[00:00:43] <kbr_google> fta: how do i install these into a temporary directory?
[00:01:22] *** cbentzel has quit IRC
[00:01:35] <fta> kbr_google, well, it's a deb, so it's best system wide. it should not hurt your system, unless you already have /usr/lib/chromium-browser/ and /usr/bin/chromium-browser, it's easy to remove cleanly afterwards.
[00:01:56] *** taf2_ has quit IRC
[00:02:15] <fta> extracting the files is not difficult but you'll have to tweak the wrapper
[00:02:19] <kbr_google> fta: ok, i don't have those directories currently. could you please give me a command line to install these packages?
[00:02:40] <fta> kbr_google, sudo dpkg -i chro*.deb
[00:02:56] *** dmaclach_ has quit IRC
[00:03:24] <fta> kbr_google, to remove them: sudo dpkg -r chromium-browser chromium-browser-inspector
[00:07:19] *** ivan^ has left #chromium
[00:09:38] *** hoverbear has joined #chromium
[00:10:18] *** kinnetica has quit IRC
[00:11:03] *** hoverbear has left #chromium
[00:18:08] <kbr_google> fta: your builds run webgl for me with --enable-webgl passed on the command line
[00:18:36] <kbr_google> fta: ps -efww | grep fta should find "/proc/self/exe --type=gpu-process" btw
[00:19:26] <kbr_google> fta: tried several of the demos from http://khronos.org/webgl/wiki/Demo_Repository ; all work
[00:20:07] <kbr_google> i haven't done testing on a wide range of hardware, so it's possible our initialization sequence needs to be made more robust
[00:20:48] <fta> kbr_google, hm, so it's something in lucid then.
[00:21:36] <fta> kbr_google, oh, i have another chromium running with a different profile
[00:23:16] <fta> kbr_google, it seems to work fine when i just have one. weird. it means --user-data-dir= & --disk-cache-dir= are not enough start an isolated browser
[00:23:21] <fta> +to
[00:24:23] <jamesr> hmmm
[00:24:29] <jamesr> since the singleton lives in there
[00:24:41] <fta> kbr_google, thanks anyway. sorry for the troubles
[00:25:18] <fta> jamesr, i thought so but apparently not for everything
[00:26:05] <kbr_google> fta: i just ran the following from /usr/lib/chromium-browser
[00:26:33] <kbr_google> ./chromium-browser --enable-webgl --user-data-dir=/tmp/data-dir-1 --disk-cache-dir=/tmp/cache-dir-1
[00:26:43] <kbr_google> ./chromium-browser --enable-webgl --user-data-dir=/tmp/data-dir-2 --disk-cache-dir=/tmp/cache-dir-2
[00:26:47] <kbr_google> in two different shell windows
[00:26:54] <kbr_google> san angeles demo runs fine in both
[00:27:20] <fta> kbr_google, no, try the 1st one without --enable-webgl, then the 2nd one with it
[00:28:19] <fta> that was my case, i had my regular browser (without --enable-webgl) that i didn't want to restart, so i started another instance with a new profile & --enable-webgl and it didn't work
[00:28:22] <kbr_google> fta: first instance: no webgl; second instance: webgl works fine
[00:28:48] <fta> weird
[00:29:21] <fta> ok, it's not a common use case anyway
[00:29:48] <kbr_google> fta: also works for me if the first instance has no command line options and the second instance has all three command line options
[00:31:40] <fta> what's preventing webgl to be enabled by default?
[00:32:09] *** dmaclach has joined #chromium
[00:33:50] <kbr_google> fta: several things. compliance with the GLSL ES shading language (the single biggest issue right now); unimplemented portions of the webgl spec; spec compliance in general
[00:34:39] <jamesr> iow it's not done yet
[00:34:44] <kbr_google> fta: ping me directly if you want to talk more
[00:35:11] <Kaosevil> https://chrome.google.com/extensions/detail/kpjhnfgfocpnmocpjdafgpkimhejfnag
[00:35:50] <fta> kbr_google, ok, thanks. good enough for me. i often get requests from users about this so now i know what to answer :)
[00:36:01] <kbr_google> fta: you're welcome
[00:36:12] <Kaosevil> hi guys
[00:36:14] *** tonikitoo has joined #chromium
[00:36:24] <Kaosevil> try my new plugin
[00:36:32] <Kaosevil> and send me a feedback
[00:36:45] <Kaosevil> and comments about upgrade
[00:37:24] *** glotov has quit IRC
[00:40:16] *** cleary has joined #chromium
[00:44:11] <stuartmorgan> Kaosevil: Chromium developers don't test every extension
[00:45:37] *** sshc has joined #chromium
[00:47:53] *** jrmuizel has quit IRC
[00:50:01] *** guests98798734 has joined #chromium
[00:50:29] *** guests98798734 has left #chromium
[00:53:25] *** glotov has joined #chromium
[00:54:28] *** dmaclach has quit IRC
[00:55:46] <oshima> sheriffs, troopers: i'd like to stop and use chromeos valgrind bot (1) to troubleshoot linkage error.
[00:55:59] <oshima> ok?
[00:56:05] <asargent> oshima: sounds good to me, I was just wondering who to ask about that
[00:56:29] <oshima> asargent: thanks.
[00:56:50] <oshima> asargent: i landed a fix last week, but seems like it's not enough.
[00:57:06] <oshima> stopping it now
[01:04:40] *** rsesek has quit IRC
[01:05:57] <willchan> anyone know why ENABLE_GPU is set in common.gypi?  it looks like only the chrome directory uses it.
[01:07:30] *** markmentovai has quit IRC
[01:08:39] <jamesr> kbr_google might
[01:08:50] *** chebfun has joined #chromium
[01:09:04] *** chebfun has left #chromium
[01:09:20] <kbr_google> willchan, jamesr: i'll ask apatrick
[01:11:06] <kbr_google> willchan: apatrick thinks that originally src/gpu/ used that #define, but not any more, so the setting could move
[01:11:30] <kbr_google> is the thinking to define that with as small scope as possible?
[01:12:45] <willchan> it just seems unclean to me to leak preprocessor defines to other targets
[01:13:05] <willchan> ENABLE_GPU is unlikely to collide, but it's more the principle
[01:13:34] <kbr_google> willchan: it's fine. if you don't mind making the change, please go ahead and make apatrick the reviewer, otherwise file a bug and assign it to me or him
[01:14:23] <willchan> kbr_google: ok, i'll file a bug.  thanks.
[01:14:41] *** icqn has joined #chromium
[01:16:55] *** tedoc2000 has quit IRC
[01:17:42] *** tedoc2000 has joined #chromium
[01:17:43] *** ibqn has quit IRC
[01:21:59] *** sundiamonde has quit IRC
[01:23:12] <fta> speaking of --user-data-dir, this just landed on planet.ubuntu a few secs ago: http://penguindroppings.wordpress.com/2010/05/17/browser-profiles-in-chromium/
[01:24:59] <thakis> fta: :-)
[01:25:01] <thakis> evmar: ^
[01:26:56] <jamesr> cute
[01:32:44] *** dale1v has joined #chromium
[01:32:44] *** dale1v has joined #chromium
[01:37:15] *** b3nn has quit IRC
[01:38:47] *** skerner has quit IRC
[01:46:39] <stuartmorgan> Wow, UI test cookies don't work at all the way I thought. Is anyone here familiar with WaitUntilCookieValue?
[01:47:20] <shoe> how can I find out if chrome's <video> element supports a rtsp:// scheme?
[01:48:34] *** Killarny has quit IRC
[01:48:56] *** sundiamonde has joined #chromium
[01:50:35] <atwilson> Looks like one of the canary bots has a problem: http://build.chromium.org/buildbot/waterfall.fyi/builders/Webkit.org%20Builder/builds/4457
[01:50:45] <atwilson> Anyone have any idea of what I can do to fix it?
[01:51:20] <atwilson> Looks like the XP Perf canary hasn't run in a week - is that a known issue?: http://build.chromium.org/buildbot/waterfall.fyi/builders/XP%20Perf%20(webkit.org)
[01:52:23] *** dglazkov has quit IRC
[01:53:11] <stuartmorgan> ananta: ping
[01:54:13] *** Kaosevil has quit IRC
[01:56:28] *** tomato has joined #chromium
[01:57:07] <tomato> hey, anybody know of a workaround for onbeforeunload not firing?
[01:57:21] *** hbono has joined #chromium
[01:58:01] *** lukas___ has joined #chromium
[01:58:10] <tomato> I've got to pass something to a flash window to unregister some stuff on the server when the user exits. Works fine under IE and Firefox
[01:58:36] *** dale1v is now known as dale1v[away]
[01:58:38] <tomato> Maybe I'll do a big button "Press this to close". That's good design, right?
[01:59:21] *** sbyer has left #chromium
[02:00:02] <tomato> https://bugs.webkit.org/show_bug.cgi?id=21699 this is the issue
[02:01:28] *** suzhe has quit IRC
[02:02:33] *** tessamac has joined #chromium
[02:07:15] *** masterov has quit IRC
[02:08:25] *** chaser has quit IRC
[02:11:03] *** chaser has joined #chromium
[02:11:06] *** bent-mozilla has quit IRC
[02:13:08] *** icqn has left #chromium
[02:13:17] <tomato> It's a pretty important, very quick task, just firing off something to a php page and then onto the database. Important for the other users' experiences. Doesn't even wait for a response. Not doing some hey don't leave my porn site we've still got viruses for you to install type thing here
[02:14:16] <stuartmorgan> tomato: This seems like a question better suited to a web developer forum
[02:14:48] *** fqian has quit IRC
[02:15:36] *** tedoc2000 has quit IRC
[02:15:48] <tomato> yeah, probably
[02:15:51] *** tomato has quit IRC
[02:16:40] *** Kustnamenkloate has quit IRC
[02:17:51] *** Kustnamenkloate has joined #chromium
[02:18:18] *** bent-mozilla has joined #chromium
[02:18:50] *** RT|Chatzilla has joined #chromium
[02:20:55] *** AaronMT has quit IRC
[02:23:12] *** eseidelDesk has quit IRC
[02:23:31] *** sebmarkbage has quit IRC
[02:23:47] *** lianj_ has joined #chromium
[02:24:51] *** vladbph has quit IRC
[02:25:26] *** cying has quit IRC
[02:26:13] *** stuartmorgan has quit IRC
[02:26:17] *** inferno-sec has quit IRC
[02:26:36] <asargent> nirnimesh: are you looking at the unit_tests failure?
[02:26:52] *** eseidelDesk has joined #chromium
[02:26:52] <nirnimesh> asargent: wasn't. looking now
[02:27:03] *** lianj has quit IRC
[02:27:06] *** cedricv has quit IRC
[02:29:46] *** victorw has left #chromium
[02:31:06] *** Kustnamenkloate has quit IRC
[02:31:59] <nirnimesh> asargent: that failing test has nothing to do with my CL
[02:32:15] <asargent> nirnimesh: ok, we can reopen and see if it was just flakiness
[02:32:27] <nirnimesh> yes
[02:32:57] <asargent> ok, repoened
[02:33:01] *** dave_levin_ has joined #chromium
[02:33:07] <asargent> I need to head out, but glotov will still be around for a little while I think. Have a good evening all.
[02:33:14] <nirnimesh> asargent: the vista failure was a hang. the linux failure has nothing to do with my test.
[02:33:15] *** asargent has left #chromium
[02:33:44] <glotov> ok
[02:34:42] *** taf2_ has joined #chromium
[02:36:07] *** chaser has quit IRC
[02:36:38] *** chaser has joined #chromium
[02:38:28] *** trungl is now known as trungl_away
[02:38:52] *** bradnelson has joined #chromium
[02:39:04] *** masterov has joined #chromium
[02:39:28] *** jrmuizel has joined #chromium
[02:40:34] *** bradnelson has quit IRC
[02:42:12] *** masterov has quit IRC
[02:42:30] *** masterov has joined #chromium
[02:43:15] *** trungl_mbp has joined #chromium
[02:44:56] *** Venom_X has joined #chromium
[02:45:30] *** Venom_X has quit IRC
[02:48:03] *** JayM has quit IRC
[02:50:00] *** lukas____ has joined #chromium
[02:52:43] *** yutak_home has quit IRC
[02:52:56] *** tessamac has left #chromium
[02:53:33] *** lukas___ has quit IRC
[02:53:51] *** masterov has quit IRC
[02:54:54] *** lukas____ has quit IRC
[03:00:39] *** trungl_mbp has quit IRC
[03:01:10] *** sundiamonde_ has joined #chromium
[03:03:00] *** sundiamonde_ has quit IRC
[03:05:02] *** sundiamonde has quit IRC
[03:05:16] *** Bleak has joined #chromium
[03:07:52] *** sundiamonde has joined #chromium
[03:14:47] *** JayM has joined #chromium
[03:15:01] *** leeight has joined #chromium
[03:15:05] *** leeight has left #chromium
[03:19:41] *** SleepingCoyote is now known as coyo
[03:21:13] <glotov> leaving now, good bye all!
[03:21:59] *** JayM has quit IRC
[03:27:08] <willchan> evmar: yt?
[03:27:59] *** JayM has joined #chromium
[03:28:46] *** thakis is now known as thakis_afk
[03:28:54] *** kerz has quit IRC
[03:34:37] *** arv has quit IRC
[03:38:58] *** dglazkov has joined #chromium
[03:38:58] *** ChanServ sets mode: +v dglazkov
[03:40:08] *** TabAtkins_ has quit IRC
[03:40:19] *** cedricv has joined #chromium
[03:49:02] *** csilv has left #chromium
[03:51:45] *** eseidel has joined #chromium
[03:52:47] *** mpcomplete has quit IRC
[03:53:27] *** coyo has quit IRC
[03:54:21] *** johnny_g has quit IRC
[03:54:48] *** abarth has joined #chromium
[03:54:48] *** ChanServ sets mode: +v abarth
[03:56:47] *** bent-mozilla has quit IRC
[03:57:27] *** coyo has joined #chromium
[03:58:24] *** eseidel_ has joined #chromium
[03:59:09] *** morrita has joined #chromium
[04:01:26] *** morrita has quit IRC
[04:01:34] *** morrita has joined #chromium
[04:02:26] *** eseidel has quit IRC
[04:02:26] *** eseidel_ is now known as eseidel
[04:05:20] *** trungl_mbp has joined #chromium
[04:06:27] *** vbabiy has joined #chromium
[04:08:24] *** rafaelw1 has left #chromium
[04:09:04] *** coyo has quit IRC
[04:12:17] *** coyo has joined #chromium
[04:19:42] *** jamesr has quit IRC
[04:19:56] *** tonyg-cr has quit IRC
[04:25:30] <shoe> is there _any_ format of video that chromium on linux can live-stream using <video>?
[04:27:26] *** dpranke has quit IRC
[04:28:06] *** newone has joined #chromium
[04:31:30] *** aaron_liuj has joined #chromium
[04:32:19] *** newone has quit IRC
[04:33:33] *** shoe` has joined #chromium
[04:33:37] *** shoe has quit IRC
[04:33:58] <aaron_liuj> where i can get the doc of build chrome
[04:34:11] <aaron_liuj> where i can get the whole doc of build chrome
[04:36:06] *** zer0her0 has quit IRC
[04:39:19] *** happygrue has quit IRC
[04:39:36] *** happygrue has joined #chromium
[04:40:46] *** jamesr has joined #chromium
[04:42:40] *** ishermandom has joined #chromium
[04:50:06] *** roc_ has joined #chromium
[04:50:47] *** inferno-sec has joined #chromium
[04:51:12] *** MikeSmithW3C has joined #chromium
[04:51:27] *** inferno-sec_ has joined #chromium
[04:51:59] *** roc has quit IRC
[04:52:00] *** roc_ is now known as roc
[04:53:40] *** dmaclach has joined #chromium
[04:55:01] *** inferno-sec has quit IRC
[04:56:04] *** dmaclach_ has joined #chromium
[04:56:05] *** dmaclach has quit IRC
[04:56:05] *** dmaclach_ is now known as dmaclach
[04:56:43] *** mrossetti has quit IRC
[05:05:11] *** jamesr has quit IRC
[05:05:20] *** _rs has quit IRC
[05:06:46] *** atwilson_ has joined #chromium
[05:07:02] *** vbabiy has quit IRC
[05:10:32] *** gospch has joined #chromium
[05:10:32] *** shoe` has quit IRC
[05:10:37] *** shoe` has joined #chromium
[05:11:04] *** sshc has quit IRC
[05:16:03] *** dmaclach_ has joined #chromium
[05:18:29] *** dmaclach has quit IRC
[05:18:29] *** dmaclach_ is now known as dmaclach
[05:19:13] *** rohitrao has quit IRC
[05:21:44] *** Bleak has quit IRC
[05:37:33] *** taf2_ has quit IRC
[06:07:56] *** occamshatchet has joined #chromium
[06:08:54] *** tyoshino has quit IRC
[06:13:41] *** michaeln has quit IRC
[06:14:15] *** tyoshino has joined #chromium
[06:19:10] *** yuzo has quit IRC
[06:24:25] *** shepazu has quit IRC
[06:24:38] *** HighBit has quit IRC
[06:26:05] *** thakis has joined #chromium
[06:26:05] *** ChanServ sets mode: +v thakis
[06:28:00] *** sshc has joined #chromium
[06:29:30] *** shepazu has joined #chromium
[06:29:46] <akalin> anyone around?
[06:29:57] <akalin> preferably familiar with net/ )
[06:29:59] <akalin> :)
[06:33:57] *** hbono has quit IRC
[06:38:06] *** ishermandom has quit IRC
[06:38:47] *** sundiamonde has quit IRC
[06:39:41] *** sundiamonde has joined #chromium
[06:40:31] *** thakis has quit IRC
[06:43:42] *** sundiamonde has quit IRC
[06:44:04] *** rubenbb1 has joined #chromium
[06:45:59] <aaron_liuj> make_chroot: i686 is not supported as a host machine architecture.
[06:46:00] *** rubenbb has quit IRC
[06:54:29] *** bweinstein has quit IRC
[06:55:46] *** hbono has joined #chromium
[06:55:49] *** rubenbb has joined #chromium
[06:56:44] *** johnny_g has joined #chromium
[06:57:52] *** gospch has quit IRC
[06:57:59] *** rubenbb1 has quit IRC
[06:58:17] *** jrmuizel has quit IRC
[06:58:43] *** johnny_g_ has joined #chromium
[07:03:07] *** johnny_g has quit IRC
[07:03:08] *** johnny_g_ is now known as johnny_g
[07:13:33] *** atwilson_ has quit IRC
[07:15:22] *** vipw has left #chromium
[07:22:28] *** bers has joined #chromium
[07:23:45] *** dglazkov has quit IRC
[07:27:48] *** rubenbb has quit IRC
[07:28:50] *** rubenbb has joined #chromium
[07:36:10] *** yuzo has joined #chromium
[07:36:23] *** hayato has quit IRC
[07:37:36] *** hayato has joined #chromium
[07:45:34] *** atwilson_ has joined #chromium
[07:45:58] *** roc has quit IRC
[07:45:59] *** shepazu has quit IRC
[07:46:25] *** shepazu has joined #chromium
[07:55:52] *** trungl_mbp_ has joined #chromium
[07:56:49] *** gospch has joined #chromium
[07:59:49] *** trungl_mbp has quit IRC
[07:59:49] *** trungl_mbp_ is now known as trungl_mbp
[08:06:55] *** Adys has quit IRC
[08:15:22] *** Adys has joined #chromium
[08:19:22] *** drusepth has quit IRC
[08:29:00] *** trungl_mbp_ has joined #chromium
[08:29:16] *** wers has joined #chromium
[08:32:40] *** trungl_mbp has quit IRC
[08:32:41] *** trungl_mbp_ is now known as trungl_mbp
[08:35:04] *** gospch has quit IRC
[08:36:51] *** bevc_work has joined #chromium
[08:44:25] *** roc has joined #chromium
[08:45:32] *** HighBit has joined #chromium
[08:45:41] *** dave_levin_ has quit IRC
[08:46:27] *** stalled has quit IRC
[08:47:06] *** Forsaken|Laptop has joined #chromium
[08:47:47] *** bers has quit IRC
[08:55:45] *** stalled has joined #chromium
[09:08:06] *** loislo has joined #chromium
[09:09:22] *** johnny_g has quit IRC
[09:12:04] *** General13372 has joined #chromium
[09:14:36] *** lnostdal_ has joined #chromium
[09:15:53] *** General1337 has quit IRC
[09:16:51] *** lnostdal has quit IRC
[09:17:22] *** ibrar has joined #chromium
[09:23:27] *** teflon has joined #chromium
[09:23:29] <eroman> build failure seems like flake
[09:23:33] <eroman> i will try re-running the vista tests
[09:26:38] *** bevc_work has quit IRC
[09:29:33] *** trungl_mbp has quit IRC
[09:29:40] *** wers has quit IRC
[09:29:45] *** eseidel has left #chromium
[09:30:34] *** Yonn has joined #chromium
[09:30:59] *** ibrar has quit IRC
[09:34:16] *** dale1v[away] is now known as dale1v
[09:35:25] *** Beetny has joined #chromium
[09:39:51] *** wers has joined #chromium
[09:40:27] *** GeekShadow has joined #chromium
[09:43:05] <eroman> ya looks ok, reopened
[09:47:27] *** adamkex has quit IRC
[09:51:21] *** eseidel has joined #chromium
[16:11:19] *** echelog-1 has joined #chromium
[16:11:24] *** thakis has joined #chromium
[16:11:24] *** ChanServ sets mode: +v thakis
[16:12:56] *** miketaylr has joined #chromium
[16:18:15] *** Bleak has joined #chromium
[16:21:35] *** trungl_mbp has quit IRC
[16:27:18] *** dmaclach has quit IRC
[16:27:40] *** Zuccace has quit IRC
[16:27:44] *** stuartmorgan has joined #chromium
[16:27:44] *** ChanServ sets mode: +v stuartmorgan
[16:27:53] *** Zucca has joined #chromium
[16:28:54] *** _rs has joined #chromium
[16:28:58] *** lukas___ has joined #chromium
[16:29:29] *** Yon has joined #chromium
[16:29:34] *** Malmis has joined #chromium
[16:30:09] *** tittiatcoke has joined #chromium
[16:30:53] *** cbentzel has quit IRC
[16:31:22] *** erikkay has joined #chromium
[16:33:01] *** Russ has joined #chromium
[16:33:33] *** jcastro has joined #chromium
[16:33:44] *** Caleb has quit IRC
[16:35:42] *** Caleb has joined #chromium
[16:36:47] *** trungl_mbp has joined #chromium
[16:38:22] *** thakis has quit IRC
[16:38:30] *** trungl_mbp is now known as trungl
[16:38:32] *** ChanServ sets mode: +v trungl
[16:43:21] *** tedoc2000 has joined #chromium
[16:46:46] <trungl> 'morning, Chromium.
[16:46:53] <rsesek> good morning, trungl
[16:53:19] <malavv> morning
[16:53:57] *** Velmont has left #chromium
[16:57:23] * BUGabundo_remote wave 0/
[16:57:35] *** tonyg-cr has joined #chromium
[16:57:35] *** ChanServ sets mode: +v tonyg-cr
[16:58:58] *** Venom_X has joined #chromium
[16:59:35] *** craigsch has joined #chromium
[17:00:57] *** dglazkov has quit IRC
[17:01:16] *** cbentzel has joined #chromium
[17:09:12] *** agl has joined #chromium
[17:09:12] *** ChanServ sets mode: +v agl
[17:10:40] <rsesek> mirandac: CL to you
[17:11:22] *** ivan has quit IRC
[17:11:58] *** tonyg-cr has quit IRC
[17:12:18] *** Forsaken|Laptop has quit IRC
[17:13:50] *** Kaosevil has quit IRC
[17:16:10] <rsesek> I've built in Debug mode and I'm hitting a CHECK(), but my stack trace isn't symbolicated (it's all just ChromeMain) ? ideas?
[17:16:45] <trungl> are you running on 10.5?
[17:16:49] <rsesek> yes
[17:17:09] <rsesek> Goog MacPro
[17:17:09] <trungl> yeah, stack traces don't work on 10.5
[17:17:13] <rsesek> :(
[17:17:24] <trungl> "I have a patch for that"
[17:17:30] <trungl> (if you really need a stack trace)
[17:17:40] <rsesek> yes plz
[17:17:59] <trungl> http://codereview.chromium.org/165224/show
[17:18:04] <trungl> or install 10.6
[17:18:24] <rsesek> I think 10.6 may the answer because I hate Xcode 3.1
[17:19:08] <trungl> I just generically hate xcode, so it's all the same to me
[17:19:10] <trungl> well, almost
[17:19:23] <trungl> I hate Visual Studio even more though
[17:19:38] <rsesek> trungl: why hasn't this been checked in?
[17:20:09] <trungl> rsesek: 'cause a) it's fixed in 10.6, b) it's a lot of complicated code to review and maintain
[17:20:25] <rsesek> okay. hopefully I can get on 10.6 and forget about all of this :)
[17:20:45] <trungl> that's my strategy, except on my laptop
[17:20:47] <trungl> :(
[17:20:52] <trungl> which is still stuck on 10.5
[17:21:01] <rsesek> did they edge you out with the sys reqs?
[17:21:24] <trungl> no, it's just a google-image laptop
[17:22:43] *** Zucca has quit IRC
[17:24:19] *** rsesek has quit IRC
[17:24:21] *** rsesek1 has joined #chromium
[17:24:22] *** rickspencer3 has joined #chromium
[17:24:33] *** rsesek1 is now known as rsesek
[17:24:41] <evmar> wow, trung, you are a madman
[17:24:57] <rickspencer3> evmar, hiya
[17:25:11] <evmar> hi rick!
[17:25:31] <rickspencer3> evmar, you're at a hack fest, right?
[17:25:32] *** Zucca has joined #chromium
[17:26:02] <evmar> no, it fell through due to behdad having visa issues
[17:26:08] <rickspencer3> :/
[17:26:17] <evmar> it's ok, i have lots of other work to do
[17:26:25] *** rsesek has left #chromium
[17:26:25] <rickspencer3> hehe
[17:26:26] <rickspencer3> I bet
[17:26:29] *** rsesek has joined #chromium
[17:27:54] *** lukas___ has quit IRC
[17:29:38] *** tonyg-cr has joined #chromium
[17:29:39] *** ChanServ sets mode: +v tonyg-cr
[17:31:19] *** steven_t has joined #chromium
[17:31:20] <steven_t> hello
[17:31:26] *** fahadsadah has quit IRC
[17:31:28] <steven_t> any devs in here?
[17:32:10] *** Zaba has quit IRC
[17:32:21] <pinkerton> yes
[17:32:26] <steven_t> i found the issue link for "cmd+ctrl+d not working in Mac OS X" and added a comment, and i would love to see this implemented
[17:32:29] *** d1b has quit IRC
[17:32:29] *** sidh has quit IRC
[17:32:29] <steven_t> http://code.google.com/p/chromium/issues/detail?id=17951
[17:32:31] *** loislo has quit IRC
[17:32:32] <pinkerton> yes, i just sent you email
[17:32:35] *** ku0522a___ has joined #chromium
[17:32:36] <steven_t> awesome
[17:32:46] <pinkerton> i think you underestimate the difficulty
[17:32:50] *** stalled has quit IRC
[17:32:50] *** paul_irish has quit IRC
[17:32:50] <pinkerton> not knowing our architecture
[17:33:08] <steven_t> i see.
[17:33:11] *** Russ has quit IRC
[17:33:11] *** RobWork has quit IRC
[17:33:16] <pinkerton> how do you avoid the synchronous call to the renderer to respond to the protocol?
[17:33:33] <steven_t> i meant to say that it is not *as* difficult as implementing full accessibility
[17:33:53] <pinkerton> you're right
[17:33:56] *** shenki__ has joined #chromium
[17:33:58] <steven_t> but im not at all familiar with your architecture so i cant help
[17:34:00] *** Russ has joined #chromium
[17:34:02] <pinkerton> :)
[17:34:21] <evmar> we have a similar issue with middle-mouse paste on linux
[17:34:33] <pinkerton> getting this working is probably 90% of the way to getting full accessibility working
[17:34:43] <steven_t> just wanted to give my two cents regarding a potential solution path
[17:34:44] <evmar> do you have to select the word for tihs key to work?  or is it just whatever is below the mouse?
[17:34:45] *** lalalalilili has quit IRC
[17:34:56] *** dglazkov has joined #chromium
[17:34:56] *** ChanServ sets mode: +v dglazkov
[17:35:18] <trungl> whatever's below the mouse, I believe
[17:35:37] *** lalalalilili has joined #chromium
[17:35:40] <evmar> oh, that is harder then
[17:35:51] <evmar> i might consider some sort of sync IPC with timeout
[17:36:20] <trungl> yeah
[17:36:21] *** BUGabundo_remote has quit IRC
[17:36:22] <pinkerton> steven_t: i understand you're trying to be helpful, but the tone of your comment in the bug was a bit condescending, esp when you tell those of us who know what we're talking about  that it's "not a difficult thing to implement"
[17:36:30] <steven_t> ok
[17:36:34] <steven_t> il delete it then
[17:36:49] <pinkerton> there is working going on to cache the contents
[17:37:06] <pinkerton> i think that's almost complete, at which point, it would be easier to get the relevant data async
[17:37:08] <rohitrao> how do we do cmd-E to copy to the pasteboard?
[17:37:22] <steven_t> if its cached its cached
[17:37:25] <steven_t> nothing i can do about that
[17:37:28] <pinkerton> steven_t: if you check the chromium-dev mailing list, there's an ongoing thread about this
[17:37:33] <steven_t> ok
[17:37:37] *** ku0522a__ has quit IRC
[17:37:38] *** chase has quit IRC
[17:37:38] *** tedoc2000_ has joined #chromium
[17:37:40] *** sidh has joined #chromium
[17:37:51] <pinkerton> the pointers that you give to the protocol are helpful to have around
[17:37:52] *** fahadsadah has joined #chromium
[17:37:56] *** tedoc2000 has quit IRC
[17:37:56] *** shenki has quit IRC
[17:37:57] *** chase_ has joined #chromium
[17:38:01] <pinkerton> as is the sample code
[17:38:05] *** tedoc2000_ is now known as tedoc2000
[17:38:11] <pinkerton> but that's not the hard part :)
[17:38:21] *** paul_irish has joined #chromium
[17:38:22] *** iPac has joined #chromium
[17:38:34] <pinkerton> safari will have to solve this same problem in webkit2 when they move their renderer to a separate process
[17:38:36] *** lukas___ has joined #chromium
[17:38:53] <rohitrao> yeah, the hard part is getting the selected text, right?
[17:39:00] *** loislo has joined #chromium
[17:39:32] <trungl> rohitrao: no text selection is required
[17:39:33] <pinkerton> getting it synchronously is, yeah
[17:39:43] <trungl> but in any case
[17:39:48] *** d1b has joined #chromium
[17:39:48] <steven_t> im sorry i came off condescending, although its not the first time, i really suck at communiacationg
[17:39:53] <rohitrao> oh, you can just right-click on a word? crazy
[17:39:55] <trungl> the sync versus async problem is the problem
[17:40:10] <trungl> you can just hover a word and press cmd-ctrl-d
[17:40:11] <steven_t> my intention was to say "hey look this in in fact easier than we thought isnt that cool?? :)"
[17:40:16] <pinkerton> :)
[17:40:16] <rohitrao> did we not solve this for cmd-e?
[17:40:30] <steven_t> anyway nothing left i can do to help i guess
[17:40:40] *** stalled has joined #chromium
[17:40:40] <pinkerton> rohitrao: i think with command-e we are in control over when we put stuff on the pasteboard
[17:40:45] <trungl> rohitrao: I assume cmd-e is done in the same way as cut-and-paste support
[17:40:46] <pinkerton> rohitrao: we're not responding to a protocol
[17:40:52] <pinkerton> right
[17:40:58] *** Zaba has joined #chromium
[17:41:02] <rohitrao> ooh, I get it, ok
[17:41:06] *** ivan has joined #chromium
[17:41:20] <pinkerton> steven_t: like i said, if you clean up your comment to include the sample code and pointers to the right protocol, that would be very helpful to have for reference
[17:41:39] <pinkerton> steven_t: when we get to the point of having all the pieces, it'll all just come together
[17:41:51] <pinkerton> so that's pretty helpful to know
[17:42:05] <steven_t> ill blog about it and then link to my blog in the comments. thatll be better
[17:42:11] <pinkerton> sounds good
[17:42:13] <trungl> if I had a penny for every time someone said that this should be easy ... well, I'd almost have a nickel
[17:42:26] * pinkerton presses his "easy button"
[17:42:27] <rohitrao> trungl: man, it should be easy to get to a dime
[17:42:31] *** arm3n has joined #chromium
[17:42:37] <steven_t> dont spend it all in one hypothetical place
[17:43:06] <pinkerton> steven_t: if you're looking for other ways to contribute, we have a ton of bugs marked HelpWanted that are good first bugs to get started
[17:43:12] <trungl> biab
[17:43:16] <steven_t> i have 0 time as it is
[17:43:19] *** trungl has quit IRC
[17:43:19] <pinkerton> lol :)
[17:43:32] <steven_t> work, family, and secret idea to get rich, all come first
[17:43:44] <pinkerton> :)
[17:43:45] <steven_t> if i get rich, ill contribute all you guys want
[17:43:54] <steven_t> ;)
[17:43:57] <pinkerton> sounds like a plan
[17:44:14] <pinkerton> if i get rich i'm buying an island. you guys can finish the browser w/out me :)
[17:44:24] <steven_t> ugh, blogging about this is gonna take a few hours. doubt ill get around tot hat.
[17:44:29] <steven_t> *tooth hat
[17:44:31] <steven_t> *to that
[17:44:56] <pinkerton> you can just take half of what you posted to the bug and post that, it should be enough
[17:44:58] <steven_t> pinkerton: its gonna be reeeeal boring on that island, trust me
[17:45:40] <pinkerton> i've watched enough reality tv to know how to make an island real not boring :D
[17:46:03] *** Zaba has quit IRC
[17:46:05] <sky___> WorkerTest.WorkerScriptError seems to be consistantly failing on mac dbg for some reason.
[17:46:29] <sky___> It started when Thiago landed his patch, but I don't think it's related (Thiago only touched gtk code).
[17:46:42] <sky___> I'm going to mark flaky.
[17:46:56] *** RobWork has joined #chromium
[17:46:59] <malavv> Sorry to interrupt such nice discussion, Is someone familiar with Chromium Network stack?
[17:47:02] *** bent-mozilla has joined #chromium
[17:47:06] <steven_t> done.
[17:48:26] *** antivirtel has quit IRC
[17:48:34] *** kjalarr has quit IRC
[17:48:43] <malavv> Legit ping for networking guys?
[17:49:16] *** m0 has quit IRC
[17:49:19] <evmar> malavv: you should just ask your question
[17:49:30] <pinkerton> steven_t: thanks. hopefully folks are working on this over the summer and we should have good progress
[17:49:37] *** gkelly has quit IRC
[17:49:54] <evmar> sky___: worker tests are notoriously flaky :(
[17:50:22] <malavv> Does someone know why, when I check about:network to watch job, it always does 2 times what I ask and cancel one?
[17:50:36] <steven_t> thatd be cool. im really digging Google Chrome over Safari
[17:50:54] <steven_t> much better UI and UX, even though its not native (which surprised me atfirst)
[17:50:56] <sky___> evmar: then I don't feel bad marking flaky:)
[17:51:18] *** sbyer has joined #chromium
[17:51:25] <malavv> + New Job * 2  and then  -Canceled and -Complete on the same url
[17:52:02] *** m0 has joined #chromium
[17:52:54] *** antivirtel has joined #chromium
[17:53:07] *** GeekShado_ has joined #chromium
[17:53:10] *** kjalarr has joined #chromium
[17:54:13] <evmar> malavv: that sounds bad to me
[17:54:28] <evmar> malavv: maybe open a bug?
[17:55:03] <pinkerton> steven_t: the ui is fully cocoa
[17:55:19] <evmar> malavv: it doesn't happen to me
[17:56:10] <steven_t> yes but it doesnt use NSMenus in a lot of places safari does
[17:56:16] *** Zaba has joined #chromium
[17:56:21] <steven_t> which i applaud you for
[17:56:21] <pinkerton> steven_t: like where?
[17:56:25] *** GeekShadow has quit IRC
[17:56:42] <steven_t> bookmark folder menus
[17:56:59] <pinkerton> ah yes. some people like that, some don't
[17:57:08] <pinkerton> we had to do it for drag and drop, which i still dont care for
[17:57:08] <evmar> those blew my mind the first time i saw them
[17:57:14] <steven_t> well its better for things like dragging stuff around
[17:57:21] <evmar> i was impressed someone took the time to do 'em
[17:57:23] <steven_t> i assume its all done inside a webview
[17:57:30] <pinkerton> no, it's a custom NSView
[17:57:34] <steven_t> ah nice.
[17:57:37] <pinkerton> er NSWindow with views inside
[17:57:41] <steven_t> right
[17:57:48] <steven_t> borderless window mask etc
[17:57:53] <steven_t> yeah i dig it.
[17:57:56] *** BUGabundo_remote has joined #chromium
[17:57:58] <pinkerton> right, but it's all cocoa. no webviews there.
[17:58:02] <steven_t> do you guys volunteer to work on this?
[17:58:02] *** kenneth_reitz has joined #chromium
[17:58:08] <evmar> i volunteer
[17:58:08] <kenneth_reitz> How do i trigger a response as a download properly?
[17:58:11] <steven_t> or do any of you get paid to
[17:58:11] <evmar> and then google pays me
[17:58:18] <kenneth_reitz> 'm doing Content-disposition: attachment; filename=test.pdf
[17:58:28] <kenneth_reitz> but chrome is saving it as test.pdf.html
[17:58:32] <pinkerton> many of us in here are paid, some volunteer because they
[17:58:33] <steven_t> id love to work on it but id need to get paid for it to justify my time
[17:58:46] <pinkerton> well, google  is hiring :)
[17:58:54] <steven_t> can i work remotely? ;)
[17:58:54] *** GeekShado_ has quit IRC
[17:58:56] <pinkerton> send me your resume, you have my email :)
[17:59:12] <steven_t> we need to live by family, our 3 kids need to be by their cousins and aunts and uncles etc
[18:00:01] *** kjalarr has quit IRC
[18:00:16] <kenneth_reitz> Anyone?
[18:00:24] <kenneth_reitz> This is a chrome-specific bug
[18:00:37] *** antivirtel has quit IRC
[18:00:37] *** m0 has quit IRC
[18:01:24] <evmar> kenneth_reitz: your best bet is to open a bug
[18:01:50] <kenneth_reitz> I'm trying to determine if i'm doing it wrong
[18:01:54] *** antivirtel has joined #chromium
[18:03:21] *** mrossetti has joined #chromium
[18:03:37] *** inferno-sec has joined #chromium
[18:03:46] *** fqian has joined #chromium
[18:03:52] <lukas___> so is google ok with working remotely or not?
[18:03:53] *** trungl_away is now known as trungl
[18:04:31] <pinkerton> heh. sort of.
[18:04:35] <pinkerton> usually not.
[18:04:41] <lukas___> or is that kind of reserved for, you know, senior engineers like pinkerton working from VA etc?
[18:05:12] <pinkerton> it's reserved for people who have pictures of vp's with goats and tiaras.
[18:05:33] *** rsesek has quit IRC
[18:05:48] <lukas___> the nearest google office is in zurich i think, that's like 2-3 hours from here by train or car.
[18:05:50] <trungl> where can I get those?
[18:05:53] <pinkerton> oh wait, they have pictures of me with the goats and tiaras.
[18:05:57] <pinkerton> dangit.
[18:06:16] <trungl> my impression was that everyone had *those*
[18:06:24] <trungl> (except for me)
[18:06:33] <pinkerton> true.
[18:07:49] * trungl searches for "pinkerton goats tiaras".
[18:08:03] <trungl> There's apparently an interesting book by a guy by the name of "John Pinkerton".
[18:10:35] *** jrmuizel has joined #chromium
[18:11:15] *** rsesek has joined #chromium
[18:11:38] <pinkerton> whew. i dont show up in the image results
[18:11:51] <lukas___> lol.
[18:14:23] <sky___> atwilson: webkit roll is causing page cycler to fail.
[18:14:32] <sky___> Can you revert.
[18:16:15] *** m0 has joined #chromium
[18:16:21] <trungl> pinkerton: searching for you specifically seems to yield a lot of images of sad tabs
[18:16:30] <trungl> pinkerton: and one really pixelated picture of you
[18:16:40] <pinkerton> heh
[18:16:55] *** kjalarr has joined #chromium
[18:17:23] * trungl thinks we should replace the sad tab image with one of pinkerton.
[18:17:33] *** gkelly has joined #chromium
[18:18:18] <pinkerton> hey now
[18:18:57] <trungl> it'd be awesome!
[18:20:15] *** dmaclach has joined #chromium
[18:20:20] *** antivirtel has quit IRC
[18:21:13] *** kjalarr has quit IRC
[18:21:28] <steven_t> are the differences between chrome and chromium worth using chrome?
[18:21:42] <rsesek> chrome will auto-update. that's the major difference
[18:21:46] *** dmaclach has quit IRC
[18:21:55] <rsesek> and it's more stable because it goes through QA
[18:22:06] *** m0 has quit IRC
[18:22:12] <trungl> chrome (beta and stable channels) are built off stabilized branches
[18:22:14] *** gkelly has quit IRC
[18:22:26] <trungl> well, you could go to the trouble of building chromium off those branches
[18:22:35] <trungl> but then you'd have to build it yourself
[18:22:36] <pinkerton> and if you use chrome, we get your crash reports so we can fix your bugs
[18:22:44] <steven_t> o
[18:22:49] <trungl> but if you use chromium, you'll fix our bugs for us, right?
[18:22:54] <steven_t> lol
[18:22:55] *** dhaza has joined #chromium
[18:22:57] <pinkerton> or whine about it :)
[18:23:05] <trungl> especially if you're masochistic enough to use a debug build
[18:23:10] *** darwin_ has joined #chromium
[18:23:35] *** GeekShadow has joined #chromium
[18:23:48] *** dave_levin has quit IRC
[18:23:48] <pinkerton> now that's just crazytalk
[18:24:08] *** dave_levin has joined #chromium
[18:24:10] <trungl> That's CrazyTalk(R) to you.
[18:24:13] <pinkerton> i wish we could build render processes in release and the browser process in debug
[18:24:18] 
[18:25:00] <trungl> We could probably build WebKit in release, and the rest in debug.
[18:25:14] <trungl> Maybe WebKit + V8.
[18:25:49] <pinkerton> given that it's all one binary, that may get tricky
[18:27:49] <trungl> it shouldn't be too hard to compile the contents of chrome/renderer to be release either
[18:28:03] <trungl> (maybe what we really want is to just compile the contents of chrome/browser to be debug)
[18:28:15] <trungl> though that'd leave the various common directories as release
[18:28:36] <trungl> but there's no easy way around that since it's all one binary
[18:28:51] <steven_t> chrome would be a lot better if it was written in qbasic 3
[18:28:53] * steven_t ducks
[18:29:04] <sky___> the first version was.
[18:29:53] <dhaza> can anyone point me towards any resources on how v8 can handle the webworkers API?
[18:30:15] *** kjalarr has joined #chromium
[18:30:18] *** gkelly has joined #chromium
[18:30:19] <dhaza> does it spawn a new process?  create a thread with no shared memory?
[18:30:33] <evmar> pinkerton: we had a gyp flag for no debug symbols for webkit
[18:30:39] <evmar> pinkerton: since that's where most of the bloat is
[18:30:50] *** m0 has joined #chromium
[18:30:57] <evmar> dhaza: you migh tneed to ask #chromium-dev
[18:30:59] *** antivirtel has joined #chromium
[18:31:01] <evmar> er, that's a mailing list
[18:31:33] <pinkerton> evmar: i'm not concerned about binary size, i'm concerned about speed. i dont run debug builds because page rendering is too slow, but i wish i could run them so i could see what's up when the ui process goes wonky
[18:32:06] *** atwilson_ has quit IRC
[18:32:11] <pinkerton> like "hey, what's this thread doign taking up 100% of my cpu?!"
[18:33:10] <dhaza> evmar, will do, thank you
[18:33:25] *** dhaza has quit IRC
[18:33:44] *** dmaclach has joined #chromium
[18:34:08] *** dmaclach has joined #chromium
[18:34:58] <trungl> I had a patch which allowed debug symbols to be included/excluded from relatively arbitrary components, but the improvement in build times (on Mac) were mostly disappointing
[18:36:48] *** sundiamonde has joined #chromium
[18:36:59] <trungl> so, in webkit, a PassRefPtr indicates that ownership (of a reference) is being transferred, right?
[18:39:15] <trungl> wow, there's even documentation about this stuff
[18:39:22] <trungl> (not in the source, naturally)
[18:39:32] *** chase_ is now known as chase
[18:39:43] *** ChanServ sets mode: +v chase
[18:40:00] *** m0 has quit IRC
[18:40:05] *** kjalarr has quit IRC
[18:40:20] * trungl is enlightened.
[18:42:05] <evmar> pinkerton: oh, yeah, that.  :)  willchan actually runs debug builds always 'cause he's a madman.  i've started doing release builds + symbols + core dumps
[18:42:14] <pinkerton> :)
[18:42:21] <evmar> that means i at least get good stack traces and some variables
[18:42:58] <pinkerton> it would be great if we had symbols uploaded for our waterfall builds
[18:42:59] *** antivirtel has quit IRC
[18:43:02] <pinkerton> since i run those
[18:43:07] <pinkerton> but i understand that's a bit impractical
[18:43:20] *** gkelly has quit IRC
[18:43:41] <pinkerton> i thought disk space was free in "the future". where's my flying car, damnit!?
[18:43:57] *** miketayl has joined #chromium
[18:44:39] *** yakub has joined #chromium
[18:44:50] <evmar> you'd think we could do some sort of delta storage or something
[18:45:00] <dglazkov> good morning, Chromium!
[18:45:16] <yakub> is there a way to turn off 'off line' mode? i know perfectly well my internet connection doesn't work!  im trying to get chromium to look at a web page served on a webserver running on my local machine
[18:45:26] *** iPac has quit IRC
[18:45:27] *** miketaylr has quit IRC
[18:50:09] *** iPac has joined #chromium
[18:50:37] *** malavv has quit IRC
[18:51:10] <evmar> there is no offline mode
[18:51:42] <pinkerton> yeah, i've not seen what you're talking about. i've used local webservers before with no problem
[18:51:57] *** lukas___ has quit IRC
[18:53:22] <yakub> did it work for you while you were offline, completely?
[18:53:50] * trungl wonders if he's the only one who finds RefPtr's .release() method misleadingly named.
[18:54:00] *** roc has quit IRC
[18:54:34] *** johnny_g has joined #chromium
[18:55:52] <pinkerton> willchan: ping?
[18:56:04] *** yakub has left #chromium
[18:56:26] <stuartmorgan> trungl: I assume it's like our scoped pointers?
[18:57:01] <trungl> stuartmorgan: yes, but ...
[18:57:23] <trungl> stuartmorgan: background: a PassRefPtr is like a RefPtr, but zeros the original pointer upon any assignment/copy
[18:57:34] *** roc has joined #chromium
[18:58:00] *** erikkay has quit IRC
[18:58:01] <trungl> stuartmorgan: RefPtr<T>::release() produces a PassRefPtr<T>, not changing the ref count, but zeroing the original RefPtr<T>
[18:58:47] <trungl> so, if have a RefPtr<T> foo and you want to return a PassRefPtr<T>, you do "return foo.release();".
[18:59:10] <trungl> maybe it's not so confusing
[18:59:11] <trungl> hmmm
[18:59:18] <trungl> maybe it is
[18:59:22] * trungl isn't sure.
[18:59:24] <stuartmorgan> That's basically how our release() works
[18:59:30] <stuartmorgan> Except ours give raw pointers
[19:00:13] <stuartmorgan> Is what confuses you that you want it to be link CFRelease?
[19:00:24] <stuartmorgan> (i.e., you expect release() to mean what reset() is?)
[19:00:42] <trungl> I'm just easily confused.
[19:00:57] *** arv has quit IRC
[19:01:22] <stuartmorgan> We're working on decorating our release()s with unused return warnings to try to prevent mistakes
[19:01:41] *** arv has joined #chromium
[19:02:20] *** oshima_ has joined #chromium
[19:02:47] *** jrforbes_ has quit IRC
[19:02:57] <pinkerton> trungl: maybe go ask in #webkit?
[19:03:15] <trungl> I'm actually enlightened.
[19:03:29] <trungl> Darin Adler actually wrote an article explaining all of this.
[19:03:37] <pinkerton> k
[19:04:36] *** up365 has quit IRC
[19:07:50] *** victorw has joined #chromium
[19:09:34] *** Yon has quit IRC
[19:16:38] *** RT|Chatzilla has quit IRC
[19:16:58] *** BUGabundo_remote has quit IRC
[19:17:06] *** sky___ has left #chromium
[19:21:31] *** abarth has quit IRC
[19:21:50] *** up365 has joined #chromium
[19:22:05] <atwilson> I'm trying to track down a perf regression with my latest webkit roll. The problem is that r59638 introduced a new test that needed to be rebaselined, and the rebaseline wasn't landed until 3 hours later at r59656. Somewhere in the middle is a CL that caused a regression.
[19:22:11] <steven_t> my hobby: switching between an empty tab in chrome and a full tab, and pressing cmd+shift+b each time, to watch the height of the window consistently grow
[19:22:24] <steven_t> (or shrink depending on when you started)
[19:22:32] <atwilson> In our new world where test_expectations are upstreamed, I don't quite know how to do incremental rolls to try to narrow down where the perf regression happened.
[19:22:37] <atwilson> Any advice for me?
[19:22:49] *** MikeSmithW3C^ has joined #chromium
[19:23:12] <atwilson> Basically, I'd like to roll to (say) r59643, but I can't because there's a broken test.
[19:23:45] *** Zaba has quit IRC
[19:24:08] <atwilson> Doesn't look like there are perf trybots...
[19:24:10] <evmar> atwilson: we have a local test_expectations.txt i think, that overlays the upstream one
[19:24:34] <evmar> atwilson: i would suggest rolling + temporary "don't care" about the test
[19:24:44] *** sky__ has joined #chromium
[19:24:54] *** Mekzholan has joined #chromium
[19:25:16] * atwilson is looking for the local test_expectations.txt
[19:25:22] *** csilv has joined #chromium
[19:26:29] <Mekzholan> Hi, directly after booting my linux box it's starting chromium - but it's allways failing and showing a page that gives me the error message ERR_NAME_NOT_RESOLVED. What's the cause and how might I fix it?
[19:26:39] *** bweinstein has joined #chromium
[19:27:02] <Mekzholan> might it be that it can't access the DNS at that time?
[19:27:02] *** MikeSmithW3C has quit IRC
[19:27:20] <evmar> willchan: ping, see above q
[19:30:01] *** rsesek has quit IRC
[19:30:12] *** rsesek has joined #chromium
[19:33:58] <rohitrao> pkasting: your change made slashdot! http://yro.slashdot.org/story/10/05/18/1523250/Chrome-Private-Mode-Not-Quite-Private?art_pos=2
[19:34:04] *** dhhwong has joined #chromium
[19:34:19] *** Zaba has joined #chromium
[19:35:12] *** malavv has joined #chromium
[19:35:23] *** dhhwong has quit IRC
[19:35:25] *** mpcomplete has joined #chromium
[19:35:25] *** ChanServ sets mode: +v mpcomplete
[19:36:49] *** loislo has quit IRC
[19:37:33] *** kenneth_reitz has quit IRC
[19:38:06] *** miketayl has quit IRC
[19:38:13] *** miketaylr has joined #chromium
[19:38:58] *** alyxuk has quit IRC
[19:41:10] *** rafaelw has joined #chromium
[19:41:56] *** kjalarr has joined #chromium
[19:42:00] *** antivirtel has joined #chromium
[19:42:15] *** Venom_X is now known as Venom_lnch
[19:43:15] *** dhw_ has joined #chromium
[19:43:30] <dhw_> Hi Tony^work
[19:46:02] *** gkelly has joined #chromium
[19:46:08] *** alyxuk has joined #chromium
[19:46:09] *** alyxuk has quit IRC
[19:46:09] *** alyxuk has joined #chromium
[19:46:14] *** m0 has joined #chromium
[19:48:48] *** blindmurray has joined #chromium
[19:51:49] *** roc_ has joined #chromium
[19:53:54] *** roc has quit IRC
[19:53:54] *** roc_ is now known as roc
[19:57:00] *** foucist has joined #chromium
[19:57:21] *** GeekShadow has quit IRC
[19:57:27] *** evmar is now known as evmar-afk
[19:57:28] <foucist> anyone know about the feasbility of a nodejs plugin, especially one that takes advantage of the internal v8 js engine?
[19:57:34] <foucist> since nodejs uses v8
[19:57:50] <foucist> wondering about reducing redundancy
[20:03:09] <willchan> mekzholan: is your internet connection up? :)
[20:03:25] *** m0 has quit IRC
[20:03:28] <willchan> pinkerton: pong?
[20:03:54] *** roc_ has joined #chromium
[20:04:11] <willchan> er, why am i a madman again?
[20:04:14] <pinkerton> willchan: so every now and then on my desktop (using a socks proxy), i get tons and tons of sockets that just seem to hang
[20:04:15] *** roc has quit IRC
[20:04:15] *** roc_ is now known as roc
[20:04:22] <pinkerton> willchan: like on pages with lots of images, or google reader
[20:04:31] *** m0 has joined #chromium
[20:04:39] <pinkerton> willchan: going to net-internals shows i have like 50 sockets open, and they're just sitting there doing nothing
[20:04:50] <pinkerton> seen anything like that?
[20:04:57] <atwilson> Does anyone here understand the perf bot output?
[20:04:58] <willchan> pinkerton: what build are you on?
[20:05:08] <pinkerton> willchan: 6.0.400.0 (46883)
[20:05:26] <atwilson> I'm trying to understand why the perf waterfall says that page_cycler_intl1 is failing, but this chart doesn't show anything bad?
[20:05:29] <willchan> lemme check if a fix i landed is in there
[20:05:32] <atwilson> http://build.chromium.org/buildbot/perf/mac-release-10.6/intl1/report.html?history=150&rev=-1
[20:05:44] <willchan> is the cpu at 100%?
[20:05:57] <atwilson> sky: ping
[20:05:59] <pinkerton> no, everyone is well behaved. it's different than the 100% cpu bug
[20:06:17] *** apavlov has quit IRC
[20:06:31] <pinkerton> if i close down the tabs, things eventually recover, until they go haywire again sometime later (unlike the 100% cpu bug)
[20:06:52] <atwilson> sky,sky__: Looks like page_cycler_intl1 is failing on macos still even after my revert?
[20:06:55] <willchan> can you repro this?
[20:07:08] <pinkerton> usually when i least have time to play with it :)
[20:07:15] <pinkerton> it comes and goes.
[20:07:21] <pinkerton> what should i do next time i see it?
[20:07:26] <willchan> if you're behind a socks proxy, you should only be allowed to have 15 sockets running
[20:07:52] <willchan> through that proxy that is
[20:07:53] <pinkerton> maybe that's all i do, but the net-internals page has like 50 things on it
[20:08:05] *** Zucca has quit IRC
[20:08:09] <willchan> can you send me the net-internals data?
[20:08:17] *** Zucca has joined #chromium
[20:08:18] <pinkerton> is there a way to easily copy what's going on from that new UI? it's all dhtml and stuff now
[20:08:24] <willchan> there's a #data tab
[20:08:33] <pinkerton> ah ok. i'll send you that next time i see it
[20:08:35] <pinkerton> thanks
[20:08:38] <willchan> ok
[20:09:00] *** craigsch has left #chromium
[20:09:18] *** eseidel has joined #chromium
[20:09:32] <willchan> when you say 50 "things" on it, what do you mean by that?  you should probably see URL_REQUEST, CONNECT_JOB, SOCKET, etc
[20:10:01] <willchan> actually, don't bother, just repro and send me the net-internals data
[20:10:26] <willchan> my guess is you're hitting the 15 socket limit, did you just start using a socks proxy?
[20:11:45] <thakis_afk> good?hrm?morning, i guess, chromium!
[20:12:41] <pinkerton> willchan: no, a couple months ago, but only recently really started having issues
[20:13:09] *** rickspencer3 has quit IRC
[20:13:45] *** rickspencer3 has joined #chromium
[20:13:48] *** GeekShadow has joined #chromium
[20:15:14] <malavv> Btw, I don't know who has been working on about:net-internals but the new UI is 1337
[20:15:34] <pinkerton> heh
[20:16:00] <willchan> that's eroman's work
[20:16:08] *** roc has quit IRC
[20:16:10] <malavv> pinkerton: New UI has tabs and everything, not just a plain dump of data
[20:16:16] <pinkerton> i know
[20:16:24] <pinkerton> that's why i couldn't figure out how to easily send the data :)
[20:16:30] *** roc has joined #chromium
[20:16:31] <pinkerton> but i know now
[20:16:37] <eroman> malavv: thanks! I am preparing a document to explain how to use it, still fixing issues
[20:17:58] *** chronarion has joined #chromium
[20:17:58] *** ChanServ sets mode: +v chronarion
[20:18:09] <malavv> eroman: You did a nice job, all information is know usefull and understandable. Did you work on chrome://about/network too?
[20:18:14] *** kenneth_reitz has joined #chromium
[20:18:25] <malavv> eroman: is now*
[20:18:27] <chronarion> hey... it looks like a random LD error, don't think it's me
[20:18:32] <chronarion> (tree breakage)
[20:19:10] <chronarion> looks like brettw? does it have to be a clean build or something?
[20:20:48] <willchan> hm, my chromium instance SIGSEGV'd last night in tcmalloc.  sounds like we have memory corruption :(
[20:20:51] <eroman> malavv: you mean "about:network" ? no, but one of my motivations with the new UI is to be able to replace "about:network" with an implementation that works on all platforms
[20:21:29] <sky__> brett's on it.
[20:21:37] *** wr| has quit IRC
[20:21:53] *** sjr has quit IRC
[20:23:38] *** loislo has joined #chromium
[20:24:55] <malavv> eroman: Sorry, yes about:network. Ok, well congrats
[20:27:53] *** foucist has left #chromium
[20:29:24] <pkasting> glotov, sky__: I just landed something with a .grd change, but I didn't clobber because it doesn't add or remove any entries, only renames a few.
[20:29:39] *** trungl is now known as trungl_lunch
[20:29:51] <pkasting> glotov, sky__: So I assumed it should be OK.  But if the tests do turn red, that may be why
[20:31:23] <sky__> ok
[20:31:57] *** bent-mozilla has quit IRC
[20:32:29] *** Kaosevil has joined #chromium
[20:38:42] *** bent-mozilla has joined #chromium
[20:39:55] *** eseidel has quit IRC
[20:40:26] <steven_t> is there any way at all to make chrome/chromium tab[key] only to fields, not to links?
[20:43:57] *** duxklr has joined #chromium
[20:45:00] <rsesek> steven_t: there's a bug for that
[20:45:12] <rsesek> crbug.com/31177
[20:46:53] *** dhw_ has quit IRC
[20:48:51] <steven_t> woot
[20:50:09] *** robarnold|eviltw is now known as robarnold
[20:50:11] <pinkerton> willchan: email sent
[20:52:27] <Mekzholan> willchan: yes. The connection is up
[20:53:09] <Mekzholan> it's only one additional client in the whole network
[20:55:08] *** BCalvignac2 has quit IRC
[20:55:25] *** trungl_lunch is now known as trungl
[20:55:27] *** Russ is now known as Russ|Out
[20:56:43] <fta> evmar-afk, http://penguindroppings.wordpress.com/2010/05/18/chromium-saved-passwords/
[20:56:47] *** BCalvignac1 has joined #chromium
[20:56:53] <maruel> sheriffs: I'm about to install stuff on linux perfs slaves, expect weird readings
[20:57:05] <maruel> glotov: ^^
[20:57:11] <sky__> maybe we should close the tree for the rest of the day.
[20:57:13] <sky__> :)
[20:58:16] *** Blockeaux has joined #chromium
[20:58:17] *** Russ|Out is now known as Russ
[20:58:34] <sky__> I've got an interview and lunch, don't let the tree go down in flames while I'm gone!
[20:59:25] <willchan> pinkerton: got it, i'll look at it soon
[20:59:36] <pinkerton> i can send another if you want, it's doing it again
[20:59:56] <maruel> Linux Perf (1) is done
[21:00:24] <willchan> pinkerton: if you can keep net-internals open while you repro, and then send me the data, that'd be great
[21:00:55] <willchan> since we log a bit more info while net-internals is open (usually we don't log it since it's expensive in terms of memory, unless we know someone is actively looking)
[21:01:20] <pinkerton> willchan: ok sent dump, net-internals sitll open
[21:02:22] <willchan> pinkerton: i see lots of STALLED_MAX_SOCKETS
[21:02:27] <pinkerton> yeah
[21:02:29] <willchan> i think you're hitting the 15 socket limit
[21:02:32] <willchan> for a proxy server
[21:03:07] <pinkerton> shouldn't the behavior be better than "stall for  minutes on end"?
[21:03:27] <maruel> Linux Perf (2) is done, I don't think it will affect any perf result significantly
[21:03:39] *** skydrome has quit IRC
[21:04:18] <willchan> pinkerton: i don't think it's completely stalled, you're probably just getting really low throughput.  probably lots of the sockets are being occupied by hanging GETs or something
[21:04:54] <willchan> you might be stalled if there are any interdependencies on HTTP transactions.  if one socket is being used, but is waiting for data from another HTTP transaction that is stalled.
[21:05:04] <pinkerton> grr
[21:05:09] <pinkerton> is the 15-limit ours?
[21:05:17] <willchan> the only easy solution i can think of is to raise the 15 limit
[21:05:21] <willchan> yes, it's ours
[21:05:30] <pinkerton> do we know what safari or firefox use?
[21:05:38] <willchan> eroman: ping
[21:05:41] <pinkerton> when i try to load the same page in safari at the same time, it usually goes right through
[21:05:55] <eroman> willchan: pong
[21:05:57] <willchan> well, safari won't have the same sockets open
[21:06:11] <willchan> eroman: ^^^ question on safari/firefox connection limits through a socks proxy
[21:06:26] <pinkerton> true, but the same page should open a similar number
[21:06:36] <willchan> do they set a limit on the max number of sockets that can be opened to a specific proxy?
[21:06:40] <willchan> and if so, how much?
[21:07:03] <pinkerton> that page finally finished loading. nice.
[21:07:08] <eroman> willchan: Firefox has the 15 limit for HTTP proxies
[21:07:22] <eroman> willchan: I am not sure if they enforce the same limit for SOCKS
[21:08:31] <willchan> as i understand it, the limit exists so we don't overload the proxy server, right?
[21:08:52] <willchan> perhaps we should raise it if pinkerton is seeing such degraded perf with 15
[21:08:58] <willchan> since it's likely others see this too
[21:08:58] *** Venom_lnch has quit IRC
[21:09:22] <willchan> and, unlike firefox, we don't provide a configuration to change this limit
[21:12:04] <akalin> willchan, eroman: when you two get some time i'd like to talk about chrome sockets!
[21:12:30] <willchan> akalin: in a bit, i'm still thinking about this proxy connection limit issue.  and then it might be lunchtime.
[21:12:36] <eroman> willchan: yeah, low proxy limits are killer
[21:12:52] <willchan> eroman: well, we _raised_ the http proxy limit :P
[21:12:52] *** General13372 has joined #chromium
[21:13:00] <eroman> especially now that all googlers are going through proxy, they hit the performance problems more often
[21:13:14] <willchan> previously, we didn't have any limit for socks proxies, besides the standard 6 connections per host
[21:13:23] <willchan> which is why pinkerton is seeing it now
[21:13:37] <pinkerton> ah, how recent was that change?
[21:13:38] <willchan> i don't see a good reason to have http proxies behave different from socks proxies in terms of connection limits
[21:13:53] <willchan> pinkerton: 2 weeks?  i forget
[21:14:01] <pinkerton> yup, that seems about right
[21:14:09] <willchan> april 29
[21:14:50] <willchan> we used to limit 6 connections _total_ through http proxies, and unlimited for socks.  i made the socks proxies share the same limit as http proxies, and raised the limit to 15
[21:15:41] <willchan> eroman: all googlers are going through the PAC resolver, which doesn't necessarily put all their connections through proxies.
[21:15:45] <pinkerton> would using the two dumps i sent you help pick a more reasonable default? they're two pretty normal sites.
[21:16:05] *** General1337 has quit IRC
[21:16:16] <pinkerton> i guess in one instance i was loading two pages, in the other just one.
[21:16:22] <willchan> i think the "right" thing to do is raise the default to something more conservative for the end users
[21:16:32] <rsesek> does anyone know how do dump a minidump file on a Mac?
[21:16:33] <willchan> and gradually experiment with lowering it
[21:16:59] <pinkerton> raise it to something more conservative? i don't follow
[21:17:04] <willchan> run A/B tests and see if there's a point where we start getting a bunch more stalls
[21:17:10] <willchan> er, raise it to 50 or something
[21:17:16] <pinkerton> ok :)
[21:17:20] <willchan> which should be "plenty"
[21:17:20] <thakis_afk> my renderer processes still "idle" at 1.5% cpu with the current dev channel
[21:17:24] <willchan> and then start lowering
[21:17:34] <thakis_afk> anyone have an idea why by now?
[21:17:35] <willchan> firefox's limit is probably arbitrary
[21:17:40] <pinkerton> willchan: LGTM!
[21:18:10] <willchan> and since power users will configure the firefox settings, if they hit issues they'll just raise their own limits.
[21:18:37] * pinkerton subscribes to the time-honored tradition of affecting change by whining louder than everyone else
[21:19:21] <willchan> i will need to discuss further with eroman and others.  i think fishd and wtc may have opinions.
[21:19:38] <pinkerton> i'm sure. what you're suggesting seems reasonable to me.
[21:19:50] <pinkerton> but i defer to the experts.
[21:21:46] *** eseidel has joined #chromium
[21:22:50] *** dpranke has joined #chromium
[21:22:50] *** ChanServ sets mode: +v dpranke
[21:22:53] * thakis_afk sighs and starts build-bisect.py
[21:27:21] *** dcheng has quit IRC
[21:27:59] *** Beetny has joined #chromium
[21:28:12] *** awong_afk has quit IRC
[21:28:33] *** Pr0jectRec0n has joined #chromium
[21:28:39] <Pr0jectRec0n> hey all
[21:28:46] <Pr0jectRec0n> I'm on windows 7
[21:29:17] *** dcheng has joined #chromium
[21:29:42] *** shepazu has quit IRC
[21:29:59] *** awong_afk has joined #chromium
[21:33:00] <Pr0jectRec0n> have a problem with google chrome .. --> this one http://imgur.com/FlkHh.jpg
[21:33:04] <Pr0jectRec0n> Any ideas abt it ?
[21:33:24] * akalin looks at the topic
[21:37:19] <thakis_afk> Pr0jectRec0n: can you open a bug at http://new.crbug.com ?
[21:37:26] *** thakis_afk is now known as thakis
[21:37:50] *** Mekzholan has left #chromium
[21:38:54] <Blockeaux> okay, I followed this and it's not working --> http://code.google.com/chrome/extensions/getstarted.html
[21:38:55] <Pr0jectRec0n> thakis, I'll look into it
[21:39:07] <Blockeaux> I've got the icon but when I click it, nothing appears
[21:39:35] *** Venom_X has joined #chromium
[21:39:38] *** thomasvl has quit IRC
[21:40:21] <Pr0jectRec0n> One more issue, I tried runnogn chrome underXP compatibilty mode - then the error code differs - its 0xc00000a5 then ..
[21:40:42] <akalin> Pr0jectRec0n: that would be good info to put in the bug
[21:41:04] <Pr0jectRec0n> akalin, okay
[21:41:14] *** eseidel has quit IRC
[21:41:27] *** drusepth has joined #chromium
[21:43:15] <Pr0jectRec0n> Also, I've tried restarting the whole system, still the error persists.. I've tried uninstalling and re-installing Chrome back as well. The error code still persists. All other browsers work fine though
[21:43:41] <akalin> Pr0jectRec0n: more good info to put in the bug
[21:44:09] *** Dataforce has joined #chromium
[21:45:30] *** cburroughs has joined #chromium
[21:45:32] <Pr0jectRec0n> akalin, One question - when I did an uninstall, I thought it was a shoddy uninstall ie. not complete removal of 'folders' (like plugins, cache, appdata etc.) Is there some software which completely rips off all data that chromes' put in, or can I do it manually?
[21:45:54] <akalin> Pr0jectRec0n: no idea, i don't run windows very often
[21:46:08] <thakis> linuxens: i get "/home/thakis/Downloads/chrome-linux/chrome: error while loading shared libraries: libdbus-glib-1.so.2: cannot open shared object file: No such file or directory" after trying to execute a waterfall build. `apt-get install libdbus-glib-1-2` claims i have the latest version already. what am i doing wrong?
[21:46:08] <Pr0jectRec0n> akalin, no issues
[21:46:23] *** eseidel has joined #chromium
[21:47:25] <akalin> thakis: does a 'libdbus-glib*' file exist?
[21:47:43] <thakis> akalin: where would i look?
[21:47:47] *** Mekzholan has joined #chromium
[21:47:53] <akalin> /usr/lib or thereabouts
[21:48:08] <thakis> akalin: yes, looks like it
[21:49:02] <thakis> akalin: oh, looks like i only have the 64bit version of that lib
[21:49:09] <akalin> oh
[21:50:53] <leiz> sheriffs: win xp/vista dbg perf are going to run a little slow. They're very low on disk space and I'm doing a little cleanup ATM.
[21:51:43] <Blockeaux> newbie question, do I place code such as chrome.tabs.create directly within the browser_action {} braces?
[21:52:48] <Mekzholan> is there a way to disable the "wating for ..." on the bottom of the window? (I'm using kiosk mode)
[21:52:48] <dcheng> Blockeaux: No. http://code.google.com/chrome/extensions/browserAction.html ought to give you an idea of what you should be doing.
[21:52:51] *** cburroughs has left #chromium
[21:55:52] <thakis> downloaded a 64bit instead, which worked
[21:59:26] *** Zaba has quit IRC
[21:59:31] <Blockeaux> okay why does my browser action get a numerical overlay which counts upwards rapidly?
[22:00:33] <Blockeaux> nm
[22:00:36] <Blockeaux> :P
[22:02:07] *** albertb has quit IRC
[22:03:46] *** KnifeySpooney has joined #chromium
[22:04:02] *** KnifeySpooney has left #chromium
[22:06:16] *** AaronMT has joined #chromium
[22:06:38] *** jianli_ has joined #chromium
[22:08:29] *** jianli has quit IRC
[22:09:49] *** Zaba has joined #chromium
[22:10:26] *** feldstein has joined #chromium
[22:10:30] *** mpcomplete has quit IRC
[22:11:07] *** trungl_mbp has joined #chromium
[22:11:41] *** mpcomplete has joined #chromium
[22:11:41] *** ChanServ sets mode: +v mpcomplete
[22:16:04] *** alyxuk| has joined #chromium
[22:16:04] *** alyxuk| has quit IRC
[22:16:04] *** alyxuk| has joined #chromium
[22:16:42] *** alyxuk has quit IRC
[22:18:21] *** chaser has quit IRC
[22:19:44] *** chaser has joined #chromium
[22:20:14] <malavv> eroman: Do you know if someone started working on the GetAuthorizationHeaderName that produced infinite Auth because of empty string?
[22:20:17] *** rohitrao_ has joined #chromium
[22:20:17] *** ChanServ sets mode: +v rohitrao_
[22:20:46] <eroman> malavv: cbentzel may have
[22:20:54] <eroman> cbentzel: ^^
[22:21:20] *** lukas___ has joined #chromium
[22:21:44] <rsesek> pinkerton: ping
[22:21:47] <pinkerton> pong
[22:21:53] <Pr0jectRec0n> Opened up a new ticket at http://new.crbug.com -- > What would be the ETA ?
[22:21:56] <akalin> pingerton
[22:21:58] <Pr0jectRec0n> for the fix?
[22:22:05] <malavv> eroman: Just to see if this is still an `Help Wanted bug`
[22:22:27] <akalin> Pr0jectRec0n: depends on the problem and whether it's reproducible
[22:22:41] <rsesek> pinkerton: re crbug.com/31177 - do we change the default behavior [to not tab to links] or keep as-is?
[22:22:52] *** Pr0jectRec0n has quit IRC
[22:23:32] <pinkerton> looking
[22:23:39] <eroman> malavv: I think it is still up for grabs. Chris plumbed through the error code, but I don't think connected a policy. I suggest pinging cbentzel if you want to work on it
[22:24:29] *** rohitrao has quit IRC
[22:24:29] *** rohitrao_ is now known as rohitrao
[22:24:41] <malavv> eroman: K, k. I'll remember to do so.
[22:25:40] <pinkerton> rsesek: we should probably have some way to configure.
[22:25:50] <pinkerton> rsesek: what does camino do? i know we agonized over that for years there
[22:26:14] <rsesek> pinkerton: idk what Camino does, but Safari offers the checkbox that I'm plubming for/adding now
[22:26:40] *** gospch has joined #chromium
[22:27:02] <pinkerton> rsesek: ah, it's a popup menu in camino, default is form controls only
[22:27:19] <rsesek> what are the options in the popup? links | forms | ?
[22:27:41] <pinkerton> just text fields, all form controls, all form controls and links
[22:27:48] <rsesek> ah okay
[22:27:59] <rsesek> I don't think webkit differentiates between text fields and other form controls
[22:28:00] *** bent-mozilla has quit IRC
[22:28:08] <malavv> Someone knows if we are gooing to see SPDY anytime soon?
[22:28:34] <pinkerton> rsesek: another dupe of your page info ssl not showing all text :)
[22:28:41] <pinkerton> rsesek: what's the status?
[22:28:45] <rsesek> pinkerton: srsly?! it's fixed!
[22:28:48] <pinkerton> ok
[22:28:50] <rsesek> as of friday
[22:29:16] <rsesek> pinkerton: I just got the bugmail. I'll dupe it
[22:29:20] <pinkerton> thx
[22:29:23] *** jamesr has joined #chromium
[22:31:13] *** tonyg-cr has quit IRC
[22:34:16] <eroman> is brett here?
[22:34:29] *** Bleak has quit IRC
[22:34:43] <leiz> no brett here
[22:36:01] <eroman> ok, pinged him via IM, I think he is looking at linux64 issue
[22:37:13] *** bers has joined #chromium
[22:40:53] <thakis> rsesek / pinkerton: could the tabbing just use the full keyboard access pref in sysprefs?
[22:41:18] <pinkerton> problem is, nobody ever thinks to look there, and people also sometimes want different behavior
[22:41:29] <pinkerton> ie, i dont want to tab through controls in dialogs, but i do in web
[22:41:38] <rsesek> what he said
[22:42:34] *** ROBOd has quit IRC
[22:43:59] <thakis> k
[22:44:12] <thakis> who doesn't love options, right? :-P
[22:44:46] <pinkerton> i do agree, but this has been a sore ui point in the past with other browsers
[22:44:48] <rsesek> a checkbox hasn't (yet) killed anyone
[22:44:53] <pinkerton> "wah i want to change it how do i change it, u suck!"
[22:45:13] <thakis> stuartmorgan: do we add .../Chromium.app/Contents/Plugins to the search path because we expect people to put plugins there, or for other reasons?
[22:45:16] <pinkerton> people don't go looking in system prefs for that checkbox
[22:48:02] <rsesek> pinkerton: does it make sense to even look at the sysprefs value if we're adding a checkbox?
[22:48:08] *** brettw has joined #chromium
[22:48:08] *** ChanServ sets mode: +v brettw
[22:48:13] <rsesek> I think that could add a layer of confusion if this is to remain a checkbox
[22:48:24] <pinkerton> i dont think so, what does safari do?
[22:48:48] <rsesek> 1sec
[22:49:36] <rsesek> looks like they just use their pref checkbox
[22:52:09] <pinkerton> for reasons i indicated above. they used to use the system pref and nobody found it.
[22:53:28] *** bent-mozilla has joined #chromium
[22:54:07] *** Beetny has quit IRC
[22:54:52] *** oshima has quit IRC
[22:55:36] *** tonikitoo has quit IRC
[22:56:22] <sky__> The tree looks quite red
[22:56:32] *** tonyg-cr has joined #chromium
[22:56:32] *** ChanServ sets mode: +v tonyg-cr
[22:56:54] <sky__> Is a fix for linux 64 in?
[22:58:00] *** mpcomplete has quit IRC
[22:59:04] *** mpcomplete has joined #chromium
[22:59:04] *** ChanServ sets mode: +v mpcomplete
[23:00:50] <akalin> willchan, eroman: free to talk about sockets now?
[23:01:05] <eroman> akalin: sure
[23:01:24] *** cbentzel has quit IRC
[23:01:41] <akalin> eroman: okay, so basically i'm looking to remove our use of libjingle sockets with chrome sockets
[23:01:57] <eroman> akalin: sweet.
[23:02:08] <willchan> akalin: oh yeah, sorry, was about to ping you
[23:02:14] <stuartmorgan> thakis: I think it's pretty much just for testing
[23:02:17] <willchan> yay for using chrome sockets
[23:02:25] <akalin> one thing the libjingle sockets do that the usual ClientSockets don't is that they always listen for data and send out a callback when there's data to be read
[23:02:40] <akalin> i was digging around for something in net/ that does that, and the closest thing i could find was socket_stream.h
[23:02:47] <akalin> although there's some websockets-specific stuff there
[23:02:56] <brettw> I'm getting this error from gcl upload, does anybody know what it means? "urllib2.URLError: <urlopen error (-2, 'Name or service not known')>"
[23:03:28] <willchan> akalin: you can do that by always calling Read() again
[23:03:33] <willchan> in your callback
[23:03:57] <akalin> willchan: hmm, that's true
[23:04:06] <akalin> does that also take care of socket closure?
[23:04:17] <eroman> yes. socket closure send 0 bytes back
[23:04:20] <willchan> akalin: that'll be a 0 byte read
[23:04:25] <akalin> mmm
[23:04:39] *** pinkerton has quit IRC
[23:04:40] <akalin> it looks like socket_stream.cc does something similar, in that the callback always exits with a pending operation
[23:05:14] <akalin> so i guess there's nothing else with a similar interface like socket_stream.h, where you just pass in a delegate with handlers for read(), close(), etc.?
[23:05:37] <akalin> if not, i was thinking of just ripping out the tcp- and ssl- related stuff out of it into a separate class
[23:05:50] <akalin> which socket_stream and the xmpp stuff would both use
[23:06:14] <eroman> akalin: not that i am aware of. however as will mentioned, it should be a simple matter of doing a Read() loop to create such an interface
[23:06:16] *** estade has quit IRC
[23:06:24] <akalin> eroman: right
[23:06:31] *** Mekzholan has quit IRC
[23:06:50] <akalin> i'm just foreseeing that it'll start off that way, and then eventually mutate into something approximating what socket_stream does, which is maintain a state machine
[23:07:18] <willchan> socket_stream is used for websockets
[23:07:21] <akalin> yeah
[23:07:27] <willchan> i think it gives us some issues
[23:07:30] <akalin> oh?
[23:07:32] <willchan> since there's a fair amount of code duplication
[23:07:42] <akalin> code duplication with what?
[23:07:57] <willchan> http_network_transaction.  i think there's a parallel copy in socket_stream
[23:08:07] <willchan> which might not do everything we do for http_network_transaction.cc
[23:08:07] <akalin> ugh
[23:08:25] <willchan> eroman: please confirm my previous statement.  that's what i recall.
[23:08:37] <akalin> yeah, looks like they're both similar
[23:08:44] <eroman> willchan: yeah. socket_stream also has some duplication with the socket pools, since it repeates the SSL  connections etc..
[23:09:16] <willchan> yeah, socket_stream.cc needs to be refactored
[23:09:16] <akalin> so, yeah, that's what i'm afraid of, if I start just writing my own XmppSocket stuff, it'll have to handle SSL connections, proxies, etc.
[23:09:35] <akalin> and I don't want to duplicate more code
[23:09:39] <willchan> i'm worried we might not be doing proxies correctly in socket_stream.cc
[23:09:51] <eroman> willchan: crbug.com/37810
[23:09:57] <akalin> OTOH, i don't want to undertake a major refactoring of these classes :)
[23:10:17] <akalin> any ideas for how i should proceed?
[23:10:22] <willchan> eroman: aha yes, i was just about to comment that SocketStream does not use our socket pools :P
[23:10:26] <willchan> so it bypasses our limits
[23:10:32] <akalin> oh
[23:10:43] <akalin> ha, i thought sync was the only one that did that :P
[23:11:21] <eroman> my recommendation is to pull sockets out of the main socket pool, and create a simple wrapper to mimic OnDataAvailable() style callback
[23:11:43] <eroman> the SocketStream /socket pool refactoring is a larger ongoing task IMO
[23:11:47] <akalin> okay
[23:11:51] *** urbanape has quit IRC
[23:12:20] <akalin> where does the main socket pool live?
[23:12:25] <akalin> iothread somewhere?
[23:12:31] <willchan> HttpNetworkSession
[23:12:40] <willchan> HttpNetworkSession::tcp_socket_pool_ or something
[23:13:10] <willchan> and the HttpNetworkSession is located within the URLRequestContext, so you should pick the main URLRequestContext
[23:13:20] *** sbyer has left #chromium
[23:13:22] <willchan> as opposed to the incognito/media/etc ones
[23:15:07] <akalin> hmm okay
[23:15:15] <akalin> it looks like the main URLRequestContext lives on the UI thread?
[23:15:26] <willchan> nope
[23:15:28] <willchan> it's on the IO thread
[23:15:30] <akalin> just the getter?
[23:15:37] <akalin> okay, i see
[23:18:01] *** yaar has joined #chromium
[23:18:14] <akalin> so sockets don't necessarily have to be used on the IO thread, right>
[23:18:20] <akalin> just any thread with an IO message loop
[23:18:50] <akalin> sockets and their related classes
[23:18:50] *** lukas___ has quit IRC
[23:18:53] <willchan> just the getter is on the UI thread
[23:19:03] <akalin> oh, well, i guess the NetworkChangeNotifier needs to live on the IO thread, but that's taken care of
[23:19:04] <willchan> since the UI thread needs to access some URLRequestContext stuff
[23:19:04] *** loislo has quit IRC
[23:19:08] <akalin> willchan: okay
[23:19:09] <willchan> that needs to be created on the IO thread
[23:19:12] <willchan> but it might not exist yet
[23:19:22] <willchan> hence, it's a proxy object for the actual URLRequestContext
[23:19:33] <willchan> yes, the sockets don't necessarily have to be used on the IO thread
[23:19:37] <akalin> okay
[23:19:41] *** arv has quit IRC
[23:19:44] <akalin> cool, i think i can get started on this
[23:19:45] <akalin> thanks!
[23:19:49] <willchan> but the socket pools which "own" the thread all live on the IO thread
[23:19:55] <willchan> and so, if you release a thread to the pool
[23:19:57] <willchan> er, a socket
[23:19:58] <akalin> oh
[23:20:05] <willchan> then you have to make sure to do it on the IO thread
[23:20:13] <willchan> so you need to do proxying or something
[23:20:19] <willchan> or need to do the work on the IO thread
[23:20:20] <akalin> so i need to interact with the socket pool on the IO thread, basically
[23:20:31] <willchan> yep
[23:20:33] <Blockeaux> can I use chrome.extension.sendRequest to pass an array of strings
[23:20:37] <akalin> okay
[23:20:44] *** GeekShadow has quit IRC
[23:21:55] *** arm3n has quit IRC
[23:26:33] *** pkasting has quit IRC
[23:27:41] <Blockeaux> everything works fine when I send chrome.extension.sendRequest({greeting: "hello", other:"blah"}, function(response) {
[23:27:50] <Blockeaux> but nothing is recieved if i change one of those strings to an array
[23:28:38] *** dmaclach has quit IRC
[23:29:02] *** dmaclach has joined #chromium
[23:30:39] *** eseidel_ has joined #chromium
[23:31:04] *** pkasting has joined #chromium
[23:31:04] *** ChanServ sets mode: +v pkasting
[23:31:20] <akalin> ugh, death tests still don't work on OS X
[23:31:50] *** tonyg-cr has quit IRC
[23:33:06] <sky__> Anyone know how to reset one of the bots?
[23:33:15] <sky__> 10.6 tests dbg seems like it needs a reboot.
[23:33:24] <malavv> Is there an easy way to check every outgoing HTTP request without actually sniffing the network?
[23:33:38] <malavv> I need to see the actual headers
[23:34:00] <willchan> about:net-internals
[23:34:23] <willchan> and sniffing the network won't work for ssl
[23:34:41] <motownavi> akalin: I've used death tests in the oom tests
[23:34:42] *** eseidel has quit IRC
[23:34:42] *** eseidel_ is now known as eseidel
[23:34:50] <willchan> unless you know how to hack ssl, in which case, please share the details with me
[23:35:13] <akalin> motownavi: every time i encounter a death test it just hangs for me :(
[23:35:50] <malavv> willchan: hahah, I then sure will in a press conference
[23:36:05] <sky__> any troppers around?
[23:36:09] *** tonyg-cr has joined #chromium
[23:36:09] *** ChanServ sets mode: +v tonyg-cr
[23:37:37] <finnur> sky__: It probably works better if you ping them directly... nsylvain, maruel, ... ^^
[23:37:49] <malavv> willchan: I'm not sure where can I see the transaction headers in about:net-internals, I can see method and url of URL_REQUEST_START_JOB but nothing else
[23:38:15] <sky__> ya, I tried.
[23:38:24] <willchan> malavv: did you build from trunk/
[23:38:35] <sky__> all the troopers must be at the donut shop:)
[23:38:36] <malavv> yes fresh build
[23:38:48] <sky__> actually, brad just responded. He's looking.
[23:38:49] <willchan> malavv: open up about:net-internals and _then_ browse to the sites you're interested in
[23:39:04] <willchan> the new URL_REQUESTs will have the request and response headers
[23:39:13] <willchan> we don't log it all the time because the headers are not small
[23:39:21] <willchan> so, we only log them when net-internals is open
[23:39:42] <malavv> willchan: Thanks a lot, works perfectly
[23:39:43] <willchan> so the passively logged ones that initially populate net-internals won't have the headers
[23:41:09] <willchan> malavv: good.  i can't say that for the rest of my code :)
[23:42:03] <nsylvain> sky___ im me if you need anything
[23:42:08] *** dmazzoni has joined #chromium
[23:42:12] <nsylvain> I'm not watching irc right now
[23:42:13] <malavv> willchan: Best of chance for that part
[23:42:28] <jamesr> who owns the webkit canary perf bots?
[23:42:43] <jamesr> XP Perf (webkit.org) and Linux Perf (webkit.org)
[23:43:58] *** alokp has joined #chromium
[23:45:10] <Blockeaux> can anyone explain why chrome.extension.sendrequest only works with string literals?
[23:45:40] *** sbyer has joined #chromium
[23:46:39] *** duxklr has quit IRC
[23:48:40] <dmazzoni> Can anyone review a tiny webkit patch?
[23:52:01] *** iPac has quit IRC
[23:53:33] *** rohitrao has quit IRC
[23:54:44] *** kenneth_reitz has quit IRC
[23:55:44] <jamesr> dmazzoni: you'd have better luck asking that in #webkit i think
[23:57:10] *** cedricv has joined #chromium
[23:57:17] <dmazzoni> jamesr: ok, thanks
[23:57:25] *** dmazzoni has quit IRC
[23:58:23] *** roc has quit IRC
[23:58:46] <Blockeaux> How can I create a new tab and generate the HTML for it on the fly?
[23:59:04] <akalin> Blockeaux: you may have better luck with these questions on the extensions list
[23:59:26] *** sebmarkbage has joined #chromium
[23:59:28] <akalin> this channel is for development on chrome itself





top