#instantbird log on 12 02 2014

All times are UTC.

00:00:01 <nhnt11> aleth: It's too saturated right?
00:00:02 <nhnt11> The blue
00:00:14 <nhnt11> (are you seeing what I'm seeing)
00:00:40 <aleth> There's less gradient on it too
00:00:56 <-- nhnt11 has quit (Quit: Instantbird 1.6a1pre -- http://www.instantbird.com)
00:01:04 --> nhnt11 has joined #instantbird
00:01:04 * ChanServ sets mode +h nhnt11 
00:01:11 <nhnt11> Hello, Helvetica Instantbird! :)
00:01:30 <aleth> 10.10 must have looked really odd to you :P
00:01:41 <nhnt11> aleth: Everything other than Fx and Ib was Helvetica :S
00:02:17 <nhnt11> I kinda like LucidaGrande, so I never mentioned anything :]
00:02:41 <nhnt11> Actually, it's not that I like LucidaGrande, it's that I really dislike Helvetica in bold in the tab titles
00:03:07 <nhnt11> Firefox doesn't bold tab titles (anymore?), I think we should stop doing it too
00:03:09 <aleth> One of these days you'll just have to port Australis ;)
00:03:13 <nhnt11> yeah :(
00:03:20 <aleth> Can't be that bad
00:03:26 <nhnt11> I have time in December :)
00:03:33 <nhnt11> I think I've got nearly 4 weeks of holidays
00:03:49 * nhnt11 wants to land log indexing
00:04:00 <aleth> yup, that would be good :-)
00:04:08 * aleth hates grepping logs
00:05:09 <nhnt11> This looks weird: http://puu.sh/ddnia/62b5323301.png
00:05:14 * nhnt11 plays a bit
00:05:49 <nhnt11> Better, I think: http://puu.sh/ddnll/acceab7671.png
00:06:46 <aleth> Just steal the size from the Australis CSS?
00:07:02 --> flo-retina has joined #instantbird
00:07:02 * ChanServ sets mode +qo flo-retina flo-retina 
00:07:09 <nhnt11> Might as well steal Australis ;)
00:07:13 <aleth> Yup!
00:07:28 <aleth> gooood morning Portland! :)
00:08:24 <flo-retina> aleth: it's 4pm already ;)
00:08:30 <aleth> nhnt11: Get rid of the title bar while you are at it :)
00:08:46 * nhnt11 isn't porting Australis right now
00:08:50 <aleth> I know.
00:09:26 <flo-retina> if I get bored during the last week of the year, I may do it
00:10:21 <clokep> :)
00:10:32 <clokep> Maybe you'll make the Windows UI not look like crap too? :P
00:11:47 <flo-retina> spending the week on Windows doesn't seem like a great Christmas present to do to myself :-D
00:12:14 <clokep> Maybe you've been bad and it's like coal? :P
00:17:11 --> clokep_work has joined #instantbird
00:17:11 * ChanServ sets mode +o clokep_work 
00:20:06 <-- clokep_work has quit (Quit: Instantbird 1.6a1pre -- http://www.instantbird.com)
00:20:14 --> clokep_work has joined #instantbird
00:20:14 * ChanServ sets mode +o clokep_work 
00:23:56 <-- arlolra has quit (Client exited)
00:25:35 <-- clokep_work has quit (Quit: Instantbird 1.6a1pre -- http://www.instantbird.com)
00:38:01 --> rosonline has joined #instantbird
00:43:50 * nhnt11 wonders if flo-retina can reproduce bug 978057
00:43:52 <instantbot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=978057 nor, --, ---, nobody, NEW, Text overlaps search icon in awesometab filterbar
00:52:14 * clokep doesn't have a non-retina screen to try on
00:55:18 <nhnt11> clokep: There's this app called RDM (I think) that let's you change your resolution to non-retina
00:55:27 <nhnt11> i.e. sans pixel doubling
00:55:42 <nhnt11> (you can choose either 2880x1800 full, or 1440x900 stretched)
00:55:43 <nhnt11> just fyi
00:58:03 <nhnt11> https://www.reddit.com/r/apple/comments/vi9yf/set_your_retina_macbook_pros_resolution_to/
00:59:51 <flo-retina> bug 791199 is really more about TB than IB, so 1.6-whatever doesn't make a terrible amount of sense
00:59:53 <instantbot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=791199 maj, --, ---, nobody, NEW, chat-buddy-add notification implicitly assumes that there will be no more than one observer
01:00:06 <flo-retina> clokep: I haven't packed a non-retina screen in my suitcase
01:01:40 <flo-retina> why does bug 978057 have a tiff picture attached? :-S
01:01:42 <instantbot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=978057 nor, --, ---, nobody, NEW, Text overlaps search icon in awesometab filterbar
01:02:42 <nhnt11> Probably because OS X's Grab tool uses TIFF as default..
01:03:54 <flo-retina> nhnt11: it has used PNG since 10.5
01:04:13 <nhnt11> flo-retina: Weird, it definitely still defaults to TIFF here
01:04:16 <flo-retina> or do you mean they regressed that in 10.10?
01:04:21 <nhnt11> I'm talking about the app named Grab btw
01:04:25 <nhnt11> not Cmd+Shift+3/4
01:04:33 <flo-retina> that app is useless :-D
01:04:35 <nhnt11> It's in /Applications/Utilites
01:04:37 <nhnt11> I agree
01:04:46 <nhnt11> But if you don't know about Cmd+Shift+3/4...
01:04:47 <flo-retina> hmm, except for timed screenshots
01:06:00 <clokep> flo-retina: I ran into that tiff issue earlier. ;)
01:06:23 --> clokep_work has joined #instantbird
01:06:23 * ChanServ sets mode +o clokep_work 
01:06:53 <-- clokep_work has quit (Quit: Instantbird 1.6a1pre -- http://www.instantbird.com)
01:07:05 * clokep hates when he adds debugging statements and they don't give him mor einfo.
01:07:10 <clokep> Cause I forgot to include important info. :-D
01:09:06 <flo-retina> clokep: so re bug 1106539, I'm hesitating
01:09:08 <instantbot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=1106539 min, --, ---, yarik.sheptykin, NEW, Extend httpRequest API to allow setting custom MIME type for POST and overriding response MIME type.
01:09:20 <clokep> Between?
01:09:47 <flo-retina> between reviewing my r? flag, saying I agree with your comments, or adding another r- reason
01:10:14 <flo-retina> I think I would prefer isContentTypeSet to be computed only if Array.isArray(POSTData)
01:10:50 <flo-retina> s/reviewing/removing/
01:11:08 <clokep> If you think it will make it clearer, please add the comment. :)
01:11:19 <clokep> Also, is it kosher for me to r- random things? :-D
01:11:58 --> clokep_work has joined #instantbird
01:11:59 * ChanServ sets mode +o clokep_work 
01:12:02 <flo-retina> yes, you know that code
01:12:16 <flo-retina> it wouldn't be for _really_ random things ;)
01:12:33 <clokep> aleth: I'm having trouble reproducing that bug. :(
01:14:34 <clokep> aleth: AH! I THINK I know what's causing it. You still here to brainstorm?
01:14:39 <clokep> (About WTF my code is trying to do. :-D)
01:15:35 <-- clokep_work has quit (A TLS packet with unexpected length was received.)
01:18:24 <-- mpmc has quit (Quit: Instantbird 1.6a1pre -- http://www.instantbird.com)
01:22:36 <clokep> flo-retina should be happy about that patch I just put up. ;)
01:22:42 <-- rosonline has quit (Quit: Instantbird 1.5 -- http://www.instantbird.com)
01:26:07 <clokep> Now that I feel like I've helped w/ the 1.6 blockers, I can do something fun.
01:26:13 <EionRobb> ooh ooh!
01:26:15 <flo-retina> :)
01:26:36 <clokep> But it won't be fun reading EionRobb's code. :'(
01:26:58 <EionRobb> :O rude
01:28:14 <clokep> EionRobb: I tease. :)
01:28:19 <EionRobb> hehe, I know
01:29:39 <clokep> Bah...I really wish I had like a template for making new protocols...
01:34:01 <flo-retina> wasn't the test prpl supposed to be that?
01:34:08 --> Hoony has joined #instantbird
01:35:06 <clokep> Maybe....
01:36:52 --> Widdershins has joined #instantbird
01:45:15 --> Hoony1 has joined #instantbird
01:45:15 <-- Hoony has quit (Connection closed)
01:47:08 --> clokep_work has joined #instantbird
01:47:08 * ChanServ sets mode +o clokep_work 
01:49:23 <clokep_work> EionRobb: The purple API for HTTP is gross. :)
01:49:29 <EionRobb> yes
01:49:54 <EionRobb> there's apparently a new api for libpurple3, but I haven't tried it yet
01:51:19 <clokep_work> Although I do wish our HTTP API used promises...
01:51:44 <EionRobb> that's easy enough to do right?
01:56:12 <clokep_work> EionRobb: What is 524288 in https://code.google.com/p/skype4pidgin/source/browse/trunk/skypeweb/skypeweb_login.c#83 I don't see it in https://developer.pidgin.im/doxygen/dev2.x.y/html/util_8h.html#af3a7be80051547d7a16fb404ab27bd3a?
01:56:37 <clokep_work> It looks like you're using _len version that has a max_len?
01:56:52 <EionRobb> there's some #define's in libskypeweb.h to make it compatibible with libpurple3 api
01:57:11 <EionRobb> otherwise there was a lot of #if !PURPLE_VERSION(3, 0, 0)  blah
01:57:24 <EionRobb> but yeah, using the with-account with-len versions of that function
01:57:33 <clokep_work> Ah you aliased stuff, yeah it's purple_util_fetch_url_request_len_with_account
01:57:42 <EionRobb> sorry about that
01:57:42 <clokep_work> Thanks. :)
01:57:59 <clokep_work> No big deal. :) Not that I care about the max length anyway.
01:58:05 <clokep_work> Just wanted to ensure I understood.
01:58:09 <-- aleth has quit (Ping timeout: 121 seconds)
01:58:14 <EionRobb> and the 512k is just the default that pidgin uses, for better or for worse.  just wanted to get rid of the warnings about '-1 length set, setting to 512k'
02:00:15 <clokep> Yep.
02:04:46 <clokep_work> That's lame that you have to encode all the parameters yourself and such.
02:07:51 <EionRobb> whee
02:07:53 <EionRobb> welcome to C
02:08:19 <clokep> Yeah I guess...
02:09:12 <-- Mook_a1 has quit (Client exited)
02:22:10 <clokep_work> EionRobb: So really I can just open that URL in a window and let it render and boom, I'm done, right? :P
02:22:18 <clokep_work> Or just eval() the JS? That seems safe.
02:22:40 <EionRobb> which url?
02:22:55 <EionRobb> oh, yes lol
02:23:00 <EionRobb> and block the elegibityCheck :)
02:24:36 <clokep> Yeah...
02:24:48 <clokep> I wonder if it's considered "unsafe" to create a DOM out of this.
02:24:54 <clokep> Or if I should just do string parsing like you did...
02:38:02 <clokep_work> EionRobb: Where's the eligibility check occur? Just curious. :)
02:43:26 <EionRobb> on https://web.skype.com/main  it does the elegibity check then
02:44:41 <clokep_work> Ah, I see.
02:45:03 <clokep_work> EionRobb: Sorry for all the questions, but the headers you're adding at https://code.google.com/p/skype4pidgin/source/browse/trunk/skypeweb/skypeweb_login.c#75 are any of those in particular important?
02:45:09 <clokep_work> I've never seen this "BehaviorOverride" before.
02:45:25 <clokep_work> (Obvously Host, ContentType, etc. are standard.)
02:46:16 <EionRobb> BehaviorOverride looks like a custom microsoft thing.  it stops the response from doing a 302 Found Location redirect thing, since there's important headers that need to be plucked before the redirect happens
02:46:31 <clokep_work> Huh, interesting.
02:46:50 <clokep_work> And how did you figure that out? :-D
02:47:21 <EionRobb> through the miracle of sound
02:48:01 <EionRobb> some of the other requests sent that header :)
02:48:06 <EionRobb> but mostly the sound one ;)
02:50:20 <EionRobb> clokep_work: don't apologise for the questions either, that's very offensive!
02:50:56 <EionRobb> ;)
03:01:31 * clokep_work is going to set you as a reviewer for this. :P
03:01:55 <EionRobb> :O
03:02:56 <EionRobb> something new in there is that it lists the 'endpoints' like all the xmpp resources for a user.  I'm gonna bring that through in the tooltips tonight and see what weird devices my buddies use
03:03:42 <EionRobb> I'm debating about whether to say that it's pidgin or to pretend to be the skype webpage
03:03:55 <EionRobb> currently it pretends to be a win10 chrome desktop
03:04:34 <clokep_work> :)
03:04:51 <clokep_work> I'm hoping to get through the login sequence today, figured that's a reasonable goal.
03:05:39 <EionRobb> yeah
03:05:43 * nhnt11 is interested
03:05:52 <EionRobb> will you include the login to the messaging system as part of that?
03:06:07 <nhnt11> clokep_work: I've got 4 holidays after my exam today, I'd like to look at it/help where I can :)
03:06:16 <EionRobb> you can make the coffee
03:06:29 <clokep_work> EionRobb: You mean that crazy SHA256 thing?
03:06:37 <clokep_work> I'm just following through your sequence.
03:06:43 <clokep_work> We'll see how far I get before I'm tired. ;)
03:06:50 <EionRobb> yeah, I think I mean that
03:07:09 * clokep_work is trying to see the deal w/ these cookies.
03:07:14 <clokep_work> (Mmmm....)
03:07:18 <clokep_work> nhnt11: Cool. :)
03:07:19 <EionRobb> no need for the cookies
03:09:30 <clokep_work> So skypeweb_update_cookies is useless? :-D
03:09:57 <EionRobb> a hangover from the steam prpl
03:10:08 <EionRobb> and part of me banging my head at why the skypetoken wasn't working
03:10:31 <EionRobb> I thought it had to come from the cookie, but it turns out the one that gets dumped out in html (only very subtly different) is the right token
03:10:37 <clokep_work> Yeah...you should remove that then. :-D
03:10:47 <EionRobb> the other token doesn't have permission to get the contacts list.  grr
03:10:57 <EionRobb> I might make it look more important as a trap for new players
03:11:10 <EionRobb> not enough easter eggs in the code, if you ask me
03:12:00 <clokep_work> Hah.
03:12:09 <clokep_work> skypeweb_do_all_the_things :-D
03:12:10 <clokep_work> Nice...
03:12:37 <EionRobb> hehe
03:18:15 --> mconley has joined #instantbird
03:23:20 <clokep_work> Authentication was successful. :)
03:24:05 <EionRobb> yay
03:24:12 <EionRobb> you make it look so easy
03:24:22 <clokep> Haha.
03:24:26 <clokep> Want to see the code so far?
03:24:30 <EionRobb> yesplz
03:24:50 <clokep_work> EionRobb: http://pastebin.instantbird.com/1077452
03:25:06 <clokep_work> A lot of boilerplate still obviously.
03:26:02 <EionRobb> hmm... is the sign for the timezone the right way around?
03:26:29 <clokep_work> EionRobb: The code is literally copied from the returned webpage.
03:27:23 <clokep_work> How often are we expected to get a registration token? Is that a one-time thing that should be cached forever or a per run type thing?
03:28:50 <EionRobb> 'should' be cached 'forever' probably (it has an expiry, I believe)
03:29:08 <clokep_work> Hm. OK. I'll be lazy for the moment and just not care.
03:35:35 <clokep_work> Looks like I got to your "fun" function.
03:36:09 <clokep_work> EionRobb: A stupid question, time(NULL) (https://code.google.com/p/skype4pidgin/source/browse/trunk/skypeweb/skypeweb_messages.c#381) is just seconds since the unix epoch?
03:36:32 <EionRobb> yeah
03:36:46 <clokep> K, time to work on skypeweb_hmac_sha256 then...
03:37:07 <EionRobb> I should change the name of that.  I thought it was hmac-sha256 to begin with
03:37:07 <clokep> I don't suppose you have a test case for it? :-D
03:37:15 <clokep> What is it actually?
03:37:18 <EionRobb> I don't test my code, I just make it work ;)
03:37:31 <EionRobb> its a magic thing that msn added to their protocol back in msnp4 to stop 3rd parties
03:37:51 <EionRobb> ... updated by swapping md5() for sha256()
03:40:40 <clokep_work> Haha, cute. :)
03:49:20 <clokep_work> Oh god, this deals with endianess and such?
03:50:02 <EionRobb> mwahaha
03:50:11 <EionRobb> if it helps, ms managed to do it in js
03:53:23 <-- Hoony1 has quit (Client exited)
03:54:55 <clokep> Yeah...well...yeah...
03:59:25 <clokep_work> EionRobb: Am I stupid or does the math in https://code.google.com/p/skype4pidgin/source/browse/trunk/skypeweb/skypeweb_util.c#127 not make sense?
03:59:44 <clokep_work> It says split it into 4 integers...but SHA 256... 256 / 32 = 8 ints?
04:00:16 <EionRobb> ahahaha, yeah its only the first half of the sha256() otherwise it'd break the rest of the algorithm
04:00:25 * clokep_work sighs.
04:00:41 <EionRobb> because sha256()/2 is more secure than md5() right? 
04:00:43 <EionRobb> right? :D
04:01:49 <EionRobb> I'm sure some manager came along and said "what‽ why is there md5() in this code, this is not secure!" :D
04:03:13 <clokep> I'm sure, haha.
04:03:20 <clokep> nhnt11: Ping
04:03:25 <nhnt11> clokep: pong
04:03:37 <clokep_work> nhnt11: Does the scratchpad have access to chrome stuff?
04:03:43 * clokep_work hates the scratchpad and finds it pretty confusing.
04:03:45 <nhnt11> Yes
04:03:50 <clokep_work> :-\
04:04:05 <clokep_work> Components.utils is undefined :-S
04:04:08 <nhnt11> clokep_work: You should see an "Environment" menu
04:04:16 <clokep_work> Ah.
04:04:17 <nhnt11> Choose "Browser"
04:14:18 <clokep_work> EionRobb: So...my C is rusty, What does & 0x7FFF.... do?
04:14:36 <clokep_work> That just chops off the top bit of the int?
04:15:14 <nhnt11> aleth, clokep_work: Fwiw the latest Fx nightly seems to have issues with images too.. with buttons though
04:15:35 <nhnt11> Sometimes buttons are just... blank, until I hover them
04:15:45 <nhnt11> So the tab thing may be an upstream issue..
04:16:24 <EionRobb> clokep_work: yes, choppy choppy
04:22:01 <instant-buildbot> build #1245 of linux-nightly-default is complete: Failure [4failed compile]  Build details are at http://buildbot.instantbird.org/builders/linux-nightly-default/builds/1245
04:37:23 <clokep_work> EionRobb: So... what is https://code.google.com/p/skype4pidgin/source/browse/trunk/skypeweb/skypeweb_util.c#156 doing?
04:37:31 <clokep_work> Taking something, converting it to little endian and then setting it back?
04:38:19 <EionRobb> yeah, ensuring it's little endian
04:38:44 <clokep_work> Hm....OK...
04:39:22 <EionRobb> the annoying bit is there's a fat wad of js just sitting around tempting you to look at how to do it, but you're not allowed to see it
04:39:52 <clokep_work> Yes. ;)
04:40:15 <clokep_work> I had written half the timezone thing...and it was identical (- minor changes in style) so I just checked against it. :)
04:40:34 <EionRobb> like someone parading around with the answers to the exam that you're presently sitting
04:40:37 <clokep_work> But copying this owuld be a little...excessive.
04:42:52 <EionRobb> I've just tried looking around to see if someone else has written any msn auth in javascript that you could borrow from :)
04:43:14 <clokep_work> I tried too, didn't see anything.
04:43:18 <clokep_work> Was searching the node modules myself.
04:43:29 * clokep_work is confused at why this is constantly converting to little endian.
04:43:38 <EionRobb> I did find http://wso2.com/project/mashup/2.0.1/docs/apidocs/org/wso2/mashup/hostobjects/im/IM.html which confused me though
04:43:47 <clokep_work> I'm pretty sure I'm going to mess one of them up. :(
04:48:42 <EionRobb> but hey, if you do it right, you'll be able to use it for an msn-js prpl :)
04:48:58 <EionRobb> I have a few test cases around if you want to try
04:49:09 <-- gerard-majax has quit (Ping timeout: 121 seconds)
04:52:53 * clokep_work is getting sleepy...
04:53:27 <EionRobb> I spent a full day trying to get that thing working, if that helps
04:53:38 --> gerard-majax has joined #instantbird
04:53:46 <EionRobb> turns out that only allocating half the necessary memory for the sha256 is kinda important
04:54:18 <EionRobb> also didn't help that I tried going down the hmac-sha256 route first
04:56:53 <instant-buildbot> build #2447 of macosx-nightly-default is complete: Success [3build successful]  Build details are at http://buildbot.instantbird.org/builders/macosx-nightly-default/builds/2447
04:58:28 <clokep_work> Uh oh...
04:58:33 <clokep_work> I might have done the HMAC...
04:59:42 <EionRobb> and that's why I need to rename that function
04:59:57 <-- sherief has quit (Ping timeout: 121 seconds)
05:02:15 <-- mconley has quit (Ping timeout: 121 seconds)
05:02:55 <clokep_work> Haha
05:03:04 <clokep_work> I'm unsure how easy it is for me to do an SHA 256...hmm..
05:04:15 <-- nhnt11 has quit (Quit: Instantbird 1.6a1pre -- http://www.instantbird.com)
05:04:25 --> nhnt11 has joined #instantbird
05:04:25 * ChanServ sets mode +h nhnt11 
05:05:19 <EionRobb> http://code.google.com/p/crypto-js/ ? :)
05:07:49 --> mconley has joined #instantbird
05:09:23 <EionRobb> well, time for me to cycle home
05:09:32 <EionRobb> good luck, brave soldier
05:09:40 <-- EionRobb has quit (Quit: Leaving.)
05:10:02 * clokep_work forgets how to make new tests actually work.
05:12:13 <-- mconley has quit (Ping timeout: 121 seconds)
05:16:52 --> mconley has joined #instantbird
05:21:27 <-- clokep has quit (Ping timeout: 121 seconds)
05:25:50 <-- mconley has quit (Ping timeout: 121 seconds)
05:49:16 <-- gerard-majax has quit (Ping timeout: 121 seconds)
05:51:09 --> mconley has joined #instantbird
05:55:52 <-- flo-retina has quit (Ping timeout: 121 seconds)
05:56:08 <-- mconley has quit (Ping timeout: 121 seconds)
05:56:37 --> mconley has joined #instantbird
06:02:51 --> Hoony has joined #instantbird
06:02:54 <-- Hoony has quit (Client exited)
06:03:02 --> Hoony has joined #instantbird
06:11:01 <-- mconley has quit (Ping timeout: 121 seconds)
06:11:04 <-- Hoony has quit (Ping timeout: 121 seconds)
06:11:43 --> Hoony has joined #instantbird
06:12:56 --> mconley has joined #instantbird
06:17:17 --> flo-retina has joined #instantbird
06:17:17 * ChanServ sets mode +qo flo-retina flo-retina 
06:32:21 --> Defman has joined #instantbird
06:39:40 --> Hoony1 has joined #instantbird
06:42:19 <-- Hoony has quit (Ping timeout: 121 seconds)
06:44:24 <-- Defman has quit (Ping timeout: 121 seconds)
06:52:41 <-- Hoony1 has quit (Ping timeout: 121 seconds)
07:04:44 <instant-buildbot> build #141 of linux64-nightly-default is complete: Success [3build successful]  Build details are at http://buildbot.instantbird.org/builders/linux64-nightly-default/builds/141
07:07:34 --> Hoony has joined #instantbird
07:14:56 <-- Widdershins has quit (Ping timeout: 121 seconds)
07:34:11 <-- Hoony has quit (Client exited)
07:44:23 <-- mconley has quit (Connection closed)
07:49:52 <-- Suiseiseki has quit (Connection closed)
07:56:34 --> Suiseiseki has joined #instantbird
07:57:45 --> MattATobin has joined #instantbird
07:59:22 <-- Tobin has quit (Ping timeout: 121 seconds)
08:05:14 * Fallen|away is now known as Fallen 
08:17:27 <-- flo-retina has quit (Connection closed)
08:21:02 * MattATobin is now known as Tobin 
08:41:54 --> Hoony has joined #instantbird
08:44:29 <-- Hoony has quit (Client exited)
09:06:17 --> hadi has joined #instantbird
09:11:43 * Tobin is now known as BinaryOutcast 
09:32:14 <-- clokep_work has quit (Connection closed)
09:32:23 --> clokep_work has joined #instantbird
09:32:23 * ChanServ sets mode +o clokep_work 
09:56:17 --> aleth has joined #instantbird
09:56:18 * ChanServ sets mode +o aleth 
09:57:37 <-- aleth has quit (Quit: :tiuQ)
09:57:40 --> aleth has joined #instantbird
09:57:41 * ChanServ sets mode +o aleth 
10:01:42 --> BWMerlin has joined #instantbird
10:25:46 <aleth> clokep_work: Does |crypto.subtle.digest("SHA-256", data).then((digest) => ...)| already work?
10:27:06 <aleth> https://developer.mozilla.org/en-US/docs/Web/API/Crypto somewhat badly documented
10:30:48 <aleth> yeah, looks like it works. data is an Arraybuffer and so is the output
10:36:13 --> mpmc has joined #instantbird
11:10:28 --> Armada has joined #instantbird
11:17:42 <clokep_work> aleth: Thanks for figuring that out. :) My plan was to check that.
11:18:07 <clokep_work> :-S
11:18:18 <clokep_work> Now if someone could tell me what to do to have a test 'registered' with the build system...
11:18:33 --> clokep has joined #instantbird
11:18:33 * ChanServ sets mode +o clokep 
11:18:59 <aleth> xpcshell.ini?
11:20:01 <-- BWMerlin has quit (Client exited)
11:20:11 --> BWMerlin has joined #instantbird
11:29:38 <clokep_work> Yeah, I have that set.
11:29:46 <clokep_work> It doesn't seem to be using it or something.
11:30:08 <clokep_work> When I try to run tests it says none are found.
11:30:18 <aleth> Have you set it in moz.build?
11:30:28 <aleth> XPCSHELL_TESTS_MANIFESTS += ['test/xpcshell.ini']
11:32:01 <clokep_work> Yes.
11:32:11 <aleth> Strange, that should work
11:32:48 <aleth> and im/test/xpcshell.ini: ?
11:33:32 * aleth is unimpressed by the number of places you have to add it
11:33:59 <clokep_work> aleth: http://pastebin.instantbird.com/1078106
11:34:12 <clokep_work> You only have to add it there if you're running all tests, Ithink.
11:34:37 <clokep_work> If you can find the issue I'd be eternally grateful. :-D
11:34:40 * clokep_work shuts down this PC.
11:37:34 <aleth> Well, I'd try adding it there.
11:38:32 <clokep> OK. I'll try when I get to work.
11:38:42 <-- clokep_work has quit (Ping timeout: 121 seconds)
11:55:12 <-- aleth has quit (Quit: :tiuQ)
11:56:40 <-- clokep has quit (Ping timeout: 121 seconds)
12:01:40 --> Hoony has joined #instantbird
12:05:16 --> aleth has joined #instantbird
12:05:16 * ChanServ sets mode +o aleth 
12:28:11 <-- BWMerlin has quit (Client exited)
12:37:13 --> clokep_work has joined #instantbird
12:37:13 * ChanServ sets mode +o clokep_work 
12:42:09 <clokep_work> We really need to fix http://dxr.mozilla.org/comm-central/source/im/test/xpcshell.ini#9
12:48:09 <clokep_work> aleth: Well you were right. :-S
12:48:44 <clokep_work> https://developer.mozilla.org/en-US/docs/Mozilla/Performance/Reporting_a_Thunderbird_Performance_Problem_with_G could be interesting to people...
13:07:54 <-- Hoony has quit (Client exited)
13:16:08 <clokep_work> nhnt11: Do we have examples of atests w/ a promise?
13:16:11 <clokep_work> A simple one. :)
13:16:28 <nhnt11> clokep_work: test_logger.js, but not sure if those are "simple"
13:16:35 <nhnt11> I mean, not sure what you mean by simple
13:17:15 <nhnt11> If you just mean async testing with Promises, check this out: https://developer.mozilla.org/en-US/docs/Mozilla/QA/Writing_xpcshell-based_unit_tests#Task-based_asynchronous_tests
13:18:07 <Fallen> is there an obs service notification or such when your nick gets highlighted?
13:18:40 <clokep_work> Fallen: I think some events are fired, yes.
13:18:43 <clokep_work> They're cancellable too.
13:19:03 <clokep_work> nhnt11: DOH! yield, of course...
13:19:57 <clokep_work> Hm....although...do I need to create my own task first then? :-S
13:20:49 <aleth> for add_task, yes
13:21:08 <Fallen> clokep_work: http://mxr.mozilla.org/comm-central/source/calendar/test/unit/test_deleted_items.js#35
13:22:52 <Fallen> make that function* test_deleted_items() for extra points
13:24:18 <clokep_work> Does anyone else hate the style of having add_test/add_task at the top level with a function immediately inside of it?
13:25:15 <clokep_work> aleth: Meh...crypto is undefined. :(
13:25:22 <clokep_work> I probably need to figure out how to import that.
13:25:45 <aleth> It's window.crypto, technically
13:25:57 <clokep_work> aleth: There's no window, it's a component.
13:25:58 <aleth> So you might need to get some window :-/
13:26:04 <clokep_work> Nah.
13:26:05 <aleth> Yes, I just realized
13:26:41 <clokep_work> (I don't think so, at least...)
13:27:00 <Fallen> clokep_work: you can use add_task in run_tests() too
13:27:21 <Fallen> There is probably an nsIDOM something to get crypto
13:27:34 <-- hadi has quit (Quit: Instantbird 1.6a1pre -- http://www.instantbird.com)
13:28:57 <clokep_work> Fallen: Yep, that's where I'm looking...although it seems to be webidl. :-S
13:29:09 <Fallen> damn webidl :-)
13:29:47 <clokep_work> Any idea if there's a way to build an instance then?
13:35:06 <Fallen> clokep_work: does nsICryptoHash do what you need?
13:35:33 <Fallen> alternatively, try using the hidden window
13:35:39 <aleth> Really there should be a Services.crypto...
13:36:29 <Fallen> clokep_work: or better yet nsICryptoHMAC
13:39:31 <clokep_work> Fallen: I apparently don't need the HMAC version...
13:39:54 <clokep_work> But maybe nsICryptoHash...
13:40:00 * clokep_work isn't entirely sure of the difference. :-[
13:43:03 <Fallen> clokep_work: this is how we use it: http://mxr.mozilla.org/comm-central/source/calendar/base/src/calAttachment.js#35
13:43:40 <clokep_work> Fallen: I have code to use it.
13:43:47 <Fallen> ok :)
13:43:48 <clokep_work> I just don't understand what I'm trying to do.
13:43:51 <Fallen> ha
13:44:31 <clokep_work> (I don't understand the difference between an HMAC and a "Hash" and I don't understand why none of the Mozilla docs say which one is being used...)
13:51:01 <Fallen> clokep_work: http://benlog.com/2008/06/19/dont-hash-secrets/ :)
13:52:24 <clokep_work> Meh...people need to be more specific when they talk then. ;) But thanks.
14:32:42 <-- DGMurdockIII has quit (Quit: Leaving)
14:41:27 <-- aleth has quit (Ping timeout: 121 seconds)
14:42:39 --> gerard-majax has joined #instantbird
14:42:49 <clokep_work> Hmm...can |(0x0E79A9C1 * <some number, e.g. 9091942890>) % 0x7FFFFFFF| be made to not overflow an int32?
14:43:52 --> aleth has joined #instantbird
14:43:52 * ChanServ sets mode +o aleth 
14:45:55 <Fallen> please mention my nick for a test
14:47:14 <aleth> Fallen
14:47:22 <Fallen> works :-)
14:48:12 <Fallen> http://mozilla.kewis.ch/ib-notify-1.0.xpi
14:48:40 <aleth> :-)
14:48:43 <Fallen> needs lots of options, just started this 20 minutes ago
14:48:44 <aleth> what does that do?
14:48:55 <Fallen> try it, its restartless :)
14:49:32 --> Hoony has joined #instantbird
14:50:31 <aleth> OK
14:50:36 <Fallen> aleth: how is this?
14:51:10 <aleth> It works, but what's the difference to the built in notifications?
14:51:16 <Fallen> I don't get them?
14:51:45 <aleth> Preferences->General doesn't work for you? Maybe there is a bug...
14:51:47 <Fallen> oh I think I did it again, write an extension for an option I don't have enabled
14:52:05 <aleth> But what I like in your notification is that it mentions the channel name etc
14:52:25 <aleth> I'm not sure the default ones do that
14:52:51 <Fallen> well, it was a good exercise
14:53:21 <aleth> and you can still add lots of options :D
14:58:11 <Fallen> also mine notifies when you have the window active, but not the tab
14:58:33 <aleth> That sounds like a good idea
15:24:18 <-- aleth has quit (Ping timeout: 121 seconds)
15:28:00 * Fallen is now known as Fallen|away 
15:31:14 --> aleth has joined #instantbird
15:31:14 * ChanServ sets mode +o aleth 
15:38:45 <aleth> VS2010 will likely be unsupported at the next uplift.
15:42:04 <-- gerard-majax has quit (Ping timeout: 121 seconds)
15:51:04 <aleth> clokep_work: To use webcrypto, try Cu.import("resource:///modules/hiddenWindow.jsm"); and then getHiddenWindow or getHiddenHTMLWindow ? We already do this elsewhere in /chat
15:51:51 <clokep_work> aleth: I don't need it.
15:51:52 <clokep_work> I think.
15:51:54 <clokep_work> But thanks.
15:56:47 <clokep_work> aleth: Current issue is dealing with Big Integer math....
15:56:49 <clokep_work> Joy joy joy
15:57:38 <aleth> free headaches :-(
16:00:02 <aleth> that kind of thing is not what JS was built for...
16:17:08 <-- mpmc has quit (Quit: Instantbird 1.6a1pre -- http://www.instantbird.com)
16:22:11 --> Defman has joined #instantbird
16:40:15 <-- Hoony has quit (Client exited)
16:56:12 --> mconley has joined #instantbird
17:03:59 --> BillBinkley1 has joined #instantbird
17:05:52 <-- BillBinkley has quit (Ping timeout: 121 seconds)
17:06:06 --> BillBinkley has joined #instantbird
17:08:01 <-- BillBinkley1 has quit (Ping timeout: 121 seconds)
17:08:13 --> BillBinkley1 has joined #instantbird
17:08:21 <-- mconley has quit (Connection closed)
17:10:08 <-- BillBinkley has quit (Ping timeout: 121 seconds)
17:18:13 <-- nhnt11 has quit (Ping timeout: 121 seconds)
17:45:49 <clokep_work> I think I took care of the big int stuff ove rlunch...but still getting a different answer so...not sure what the deal is.
17:46:30 <clokep_work> Might need to get a dump of intermediate steps. :( Which is a pain.
17:48:22 --> arlolra has joined #instantbird
17:49:24 <aleth> Some overflow issue because JS uses floats?
17:50:53 <clokep_work> ...didn't I just say I took care of the big int stuff? :P
17:52:01 <aleth> alright then ;)
17:53:32 <aleth> If it's any consolation, I was just looking at the first part of bug 1105889 and I'm stumped
17:53:34 <instantbot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=1105889 nor, --, ---, nobody, NEW, Libpurple proxy preferences are broken
17:59:21 <clokep_work> :-\
18:01:14 --> Mook_as has joined #instantbird
18:21:22 * Fallen|away is now known as Fallen 
18:55:53 <-- Mook_as has quit (Client exited)
18:56:14 --> Mook_as has joined #instantbird
19:38:38 --> mconley has joined #instantbird
19:40:55 <-- mconley has quit (Connection closed)
19:41:00 --> mconley_ has joined #instantbird
19:43:11 --> mconley has joined #instantbird
19:46:14 <-- mconley_ has quit (Ping timeout: 121 seconds)
19:49:15 <clokep_work> arlolra: So I never responded because it seemed like you got enough feedback. :)
19:49:26 <clokep_work> Didn't have much to say beyond what those guys replied.
19:50:05 <arlolra> no problem. i think we got a few good suggestions out of it
19:53:05 <clokep_work> aleth: I threw some stuff on https://thunderbird.etherpad.mozilla.org/status-meeting-minute-taking, not sure if other stuff is used tlaking about.
19:56:30 --> EionRobb has joined #instantbird
20:01:16 <clokep_work> Hello EionRobb!
20:01:23 <EionRobb> oh, not you again
20:01:24 <clokep_work> So I wrote the algorithm and it doesn't work. :P
20:01:32 <clokep_work> But at _work so can't debug it yet...
20:01:32 <EionRobb> \o/
20:01:39 <EionRobb> can i see it?
20:04:37 <clokep_work> Umm...it's a total mess right now.
20:07:21 <EionRobb> its a bit runnier than you'd like it sier
20:10:18 <clokep_work> EionRobb: http://pastebin.instantbird.com/1078817
20:12:33 --> Mook_astb has joined #instantbird
20:13:10 * Mook_astb is now known as Mook_a1 
20:16:10 <-- Mook_as has quit (Ping timeout: 121 seconds)
20:16:14 <clokep_work> EionRobb: If you can find the bug I'll get you a burrito. :)
20:16:31 <EionRobb> now you have my attention
20:18:07 <EionRobb> what's divmod() do?  dividing?
20:18:15 <aleth> Why are you rolling your own crypto?
20:18:36 <aleth> Is it some bizarre MS variant of SHA?
20:18:55 <EionRobb> its not crypto, its just a shared hash
20:19:07 <aleth> still...
20:19:20 <EionRobb> he's copying https://code.google.com/p/skype4pidgin/source/browse/trunk/skypeweb/skypeweb_util.c#94
20:19:49 <EionRobb> clokep_work: you're probably better off with using .xor()
20:20:01 <aleth> EionRobb: You're right, I was reading it incorrectly
20:20:27 <aleth> EionRobb: He's using divmod so the remainder is stored
20:20:30 <clokep_work> EionRobb: Hmm...I could probably xor, you're right.
20:20:41 <clokep_work> EionRobb: The code you have uses % so ...
20:20:55 <clokep_work> aleth: It's not "rolling your own"...I'm not sure what you're saying?
20:21:08 <aleth> clokep_work: ignore that, I totally misread the code
20:21:24 <EionRobb> oh true, it does do that
20:21:40 <EionRobb> I keep getting flashbacks of the 'other' js code that I saw
20:22:55 <clokep_work> |long long|s are fun in JS...
20:23:54 <EionRobb> is view = new DataView(sha256Buffer, 0, 16); just taking the first 16 chars of the sha256Buffer?
20:24:10 <EionRobb> or is it different once its an array buffer?
20:24:52 <clokep_work> EionRobb: It's a 'view' to the first 16.
20:24:59 <clokep_work> So modifications to that change the underlying ArrayBuffer.
20:26:55 <aleth> That's a strange algorithm. How did you figure that one out EionRobb?
20:27:47 <EionRobb> did the function comment not give it away? :)
20:28:20 --> DGMurdockIII has joined #instantbird
20:29:21 <aleth> EionRobb: an educated guess?
20:30:43 <clokep_work> EionRobb: If you had intermediate results of various variables that'd probably help me debug it in a jiff.
20:32:52 --> Bollebib has joined #instantbird
20:36:20 <EionRobb> hmmm
20:37:34 <EionRobb> which variables?
20:38:57 <clokep_work> EionRobb: sha256Hash, sha256Parts and newHashParts at https://code.google.com/p/skype4pidgin/source/browse/trunk/skypeweb/skypeweb_util.c#138; buf at https://code.google.com/p/skype4pidgin/source/browse/trunk/skypeweb/skypeweb_util.c#149; nLow/nHigh at https://code.google.com/p/skype4pidgin/source/browse/trunk/skypeweb/skypeweb_util.c#170 ;)
20:39:12 <clokep_work> And newHash on line 180.
20:40:49 <EionRobb> :O buf is the easiest one though :)
20:47:38 <clokep_work> Haha.
20:47:50 <clokep_work> I'm mostly concerned about sha256Parts/newHashParts and then nLow/nHigh.
20:48:49 <EionRobb> just trying to work out what kind of printf to get for you.  %#x %#x %#x %#x maybe?
20:53:27 <clokep_work> Hmm...
20:53:30 <clokep_work> I forget what # does, one seond.
20:54:20 <clokep_work> Yes, that's likely useful.
20:54:25 <clokep_work> I'm going to leave in a few minutes though.
20:54:32 <EionRobb> got it now
20:55:01 <clokep_work> aleth: WTF....seriously that's how you fixed that? :-S
20:55:08 * clokep_work is looking at https://bugzilla.mozilla.org/page.cgi?id=splinter.html&bug=1105889&attachment=8531304
20:55:12 <EionRobb> for 1416264993  its  sha256Parts: 0x18999354 0x7408aeda 0x14aaf653 0xb5446c1     newHashParts: 0x18999354 0xf408aeda 0x14aaf653 0x8b5446c1    nLow: 0x5f35a06e nHigh: 0x57c98024
20:55:12 <clokep_work> I would *not* have guessed that
20:55:24 <aleth> It took me a while to trace it
20:56:30 <aleth> I hope it's correct (my knowledge of all these NS_INTERFACE macros is a bit rough)
20:56:50 <aleth> Makes sense though.
21:21:10 <clokep_work> Mook_a1: might know.
21:21:17 <clokep_work> (o_O)
21:21:28 * clokep_work wonders how he ended up as a1 instead of as.
21:22:10 <Mook_a1> I accidentally closed the window at some point...
21:22:33 <clokep_work> Mook_a1: Well...newer builds should automatically grab your nick back. :)
21:23:13 <Mook_a1> 2014-12-02? Of course, this is thunderbird...
21:23:26 <clokep_work> Oh...hmm...yeah
21:23:33 <Mook_a1> where is purpleICoreService defined?
21:23:51 <clokep_work> Mook_a1: http://hg.mozilla.org/users/florian_queze.net/purple/
21:24:06 <clokep_work> Hmm...I think your nick should have been gotten back, aleth should know. ;)
21:24:22 <aleth> Mook_a1: It should work in Thunderbird too
21:24:25 * clokep_work goes home
21:24:46 <Mook_a1> Yeah, that patch looks right
21:25:09 <Mook_a1> your comment on line 12 even says it should be there :)
21:26:17 <aleth> Mook_a1: right! That's why it was a regression :-)
21:26:41 <aleth> Thanks for taking a look
21:27:44 <Mook_a1> You already did the work, I'm just nodding my head and pretending I know what I'm doing
21:28:27 <-- clokep_work has quit (Ping timeout: 121 seconds)
21:32:27 --> flo-retina has joined #instantbird
21:32:27 * ChanServ sets mode +qo flo-retina flo-retina 
22:00:07 <-- EionRobb has quit (Connection closed)
22:00:29 --> EionRobb has joined #instantbird
22:11:14 --> nhnt11 has joined #instantbird
22:11:14 * ChanServ sets mode +h nhnt11 
22:13:22 --> Hoony has joined #instantbird
22:14:53 --> Hoony1 has joined #instantbird
22:17:26 <-- Hoony has quit (Ping timeout: 121 seconds)
22:23:41 --> gerard-majax has joined #instantbird
22:30:58 <-- Defman has quit (Quit: My Double-Powered Atomic Reactor has gone to sleep. caaBOOOOM! ZZZzzz....)
22:50:33 <-- flo-retina has quit (Quit: Instantbird 1.6a1pre -- http://www.instantbird.com)
22:54:04 <-- DGMurdockIII has quit (Connection closed)
22:55:20 * Fallen is now known as Fallen|away 
23:11:17 --> clokep has joined #instantbird
23:11:17 * ChanServ sets mode +o clokep 
23:12:56 <-- gerard-majax has quit (Ping timeout: 121 seconds)
23:13:05 <-- mconley has quit (Ping timeout: 121 seconds)
23:17:43 * BinaryOutcast is now known as Tobin 
23:24:51 <clokep> aleth: Those are the same patch..?
23:25:03 <aleth> clokep: oops, my bad
23:25:25 <clokep> :)
23:38:12 --> gerard-majax_ has joined #instantbird
23:38:24 <clokep> aleth: So is that second patch a real issue or?
23:38:51 <aleth> A user complained about it a couple days ago, so yes
23:39:14 <aleth> If the initialization is not done, everything in the proxy dialog breaks and you can't click OK
23:39:54 <clokep> Huh, interesting.
23:40:07 <clokep> aleth: So...that'll happen if you have no purple accounts?
23:40:13 <aleth> Yes
23:40:34 <clokep> That dialogue will likely break hard if you don't have libpurple...
23:40:42 <aleth> That's true.
23:40:48 <clokep> :)
23:40:48 --> mconley_ has joined #instantbird
23:40:49 <aleth> So will the pref pane.
23:40:52 <clokep> Well thanks for fixing a blocker!
23:41:16 <aleth> a bit edge casey, but it's gone.
23:42:18 <aleth> arlolra: Maybe something for the Tor-IB todo list? Remove or disable the libpurple proxy prefs in Preferences->Advanced->Network
23:42:28 <aleth> clokep: thanks for the reviews
23:44:27 <aleth> arlolra: that's if you're not going to ship libpurple
23:47:00 <arlolra> aleth: we most definitely aren't :)
23:47:12 <aleth> :-)
23:48:09 <arlolra> you're on your own
23:49:14 <clokep> aleth: Or get rid of them like flo wants? :-D
23:49:30 <aleth> Get rid of ...?
23:50:12 <aleth> You mean, unify those two proxy panes? I bet there's an old bug for that open somewhere ;)
23:51:58 <clokep> bug 953602
23:52:00 <instantbot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=953602 enh, --, ---, nobody, NEW, Unify proxy settings
23:52:10 <clokep> aleth: Not the panes. Get rid of the purple proxy implementation and use the SAME proxy for both.
23:52:19 * clokep got to https://bugzilla.instantbird.org/show_bug.cgi?id=155 from his awesomebar. :-D
23:52:43 <aleth> wow, does your history never expire? ;)
23:53:19 <aleth> I don't think I have any BIO bits floating around anymore
23:54:18 <clokep> I thought I had removed bugzilla.instantbird.org, but apparently not...
23:54:18 <aleth> clokep: how goes the battle with multiendian long longs?
23:54:27 <clokep> aleth: The battle right now is that I'm cooking fish.
23:54:36 <aleth> eminently more tasty.
23:54:50 <clokep> But what EionRobb gave early should help me verify a lot. :-D
23:58:09 --> aleth-build has joined #instantbird
23:59:04 <-- aleth-build has quit (Quit: Instantbird 1.6a1pre -- http://www.instantbird.com)
23:59:31 <-- gerard-majax_ has quit (Ping timeout: 121 seconds)
23:59:47 <clokep> Sorry I hvaen't gotten to that SASL patch btw.
23:59:52 <clokep> I hvaen't wanted to read the spec. :-[