#instantbird log on 09 03 2013

All times are UTC.

00:04:42 <-- dionisos has quit (Ping timeout)
00:57:00 --> clokep has joined #instantbird
00:57:01 * ChanServ sets mode +o clokep 
01:06:57 <clokep> qheaden: Does the code in socket.jsm really not do what you want it to do?
01:07:05 <clokep> I have a feeling you ARE reinventing stuff that's already been done.
01:31:52 <-- qheaden has quit (Quit: ZNC - http://znc.in)
01:38:26 --> dew1 has joined #instantbird
01:39:05 <-- dew has quit (Ping timeout)
01:56:22 <-- wnayes has quit (Ping timeout)
01:58:57 --> wnayes has joined #instantbird
02:27:16 <-- clokep has quit (Quit: Instantbird 1.5a1pre -- http://www.instantbird.com)
03:56:26 <-- wnayes has quit (Quit: wnayes)
03:57:03 --> FireFly_TB has joined #instantbird
03:57:39 <instant-buildbot> build #975 of macosx-nightly-default is complete: Success [build successful]  Build details are at http://buildbot.instantbird.org/builders/macosx-nightly-default/builds/975
04:17:44 <instant-buildbot> build #959 of linux-nightly-default is complete: Success [build successful]  Build details are at http://buildbot.instantbird.org/builders/linux-nightly-default/builds/959
04:48:29 --> qheaden has joined #instantbird
05:14:48 <instant-buildbot> build #1058 of win32-nightly-default is complete: Failure [failed compile]  Build details are at http://buildbot.instantbird.org/builders/win32-nightly-default/builds/1058
05:37:39 <-- FireFly_TB has quit (Ping timeout)
05:38:05 <-- EionRobb has quit (Quit: Leaving.)
06:44:54 <-- jamesw has quit (Ping timeout)
07:04:38 --> Mic has joined #instantbird
07:04:38 * ChanServ sets mode +h Mic 
07:07:18 <-- Mic has quit (Quit: Instantbird -- http://www.instantbird.com)
07:11:09 --> chrisccoulson has joined #instantbird
07:17:02 --> jb has joined #instantbird
07:18:01 <-- jb has quit (Quit: jb)
07:19:16 --> jb has joined #instantbird
07:22:06 <-- chrisccoulson has quit (Quit: OSError: [Errno 130] Owner died)
08:07:13 --> dionisos has joined #instantbird
08:08:01 --> EionRobb has joined #instantbird
08:30:20 --> chrisccoulson has joined #instantbird
09:17:48 --> Mic has joined #instantbird
09:17:48 * ChanServ sets mode +h Mic 
09:18:53 <-- Gizmokid2005 has quit (Ping timeout)
09:22:37 <Mic> "Nick iLobster isn't registered." :(
09:22:46 <Mic> Has anyone seen him recently?
09:23:31 <-- Even has quit (Ping timeout)
09:24:49 <Mic> He had problems with GMX and their XMPP server, I want to suggest to try js-xmpp as this is possible now.
09:26:12 --> Gizmokid2005 has joined #instantbird
09:33:05 <Mic> Nevermind, I found his email on BIO.
09:43:36 --> gerard-majax has joined #instantbird
09:46:45 --> aleth has joined #instantbird
09:46:46 * ChanServ sets mode +h aleth 
09:49:33 <-- flo-retina has quit (Quit: Instantbird 1.5a1pre -- http://www.instantbird.com)
09:50:45 <-- florian has quit (Ping timeout)
10:04:41 --> clokep has joined #instantbird
10:04:41 * ChanServ sets mode +o clokep 
10:15:23 <-- chrisccoulson has quit (Client exited)
10:17:30 --> florian has joined #instantbird
10:17:37 --> chrisccoulson has joined #instantbird
10:25:15 --> flo-retina has joined #instantbird
10:25:15 * ChanServ sets mode +qo flo-retina flo-retina 
10:33:47 <instantbot> aleth@instantbird.org granted review for attachment 2807 on bug 2135.
10:33:54 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=2135 nor, --, ---, qheaden, ASSI, Enable JS Yahoo By Default
10:34:01 <aleth> Hi Mic :)
10:34:04 <aleth> wb :)
10:35:05 <clokep> aleth: Don't mark checkin-needed on that.
10:35:27 <aleth> clokep: Why?
10:35:38 <aleth> Don't you think it's ready?
10:35:38 <flo-retina> because we aren't ready ;)
10:36:22 <flo-retina> aleth: I think we wanted to try JS-Yahoo as a #chatdev replacement for a week before switching it on by default for all nightly users
10:36:38 <clokep> aleth: No.
10:36:42 <clokep> Well, maybe.
10:36:45 <aleth> flo-retina: We could do both at the same time
10:36:50 <clokep> But qheaden still hasn't come up with the list I asked him to.
10:37:07 * flo-retina would like to review all the strings before we preff'on
10:37:36 <aleth> That makes sense, so that translators don't jump in/
10:44:31 <-- EionRobb has quit (Quit: Leaving.)
10:51:24 <-- clokep has quit (Quit: Instantbird 1.5a1pre -- http://www.instantbird.com)
10:51:45 * flo-retina will go to the Paris office tomorrow, and expects to have 2 train rides to do reviews ;).
10:55:04 <-- jb has quit (Ping timeout)
11:26:21 --> jb has joined #instantbird
11:36:25 --> clokep_ has joined #instantbird
11:56:57 <clokep_> flo-retina: Did you see jcranmer's response on the OTR bug? https://bugzilla.mozilla.org/show_bug.cgi?id=779052 also our conversation last night: http://logbot.glob.com.au/?c=mozilla%23maildev&s=2+Sep+2013&e=3+Sep+2013#c95469
12:05:26 <-- jb has quit (Quit: jb)
12:06:42 --> jb has joined #instantbird
12:14:46 <-- jb has quit (Quit: jb)
12:15:53 --> jb has joined #instantbird
12:26:19 <flo-retina> clokep_: I saw the bugmail and wondered if something happened in #maildev :-D
12:29:48 <clokep_> flo-retina: Sounds like overall jcranmer thinks re-doing it from scratch would be best or shimming libotr.
12:30:14 <flo-retina> that's what I just finished reading
12:30:27 <flo-retina> + that other guy saying non sense about "not using IM"
12:31:55 <flo-retina> clokep_: btw, why do Americans always say "Social Security number" where French people would say "credit card number" in security discussions? Is there anything particularly interesting about knowing someone's US social security number?
12:32:32 <clokep_> flo-retina: A Social Security Number is pretty much a persons UUID to the US Government.
12:32:43 <clokep_> (Everyone is issued one at birth and it is yours for life, etc.)
12:32:46 <flo-retina> here too. But then what?
12:32:54 <clokep_> They do re-use them when people die though...
12:32:59 <clokep_> I'm not sure what you mean.
12:33:05 <clokep_> "then what"?
12:33:20 <flo-retina> how is it important to not share a number that uniquely identifies you?
12:34:29 <clokep_> Identity theft, if you have someone's social security number it's a lot easier to pretend to be them.
12:34:35 <flo-retina> a license plate on a car's photo also uniquely identifies, you, doesn't it? But you don't hide your license plate when you park your car on the street.
12:35:44 * flo-retina doesn't know how long it would take him to find his US social security number again.
12:36:02 <clokep_> flo-retina: http://en.wikipedia.org/wiki/Social_Security_number#Identity_theft
12:36:21 <clokep_> Mostly the sentence "The SSN is frequently used by those involved in identity theft, since it is interconnected with so many other forms of identification, and because people asking for it treat it as an authenticator.", I believe.
12:37:08 <flo-retina> "treat it as an authenticator." ah!
12:38:00 <clokep_> It's like giving someone your password. ;)
12:38:16 <flo-retina> "it is assumed that no one except the person to whom it was issued will know it." I expect every employer, every insurance company, every bank I had an account in, every school I attended to know mine (the French one I mean), so I would really not see it as an authenticator.
12:38:44 <flo-retina> a password you can't change? What a strange idea :-S.
12:38:55 <Mic> bug 53
12:38:59 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=53 nor, --, ---, nobody, NEW, Prompt to add buddy to the buddy list after an authorization request
12:39:22 <aleth> So it seems sometimes yahoo does require the full email address of a contact, not just the username
12:39:50 <aleth> "ymail is different" maybe ;)
12:41:45 <Mic> I'd expect that to be a yahoo.com, !yahoo.com difference?
12:41:46 <instantbot> New Core - Yahoo! Messenger bug 2149 filed by aleth@instantbird.org.
12:41:49 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=2149 nor, --, ---, nobody, NEW, Attempting to store duplicate buddy error after failed auth request
12:42:08 <aleth> Mic: You'd think so, but yahoo.co.uk does not cause any problems
12:42:15 <Mic> I can log into their webmail with only the username (local part) of the email adress for my @yahoo.com adress but not with the ymail.com
12:42:17 <Mic>  one.
12:47:03 <aleth> Looks like all my attempts to invite Mic to a yahoo conference fail :(
12:47:37 <clokep_> aleth: I fully expect ymail.com emails to need the ymail.com part...
12:47:44 <clokep_> It's not a yahoo.* domain.
12:48:10 <aleth> clokep_: Inviting Mic with the full ymail address fails as well.
12:48:26 <clokep_> aleth: That's not expected. ;)
12:52:15 <flo-retina> aleth: at least you now know why we aren't preff'ing on right now :)
12:53:29 <aleth> Maybe Mic can invite me ;)
13:03:18 <-- GeekShadow has quit (Ping timeout)
13:03:21 --> GeekShadow has joined #instantbird
13:28:23 <Mic> aleth: I don't have your Yahoo account name.
13:28:28 <instantbot> clokep@gmail.com denied review for attachment 2814 on bug 2102.
13:28:31 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=2102 nor, --, ---, qheaden, ASSI, Support Server Pings
13:28:42 * clokep_ feels slightly bad since he only read half that patch. :-/
13:33:49 <-- jb has quit (Ping timeout)
13:35:58 <instantbot> New Core - Yahoo! Messenger bug 2150 filed by aleth@instantbird.org.
13:36:01 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=2150 maj, --, ---, nobody, NEW, Conference invites don't work
13:39:23 <Mic> Opening several Yahoo! chatrooms increments a number at the end of their name.
13:39:39 <Mic> It's incrementing in steps of two ...?
13:39:58 <aleth> :D
13:42:00 <Mic> Doesn't make a difference - as there's still denumerably infinite many numbers? Too bad we're dealing with computers here ;)
13:50:31 --> jb has joined #instantbird
14:09:26 <clokep_> We have failing Windows nightlies still? I thought we fixed that?
14:09:42 <flo-retina> clokep_: we need a new windows vm.
14:09:54 --> wnayes has joined #instantbird
14:10:04 <flo-retina> moz23 won't compile on a 32bit Windows
14:10:39 <clokep_> Oh. :(
14:10:46 * clokep_ wonders if he missed this news somewhere.
14:12:23 <-- Mic has quit (Quit: Instantbird -- http://www.instantbird.com)
14:12:36 <flo-retina> maybe it saved you from being disappointed during the week-end? ;)
14:13:58 --> mconley has joined #instantbird
14:29:46 --> nhnt11 has joined #instantbird
14:34:48 <-- nhnt11 has quit (Quit: Instantbird 1.5a1pre -- http://www.instantbird.com)
14:34:58 --> nhnt11 has joined #instantbird
14:48:41 <clokep_> Maybe. :)
14:48:45 <clokep_> I'm disappointed now though. :P
14:51:39 <-- nhnt11 has quit (Ping timeout)
14:51:58 <-- aleth has quit (Quit: Ciao)
14:52:16 --> aleth has joined #instantbird
14:52:16 * ChanServ sets mode +h aleth 
15:48:36 --> nhnt11 has joined #instantbird
15:52:52 <-- dionisos has quit (Ping timeout)
15:53:56 <-- jb has quit (Quit: jb)
15:54:04 <-- nhnt11 has quit (Quit: Instantbird 1.5a1pre -- http://www.instantbird.com)
15:54:13 --> nhnt11 has joined #instantbird
15:57:01 --> jb has joined #instantbird
16:22:08 <-- gerard-majax has quit (Ping timeout)
16:26:07 <qheaden> Oops. Said hello in the wrong channel. :)
16:27:36 * qheaden is very sad to hear about bug 2150
16:27:39 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=2150 maj, --, ---, nobody, NEW, Conference invites don't work
16:29:59 <qheaden> aleth: When using the /invite command, you guys did use the Yahoo ID, and not the display name right?
16:30:13 <aleth> qheaden: Sure
16:30:29 <aleth> qheaden: Tried making each other buddies first too...
16:30:37 <qheaden> How strange. Conference invites were working for me. :-S
16:30:51 <aleth> Try inviting me (alethtest) ?
16:32:09 <qheaden> aleth: You get an invite?
16:32:13 <aleth> qheaden: No
16:32:20 <qheaden> Let me run wireshark
16:32:47 * aleth thinks qheaden should use debug logs ;)
16:33:03 <aleth> Can't get feedback from users if they need to run wireshark ;)
16:33:38 <qheaden> I see that the packet is being sent.
16:33:47 <aleth> OK, try making me a buddy first.
16:33:51 <qheaden> Okay.
16:34:28 <qheaden> aleth: You get the buddy invite?
16:34:56 <aleth> qheaden: Yes
16:36:27 <aleth> qheaden: Do you do something with the auto-accept-invite pref?
16:37:08 <qheaden> aleth: Is your pref set to false?
16:37:23 <aleth> qheaden: No
16:37:52 <aleth> Your debug logs log no incoming data at all :-S
16:38:10 <aleth> There's only the onTransportStatus(STATUS_RECEIVING_FROM) from the socket
16:38:42 <qheaden> aleth: Looking here http://lxr.instantbird.org/instantbird/source/chat/protocols/yahoo/yahoo.js#252, if you have the auto accept invite pref disabled, no conference windows are opened.
16:39:34 <aleth> Mine is set to 1
16:40:12 <qheaden> aleth: Is ignore invites turned off in the account settings for your Yahoo account?
16:40:31 <aleth> No
16:40:40 <qheaden> Uncheck that
16:40:42 <qheaden> If it is checked
16:40:46 <aleth> I mean, it's not checked.
16:40:55 <qheaden> oh
16:41:17 <qheaden> Let me try inviting my other account from a second instance
16:46:33 <qheaden> aleth: Okay, so something is broken with my join conference code. I was able to successfully invite Pidgin, but not Instantbird. :(
16:46:41 <aleth> :(
16:46:47 <qheaden> Not sure what is broken, but I will try to get a fix out today or tomorrow.
16:46:49 <aleth> Well, at least the bug is discovered ;)
16:47:02 <qheaden> aleth: You and your bug finding techniques! :P
16:47:12 <aleth> With the help of Mic :D
16:47:16 --> Mook_as has joined #instantbird
16:47:21 <qheaden> :)
16:47:30 <aleth> We were hoping to move #chatdev to yahoo... but it was a bit lonely.
16:47:35 * nhnt11 realizes there's a comment on bug 2066 he needs to address
16:47:40 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=2066 enh, --, ---, nhnt11, NEW, New conversation tab should suggest chat rooms
16:47:58 <aleth> nhnt11: I think that bug is waiting mainly for working Windows nightlies.
16:48:07 <qheaden> aleth: I'm on my between-class break at school right now. When I get home, I'll look into this further.
16:48:19 <nhnt11> aleth: You said to change it to 100ms, and also I found a bug :)
16:50:53 --> Mic has joined #instantbird
16:50:53 * ChanServ sets mode +h Mic 
16:52:30 <-- Mic has quit (Quit: Instantbird -- http://www.instantbird.com)
16:52:44 --> Mic has joined #instantbird
16:52:44 <instantbot> nhnt11@gmail.com cancelled review?(florian@instantbird .org) for attachment 2811 on bug 2066.
16:52:45 * ChanServ sets mode +h Mic 
16:52:46 <instantbot> nhnt11@gmail.com requested review from florian@instantbird .org for attachment 2815 on bug 2066.
16:52:49 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=2066 enh, --, ---, nhnt11, NEW, New conversation tab should suggest chat rooms
16:52:54 * qheaden wonders if the the bug is being caused by http://lxr.instantbird.org/instantbird/source/chat/protocols/yahoo/yahoo-session.jsm#844
16:52:59 <-- florian has quit (Quit: Instantbird 1.5a1pre -- http://www.instantbird.com)
16:53:07 <Mic> :(
16:53:22 <nhnt11> aleth: I know you already r+'d that, but I r?'d you again anyway for the small changes I made
16:53:31 <aleth> qheaden: If you add logging for incoming packets you might find out ;)
16:53:39 <Mic> My Yahoo chatroom <name>-3 was restored after restarting. As <name>-1 :(
16:53:43 * aleth realizes he is nagging
16:54:12 <nhnt11> One is the 100ms change, the other is clearing the chat attribute in newtab.xml (line 415 in the diff view)
16:55:25 <aleth> nhnt11: What does "rebuilding to a non-MUC" mean?
16:55:51 <-- mconley has quit (Connection reset by peer)
16:56:03 --> mconley has joined #instantbird
16:56:20 <nhnt11> aleth: Well if isChat was true, the chat attribute was set to true - so the MUC icon was used for the buddy icon. Since the items are reused, when rebuilding to a possible conv with isChat=false, the attribute was never cleared so it kept using the MUC icon
16:56:33 <aleth> Oh, I see
16:56:35 <Mic> aleth: logging the binary data as hex string or what are you thinking about?
16:56:53 <aleth> Mic: Something more readable, hopefully
16:57:03 * qheaden sees he needs to work on conferences and logging this week
16:57:04 <aleth> But currently there is no logging at all for incoming data afaik
17:00:31 --> qlum has joined #instantbird
17:01:45 <qheaden> aleth: Could you invite me to a conference
17:01:50 <flo-retina> aleth: you can add an r- to the "enabled by default" patch saying that debug logging should be useful. And I meant it: we will get no benefit from testing with a wider set of people if they can't usefully report bugs.
17:01:53 <qheaden> I made a quick code change.
17:02:12 <aleth> flo-retina: I totally agree. I didn't realize until today there was no logging at all for incoming data
17:02:41 <qheaden> flo-retina, aleth: How should it be displayed? Hex data, or as packet information with key/value pairs?
17:02:44 <qheaden> I vote for the latter.
17:02:54 --> Optimizer has joined #instantbird
17:03:13 <Optimizer> IB nightly no longer opens a closed conversation if someone pinged me :(
17:03:52 <instantbot> aleth@instantbird.org cancelled review+ for attachment 2807 on bug 2135.
17:03:56 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=2135 nor, --, ---, qheaden, ASSI, Enable JS Yahoo By Default
17:05:30 <Mic> Optimizer: what is "closed" here?
17:06:02 <aleth> Optimizer: Try updating your nightly to get http://hg.instantbird.org/instantbird/rev/07cc8479483a
17:06:06 <Optimizer> like in this window, I press Ctrl W
17:06:25 <Mic> So they're not completely closed but only put on hold in the contact list?
17:06:30 <Optimizer> yes
17:06:38 <Mic> Sounds like bug 2138 should have fixed that.
17:06:40 <nhnt11> aleth, flo-retina, Mic: If, instead of storing the total message count for conversation stats, I used a getter that added the incoming and outgoing counts, would the small performance loss be justified by the decrease in memory usage?
17:06:42 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=2138 nor, --, 1.5, florian, RESO FIXED, A conversation on hold should open itself in a tab when the user is pinged
17:06:52 <Optimizer> so was it a regression ?
17:07:00 <Optimizer> because this used to work couple of days ago
17:07:26 <-- Optimizer has quit (Quit: Time to optimize myself)
17:07:55 <qheaden> Okay. Time for class. I'll be on later today.
17:08:38 <flo-retina> Optimizer: yes, it was a regression, and we fixed it (or think so at least).
17:09:10 <-- nhnt11 has quit (Quit: Instantbird 1.5a1pre -- http://www.instantbird.com)
17:09:15 --> nhnt11 has joined #instantbird
17:10:08 <aleth> nhnt11: I have no idea. Maybe leave it for after you get some feedback on a commented patch, if you are not sure?
17:10:21 <nhnt11> Ok.
17:10:46 <Mic> nhnt11: you're talking about having a property with a value (number) vs having a getter that adds the values of two other properties?
17:10:50 <nhnt11> (I'm working on that new patch btw)
17:10:56 <nhnt11> Mic: Yes
17:11:20 <Mic> I've no idea how the JS engine works but I wouldn't try to optimize on something like this. Maybe having a getter is much more expensive memory-wise?
17:11:23 <Mic> Who knows...
17:11:36 <aleth> It seems premature.
17:13:56 <-- nhnt11 has quit (Ping timeout)
17:14:05 <clokep_> qheaden: THe more you can decode in the packet the better, I'm not sure how easily / generically you can do this. :)
17:14:57 <flo-retina> qheaden: if you think your code may fail to decode the packet, then you need to include both the decoded and the hex versions. If you are confident the decoding code can never fail, then just the decoded version is enough.
17:16:07 --> nhnt11 has joined #instantbird
17:16:32 <nhnt11> Er, magic copy is broken??
17:17:09 <nhnt11> 22:43:38 - nhnt11: Ok, I'll leave it as it is.
17:17:09 <nhnt11> 22:43:45 - nhnt11: It was just a thought I had :P
17:17:09 <nhnt11> 22:44:24 - nhnt11: aleth: Where did you find my code vague-est?
17:17:09 <nhnt11> 22:44:30 - nhnt11: Er, most vague* :P
17:17:13 <aleth> nhnt11: https://bugzilla.instantbird.org/buglist.cgi?quicksearch=checkin-need
17:17:41 <nhnt11> aleth: Cool, I hadn't noticed it was broken till now
17:18:06 <aleth> nhnt11: Well, I asked you some questions like "why are your ids so long and why are they what they are" and you said "wait for comments"...
17:18:32 <nhnt11> Ah okay. That needs comments yeah
17:20:12 <nhnt11> aleth: The answer btw is that it's the easiest way to map log entries to PossibleConversations
17:21:08 <aleth> My larger design question was "does the indexing of the db allow for the searches we want to make"
17:23:08 <aleth> A possible answer is that maybe this is just a db for stats and not a db for searches
17:23:17 <aleth> But I'd be interested in your strategy.
17:23:29 <nhnt11> aleth: I'm doing the filtering test right now
17:24:03 <nhnt11> aleth: IndexedDB seemed an easy and convenient way to store stats, but when I chose to use it I hadn't thought about fast searches
17:24:16 <aleth> There's the test, and there's the question of how the awesomebar does it (just to get a point of comparison which maybe you shouldn't follow...)
17:24:34 <aleth> But it seems worth being conscious of the choice being made and the tradeoffs
17:24:36 <nhnt11> I know we want to search by topic for IRC channels, but I hadn't really thought about implementing a proper indexed search just for this
17:24:57 <aleth> It's not just the topic, it's a whole range of fields I think
17:25:12 <nhnt11> aleth: statusText.indexOf(aFilterStr) != -1 is pretty fast
17:25:16 <aleth> Certainly display name, possibly tags...
17:25:20 <nhnt11> I see no difference in snappiness, at least
17:25:29 <nhnt11> aleth: but for tags, I don't think we need to match substrings
17:25:40 <aleth> nhnt11: case-independent?
17:25:51 <nhnt11> I think MUC topics may be the only case we need to care about substrings
17:25:52 <nhnt11> yeah
17:27:06 <flo-retina> nhnt11: how are muc topics different from display names?
17:27:19 <flo-retina> in both case you want to match the beginning of any word it contains, right?
17:27:37 <aleth> flo-retina: Don't you also want to match substrings?
17:27:47 <flo-retina> aleth: example use case?
17:27:47 <aleth> The awesomebar does...
17:27:54 <flo-retina> aleth: really?
17:28:04 <nhnt11> yeah the awesomebar searches for substrings
17:28:31 <aleth> flo-retina: Stupid example, "bird"
17:28:38 <nhnt11> flo-retina: The idea is that if you're looking for a channel with a topic (let's say cake?) and you search for "cake", it's unlikely "cake" is the first word in the topic string
17:28:42 <nhnt11> Bird works too :P
17:29:30 <aleth> Also e.g. usernames which include a surname, without spaces, should be found using the surname as the filterterm
17:29:38 <flo-retina> nhnt11: cake is still a word. I said "the beginning of any word it contains"
17:29:52 <nhnt11> Ah
17:30:53 <nhnt11> Sorry, I read that wrong
17:30:57 <aleth> Anyway, beginning of word vs substr is really distinct from the question of indexing
17:32:25 <flo-retina> aleth: well, you can index a bit more efficiently if you only care about words
17:32:33 <nhnt11> aleth: I'm connected to freenode as well as moznet, and matching topics the same way we match the display name has no visible impact on performance
17:33:14 <aleth> flo-retina: You could index words and then look for exact word matches first and then substrings of words... My suggestion was to follow the awesomebar strategy
17:33:26 <flo-retina> aleth: I suspect substring match algorithms look first for words that match the string, and then search entries containing any of these words.
17:33:46 * flo-retina hasn't looked at the awesomebar's code recently
17:33:48 <aleth> flo-retina: Right
17:33:51 <flo-retina> do you suggest using sqlite?
17:34:01 <nhnt11> I searched for how the awesomebar does it for a few minutes the other day, but couldn't find it, I'll try finding the code again now
17:34:03 <aleth> flo-retina: Not necessarily, I just wanted an informed decision ;)
17:34:27 <flo-retina> I gotta run. Will likely be back online in less than an hour
17:34:28 <-- flo-retina has quit (Quit: Instantbird 1.5a1pre -- http://www.instantbird.com)
17:35:31 <aleth> nhnt11: If you're sure the performance of filter() is good enough, then maybe improve the filter function in the existing patch along the lines we discussed here and on the etherpad and then we will see when we land that patch ;)
17:35:49 <nhnt11> aleth: So in all, I really don't think IndexedDB is great for substring searching
17:36:07 <aleth> nhnt11: iirc more important than the substr issue is the "term1 term2" = term1 AND term2 issue
17:36:17 <nhnt11> Right but that's still a form of substring searching
17:36:45 <aleth> Well, it's an AND, i.e. two successive filters
17:37:01 <nhnt11> I just read what you said completely wrong again
17:37:59 <aleth> Or file a separate patch with a better filter function etc and we'll land that with/directly after 2066 which should be as soon as nightlies are back
17:38:28 <aleth> Then we can all see the performance in nightlies
17:38:53 <nhnt11> Ok
17:40:24 <aleth> You can leave out the substr for now and we will see if we miss it ;)
17:41:03 <nhnt11> aleth: You mean only match words right?
17:41:13 <aleth> Right, match the beginning of words.
17:41:26 <aleth> I think that's the current behaviour.
17:41:37 <aleth> But match on all the fields of interest, and allow for AND
17:41:40 <nhnt11> For the display names, yes
17:41:49 <nhnt11> ok
17:43:28 <nhnt11> aleth: I'm going to use bug 2146
17:43:32 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=2146 nor, --, ---, nobody, NEW, Awesometab should have smarter filtering
17:43:51 <aleth> An easy optimization could be to check if during typing we are just adding to the length of the filter string and use the existing filter result if there hasn't been a change in data
17:44:05 <aleth> Sounds good :)
17:48:06 <clokep_> nhnt11: Any reason you added an f? for me on bug 2143?
17:48:10 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=2143 enh, --, ---, nhnt11, NEW, Stats service should maintain statistical data for conversations and use it for sorting.
17:48:12 * clokep_ doesn't seem to be a good reviewer of that.
17:48:56 <nhnt11> clokep_: No reason, just figured if you could provide any feedback, it would be useful :)
17:49:01 <nhnt11> If it's not your thing, I'll cancel it
17:49:07 <nhnt11> (It isn't UI stuff :P )
17:49:26 <clokep_> nhnt11: I just don't know that code at all. :)
17:49:45 <nhnt11> Sure.
17:50:02 <instantbot> clokep@gmail.com cancelled feedback?(clokep@gmail.com) for attachment 2812 on bug 2143.
17:53:49 <-- jb has quit (Ping timeout)
17:59:57 --> flo-retina has joined #instantbird
17:59:58 * ChanServ sets mode +qo flo-retina flo-retina 
18:03:29 <instantbot> clokep@gmail.com granted review for attachment 2807 on bug 2135.
18:03:31 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=2135 nor, --, ---, qheaden, ASSI, Enable JS Yahoo By Default
18:05:40 <clokep_> bugspam for eveyrone!
18:13:01 --> unghost has joined #instantbird
18:24:35 <nhnt11> How do I declare in an IDL file that a function accepts an array as a parameter?
18:25:09 <nhnt11> I'm trying |array in ibIPossibleConversation aConvsToFilter|, not sure if it'll work
18:26:05 <nhnt11> Nope, didn't work as expected... maybe I need [array]
18:27:22 <nhnt11> bah, that gives me |struct.error: cannot convert argument to integer|
18:27:30 <nhnt11> I wonder if I need to have a size_is there...
18:28:11 <nhnt11> Or if I need to use nsIArray..
18:29:33 --> jb has joined #instantbird
18:32:27 <flo-retina> nhnt11: what are you trying to do?
18:33:58 <nhnt11> flo-retina: I'm trying to give the stats service an array of PossibleConvs filtered with filter string A, which need to be filtered again with filter string B, where B starts with A.
18:34:13 <nhnt11> That was a really weird explanation, sorry.
18:35:00 <nhnt11> Basically I'm trying to figure out how to give an array parameter to a function in stats service from the newtab binding..
18:35:47 <nhnt11> It appears the only way I can avoid a size_is property is to use nsIArray..
18:35:51 <flo-retina> so why are you moving possible conv instances back and forth between the UI and the service?
18:36:18 <nhnt11> I would have to keep references to lists of filtered conversations per newtab in the stats service otherwise.
18:36:47 <flo-retina> s/per newtab/per observer/ ?
18:36:53 <nhnt11> yeah.
18:37:24 <nhnt11> flo-retina: Actually a consumer needn't be an observer to use the getFilteredConvs function...
18:42:59 <instantbot> New Core - Yahoo! Messenger bug 2151 filed by aleth@instantbird.org.
18:42:59 <nhnt11> Aha
18:43:02 <instantbot> Bug https://bugzilla.instantbird.org/show_bug.cgi?id=2151 nor, --, ---, nobody, NEW, Missing packet handlers
18:43:20 <nhnt11> I forgot to add the optional parameter I was using to specify size_is in the stats service code
18:43:54 <nhnt11> Looks like it's all working now
18:45:24 <aleth> It might slow it down quite a bit if you have to pass wrapped arrays around though...
18:45:39 <-- aleth has quit (Quit: Ciao)
18:47:00 <flo-retina> aleth: passing arrays back and forth through xpconnect feels r-, but I'll wait and look at the code first; to see if I can suggest something better ;).
18:48:54 <nhnt11> flo-retina: I'm r-
18:48:59 <nhnt11> I'm r-'ing this myself
18:49:03 <nhnt11> Because it kinda sucks
18:49:24 <nhnt11> The objects lose any properties not defined in the interface
18:49:35 <nhnt11> and overall it just kinda sucks :P There must be a better way...
18:55:09 --> Even has joined #instantbird
18:55:09 * ChanServ sets mode +o Even 
18:59:31 <-- mconley has quit (Connection reset by peer)
18:59:48 --> mconley has joined #instantbird
19:02:09 <-- jb has quit (Ping timeout)
19:26:30 <-- chrisccoulson has quit (Quit: OSError: [Errno 130] Owner died)
19:27:52 <nhnt11> Hmm, filtering is taking between 100-120ms
19:28:11 <nhnt11> (with the status text search-by-word stuff)
19:28:26 <nhnt11> this is.. way too much. :/
19:29:48 <nhnt11> flo-retina: The only real way to speed this up would be to index every chat room we receive. Do we want to do this?
19:30:14 <flo-retina> any reason for not doing it?
19:30:43 <nhnt11> Indexing only chats we've already opened wouldn't work for the case where you're searching for "bird"
19:31:00 <nhnt11> flo-retina: No, but then the stats service would have to be renamed :P
19:31:15 <flo-retina> I'm not searching for 'bird'.
19:31:37 <nhnt11> Um..
19:32:33 <nhnt11> Are you saying you don't care about channel discovery, or simply picking at my use of the word "you", or something else entirely? :P
19:32:54 <flo-retina> nhnt11: I don't care about searching in the middle of words.
19:33:01 <flo-retina> I care about cakes though :-P.
19:34:21 <nhnt11> flo-retina: I didn't mean bird as in thunderbird, just bird :P
19:34:23 <nhnt11> ok..
19:34:28 <flo-retina> ah :)
19:37:57 <nhnt11> flo-retina: OK so I guess I'll have to move to using sqlite if we want proper indexing
19:38:12 <nhnt11> IndexedDB doesn't seem very well suited..
19:38:20 <nhnt11> I used it because it was very convenient for storing stats
19:45:09 <nhnt11> Hmm, actually it can be done with IndexedDb...
19:45:29 <nhnt11> I'd have to store a stats object multiple times for each word in its status/topic/name/whatever
19:45:43 --> FireFly_TB has joined #instantbird
19:46:51 <nhnt11> Or at least, store the conversation id, and use that to get the actual conv.
19:49:14 --> chrisccoulson has joined #instantbird
19:52:43 --> jb has joined #instantbird
19:53:53 <nhnt11> flo-retina: This requires more thought, but I think it can be done with IndexedDb. I would have to store a conversation id multiple times in a given index, once for each word in the indexed property. Then open a cursor with the lower bound as the filter string (say "cake") and the upper bound as the first string in lexicographical order that wouldn't match (in this case "cakf"). Then sort the matching conversations by rank and return
19:53:54 <nhnt11> them. If this seems ok, I'd like to complete the stats patch first and then work on this.
19:57:02 * nhnt11 doesn't have much experience (none really) with databases and doesn't know if this is ok or gross.
20:00:22 * nhnt11 is reading up on sqlite.
20:02:37 <nhnt11> Looks like it would be a lot easier with sqlite..
20:05:49 <clokep_> sqlite is pretty easy to use.
20:07:29 <nhnt11> clokep_: More stuff to learn :D
20:14:09 <clokep_> Databases are a good thing to understand to a degree.
20:17:49 <nhnt11> Hmm, the contacts service seems to do something similar to what I had in mind
20:18:34 <nhnt11> Tags for example, are kept in memory by id, and getTagByName uses an sql query to get the id for a tag given its name, and then uses that to get the tag from memory.
20:19:59 <nhnt11> So all this reading is telling me sqlite would be pretty convenient and I should use it.
20:20:38 <nhnt11> Right now though I need to go to bed. I'll be back tomorrow after lunch to code this up.
20:20:43 <nhnt11> Good night
20:20:45 <-- nhnt11 has quit (Quit: Instantbird 1.5a1pre -- http://www.instantbird.com)
20:25:17 <Mic> Good night
20:26:57 <-- Mic has quit (Quit: Instantbird -- http://www.instantbird.com)
20:28:53 <-- jb has quit (Quit: jb)
20:32:37 --> jb has joined #instantbird
20:40:17 <flo-retina> nhnt11: FYI, I'm pretty unhappy with the content of imContacts.js. I don't think it has any good reason to use sqlite, as we basically read all the database and put it in memory at startup.
20:40:35 <flo-retina> that doesn't mean you shouldn't use sqlite :). Just that imContacts.js may not be the best example
20:50:25 --> EionRobb has joined #instantbird
21:03:22 --> jb1 has joined #instantbird
21:03:39 <-- jb has quit (Connection reset by peer)
21:08:49 <-- clokep_ has quit (Quit: http://www.mibbit.com ajax IRC Client)
21:27:12 <-- mconley has quit (Input/output error)
21:31:07 --> dionisos has joined #instantbird
21:37:12 --> rosonline has joined #instantbird
21:37:19 <-- unghost has quit (Quit: Ухожу я от вас (xchat 2.4.5 или старше))
21:49:24 <-- jb1 has quit (Ping timeout)
21:55:17 <-- dionisos has quit (Ping timeout)
22:03:05 --> clokep has joined #instantbird
22:03:05 * ChanServ sets mode +o clokep 
22:22:26 <-- qlum has quit (Quit: Getting the <censored> out.)
22:30:22 <clokep> flo-retina: You have a few minutes?
22:30:56 <flo-retina> ask the question to let me decide if I've a few minutes then? :-D
22:31:07 <clokep> I want to talk about https://bugzilla.instantbird.org/show_bug.cgi?id=1587#c13
22:31:11 <instantbot> Bug 1587 min, --, 1.5, clokep, REOP, Re-sync build scripts with comm-central
22:31:19 <clokep> All your "whys?" are "Because I synced the scripts with c-c"
22:31:22 <clokep> (Well, c-r)
22:31:56 <flo-retina> there are only 2
22:32:07 <clokep> Yes.
22:32:17 <flo-retina> and I think these 2 change would be regressions
22:32:45 <clokep> OK.
22:33:05 <clokep> It's very frustrating not knowing what WE added and what was part of the stuff we got from comm-* btw. :(
22:35:14 <clokep> flo-retina: Also I don't understand this sentence "some of these
22:35:14 <clokep> changesets undo what was done in previous changesets, so I think we would be
22:35:14 <clokep> better off merging everything into the other part."
22:35:23 <flo-retina> can't the hg blame help with that?
22:36:50 <flo-retina> clokep: the patches you attached contain several changesets. One of the changeset does something that another changeset reverts it.
22:39:22 <clokep> flo-retina: OK. So you want all those changes as one changeset and not ports of the comm-* changes?
22:39:58 <flo-retina> yeah
22:40:03 <flo-retina> that's the meaning of that sentence
22:40:09 <flo-retina> I don't remember what's exactly in there
22:41:40 <clokep> OK, that's fine.
22:41:51 <clokep> I always am unsure which way we prefer doing things like that. :)
22:52:59 <flo-retina> I think we prefer having the most meaningful hg history
22:53:22 <flo-retina> so when the separate changesets add value because they explain the reason for different changes, it's interesting to keep them separate.
22:53:31 <flo-retina> if they are just "adding noise", we can as well merge them
22:56:30 <clokep> OK.
22:56:35 * clokep is working on some comments...
23:03:33 <clokep> flo-retina: Is there any trick to finding when lines were removed from a file using hg annotate?
23:03:58 <flo-retina> I don't know any
23:04:42 <clokep> :(
23:10:41 --> dionisos has joined #instantbird
23:13:16 <flo-retina> Good night
23:34:27 <-- qheaden has quit (Ping timeout)
23:42:00 --> qheaden has joined #instantbird