Yes.
01:07:27 <qheaden> Okay
01:07:35 <clokep> qheaden: If you can get a build working by including the chat portions too, do that and we'll commit it before them.
01:08:34 <qheaden> clokep: I've got a build running now. It was a configure-level issue, so I think my patch solved it. But I will wait until the build completes before positing the patch.
01:09:10 <qheaden> clokep: Should I put the original bug number in the title, perhaps saying we are porting changes?
01:09:39 <clokep> qheaden: If you can find the original m-c bug, yes.
01:09:48 <qheaden> Okay.
01:12:03 <clokep> :)
01:12:31 <instantbot> New Chat Core - General bug 1045947 filed by qheaden@phaseshiftsoftware.com.
01:12:32 <instantbot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=1045947 nor, --, ---, nobody, NEW, Fix configure failure caused by using PARALLEL_DIRS in moz.buld files
01:14:11 <qheaden> I found the original bug and updated the title.
01:14:13 <qheaden> Bug 1045484
01:14:15 <instantbot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=1045484 nor, --, ---, nobody, NEW, TOOL_DIRS, TEST_TOOL_DIRS and PARALLEL_DIRS are no more
01:14:20 <qheaden> That's the original.
01:15:19 <-- Mook_as has quit (Quit: Mook_as)
01:17:17 <clokep> Hah
01:18:02 <clokep> qheaden: I updated it to the way we normally do it fyi. :)
01:36:13 <qheaden> clokep: Okay, thanks.
01:37:04 <clokep> qheaden: Sure, let me know if you have questions.
01:44:34 <qheaden> clokep: As for the patch message, should I just copy the bug title verbatim?
01:45:01 <clokep> qheaden: Yes.
01:53:14 <clokep> qheaden: Most likely r? flo as I'll probably be in bed by the time you rbuild finishes.
01:53:25 <clokep> But If I'm still awake, feel free to ask me.
01:53:57 <qheaden> clokep: Oh yeah, the build failed. :P
01:54:04 <qheaden> but not as a result of the bug I was fixing.
01:54:08 <qheaden> It was some C++ bug.
01:54:35 <qheaden> The bug i fixed stopped a failure during the configure process.
01:56:49 <clokep> qheaden: OK. There was another thing flo-retina was looking at...
01:56:53 <clokep> I can't seem to find it though.
01:57:33 <clokep> qheaden: https://bugzilla.mozilla.org/show_bug.cgi?id=1045329
01:57:36 <instantbot> Bug 1045329 cri, --, ---, mh+mozilla, ASSI, compile_targets needs to take into account comm-central
02:01:28 <-- rosonline has quit (Quit: Instantbird 1.5 -- http://www.instantbird.com)
02:05:07 <-- clokep has quit (Quit: Instantbird 1.5 -- http://www.instantbird.com)
02:08:37 <qheaden> clokep: Should I make a separate patch for changes to the purple extension since it is in a separate repository?
02:08:52 * qheaden didn't realize clokep went offline. :P
02:09:06 <qheaden> I'll make it separate.
02:15:04 <instant-buildbot> build #2288 of macosx-nightly-default is complete: Failure [4failed compile]  Build details are at http://buildbot.instantbird.org/builders/macosx-nightly-default/builds/2288
02:17:51 <instant-buildbot> build #1465 of win32-nightly-default is complete: Failure [4failed compile]  Build details are at http://buildbot.instantbird.org/builders/win32-nightly-default/builds/1465
02:18:11 <-- Suiseiseki has quit (Ping timeout)
02:23:03 <-- qheaden has quit (Quit: Instantbird 1.6a1pre -- http://www.instantbird.com)
02:36:10 --> Suiseiseki has joined #instantbird
02:56:35 --> mconley has joined #instantbird
03:28:05 --> Mook has joined #instantbird
03:41:04 <-- mconley has quit (Input/output error)
03:41:33 --> mconley has joined #instantbird
03:43:25 <-- mconley has quit (Ping timeout)
04:21:53 <-- gerard-majax_ has quit (Ping timeout)
04:50:43 --> jb has joined #instantbird
05:01:34 <-- Rym has quit (Ping timeout)
05:25:15 <-- jb has quit (Ping timeout)
05:39:54 --> CAKCy has joined #instantbird
05:41:04 <-- CAKCy has left #instantbird ()
05:43:38 --> Rym has joined #instantbird
05:45:23 --> gerard-majax_ has joined #instantbird
05:45:46 <-- Rym has quit (Ping timeout)
05:49:29 --> mpmc has joined #instantbird
05:55:16 <-- EionRobb has quit (Quit: Leaving.)
06:02:32 <-- Mook has quit (Ping timeout)
06:06:21 --> Mook has joined #instantbird
06:11:35 <-- Mook has quit (Ping timeout)
06:13:38 --> Mook has joined #instantbird
06:15:45 --> jb has joined #instantbird
06:34:11 --> nhnt11 has joined #instantbird
06:58:58 --> Rym has joined #instantbird
06:59:42 <-- gerard-majax_ has quit (Ping timeout)
07:27:10 --> Bollebib has joined #instantbird
07:28:56 <flo-retina> bah, bug 1045969
07:28:59 <instantbot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=1045969 nor, --, ---, mh+mozilla, ASSI, export_dirs and libs_dir need to take into account comm-central
07:29:05 <flo-retina> our build will probably stay busted for at least another day
07:32:00 * Fallen|away is now known as Fallen
07:40:25 <-- nhnt11 has quit (Ping timeout)
07:48:44 --> EionRobb has joined #instantbird
07:50:39 <-- chrisccoulson has quit (Quit: OSError: [Errno 130] Owner died)
08:01:54 --> chrisccoulson has joined #instantbird
08:26:10 --> aleth has joined #instantbird
08:26:10 * ChanServ sets mode +o aleth 
08:44:49 <-- aleth has quit (Ping timeout)
08:48:37 --> aleth has joined #instantbird
08:48:37 * ChanServ sets mode +o aleth 
08:53:57 <-- flo-retina has quit (Ping timeout)
08:56:34 <-- jb has quit (Ping timeout)
09:01:57 --> Armada has joined #instantbird
09:41:51 --> flo-retina has joined #instantbird
09:41:51 * ChanServ sets mode +qo flo-retina flo-retina 
09:47:27 --> gerard-majax_ has joined #instantbird
10:06:41 --> BWMerlin has joined #instantbird
10:13:14 --> clokep has joined #instantbird
10:13:15 * ChanServ sets mode +o clokep 
10:17:07 <instantbot> aleth@instantbird.org changed the Resolution on bug 1045622 from --- to WORKSFORME.
10:17:08 <instantbot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=1045622 nor, --, ---, nobody, RESO WORKSFORME, /back command should reconnect immediately even if a reconnect timer is active
10:17:54 <aleth> RESOLVED BADCONNECTIONISBADGETUSEDTOIT
10:22:38 <flo-retina> aleth: why are we stuck on "connecting…", aren't we supposed to have a timeout?
10:23:17 <aleth> There is a timeout, but it seems to be quite a long one.
10:23:27 <flo-retina> I thought it was either 30s or 2minutes
10:24:11 <aleth> And when the timeout is over, an immediate reconnect (no timer) usually just ends up in the same state
10:24:34 <aleth> So I don't see much we can do to improve what we have
10:30:37 <aleth> I think what might have caused me to file it is that we have throbbers signalling "attempting to join", but they don't work before the account itself is connected, so there is no visual feedback to /back in that case
10:31:20 * aleth wonders if he can fix that
10:31:50 <clokep> I've gotten stuck in connecting before...which is always weird to me. :(
10:32:58 <flo-retina> it's even stranger to me when it fails immediately without even attempting to connect
10:33:06 <flo-retina> but I suspect that bug occurs only when using a socks proxy
10:33:30 <flo-retina> and after several reconnections attempts while the connection was down.
10:33:42 <aleth> clokep: It may be when one runs into server-side issues or throttling
10:55:49 --> jb has joined #instantbird
10:59:55 <-- clokep has quit (Quit: Instantbird 1.5 -- http://www.instantbird.com)
11:16:42 <-- EionRobb has quit (Quit: Leaving.)
11:37:24 * Fallen is now known as Fallen|away
11:41:20 --> clokep_work has joined #instantbird
11:41:21 * ChanServ sets mode +o clokep_work 
11:48:54 --> sonny has joined #instantbird
12:01:07 <clokep_work> flo-retina: Well one of those bugs landed in m-c finally! :0
12:09:50 <-- BWMerlin has quit (Quit: BWMerlin)
12:14:36 <flo-retina> clokep_work: yeah...
12:14:54 <flo-retina> bug 1045969 looks similarly annoying though
12:14:56 <instantbot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=1045969 nor, --, ---, mh+mozilla, ASSI, export_dirs and libs_dir need to take into account comm-central
12:45:54 --> qheaden has joined #instantbird
12:47:31 * aleth is now known as alias
12:47:38 <-- jb has quit (Ping timeout)
12:50:50 <alias> qheaden! :)
12:50:59 * alias is now known as aleth
12:51:08 <qheaden> Hey aleth!
12:51:50 <aleth> Do you want to do that yahoo MUC test one of these days or is it all working already? 
12:52:23 * aleth can't remember
12:52:40 <qheaden> aleth: We've fixed the group chat issues (I hope all of them are gone). clokep_work and nhnt11 helped me test it out once.
12:52:50 <aleth> great! :) ship it!
12:53:05 <qheaden> :)
12:53:43 <clokep_work> qheaden: More seriously...what is left before we can ship it?
12:53:52 <clokep_work> qheaden: Is it worth at least turning it on for TB? (As that has no Yahoo support.)
12:53:59 * clokep_work thinks the answer is "Yes". ;)
12:54:29 <qheaden> clokep_work: I think we are good to go for enabling it. The next thing I was going to work on was aliases, but that isn't something that is essential.
12:54:37 <qheaden> I can add that on after.
12:54:55 <clokep_work> qheaden: I agree. :)
12:55:12 <qheaden> \o/
12:55:15 <clokep_work> qheaden: Maybe you can revert your m-c (if you have time to do this sort of thing) so you can build TB and enable it in there?
12:55:25 <aleth> Flick the pref!
12:55:49 <qheaden> clokep_work: Yeah. I plan on doing that. I'm kinda busy at work today, so when I get home, I'll try to revert to a working build and test it in TB.
12:56:13 * qheaden will brb
12:56:38 <clokep_work> aleth: More complicated. :) We have to package stuff in TB that we're not right now.
12:56:55 * aleth suspects this is what qheaden needs for the alias, but as there is no useful comment there he is not sure http://mxr.mozilla.org/comm-central/source/chat/components/public/imIContactsService.idl#236
12:59:45 <aleth> I don't see why we shouldn't turn on JS-Yahoo for IB nightlies too. It's nightlies, after all.
13:05:54 <clokep_work> aleth: That's just flipping a pref. :)
13:06:22 <flo-retina> clokep_work: what, you think aliases aren't essential? :-o
13:06:26 <flo-retina> (just kidding)
13:11:18 <clokep_work> flo-retina: I think some TB users might want support for Yahoo without aliases!
13:12:31 <aleth> If the line I pointed to is the right one, it shouldn't be too hard to support them anyway
13:12:54 <clokep_work> aleth: I think the difficulty is /getting/ them from Yahoo.
13:13:14 <aleth> Ah OK
13:14:42 <qheaden> aleth: What clokep_work said.
13:15:03 <qheaden> Aliases are really easy to implement in the client side, but handling it on the server side is complex.
13:15:30 <qheaden> It will involve me checking how the official client handles it with WireShark.
13:15:40 <qheaden> Or I need to look into how libpurple handles it, if it does.
13:16:09 * flo-retina refrains from asking alias-related questons
13:16:10 <aleth> Lets flip the pref first ;)
13:16:11 <flo-retina> *questions
13:39:21 <-- aleth has quit (Ping timeout)
13:41:11 --> aleth has joined #instantbird
13:41:11 * ChanServ sets mode +o aleth 
13:45:20 <-- sonny has quit (Ping timeout)
13:50:03 --> Hadi has joined #instantbird
13:59:46 * Fallen|away is now known as Fallen
13:59:51 <-- qheaden has quit (Quit: Instantbird 1.6a1pre -- http://www.instantbird.com)
14:01:34 --> sonny has joined #instantbird
14:02:23 <-- sonny has quit (Client exited)
14:02:29 --> qheaden has joined #instantbird
14:04:42 <qheaden> Random question. Is can Mozilla's sync framework be used by other Mozilla products other than Firefox?
14:04:52 <qheaden> s/Is can/Can
14:05:24 <qheaden> It would be cool if you could somehow sync account info and message logs within Instantbird.
14:05:33 <clokep_work> qheaden: I believe so, yes.
14:05:50 <clokep_work> (From a technical standpoint.)
14:05:59 <clokep_work> Idk if there are ToS involved or anyhting of that sort.
14:10:15 <clokep_work> qheaden: And yeah, we've talked about that for years. :-D
14:10:29 <-- aleth has quit (Ping timeout)
14:11:15 --> aleth has joined #instantbird
14:11:15 * ChanServ sets mode +o aleth 
14:11:19 <qheaden> :P
14:14:42 --> jb has joined #instantbird
14:26:25 --> mconley has joined #instantbird
14:35:51 <-- aleth has quit (Ping timeout)
14:37:10 --> aleth has joined #instantbird
14:37:10 * ChanServ sets mode +o aleth 
14:58:49 <-- aleth has quit (Ping timeout)
15:00:37 --> aleth has joined #instantbird
15:00:37 * ChanServ sets mode +o aleth 
15:07:56 <-- aleth has quit (Ping timeout)
15:28:07 --> iamjayakumars has joined #instantbird
15:40:43 --> aleth has joined #instantbird
15:40:43 * ChanServ sets mode +o aleth 
15:51:05 <-- aleth has quit (Ping timeout)
15:51:38 --> aleth has joined #instantbird
15:51:38 * ChanServ sets mode +o aleth 
15:54:33 <-- chrisccoulson has quit (Ping timeout)
15:59:22 <-- gerard-majax_ has quit (Ping timeout)
16:10:15 --> chrisccoulson has joined #instantbird
16:29:08 --> gerard-majax_ has joined #instantbird
16:34:58 <-- gerard-majax_ has quit (Ping timeout)
16:36:38 <-- aleth has quit (Ping timeout)
16:37:07 --> gerard-majax_ has joined #instantbird
16:38:38 --> Mook_as has joined #instantbird
16:39:10 --> aleth has joined #instantbird
16:39:10 * ChanServ sets mode +o aleth 
16:44:10 <-- gerard-majax_ has quit (Ping timeout)
16:52:45 <aleth> flo-retina: "Failed to load native module at path '/Applications/Instantbird.app/Contents/MacOS/components/libxpcomsample.dylib': (80004005) <unknown; can't get error from NSPR>" in current nightlies, is that similar to the bonjour issue you spotted on 64b Linux?
16:53:31 <aleth> Hmm, there's also some warnings from toolkit code that may be unmaintained now
16:53:54 <flo-retina> aleth: I also pasted that one ;)
16:54:31 <flo-retina> I don't see why that would be packaged
16:54:37 <flo-retina> afaik it's not in our package-manifest.in
16:54:58 <aleth> We shouldn't be trying to load it afaik
16:55:12 <aleth> Or is it loaded because it's packaged?
16:55:35 <flo-retina> also, according to http://buildbot.instantbird.org/builders/linux64-nightly-default/builds/4/steps/shell_5/logs/stdio we aren't packaging that file
16:55:41 <flo-retina> I've got no clue where it is coming from
16:55:51 <flo-retina> hmm
16:56:02 <flo-retina> I guess the file isn't packaged, but a manifest file that mentions it is!
16:56:21 <aleth> That could be it!
16:56:48 <aleth> But I can't find a manifest file that mentions it...
16:57:04 <flo-retina> it could be inside omni.ja
16:57:56 <aleth> something has to put it there
16:58:38 <flo-retina> it would be more interesting to figure out what's wrong about bonjour though
16:59:12 <aleth> I wonder if this is a m-c error anyway
16:59:55 <aleth> There are only two references to 'xpcomsample', in mozilla/python/mozbuild/mozbuild/frontend/sandbox_symbols.py and mozilla/xpcom/sample/moz.build
17:05:38 <-- qheaden has quit (Quit: Instantbird 1.6a1pre -- http://www.instantbird.com)
17:19:16 <-- Tonnes has quit (Quit: ChatZilla 0.9.90.1 [Firefox 31.0/20140716183446])
17:34:42 --> Mnyromyr has joined #instantbird
17:37:30 <-- jb has quit (Ping timeout)
17:46:38 <-- flo-retina has quit (Ping timeout)
17:50:13 <-- mpmc has quit (Quit: Instantbird 1.6a1pre -- http://www.instantbird.com)
17:51:51 --> aleth-mac has joined #instantbird
17:51:58 <-- aleth-mac has quit (Quit: aleth-mac)
17:53:49 <-- Rym has quit (Ping timeout)
17:54:58 <-- Mook_as has quit (Ping timeout)
17:58:16 --> Mook_as has joined #instantbird
18:11:35 --> Rym has joined #instantbird
18:15:00 <-- Rym has quit (Ping timeout)
18:17:10 --> flo-retina has joined #instantbird
18:17:10 * ChanServ sets mode +qo flo-retina flo-retina 
18:28:46 <-- iamjayakumars has quit (Client exited)
18:29:16 --> iamjayakumars has joined #instantbird
18:30:56 <-- iamjayakumars has quit (Ping timeout)
18:32:52 --> Rym has joined #instantbird
18:33:36 --> Tonnes has joined #instantbird
18:55:06 <-- clokep_work has quit (Ping timeout)
19:01:15 --> sonny has joined #instantbird
19:01:46 --> sonny1 has joined #instantbird
19:03:00 <-- sonny has quit (Ping timeout)
19:03:17 * sonny1 is now known as sonny
19:07:40 --> clokep_work has joined #instantbird
19:07:41 * ChanServ sets mode +o clokep_work 
19:52:36 <-- aleth has quit (Ping timeout)
19:54:40 --> rosonline has joined #instantbird
19:58:48 <-- clokep_work has quit (Ping timeout)
19:59:53 --> clokep_work has joined #instantbird
19:59:53 * ChanServ sets mode +o clokep_work 
20:04:36 <-- Mook_as has quit (Ping timeout)
20:06:00 --> Mook_as has joined #instantbird
20:22:10 <-- chrisccoulson has quit (Quit: OSError: [Errno 130] Owner died)
20:25:39 --> chrisccoulson has joined #instantbird
20:26:02 <-- rosonline has quit (Ping timeout)
20:26:53 --> rosonline has joined #instantbird
20:41:16 --> EionRobb has joined #instantbird
20:44:07 <-- clokep_work has quit (Ping timeout)
20:47:07 <flo-retina> "This is the full JID without the resource" ... 'full JID' means 'with the resource' ;)
20:58:46 --> aleth has joined #instantbird
20:58:46 * ChanServ sets mode +o aleth 
21:01:19 <-- aleth has quit (Ping timeout)
21:02:55 --> aleth has joined #instantbird
21:02:55 * ChanServ sets mode +o aleth 
21:03:47 <aleth> sawrubh: The UI already has ways to get the name of the buddy we're talking to, after all it's displayed in the conv-top
21:10:03 * aleth is trying to track down the last remaining source of "this._account is undefined" IRC errors
21:10:16 <aleth> but of course since I added dumps, it doesn't happen any more.
21:26:10 * flo-retina wonders if there's any progress on the bustage fixes
21:26:33 * aleth hasn't tried anything since Saturday
21:30:51 <aleth> "Replace double quotes with single quotes in Makefiles (or remove them when it makes sense) - to try and fix L10n build bustage" looks like someone else ran into that problem ;)
21:31:11 <flo-retina> did you have that problem?
21:33:43 <aleth> No, I was thinking of this https://hg.mozilla.org/comm-central/rev/6358e2178d99
21:34:16 <aleth> Different context.
21:41:05 <flo-retina> I have 3600 unread messages in #developers
21:41:16 <flo-retina> 2000+ in #tor-dev
21:41:37 <aleth> Time for nhnt11's message prepending patch ;)
21:42:05 <aleth> What happened to aororla
21:42:05 <flo-retina> I have almost 10k unread messages in convs on hold
21:42:12 <flo-retina> aleth: vacations in Europe I think
21:42:45 <flo-retina> aleth: well, I don't have that many unread messages if we have a nightly everyday ;)
21:43:02 <flo-retina> using 737MB of ram
21:45:36 <aleth> Nightlies every day would be nice ;)
21:47:08 <flo-retina> │  │  │  │       ├──0.02 MB (00.00%) ── string(length=15, copies=698, "[object Object]")/gc-heap/latin1
21:47:21 <flo-retina> I wonder why we have so many copies of that string :-S
21:47:26 <flo-retina> it doesn't seem useful at all
21:47:34 <flo-retina> probably some code doing .toString excessively
21:47:53 <flo-retina> and I wonder why this is a two-bytes string:
21:47:54 <flo-retina> │  │  │  │       ├──0.02 MB (00.00%) ── string(length=8, copies=584, "ChanServ")/gc-heap/two-byte
21:48:25 <aleth> So do we wait for bug 1043869 to be ported, or what's the next best step?
21:48:28 <instantbot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=1043869 nor, --, mozilla34, mh+mozilla, RESO FIXED, compile tier rebuilds everything that depend on libxul on windows when libxul is changed
21:48:51 <flo-retina> possibly
21:48:56 <aleth> Those are good questions
21:49:13 <flo-retina> aleth: I would like to spend some real time investigating memory usage
21:49:24 <flo-retina> I suspect there's plenty of interesting stuff to "discover" there
21:49:33 <aleth> No doubt.
21:50:01 <flo-retina> js-non-window is 204MB
21:50:13 <flo-retina> out of which, we have 30MB of strings
21:50:32 <flo-retina> why would an IM client have so much non-string data to keep around?
21:51:06 <aleth> container object overhead?
21:51:25 <flo-retina> the largest compartment (imConversations.js) weights 45MB
21:51:43 <flo-retina> it has 13MB of "compartment tables"
21:52:07 <aleth> I wonder how much all the xpconnect wrapping costs
21:52:08 <flo-retina> this makes me suspect that there are two many cross compartment references being kept around
21:53:06 <flo-retina> then there's also all the LIST data staying around for no obvious reason
21:53:08 <aleth> I'm never sure if each message in the DOM doesn't still refer back to a prpl message
21:53:22 <-- mconley has quit (Input/output error)
21:53:26 <flo-retina> why are you never sure of that?
21:53:33 <flo-retina> I'm sure they do, that's how magic copy works
21:54:06 <aleth> Yes, but is there a cross compartment wrapper around for each one or not?
21:54:27 <flo-retina> my question is more "is there _only one_" ;)
21:54:39 <aleth> heh
21:54:58 <aleth> I've been meaning to look into the LIST/requestroominfo memory issue for months :-S
21:55:05 <flo-retina> I suspect we are wrapping between the HTML document and conversation.xml, then between conversation.xml and imConversations.js, then with irc.js/xmpp.js, and then with jsProtoHelper
21:55:44 <aleth> Originally I was hoping some of that would get tidied up in the process of implementing infinite scroll, but now time seems very short :-/
21:55:55 <flo-retina> I would like to make the imConversations service keep only a JSON structure for each message, and drop the prplMessage reference
21:56:24 <aleth> How does that intersect with aororla's patch?
21:56:38 <flo-retina> aleth: if you are removing messages from the DOM and from the arrays held by the conv service, it should drop the messages from memory.
21:57:00 <flo-retina> aleth: I'm hoping arlora's patch would be a (small) first step toward that plan
21:58:33 <aleth> Hmm, I can see that should save quite a bit of memory
21:59:22 <flo-retina> "4.08 MB (00.71%) -- compartment([System Principal], resource://gre/modules/Promise.jsm)" is the 4th largest compartment
21:59:29 <flo-retina> are we leaking promises everywhere?
22:00:05 <aleth> We don't use Promise.jsm explicitly, so if we do, we or OS.File or Task.jsm or gecko code is leaking.
22:00:41 <aleth> ie. we don't use Promise.jsm promises directly, only via those APIs
22:01:14 <aleth> Our code uses DOM promises.
22:01:31 <flo-retina> really?
22:02:15 <aleth> Yes (Promise.jsm should generally be going away, hopefully)
22:02:57 <flo-retina> I wouldn't be surprised if the logging code was keeping promises alive for longer than actually needed
22:03:20 <flo-retina> with all the queues that were added, it seems that leaking a reference somewhere would be easy
22:03:23 <aleth> The only way I could see it happening is if references to *tasks* were kept around.
22:03:41 <aleth> Certainly not impossible.
22:03:41 <flo-retina> aleth: that's what I mean ;)
22:04:39 <flo-retina> aleth: I wonder if we have cross compartment wrappers for each item we put in the debug logs
22:05:18 <aleth> OK. So *how could we find out if we do* ?
22:05:44 * aleth would like to measure things.
22:06:09 <aleth> It seems a very roundabout way to stare at the code...
22:06:16 <flo-retina> yeah, the only idea I have is to do very specific benchmarks, and see how much the memory usage increases if we run the thing 1M times instead of 10 times.
22:06:37 <flo-retina> aleth: maybe adding support for the title attribute in about:memory would be helpful btw ;)
22:07:08 <aleth> The title of the compartment, or...?
22:07:23 <flo-retina> aleth: the human readable descriptions of what each line means ;)
22:07:27 <flo-retina> in tooltips
22:07:28 <aleth> Oh, you mean the tooltips
22:07:43 <aleth> Yes, let's start with that :)
22:08:01 <aleth> I'll look into it in a minute.
22:09:24 <flo-retina> at least that part seems actionable :)
22:10:47 <aleth> Yup.
22:11:40 <aleth> It would be nice if the new FX debugger could display the wrappers around a given variable when stepping through a code.
22:11:48 <aleth> I wonder if that kind of thing is on the radar.
22:12:03 <aleth> Doesn't seem very webdev-useful.
22:12:21 <-- Bollebib has quit (Client exited)
22:12:39 * Fallen is now known as Fallen|away
22:13:59 <flo-retina> aleth: the thing I would really like is a way for about:memory to dump to a file a list of all the JS objects currently in memory, with the place in the code where the object was created, and if possible how much the object weights in memory
22:14:31 <flo-retina> seeing which allocation locations grow would point directly at suspicious areas of the code
22:15:13 <flo-retina> that's more or less what valgrind provides for lower level memory allocations. I would like to have the same thing for JS objects
22:16:07 <aleth> That would be a great debug build feature
22:16:19 <flo-retina> yeah, I don't mind if it requires a debug build
22:16:37 <aleth> Probably too heavy to be around all the time
22:17:07 <flo-retina> I even don't mind if it dumps the allocation right away, and dumps object destructions during gc with the pointer, and then we need to postprocess huge logs with a script.
22:19:07 <flo-retina> "│  │   │  │  ├───0.15 MB (00.02%) ++ compartment([System Principal], resource://gre/modules/Promise.jsm)" is the size on my default Firefox profile
22:19:42 <flo-retina> are we sure we are clearing references to task queues once the last one resolves?
22:20:13 <aleth> Patch for tooltips coming up...
22:20:56 <flo-retina> already got something that works?!? :)
22:21:09 <aleth> Sure.
22:21:26 <aleth> Ah, I should file a bug ;)
22:21:44 <flo-retina> must be refreshing to write some code :)
22:22:24 <aleth> There's not much code involved ;)
22:23:13 <instantbot> New Instantbird - Other bug 1046414 filed by aleth@instantbird.org.
22:23:14 <instantbot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=1046414 nor, --, ---, aleth, ASSI, Add tooltips to the about panel
22:23:23 <aleth> But that was a really good idea, those tooltips are useful.
22:26:00 <aleth> flo-retina: The comparison to FX does make it look like there is a leak.
22:26:32 <flo-retina> these things are a lot more visible when the nightly has been running for a week or so
22:26:51 <aleth> Yes, I don't see it at all in my instance.
22:27:26 <-- Rym has quit (Ping timeout)
22:30:19 * aleth is looking at the logger queue
22:30:53 <flo-retina> wow, is that really all we needed for the tooltip to work?
22:30:53 <aleth> I'm pretty sure I tried to check for that on review though, so it's probably something more subtle :-/
22:31:05 <flo-retina> if I had know that, I would have patched that at least 6 months ago
22:31:23 <aleth> Yes (the logic is all in the imTooltip code)
22:31:56 <flo-retina> http://mxr.mozilla.org/comm-central/source/chat/components/src/logger.js#39 seems indeed to be doing some reasonable attempt at cleaning up
22:34:49 <-- Hadi has quit (Ping timeout)
22:35:28 --> Hadi has joined #instantbird
22:38:22 --> Rym has joined #instantbird
22:43:40 <-- rosonline has quit (Ping timeout)
22:45:51 --> rosonline has joined #instantbird
22:58:04 <-- Rym has quit (Ping timeout)
22:58:44 <-- Mnyromyr has quit (Quit: ChatZilla 0.9.86 [SeaMonkey 1.1.19/2010030105])
22:59:42 --> Rym has joined #instantbird
23:12:10 <-- Hadi has quit (Quit: Instantbird 1.6a1pre -- http://www.instantbird.com)
23:16:51 <-- rosonline has quit (Client exited)
23:19:11 <-- sonny has quit (Ping timeout)
23:21:45 --> rosonline has joined #instantbird
23:27:50 <-- aleth has quit (Ping timeout)
23:49:01 --> clokep has joined #instantbird
23:49:01 * ChanServ sets mode +o clokep 
23:54:29 --> sonny has joined #instantbird
23:59:44 <clokep> aleth++++