#instantbird log on 11 27 2011

All times are UTC.

00:29:07 --> billysanca has joined #instantbird
00:29:36 <-- billysanca has quit (Quit: Instantbird 1.2a1pre)
00:29:42 --> billysanca has joined #instantbird
00:46:28 <-- Tomek has quit (Quit: Instantbird 1.1)
00:55:44 * flo is testing a simple idea to speed up restoring convs on hold
01:12:42 <flo> hmm, I don't understand. I made some potentially slow code be executed once after displaying all messages instead of once per message... and it's now even slower :-/
01:14:16 --> clokep has joined #instantbird
01:14:16 * ChanServ sets mode +h clokep 
01:14:50 <-- clokep has quit (Quit: Instantbird 1.2a1pre)
01:14:58 --> clokep has joined #instantbird
01:14:58 * ChanServ sets mode +h clokep 
01:16:02 <clokep> aleth: Btw I will review that code tomorrow. :)
01:36:24 <flo> Good night :)
02:15:26 <-- aleth has quit (Input/output error)
02:20:00 <-- clokep has quit (Quit: Instantbird 1.2a1pre)
02:56:39 <-- billysanca has quit (Quit: Instantbird 1.2a1pre)
04:15:28 --> habr100 has joined #instantbird
04:47:10 <-- habr100 has quit (Quit: Instantbird 1.2a1pre)
04:47:15 --> habr100 has joined #instantbird
05:48:40 --> clokep has joined #instantbird
05:48:40 * ChanServ sets mode +h clokep 
05:52:58 <-- clokep has quit (Quit: Instantbird 1.2a1pre)
06:21:33 <-- habr100 has quit (Quit: Instantbird 1.2a1pre)
07:23:44 --> Mad_Maks has joined #instantbird
07:27:31 --> habr100 has joined #instantbird
08:11:06 <-- habr100 has quit (Ping timeout)
08:12:44 --> habr100 has joined #instantbird
09:20:27 --> jb has joined #instantbird
09:40:07 <-- habr100 has quit (Quit: Instantbird 1.2a1pre)
09:46:38 <-- jb has quit (Ping timeout)
10:08:57 --> Mic has joined #instantbird
10:08:57 * ChanServ sets mode +h Mic 
10:09:57 --> jb has joined #instantbird
10:11:48 --> jb1 has joined #instantbird
10:12:23 <-- jb has quit (Ping timeout)
10:12:55 <-- jb1 has quit (Connection reset by peer)
10:21:59 --> aleth has joined #instantbird
10:22:53 <-- aleth has quit (Quit: Instantbird 1.2a1pre)
10:23:02 --> aleth has joined #instantbird
10:24:29 <flo> when restoring a very large conversation (= takes 5s to appear) with the Simple message theme, Shark tells me that we spend 46% of the time executing JS regexp
10:33:30 <-- Mad_Maks has quit (Quit: Instantbird 1.2a1pre)
10:38:21 <aleth> Whoa. Do you know which, or what that encompasses?
10:38:49 <flo> I would love to know which regexp spends the time ;)
10:44:08 <flo> if I was doing a blind guess, I would bet on http://lxr.instantbird.org/instantbird/source/chat/modules/imThemes.jsm#468
10:45:00 <flo> but without the profiler, I wouldn't have guessed that the time is spent executing regexps, so (my) guesses are usually wrong for performance stuff ;)
10:46:07 <-- Mic has quit (Quit: Instantbird 1.2a1pre)
10:46:14 <aleth> Looks plausible though.
10:46:45 <aleth> Where did you find the profiler? Is it a mac-specific thing?
10:46:54 <flo> yes, it's a Mac tool
10:47:18 <flo> aleth: http://www.rotateright.com/index.html seems the equivalent for linux
10:48:02 <aleth> Ah, it's not JS specific then.
10:48:27 <flo> no :(
10:48:34 <flo> I don't have any JS profiler these days
10:49:04 <flo> if I had a JS profiler, I would know in which JS function the time is spent, so guessing which regexp is faulty would be trivial
10:49:25 <aleth> Yeah, that's why I asked initially.
10:50:47 <aleth> Thought you might have got some useful part of venkman working...
10:50:51 <flo> if only venkman still worked... ;)
10:51:01 <flo> heh, same idea ;)
10:57:02 <aleth> I don't understand the while loop in http://lxr.instantbird.org/instantbird/source/chat/modules/imThemes.jsm#476. Might it be optimized by pulling out the regex?
10:58:33 <flo> aleth: at each iteration the regexp matches the first %replacement% after the index replacementRegExp.lastIndex
11:00:58 <flo> if this regexp is actually where we are spending the time, we could optimize by matching the replacements in message style files only once per file and caching the result, instead of once per message as we currently do
11:01:12 <flo> the code would be more complicated though, so I would rather not do that if it's not where the time is spent
11:02:23 <flo> we have lots of other regexps that are executed for each message too. For each message we match emoticons, /me, nicks (I have the show nick add-on installed)...
11:04:28 <aleth> Yes, there are quite a lot once you start thinking about it...
11:05:23 <flo> the reason I thought about the message theme regexp is that it's the one that receives the largest input (as it doesn't get the (small) message text, but the whole HTML that's about to go into the conversation document
11:07:24 * flo tries venkman (without really expecting anything useful)
11:08:09 <aleth> If you disable that regex (which will mess things up), what's the performance increase? Or does it break things too much
11:09:12 <aleth> Probably just throws CSS errors all the time and takes much longer
11:09:26 <flo> you mean if I disable displaying messages?
11:10:21 <aleth> Just the parsing of %...% from the message style, which that regex seems to be doing
11:11:45 <flo> that would display %username% instead of nicks, %messageText% instead of the actual messages, etc...
11:11:48 <aleth> Oh that won't work
11:12:04 <aleth> Yes I just remembered about the message text also being a %..%
11:12:36 <flo> that may still "work" though (= lots of lines of HTML would still go in the conv window)
11:13:25 <aleth> Am not sure it would tell anything interesting anymore though
11:18:04 <flo> so I got some profiling data out of valgrind
11:18:34 <flo> but it spent as much time generating error while profiling as the executing the code I was trying to profile, so times may be completely off
11:19:05 <flo> and it has the very strange idea of sorting by file, then by time spent
11:20:35 <flo> s/valgrind/venkman/g
11:21:28 <flo> it thinks out of 5 seconds spent, I spent 775.33ms in replaceKeywordsInHTML
11:26:03 <aleth> about 30% of the regex time then, if the numbers are right
11:27:18 <aleth> Is it consistent? How much time does venkman think it is doing regex calls?
11:28:19 <flo> venkman won't tell anything about regexps
11:28:25 <flo> they are native code, not JS functions
11:29:29 <flo> aleth: if you want to look at Venkman's profile data: http://queze.net/goinfre/ib-profile.txt http://queze.net/goinfre/ib-profile.html
11:30:06 <flo> (I've never really been able to decide with of the txt and html format is the most usable (least unusable?) of the two)
11:31:43 <flo> my profile was apparently done on a conversation with ~ 200 messages
11:31:58 <flo> and 200 participants in the nicklist (it's also very slow to fill in that list :()
11:40:08 <aleth> Yes, nicklist filling takes a long time (joining #ubuntu is a useful demo). I doubt there are any easy wins there though, unless somehow filling the listbox "all at once" would save time.
11:40:41 <aleth> The listbox scrolling bug suggests to me the underlying code might just be slow.
11:42:14 <flo> "filling the listbox "all at once" would save time" what about the opposite?
11:42:58 <flo> handling the list of nicks in a simple JS object without the listbox, and putting in the listbox only the 30 first nicks, all at once, but after messages have appeared?
11:45:11 <flo> well, 30 would rather be as many as we can in less than 100ms
11:45:32 <flo> then continue on the next event loop spin
11:46:53 <flo> for the messages area, if we can't make it faster but can at least predict how much time it's going to take, we could do the work in small chunks to make the UI responsive, and display a progress bar
11:47:21 <flo> I would really prefer finding how to optimize it so that it can be fast though :)
11:51:03 <flo> hmm, OOo is unable to open CSV files containing decimal values, that's... interesting :-/
11:53:12 <flo> hmm, it works better if one forces the locale to en-US :)
11:53:59 <aleth> Oh, do you use the 3.440,00 convention in France?
11:54:30 <aleth> For the message area, the 'infinite scroll' idea would eventually fix that (the html would only contain a limited number of messages at any time)
11:55:22 <aleth> Interesting idea regarding the listbox. 
11:55:44 <flo> yes
11:56:04 <flo> but it's pretty obvious that if , is the field separator, it can't also be the decimal separator in the same file...
11:56:19 <aleth> But yeah, optimizing would be better.
11:56:22 <flo> taking most the values as strings, and some (2.02) as dates, is crazy too...
11:56:48 <aleth> Dates... bizarre.
11:56:56 <flo> especially as a date would certainly not be written this way in French
12:02:07 <aleth> Btw, for the Mac gestures, is creating a corresponding keyboard event in the gesture handler and dispatching it a reasonable solution there to avoid code duplication? Or is it better to get the conversation object via the DOM and call a shared 'scrolltotop' method there (which the keyboard event handler there then also calls).
12:04:11 <flo> what about adding scrollToNextSection/scrollToPreviousSection in convbrowser, and calling that for everything?
12:05:46 <aleth> I think scrollTo should be in convbrowser, but not scrollToFirstUnreadMessage, as convbrowser doesn't know where that is (that's in conversation)
12:07:57 <aleth> Of course one could move the requisite variable across to convbrowser too but that doesn't seem cleaner.
12:09:06 <aleth> But you have given me a different idea.
12:09:07 <flo> why? It's convbrowser that knows if the previous message was a context message
12:10:53 <aleth> You really think it's more efficient for convbrowser to go through some loop to find the first context message than the way the patch does it at the moment?
12:11:35 <aleth> That was the reason I put the code in conversation.xml in the first place. There's an obvious place there already where the last context message is added.
12:12:28 <aleth> But scrollToNextSection actually is a good idea, you could then press Alt-PgUp twice to get to the top of the messages
12:12:43 <aleth> That would be ideal.
12:13:52 <flo> "go through some loop to find the first context message" uh??
12:15:28 <aleth> Or update some variable each time a context message is added, whatever
12:17:25 <instantbot> aletheia2@fastmail.fm cancelled review?(clokep@gmail.com) for attachment 1032 on bug 1032.
12:17:27 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=1032 enh, --, ---, nobody, NEW, Scroll to first unread message when opening hidden conversation
12:17:35 <flo> or do it only once, the first time a non-context message is added?
12:19:37 <aleth> Either way it seems smarter to me to do it in conversation.xml, where there is _already_ a variable tracking just that. But I've come around to storing the info in convbrowser, because the 'jump to next/previous' section idea is clearly much better. 
12:20:30 <flo> aleth: the info is already in convbrowser
12:20:47 <flo> see the this._lastMessageIsContext thing
12:21:11 <flo> (it needs to become true by default, btw)
12:21:53 <aleth> OK, I'll take another look when I redo it.
12:25:35 <aleth> Btw, does the getBrowser() call in the mouse gesture handler return the tabbrowser or the message browser?
12:27:50 --> igorko has joined #instantbird
12:31:07 --> igorko1 has joined #instantbird
12:31:08 <aleth> (i.e. is it the getBrowser() in instantbird.js)
12:31:16 <-- igorko has quit (Ping timeout)
12:35:47 <flo> tabbrowser I think
12:36:00 <flo> that getBrowser is implemented in instantbird.js so you can check :)
12:50:57 <-- igorko1 has quit (Quit: Instantbird 1.1)
12:56:03 <instantbot> New Instantbird (UI) bug 1187 filed by kevin.gadd@gmail.com.
12:56:06 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=1187 cri, --, ---, nobody, UNCO, Severe pauses/hang when dealing with large conversation backlogs
13:20:45 <aleth> Looking at your venkman output, it's really noticeable how slow the listbox calls are, considering. I wonder why? e.g. 500ms spent on getItemAtIndex... why does that take so long?
13:35:26 <aleth> maybe of interest http://blog.tuenti.com/dev/tips-and-tricks-with-regular-expressions-in-js/
13:37:04 <aleth> e.g. |replacementRegExp = /%([a-zA-Z]*)(\{([^\}]*)\})?%/g;| -> |replacementRegExp = /%([A-Z]*)(\{([^\}]*)\})?%/gi;| might be faster
13:46:17 --> mmkmou has joined #instantbird
13:48:08 <-- mmkmou has quit (Ping timeout)
13:50:56 <-- aleth has quit (Quit: Instantbird 1.2a1pre)
14:00:36 --> aleth has joined #instantbird
14:04:15 --> mmkmou has joined #instantbird
14:11:57 <aleth> plus, maybe using the y flag rather than g?
14:12:19 <aleth> (can't test these myself due to lack of profiler etc)
14:13:57 <flo> aleth: the profiler doesn't really help to compare the different times either
14:14:05 <flo> as you need to have the same conversation to test with, etc...
14:16:56 <-- mmkmou has quit (Ping timeout)
14:17:23 --> mmkmou has joined #instantbird
14:17:23 <aleth> Hmm, yes, that makes it tricky, as you don't have a stable test case.
14:33:16 <-- chrisccoulson has quit (Ping timeout)
14:52:21 --> Plop has joined #instantbird
14:55:38 <-- BlueMaxima has quit (Quit: Instantbird 1.1)
15:47:56 --> Mic has joined #instantbird
15:47:57 * ChanServ sets mode +h Mic 
16:08:37 <-- Even1 has quit (Ping timeout)
16:18:42 --> mikk_s has joined #instantbird
16:18:46 <-- mikk_s has quit (Quit: mikk_s)
16:24:56 --> Tomek has joined #instantbird
16:28:24 <-- mmkmou has quit (Ping timeout)
16:44:05 --> mmkmou has joined #instantbird
16:45:15 --> devfil has joined #instantbird
16:59:40 <flo> "InstantBird's UI thread hangs for around 2-3 minutes while repopulating the
16:59:41 <flo> backlog." (bug 1187). Wow!
16:59:44 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=1187 cri, --, ---, nobody, UNCO, Severe pauses/hang when dealing with large conversation backlogs
17:08:10 <Mic> The reporter is Mozilla employee btw
18:05:53 <flo> aleth: ah, I didn't think of that detail about resizing. (I did think that keeping the position rather than the element we want to scroll into view wasn't a great idea though!)
18:06:45 <aleth> It didn't matter for the original goal of just scrolling there once on opening ;)
18:07:35 <flo> right
18:07:36 <flo> :)
18:07:58 <aleth> feature creep ;)
18:09:42 <aleth> In a good way...
18:12:25 <aleth> Actually it's probably a good sign people are using IB for IRC to the extent that they accrue backlogs that take 2-3 minutes to open :)
18:12:53 <aleth> and wasn't that reported before, for twitter searches
18:15:32 <-- mmkmou has left #instantbird ()
18:19:40 <flo> i think a solution is to advise these people to use nightlies :-P
18:19:57 <flo> daily update (-> restart), no more 2-3 days backlogs :-P
18:22:53 <aleth> perfect :)
18:23:33 <Mic> lol
18:25:22 <flo> It's what I do, by the way.
18:25:36 <flo> when we don't build nightlies for a few days, I'm more annoyed by the memory usage ;)
18:26:12 <Mic> I don't mind starting and shutting down the computer when I need or don't need it.
18:26:44 <flo> I don't see how it could be a useful action
18:27:49 <flo> I just close my laptop's screen to put it to sleep when I don't use it. When I reopen the screen it's ready to use again (to be fair, there's probably a 2-3 seconds delay here).
18:28:09 <flo> it's a bit different with desktop PCs of course
18:28:27 <aleth> Having things on standby draws quite a bit of power in aggregate
18:29:03 <flo> it can stay on the battery during that time
18:29:08 <aleth> Laptops are more optimized for that kind of thing I guess
18:29:43 <Mic> If you need it that quickly then OK. I tend to know in advance when I'll need it and just press the power button first thing when I come to the room.
18:29:51 <aleth> But generally speaking, devices that are not in use should not be consuming power, just on general principle
18:30:00 <aleth> Everything else is just bad engineering
18:30:27 <Mic> flo: power drawn from the battery doesn't count as power consumption? ;)
18:30:59 <flo> Mic: I just meant it's a minimal amount
18:31:08 <Mic> For me it's just natural to start and shutdown as I need it. Like pressing the light switch when I leave the room
18:31:18 <aleth> Same here
18:31:30 <flo> Mic: it's natural for me too
18:31:31 <Mic> OK, that one won't be that necessary now that we have the power saving lamps
18:31:45 <flo> but it's also natural than when I "tell" it I'm ready to use it, it should be ready to be used.
18:32:01 <aleth> flo: yes, that's what I meant by it being an engineering problem
18:32:03 <flo> The light switch doesn't have a progress bar telling me it will be ready to light the room in a minute
18:32:31 <flo> closing/opening the screen feels a lot more like flipping the light switch, by the way
18:32:51 <aleth> that's not the part that's bad design ;)
18:33:40 <flo> yeah
18:33:53 <-- Mic has quit (Ping timeout)
18:34:24 <flo> the poor design is that a PC (even a relatively fast one with a decent operating system) takes more time to turn itself on and be ready to use than I need time to either forget what I wanted to do, or grab my laptop and starting doing something else.
18:36:18 <flo> I think it would make sense to have something like some flash memory of the same size as the RAM that could receive the content of the RAM when the machine shutdown (quickly!), and then restarting would just turn the screen on. Any RAM access would lazily fetch the relevant data in the flash memory
19:19:38 <aleth> It's actually surprising that hasn't been done, considering how relatively cheap flash memory is
19:20:09 --> chrisccoulson has joined #instantbird
19:47:03 <-- micahg_ has quit (Connection refused)
19:47:13 --> micahg_ has joined #instantbird
19:47:58 --> EionRobb has joined #instantbird
20:07:53 --> billysanca has joined #instantbird
20:23:27 --> franksouza183 has joined #instantbird
20:23:41 <franksouza183> aleth
20:24:15 <franksouza183> someone online?
20:24:50 <aleth> yes
20:24:55 <aleth> hi
20:24:55 <franksouza183> ow
20:24:59 <franksouza183> hi
20:25:41 <franksouza183> Well I'm finishing that theme for adium instantbird, but I can not register on the site of addons
20:26:01 <franksouza183> the site does not send me the confirmation email
20:26:28 <aleth> That shouldn't happen :(
20:26:29 <franksouza183> I tried several times already waited hours, but so far nothing
20:26:55 <aleth> need to ping Even or flo who know more about that site's internals
20:27:20 <aleth> It worked for me...
20:28:57 <franksouza183> well, I'am getting porting that theme successfully to instantbird, and eventually contribute to a own theme for instantbird
20:29:07 <aleth> sounds great! :)
20:30:00 <franksouza183> I will wait to resolve the site issue
20:30:06 <aleth> You may also like to try it with the current nightly
20:30:25 <aleth> Stay in this room for a bit and someone will show up to help
20:30:33 <franksouza183> ok
20:30:37 <franksouza183> thanx
20:31:17 <franksouza183> otherwise
20:31:29 <franksouza183> I'm not using that script
20:31:40 <franksouza183> it fails always here
20:31:58 <aleth> What script?
20:32:01 <franksouza183> estou fazendo e testando manualmente
20:32:18 <franksouza183> one that converts adium themes
20:32:20 <aleth> Doing it manually is easier anyway I think...
20:32:41 <aleth> (Guessing from the message style I edited the conversion does not polish things for you)
20:33:04 <franksouza183> exactly
20:33:32 <franksouza183> a more positive point to doing manually
20:33:40 <aleth> One thing it doesn't add for example is support for nick colors in MUCs
20:34:23 <franksouza183> have any theme available that use this feature?
20:34:57 <franksouza183> I will study it
20:35:01 <aleth> The standard ones? And minimal2 for example?
20:35:10 <franksouza183> hum
20:35:13 <franksouza183> ok
20:35:33 <aleth> For the bubbles theme, the bubble color corresponds to the nick color
20:35:49 <franksouza183> huu
20:35:51 <aleth> It's more obvious with the new nightly nicklist
20:35:52 <franksouza183> o.O
20:36:15 <aleth> For the other default themes, they also do something with the color info
20:36:42 <franksouza183> ok, I'll check this
20:37:03 <aleth> It's just a matter of sticking %sendercolor% in the right place as far as I remember
20:37:25 <aleth> Though the Bubbles theme does something more elaborate with javascript
20:38:24 <franksouza183> what color format?
20:38:27 <franksouza183> hex ??
20:38:29 <franksouza183> #000000
20:38:41 <franksouza183> that's it ?
20:39:00 <aleth> I think it's passed as a CSS style
20:39:09 <aleth> color: rgb(...) 
20:39:10 <aleth> something like that
20:39:16 <franksouza183> huj
20:39:24 <franksouza183> good
20:39:53 <franksouza183> otherwise
20:40:15 <franksouza183> This version of gecko is used by which version of firefox?
20:40:47 <aleth> Have a look in the about box
20:40:57 <franksouza183> yeahh
20:41:00 <aleth> Gecko 9 = Firefox 9
20:41:08 <franksouza183> ow
20:41:10 <franksouza183> ok
20:41:16 <aleth> I think 1.1 is Gecko 7
20:41:27 <franksouza183> yes
20:41:35 <aleth> 1.2 is Gecko 9 atm
20:42:07 <franksouza183> so it is easier to test on-firefox
20:42:24 <franksouza183> nightly ?
20:42:25 <franksouza183> 9 ?
20:42:28 <-- billysanca has quit (Quit: Instantbird 1.2a1pre)
20:42:29 <aleth> If it helps you...
20:42:35 --> billysanca has joined #instantbird
20:42:42 <aleth> Firefox beta is 9 at the moment I think
20:42:57 <franksouza183> no, sorry
20:43:05 <franksouza183> IB nightly is 9 ??
20:43:11 <aleth> Yes
20:43:14 <franksouza183> ok
20:43:57 <aleth> Are you using Firefox to test message styles?
20:44:09 <aleth> I always did it directly in IB
20:44:33 <franksouza183> aways on firefox
20:44:46 <franksouza183> 7
20:44:52 <franksouza183> ops
20:44:54 <franksouza183> sorry
20:44:56 <franksouza183> 8
20:45:05 <aleth> Shouldn't matter really
20:45:15 <aleth> Unless you are using some really new features ;)
20:45:26 <franksouza183> exactly
20:45:53 <franksouza183> as svg masking and clipping
20:46:19 <aleth> Did that change from 7 to 8?
20:46:38 <franksouza183> no no 
20:46:53 <franksouza183> it's since 3.5 ++
20:47:30 <franksouza183> sorry
20:47:34 <franksouza183> "Unless you are using some really new features "
20:47:46 <franksouza183> I understand now
20:47:52 <franksouza183> :)
20:48:46 <franksouza183> I see other new features introduced since version 4
20:49:16 <aleth> Yes you can use all new the HTML5 stuff for example
20:49:27 <franksouza183> I will try to introduce on theme
20:49:30 <aleth> The bubbles message style does some clever things by the way, like the collapsing system messages
20:49:39 <franksouza183> Yes, I thought that too
20:50:30 <aleth> If you use something really cutting edge, you just have to mark the compatibility of the add-on accordingly
20:50:53 <franksouza183> yes
20:51:05 <franksouza183> I see that
21:31:39 <-- aleth has quit (Quit: Instantbird 1.2a1pre)
21:32:51 --> aleth has joined #instantbird
21:33:09 <-- Tobin has quit (Quit: Everything you do is to hurt me, leave me alone.)
21:33:25 --> Tobin has joined #instantbird
21:34:24 <-- aleth has quit (Quit: Instantbird 1.2a1pre)
21:35:24 --> aleth has joined #instantbird
21:38:07 --> Even1 has joined #instantbird
21:57:28 <flo> franksouza183: have you checked if the email is in your spam folder?
21:57:40 <franksouza183> ops
21:57:44 <franksouza183> one minute
21:58:03 <franksouza183> yes
21:58:05 <franksouza183> its here
21:58:09 <franksouza183> thenx
21:58:11 <franksouza183> thanx
21:59:47 <-- billysanca has quit (Quit: Instantbird 1.2a1pre)
21:59:57 --> billysanca has joined #instantbird
22:01:47 <-- Plop has quit (Ping timeout)
22:02:07 --> Plop has joined #instantbird
22:02:25 --> clokep has joined #instantbird
22:02:25 * ChanServ sets mode +h clokep 
22:16:35 <-- aleth has quit (Quit: Instantbird 1.2a1pre)
22:16:37 --> aleth has joined #instantbird
22:18:12 <clokep> You midaired my flo. :P
22:18:18 <clokep> *me rather.
22:18:40 * flo wondered who was "your flo" that I could be mid airing with
22:19:17 <-- aleth has quit (Quit: Instantbird 1.2a1pre)
22:19:23 --> aleth has joined #instantbird
22:20:53 <flo> clokep: I think this will also depend on bug 692
22:20:56 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=692 enh, --, ---, nobody, NEW, redesign the purpleIMessage xpcom interface
22:21:09 <clokep> flo: Yes, that would probably help!
22:21:45 <flo> if we separate the prplIMessage interface that protocol plugins need to implement from an imIMessage interface that the UI uses, we can get smarter and cache some things
22:22:43 --> hunsly has joined #instantbird
22:23:23 <franksouza183> aleth: I would like to test the theme?
22:23:49 <flo> the difficulty with caches though, is that we need to be able to invalidate them when some preferences change
22:23:55 <aleth> franksouza183: Can't you just install it from the add-ons manager?
22:24:13 <aleth> Click the little toolbox and "install from file"
22:24:37 <aleth> (Experimental add-ons can't be found via the search box)
22:24:48 <franksouza183> no, I'm asking if you want to test the theme
22:24:55 <aleth> Oh ok :)
22:24:59 <aleth> link please?
22:25:07 <franksouza183> I'll send to you
22:25:09 <franksouza183> 1 min
22:26:50 <franksouza183> sent
22:27:02 <aleth> OK, I thought you had it on the add-on site
22:28:14 <franksouza183> not yet, waiting to release a more stable
22:28:25 <flo> has someone ever tried to reproduce bug 1140?
22:28:32 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=1140 maj, --, ---, nobody, UNCO, Ctrl+F search in a conversation window with a long backlog hangs Instantbird
22:28:52 <aleth> franksouza183: It would be nice if it was marked compatible with 1.2a1pre
22:29:00 <aleth> (I am doing that by hand)
22:29:23 <franksouza183> ok
22:29:39 <aleth> flo: bug 1187?
22:29:44 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=1187 cri, --, ---, nobody, NEW, Severe pauses/hang when dealing with large conversation backlogs
22:29:51 <aleth> (comment 1)
22:30:01 <aleth> sorry, description
22:30:10 <-- aleth has quit (Quit: Instantbird 1.2a1pre)
22:30:15 --> aleth has joined #instantbird
22:30:16 <flo> aleth: both bugs have been filed by the same person.
22:30:21 <aleth> oops, my bad
22:31:05 --> Mic|web has joined #instantbird
22:31:13 <Mic|web> Good evening
22:31:31 <franksouza183> aleth: <em:minVersion>0.2a1pre</em:minVersion>
22:31:33 <franksouza183>     <em:maxVersion>1.2a1pre</em:maxVersion>
22:31:37 <franksouza183> thats it ?
22:32:03 <franksouza183> sorry
22:32:05 <aleth> You probably want minVersion 1.0, better 1.1 if you don't actually test it on 1.0
22:32:07 <franksouza183> is this?
22:32:12 <franksouza183> ok
22:32:43 <franksouza183> <em:minVersion>1.1</em:minVersion>
22:32:45 <franksouza183>     <em:maxVersion>1.2a1pre</em:maxVersion>
22:32:46 <franksouza183> ok?
22:33:10 <aleth> It should probably be 1.1a1pre
22:33:40 <franksouza183> <em:minVersion>1.1a1pre</em:minVersion>
22:33:55 <aleth> OK
22:34:00 <franksouza183> ok
22:36:08 <Mic|web> flo: is there a sure way to know when a conversation is fully restored and reactive?
22:36:29 <Mic|web> Automatically, I mean?
22:42:41 <Mic|web> hmm, gtg. Good night :)
22:42:48 <-- Mic|web has quit (Quit: http://www.mibbit.com ajax IRC Client)
22:50:37 <-- aleth has quit (Quit: Instantbird 1.2a1pre)
22:50:42 --> aleth has joined #instantbird
23:05:48 <franksouza183> aleth: thanks
23:11:48 <aleth> I just looked at the stats for the minimal2 message style. It seems there are still daily users of IB on version 0.2 (around 1%), and a slightly larger though declining number on 1.0.
23:13:06 <franksouza183> on version 0.2, what gecko version 
23:13:10 <franksouza183> ?
23:13:53 <aleth> franksouza183: Nobody should still be using Instantbird 0.2, that's why it's surprising
23:14:06 <franksouza183> hum
23:14:42 <-- devfil has quit (Client exited)
23:15:28 <franksouza183> then it is good idea make it impossible to 0.2
23:15:35 <franksouza183> upgrade forced
23:15:40 <franksouza183> :)
23:16:00 <aleth> It's not trivial to make that happen, but I think it is being worked on
23:53:08 <flo> "23:36:08 - Mic|web: flo: is there a sure way to know when a conversation is fully restored and reactive?" I don't know what you mean. A sure way from what?
23:53:13 <flo> from the code in conversation.xml, it's trivial
23:53:48 <flo> aleth: I found today that there was still 0.2b2 on my mother's laptop
23:54:01 <flo> built in March 2010...
23:54:16 <flo> and Firefox 3.6.4, from April 1st (2010)!
23:54:26 <aleth> Mic: Did you mean "is there an event that fires after all the existing messages have been displayed"?
23:54:40 <aleth> flo: vintage!
23:54:58 <flo> a few months ago I found 0.1.2 on my father's computer ;)
23:55:27 <aleth> I think nothing on my parents computer updates unless it is automatic ;)
23:55:55 <flo> aleth: isn't _textSelected a boolean?
23:56:19 <aleth> where was that?
23:56:19 <flo> she didn't want to update Firefox
23:56:29 <flo> aleth: https://bugzilla.instantbird.org/attachment.cgi?id=1026
23:57:33 <aleth> flo: If I remember correctly, it's a boolean *after* it's been initialized. null means it has not been initialized yet. The original code checks for existence of the property instead.
23:57:37 <flo> because the download button said "Windows • 8.0.1 • Français" and she said "it's not for me, I don't have Windows 8" and closed the window
23:58:11 <aleth> flo: oh dear :( great story though!
23:58:23 <flo> |false != null| -> false
23:58:36 <aleth> Just goes to show how details matter on these messages
23:58:36 <flo> aleth: yeah, I think I'll tell about it to some Mozilla people ;)
23:59:44 <flo> oop
23:59:49 <aleth> flo: |false != null| is true in the error console at least
23:59:52 <flo> "|false != null| -> false" forget that line please :)
23:59:54 <aleth> I did check..