/irc-logs / mozilla / #developers / 2011-12-14 / end
Options:
- # Session Start: Wed Dec 14 00:00:00 2011
- # Session Ident: #developers
- # [00:00] * Quits: dRdR (dRdR@moz-7B0110AD.mv.mozilla.com) (Ping timeout)
- # [00:01] <firebot> Check-in: http://hg.mozilla.org/mozilla-central/rev/b0165c0d85e1 - Brad Lassey - bug 710328 - null pointer when attaching android idle handler r=mfinkle a=java-only
- # [00:01] <blassey> note that kats is watching the tree for the 3 patches I just pushed
- # [00:01] <firebot> http://hg.mozilla.org/mozilla-central/rev/1f03a95b661a - Brad Lassey - bug 707929 - java.lang.NullPointerException @ ZipFile.getInputStream(ZipFile.java:256) r=mfinke a=java-only
- # [00:01] <firebot> http://hg.mozilla.org/mozilla-central/rev/923dab6343f9 - Brad Lassey - bug 708961 - We still use ByteBuffer.allocateDirect for large buffers r=pcwalton a=java-only
- # [00:02] * Joins: wlach_ (wlach@moz-C5B44C21.vif.net)
- # [00:02] * wlach_ is now known as wlach|mac
- # [00:02] <zpao> dolske: so i could set mFullScreen myself & call MakeFullScreen directly, but it looks like SetFullScreen is a bit more complete & makes sure to handle domfullscreen stuff exiting (again, i may be able to duplicate that)
- # [00:02] * Quits: dao (dao@moz-7C5E3898.superkabel.de) (Quit: Leaving.)
- # [00:02] * Quits: kaie (kaie@moz-2A92CE45.dip.t-dialin.net) (Quit: Leaving)
- # [00:03] * Quits: eflores (AndChat@538BABFE.A073F3E.97BBD552.IP) (Quit: Bye)
- # [00:03] * Joins: alex (alex@moz-BD8D0A09.lightspeed.sntcca.sbcglobal.net)
- # [00:03] * Joins: josh (josh@moz-2EE66546.nyc.res.rr.com)
- # [00:04] * Quits: Sander (chatzilla@moz-B871F4D3.direct-adsl.nl) (Quit: And back he spurred like a madman, shrieking a curse to the sky.)
- # [00:04] * Quits: Standard8 (Standard8@B7F1AE36.48015583.54C3481B.IP) (Quit: ZNC - http://znc.in)
- # [00:05] * Quits: clee (clee@moz-BBE3ABD.mv.mozilla.com) (Quit: clee)
- # [00:06] * Quits: jhk_ (jiggy@8E6C34C1.A3F9767A.1C37C358.IP) (Ping timeout)
- # [00:06] * philor|away is now known as philor
- # [00:06] * ashughes|meeting is now known as ashughes
- # [00:06] * Joins: smaug (chatzilla@moz-18AB320D.gprs.sl-laajakaista.fi)
- # [00:07] * khuey is now known as khuey|away
- # [00:08] * Quits: klugefoo (kahr@moz-F88DFDF5.lightspeed.sntcca.sbcglobal.net) (Ping timeout)
- # [00:08] * Quits: sheppy (sheppy@moz-F39D62DA.dhcp.kgpt.tn.charter.com) (Quit: sheppy)
- # [00:08] * Joins: mike5w3c (MikeS@moz-3F30ADA0.pool.e-mobile.ne.jp)
- # [00:08] * Quits: jprmc (jprmc@F2D29657.F60B0462.67AC9B1.IP) (Ping timeout)
- # [00:09] * Quits: Mossop (mossop@moz-C03D0C61.vlan426.asr1.sfo1.gblx.net) (Ping timeout)
- # [00:09] * Quits: masayuki (masayuki@moz-911CC660.zaq.ne.jp) (Ping timeout)
- # [00:09] * Joins: masayuki (masayuki@moz-911CC660.zaq.ne.jp)
- # [00:09] * Quits: Honza (chatzilla@E07BF19C.5BB5597D.D0083327.IP) (Ping timeout)
- # [00:10] <philor> coop|buildduty: it's not possible to see any more about what happened in https://tbpl.mozilla.org/php/getParsedLog.php?id=7912776&tree=Firefox than what made it into the log, is it?
- # [00:10] * Parts: mcmanus (mcmanus@moz-FE9B5BFD.twcny.res.rr.com)
- # [00:11] * Joins: masayuki_ (masayuki@moz-911CC660.zaq.ne.jp)
- # [00:11] * Joins: klugefoo (kahr@moz-F88DFDF5.lightspeed.sntcca.sbcglobal.net)
- # [00:11] * Quits: ericjung (Mibbit@5210CFD5.1A5EA44.72B23B3D.IP) (Quit: http://www.mibbit.com ajax IRC Client)
- # [00:11] * Quits: gerv (gerv@moz-8E68CF56.in-addr.arpa) (Quit: Leaving.)
- # [00:11] * Quits: masayuki (masayuki@moz-911CC660.zaq.ne.jp) (Ping timeout)
- # [00:11] * Joins: clee (clee@moz-BBE3ABD.mv.mozilla.com)
- # [00:12] * masayuki_ is now known as masayuki
- # [00:12] <coop|buildduty> philor: not really. i can connect to it and see if there's some obvious state
- # [00:13] * Joins: bsmith (bsmith@A2413EEC.695C1090.1F72B910.IP)
- # [00:13] <coop|buildduty> ...and there's not :/
- # [00:13] * catlee is now known as catlee-away
- # [00:13] <philor> yeah, my theory is log overflow, so it's not likely to leave much behind
- # [00:14] <dolske> zpao: I mean just looking at how that button and our code currently interact, and then cut'n'paste your way to success.
- # [00:14] * Joins: bholley (bholley@moz-FCAF9AAB.hsd1.ca.comcast.net)
- # [00:14] * Joins: bmoss (bmoss@moz-BBE3ABD.mv.mozilla.com)
- # [00:14] * bholley just realized he wasn't on IRC
- # [00:14] <dolske> since I'd assume both would want to work in similar ways. maybe I'm just being confusing, I've never looked at that code.
- # [00:14] <bholley> no wonder I was so productive
- # [00:14] <dolske> bholley: welcome back to IRC!
- # [00:15] * Quits: rillian (rilllian@moz-F023CC1E.thaumas.net) (Connection reset by peer)
- # [00:15] <zpao> dolske: that's what i was looking at. we literally just dispatch events from there
- # [00:15] * Quits: ajuma (ajuma@F2D29657.F60B0462.67AC9B1.IP) (Quit: ajuma)
- # [00:15] <zpao> dolske: we can make it work that way, just with a bit of duplication i think
- # [00:15] <firebot> Check-in: http://hg.mozilla.org/mozilla-central/rev/58a278555680 - Patrick Walton - Bug 710096 - Hold the monitor on the layer controller when adjusting and rendering the viewport metrics. r=Cwiiis a=java-only
- # [00:16] * Quits: TheLink (TheLink@moz-4FAE5288.pools.arcor-ip.net) (Client exited)
- # [00:16] * Quits: masayuki (masayuki@moz-911CC660.zaq.ne.jp) (Ping timeout)
- # [00:17] * Joins: dbaron (dbaron@moz-BBE3ABD.mv.mozilla.com)
- # [00:17] * ChanServ sets mode: +o dbaron
- # [00:17] * Quits: mdas (mdas@8CB764BC.D30B51A1.412CF160.IP) (Ping timeout)
- # [00:17] <philor> I'll just blame dbaron again
- # [00:17] <@dbaron> for what?
- # [00:18] <philor> Win M4 "disconnects"
- # [00:18] <philor> for which I still-or-again like blaming the size of the logs
- # [00:18] * Quits: stransky (stransky@moz-6F10C81D.net.upcbroadband.cz) (Quit: Connection reset by beer)
- # [00:19] <@dbaron> philor, didn't a patch substantially reduce the size of the logs a few weeks ago?
- # [00:20] <philor> it did indeed, and then they went away and we just had the M1 ones to deal with, until today
- # [00:20] <@dbaron> philor, did the logs recently get bigger?
- # [00:22] * shorlander is now known as shorlander-away
- # [00:23] <philor> dbaron: successful ones seem to still be the same compressed size, have to grab a couple to see about uncompressed, but I don't know what the unsuccessful ones look like, since they look like the master saying "hey, wasn't there a slave around here somewhere talking to me a while back?"
- # [00:23] * Joins: birtles (chatzilla@moz-348F61F0.mozilla.or.jp)
- # [00:24] * bnicholson is now known as bnicholson|afk
- # [00:24] * Quits: @mkaply (chatzilla@moz-92EDDD02.lightspeed.austtx.sbcglobal.net) (Ping timeout)
- # [00:24] * Joins: jwq (jwq@moz-8238A1D6.irl.cri.nz)
- # [00:25] <firebot> Check-in: http://hg.mozilla.org/mozilla-central/rev/9fae755f5103 - Brian R. Bondy - Bug 710079 - ICO files that contain PNG would end in some uninitialized memory. r=joe. a=joe
- # [00:25] <smaug> um
- # [00:25] <smaug> apparently geolocation spec is not reviewed
- # [00:25] * Quits: madhava (madhava@F2D29657.F60B0462.67AC9B1.IP) (Quit: madhava)
- # [00:25] <smaug> and it is only W3C CR :/
- # [00:25] <philor> blah, tinderbox-builds retention is four days, that won't tell me much
- # [00:26] <nthomas> eta for improving that is end of the week
- # [00:26] <nthomas> plus time to build history back up
- # [00:26] * Ziggy_Maes is now known as Ziggy|AWAY
- # [00:27] * Joins: dRdR (dRdR@moz-BBE3ABD.mv.mozilla.com)
- # [00:27] * Quits: dRdR_ (dRdR@moz-BBE3ABD.mv.mozilla.com) (Ping timeout)
- # [00:27] * jcranmer|away is now known as jcranmer
- # [00:28] * Joins: jhammink (textual@moz-BBE3ABD.mv.mozilla.com)
- # [00:28] <philor> just over half the size where they would get truncated, but I dunno either where the size where they'll start hanging Windows slaves is, or who might be intermittently spewing in a loop
- # [00:29] <philor> I still think 28MB is a bit heavy for a test log, but then my connection at home and my computer at work are probably the two weakest things around
- # [00:31] <Unfocused> nthomas: whats the retention changing to?
- # [00:31] <espindola> dolske, fell free to ping me if you need a more detailed description or have other questions about 702848
- # [00:31] * coop|buildduty is now known as coop|afk
- # [00:31] <nthomas> Unfocused: we had 30 days for tinderbox-builds, 14 for try before the array where dump the older stuff died
- # [00:32] * Joins: mkaply (chatzilla@moz-92EDDD02.lightspeed.austtx.sbcglobal.net)
- # [00:32] * ChanServ sets mode: +o mkaply
- # [00:32] <espindola> s/fell/feel/
- # [00:32] * Joins: gwagner (idefix2@moz-C03D0C61.vlan426.asr1.sfo1.gblx.net)
- # [00:32] <Unfocused> ah, right
- # [00:33] * Quits: not_gavin (gavin@D18F6DDB.6A393516.2321E71E.IP) (Connection reset by peer)
- # [00:33] * Joins: not_gavin (gavin@D18F6DDB.6A393516.2321E71E.IP)
- # [00:34] * Quits: bjarne (bjarne@moz-B436FB98.nextgentel.com) (Quit: Leaving.)
- # [00:34] * northWind is now known as northAway
- # [00:35] * Quits: mcsmurf (mcsmurf@moz-ACAF2E55.dip.t-dialin.net) (Quit: )
- # [00:35] <philor> and test_transitions_per_property is only 14MB of that 28MB
- # [00:35] <philor> alone
- # [00:35] <philor> for that single test
- # [00:38] * Quits: bmoss (bmoss@moz-BBE3ABD.mv.mozilla.com) (Ping timeout)
- # [00:38] <edmorley> yikes
- # [00:39] * Joins: mdas (mdas@8CB764BC.D30B51A1.412CF160.IP)
- # [00:39] <philor> it compresses well
- # [00:39] <philor> it's just big boned
- # [00:39] * Quits: sewardj (sewardj@moz-908CCEC7.dip.t-dialin.net) (Ping timeout)
- # [00:39] * jhopkins is now known as jhopkins|away
- # [00:40] * Quits: Hendikins (wolfox@moz-9A361C95.static.internode.on.net) (Ping timeout)
- # [00:40] * Joins: Hendikins (wolfox@moz-9A361C95.static.internode.on.net)
- # [00:41] * Quits: squib (squib-@moz-415BAA34.engr.wisc.edu) (Quit: Leaving)
- # [00:42] * gregglind is now known as gregglind_away
- # [00:44] * Quits: @mkaply (chatzilla@moz-92EDDD02.lightspeed.austtx.sbcglobal.net) (Ping timeout)
- # [00:45] * Quits: florian (florian@moz-87C33FDA.kimsufi.com) (Quit: Instantbird 1.2a1pre)
- # [00:45] * bwinton is now known as bwinton_away
- # [00:45] * Quits: gwagner (idefix2@moz-C03D0C61.vlan426.asr1.sfo1.gblx.net) (Quit: gwagner)
- # [00:45] * Quits: pnemsak (Miranda@moz-7409BD53.orange.sk) (Quit: pnemsak)
- # [00:48] * joduinn is now known as joduinn-mtg
- # [00:50] * Quits: sgimeno (santiago.g@A922400D.48362F18.BCC03455.IP) (Ping timeout)
- # [00:50] * jwir3 is now known as jwir3|away
- # [00:50] * Joins: sgimeno (santiago.g@A922400D.48362F18.BCC03455.IP)
- # [00:50] * Joins: larfdesk (Adam_Hinke@moz-F92153ED.longlines.com)
- # [00:50] * Joins: davidb (davidb@moz-976CD2.dsl.bell.ca)
- # [00:51] * Quits: gandalf (zbraniecki@moz-78F2D0A3.neoplus.adsl.tpnet.pl) (Quit: Textual IRC Client: http://www.textualapp.com/)
- # [00:51] * Quits: davidb (davidb@moz-976CD2.dsl.bell.ca) (Input/output error)
- # [00:55] * Parts: cpeterson (cpeterson@moz-C03D0C61.vlan426.asr1.sfo1.gblx.net)
- # [00:56] <gps> for those doing the HG <=> Git dance, Git 1.7.8 now supports direct importing of HG syntax patches
- # [00:57] <hub> nice
- # [00:58] * Joins: jprmc (jprmc@CA108B7E.5BCEC6DB.DA78B690.IP)
- # [00:59] * Quits: smaug (chatzilla@moz-18AB320D.gprs.sl-laajakaista.fi) (Ping timeout)
- # [01:00] * Quits: Steffen (chatzilla@moz-83564970.adsl.alicedsl.de) (Quit: ChatZilla 0.9.87 [Firefox 9.0a2/20110929020842])
- # [01:01] * Joins: jgoulie (jgoulie@moz-BBE3ABD.mv.mozilla.com)
- # [01:01] * Joins: bmoss (bmoss@moz-BBE3ABD.mv.mozilla.com)
- # [01:02] * Joins: bmoss|2 (bmoss@moz-BBE3ABD.mv.mozilla.com)
- # [01:03] * Quits: bmoss (bmoss@moz-BBE3ABD.mv.mozilla.com) (Ping timeout)
- # [01:05] * Quits: bmoss|2 (bmoss@moz-BBE3ABD.mv.mozilla.com) (Quit: bmoss|2)
- # [01:05] * Joins: bmoss (bmoss@moz-BBE3ABD.mv.mozilla.com)
- # [01:06] * joe|sheriff is now known as joe
- # [01:06] * Quits: mdas (mdas@8CB764BC.D30B51A1.412CF160.IP) (Quit: mdas)
- # [01:06] * Quits: jprmc (jprmc@CA108B7E.5BCEC6DB.DA78B690.IP) (Quit: Leaving)
- # [01:07] * Quits: smooney (smooney@moz-BBE3ABD.mv.mozilla.com) (Quit: smooney)
- # [01:09] * Quits: KaiRo_away (robert@moz-DB7D336E.adsl.highway.telekom.at) (Ping timeout)
- # [01:09] * Joins: KaIRC (robert@moz-8C499365.adsl.highway.telekom.at)
- # [01:10] * Quits: wlach|mac (wlach@moz-C5B44C21.vif.net) (Quit: Computer has gone to sleep.)
- # [01:10] * Joins: bmoss|2 (bmoss@moz-BBE3ABD.mv.mozilla.com)
- # [01:10] * Quits: bmoss|2 (bmoss@moz-BBE3ABD.mv.mozilla.com) (Quit: bmoss|2)
- # [01:11] * Quits: bmoss (bmoss@moz-BBE3ABD.mv.mozilla.com) (Quit: bmoss)
- # [01:11] * Joins: bmoss (bmoss@moz-BBE3ABD.mv.mozilla.com)
- # [01:11] * Quits: jhammel (jhammel@moz-C03D0C61.vlan426.asr1.sfo1.gblx.net) (Quit: leaving)
- # [01:14] * Quits: tH (Rob@cpc16-seac19-2-0-cust234.7-2.cable.virginmedia.com) (Ping timeout)
- # [01:14] * Quits: bmoss (bmoss@moz-BBE3ABD.mv.mozilla.com) (Quit: bmoss)
- # [01:15] * Joins: tH (Rob@cpc16-seac19-2-0-cust234.7-2.cable.virginmedia.com)
- # [01:15] * Quits: birtles (chatzilla@moz-348F61F0.mozilla.or.jp) (Connection reset by peer)
- # [01:16] * Joins: jhammel (jhammel@moz-C03D0C61.vlan426.asr1.sfo1.gblx.net)
- # [01:17] * Quits: AaronMT|dinner (AaronMT@moz-DEDA283.cpe.net.cable.rogers.com) (Quit: Textual IRC Client: http://www.textualapp.com/)
- # [01:18] * Quits: johnath_home (noyb@moz-5A10C3E8.dsl.teksavvy.com) (Ping timeout)
- # [01:18] * Joins: johnath_home (noyb@moz-5A10C3E8.dsl.teksavvy.com)
- # [01:18] * Quits: jfkthame (jfkthame@DBA4BEB5.90783722.9542EC20.IP) (Quit: jfkthame)
- # [01:20] * Joins: Honza (chatzilla@E07BF19C.5BB5597D.D0083327.IP)
- # [01:20] * zpao is now known as zpao|detached
- # [01:22] * Joins: smooney (smooney@moz-BBE3ABD.mv.mozilla.com)
- # [01:24] * Quits: robhawkes (robhawkes@moz-33A339B7.dsl.cnl.uk.net) (Quit: Leaving...)
- # [01:24] * Joins: clokep (clokep@moz-69FB3955.c3-0.smr-ubr1.sbo-smr.ma.cable.rcn.com)
- # [01:25] <firebot> Check-in: http://hg.mozilla.org/mozilla-central/rev/0c46aac6a23f - Matt Brubeck - Bug 710302 - going to about: pages does a 'feeling lucky' search [r=mfinkle a=java-only]
- # [01:25] <firebot> http://hg.mozilla.org/mozilla-central/rev/96a644cb3158 - Wes Johnston - Bug 697265 - Support clipboard operations in edit fields [r=mfinkle a=android-only]
- # [01:26] * Quits: ejpbruel (ejpbruel@moz-5EE20326.adsl2.static.versatel.nl) (Ping timeout)
- # [01:26] * Joins: jgilbert (jgilbert@moz-2B3CF81C.hsd1.ca.comcast.net)
- # [01:27] * alex is now known as alex|afk
- # [01:27] * Quits: jimm (jmathies@moz-7F164CA1.pn.at.cox.net) (Quit: )
- # [01:28] * ctalbert is now known as ctalbert|afk
- # [01:28] * Joins: lsumar (lsumar@538BABFE.A073F3E.97BBD552.IP)
- # [01:29] * Joins: ejpbruel (ejpbruel@933E3A2E.83A511AD.F413BF4D.IP)
- # [01:30] * Quits: anant (anant@moz-E31CD2CB.mozilla.org) (Ping timeout)
- # [01:31] * Quits: Pike (Pike@moz-AA98C591.pool.mediaways.net) (Quit: ChatZilla 0.9.87 [Firefox 8.0.1/20111120135848])
- # [01:31] * Quits: beaufour (beaufour@18D5CC88.C7EE4FB2.ECED8BE3.IP) (Quit: beaufour)
- # [01:32] <gps> am I on crack, or are my xpcshell test suites now printing process arguments for every test?
- # [01:32] <jcranmer> gps: no, khuey left debug in a patch
- # [01:33] <gps> jcranmer: you have a bug or file? I really want to nuke this locally
- # [01:34] <firebot> Check-in: http://hg.mozilla.org/mozilla-central/rev/ceffbd680e2c - Margaret Leibovic - Bug 707665 - Save as PDF option should be disabled for about:home and any XUL documents. r=mfinkle a=android-only
- # [01:36] * Quits: kumar (kmcmillan@moz-E865F8D8.hsd1.il.comcast.net) (Ping timeout)
- # [01:37] * Quits: ejpbruel (ejpbruel@933E3A2E.83A511AD.F413BF4D.IP) (Ping timeout)
- # [01:37] * Joins: kumar (kmcmillan@moz-E865F8D8.hsd1.il.comcast.net)
- # [01:37] * Joins: ejpbruel (ejpbruel@BE7AA7B8.AC20774A.78FB44E5.IP)
- # [01:38] * Joins: timeless_xchat (timeless@moz-5E2C9899.eng.wind.ca)
- # [01:38] * ewong|away is now known as ewong
- # [01:38] * Quits: tH (Rob@cpc16-seac19-2-0-cust234.7-2.cable.virginmedia.com) (Ping timeout)
- # [01:39] * Joins: tH (Rob@cpc16-seac19-2-0-cust234.7-2.cable.virginmedia.com)
- # [01:39] <jcranmer> just find 'print cmd' in testing/xpcshell/runxpcshelltests.py and delete them
- # [01:39] <NeilAway> gps: how does hg syntax differ from git syntax, particularly with reference to the --git option to hg diff...
- # [01:40] * Quits: timeless_xchat (timeless@moz-5E2C9899.eng.wind.ca) (Ping timeout)
- # [01:41] * Quits: ehsan (ehsan@F2D29657.F60B0462.67AC9B1.IP) (Input/output error)
- # [01:41] <gps> NeilAway: I think git will now accept HG patches w/o —git
- # [01:41] * Joins: timeless_xchat (timeless@moz-5E2C9899.eng.wind.ca)
- # [01:41] * Quits: akeybl (akeybl@moz-BBE3ABD.mv.mozilla.com) (Quit: Leaving...)
- # [01:42] * jcranmer is now known as jcranmer|away
- # [01:42] * Joins: akeybl (akeybl@moz-BBE3ABD.mv.mozilla.com)
- # [01:42] <timeless_xchat> neil: --git changes how new and binary files are handled
- # [01:42] * Quits: akeybl (akeybl@moz-BBE3ABD.mv.mozilla.com) (Input/output error)
- # [01:43] <ejpbruel> does this html snippet crash nightly for anyone else? http://pastebin.mozilla.org/1404656
- # [01:43] <NeilAway> timeless_xchat: sure, but what I was confused about was git apparently being unable to accept --git patches
- # [01:43] <timeless_xchat> !!
- # [01:44] <timeless_xchat> that really shouldn't happen
- # [01:44] <timeless_xchat> visit freenode / #mercurial?
- # [01:45] * Joins: wg9s (bill@moz-7A06A043.hsd1.ma.comcast.net)
- # [01:46] * Quits: mak (chatzilla@moz-D2B42BA6.retail.telecomitalia.it) (Quit: ChatZilla 0.9.86.1-rdmsoft [XULRunner 2.0/20110318052756])
- # [01:46] * Quits: jrmuizel (jrmuizel@F2D29657.F60B0462.67AC9B1.IP) (Input/output error)
- # [01:46] * Joins: mjschranz (mjschranz@C557B26E.531FD64C.630E4E47.IP)
- # [01:49] * Quits: mjschranz (mjschranz@C557B26E.531FD64C.630E4E47.IP) (Ping timeout)
- # [01:49] <ejpbruel> apparently its not possible to load news.google.com in an iframe?
- # [01:50] * Joins: ejpbruel_ (ejpbruel@933E3A2E.83A511AD.F413BF4D.IP)
- # [01:51] * Quits: ejpbruel (ejpbruel@BE7AA7B8.AC20774A.78FB44E5.IP) (Ping timeout)
- # [01:51] * ejpbruel_ is now known as ejpbruel
- # [01:53] <timeless_xchat> sounds right
- # [01:54] <timeless_xchat> one of greyhound or megabus tried to reframe sites for advertising
- # [01:54] <timeless_xchat> and news.google.com looped forever fighting it
- # [01:54] <timeless_xchat> probably greyhound
- # [01:54] * Quits: ewong (chatzilla@moz-E5D50C2E.static.netvigator.com) (Quit: Reboot)
- # [01:55] * Joins: bmoss (bmoss@moz-BBE3ABD.mv.mozilla.com)
- # [01:57] * Joins: mkaply (chatzilla@moz-92EDDD02.lightspeed.austtx.sbcglobal.net)
- # [01:57] * ChanServ sets mode: +o mkaply
- # [01:57] * Joins: Mossop (mossop@moz-C03D0C61.vlan426.asr1.sfo1.gblx.net)
- # [01:58] * Joins: mjschranz (mjschranz@117F2F8F.531FD64C.630E4E47.IP)
- # [01:58] * Quits: timeless_xchat (timeless@moz-5E2C9899.eng.wind.ca) (Ping timeout)
- # [01:59] * Joins: ehsan (ehsan@F0B20A8D.8458880F.57F33CED.IP)
- # [02:01] * Joins: srinivas (chatzilla@moz-2790C4FB.hsd1.ca.comcast.net)
- # [02:02] * Quits: Mossop (mossop@moz-C03D0C61.vlan426.asr1.sfo1.gblx.net) (Ping timeout)
- # [02:02] <ehsan> who wants to review a patch which lets us reopen the tree? ;)
- # [02:02] <ehsan> bsmedberg: ^
- # [02:02] <ehsan> khuey|away: ^
- # [02:02] <roc> I would if I could
- # [02:03] <gavin> what's the patch?
- # [02:03] * Joins: Enn (enn@moz-DB6467E3.cpe.net.cable.rogers.com)
- # [02:04] <darktrojan> he didn't actually claim the patch exists
- # [02:04] * Quits: mike5w3c (MikeS@moz-3F30ADA0.pool.e-mobile.ne.jp) (Ping timeout)
- # [02:04] * Joins: ewong (chatzilla@moz-E5D50C2E.static.netvigator.com)
- # [02:04] * Joins: mike5w3c (MikeS@moz-98B1BB1D.pool.e-mobile.ne.jp)
- # [02:05] * Joins: raccettura (raccettura@moz-660B8F4B.hsd1.nj.comcast.net)
- # [02:07] <ehsan> roc: https://bugzilla.mozilla.org/show_bug.cgi?id=710473
- # [02:07] <ehsan> roc: I'm gonna prepare to land it
- # [02:07] <ehsan> doubt that anyone would object ;)
- # [02:07] <roc> darktrojan: oh yeah. "who wants to review a patch to do XYZ?" "Me!" "OK, I guess I'd better write one then"
- # [02:07] * Joins: Pike (Pike@moz-E31CD2CB.mozilla.org)
- # [02:08] <darktrojan> exactly
- # [02:08] <ehsan> heh, I was filing the bug...
- # [02:08] <roc> is splinter broken for everyone or just me?
- # [02:08] <bent> wait, i thought the tree closure was due to PGOing too much code?
- # [02:08] <ehsan> bent: yes, I'm removing lots of code
- # [02:08] <derf> roc: Splinter's been working fine for me.
- # [02:08] <roc> bent: yeah it is
- # [02:09] <ehsan> now, who knows how we can force pgo builds using the selfserve API?
- # [02:10] * mcote|afk is now known as mcote
- # [02:10] <firebot> Check-in: http://hg.mozilla.org/mozilla-central/rev/221eccfa6a3f - Ehsan Akhgari - Bug 710473 - Backout the nsDerivedSafe part from bug 666414 to make PGO builds on Windows work again; r=roc (cause he volunteered) a=xul-diet
- # [02:10] <bent> it's at the bottom of the self serve page
- # [02:11] <ehsan> ah yes
- # [02:11] <ehsan> thanks
- # [02:11] <ehsan> who wants to volunteer to watch the tree and do the chores of reopening etc?
- # [02:11] * nthomas is now known as nthomas|away
- # [02:11] <bent> this makes me all sorts of sad
- # [02:12] <bent> seems to me that the linker should remove most of that derivedsafe stuff on the first link
- # [02:13] <ehsan> well, it probably needs to generate some internal data structures to analyze things etc
- # [02:13] * Quits: rwaldron (rwaldron@moz-BDCCF091.hfc.comcastbusiness.net) (NickServ (GHOST command used by rick))
- # [02:14] * Joins: rick (rwaldron@moz-CE27855E.hfc.comcastbusiness.net)
- # [02:14] * rick is now known as rwaldron
- # [02:14] * Quits: rwaldron (rwaldron@moz-CE27855E.hfc.comcastbusiness.net) (Quit: Leaving...)
- # [02:15] * Joins: aja_ (chatzilla@A8F7EE0.72086529.7880DB15.IP)
- # [02:15] * Quits: aja (chatzilla@A8F7EE0.72086529.7880DB15.IP) (Ping timeout)
- # [02:15] * aja_ is now known as aja
- # [02:16] * Quits: myk (myk@moz-C03D0C61.vlan426.asr1.sfo1.gblx.net) (Quit: Leaving.)
- # [02:17] * aja is now known as IRCMonkey36567
- # [02:17] * Quits: bmoss (bmoss@moz-BBE3ABD.mv.mozilla.com) (Ping timeout)
- # [02:17] * Quits: lsumar (lsumar@538BABFE.A073F3E.97BBD552.IP) (Quit: This computer has gone to sleep)
- # [02:17] * Quits: redfive (chatzilla@moz-B94E1A81.ipnetworksinc.net) (Ping timeout)
- # [02:18] <@dbaron> Hmmm... it seems like I can't load about: URLs in today's Fennec nightly... I end up at a google search for about:crashes or about:fennec instead.
- # [02:18] <gavin> known bug
- # [02:18] <gavin> dbaron: bug 710302
- # [02:19] <@dbaron> I wonder if there's a workaround...
- # [02:19] * Quits: jgilbert (jgilbert@moz-2B3CF81C.hsd1.ca.comcast.net) (Connection reset by peer)
- # [02:19] <darktrojan> figure out the chrome url and go there?
- # [02:20] <@dbaron> well, view-source:about:crashes also goes to google
- # [02:20] <@dbaron> I guess I'll find my crash tomorrow
- # [02:20] <@dbaron> on the off chance that I actually got it reported
- # [02:20] * Joins: jgilbert (jgilbert@moz-2B3CF81C.hsd1.ca.comcast.net)
- # [02:21] <gavin> dbaron: chrome://global/content/crashes.xhtml ?
- # [02:21] <@dbaron> gavin, thanks
- # [02:22] <edmorley> !seen bjacob
- # [02:22] <firebot> bjacob was last seen 3 hours, 15 minutes and 15 seconds ago, saying 'dRdR: actually LoadSymbols is just at the beginning of GLContext.cpp... it doesn't seem to want to stop on null function name. maybe step through it' in #gfx.
- # [02:23] <espindola> interesting that removing some template magic makes xul link again
- # [02:23] <espindola> shows that the IL used by msvc is fairly high level
- # [02:24] * Quits: jgoulie (jgoulie@moz-BBE3ABD.mv.mozilla.com) (Quit: jgoulie)
- # [02:24] <espindola> bsmedberg, ehsan: congratulations!
- # [02:24] <ehsan> the credit should go to bsmedberg really
- # [02:25] * Quits: ashughes (ashughes@21B7B9F2.B87E9213.6E712CE2.IP) (Quit: ashughes)
- # [02:25] * Quits: nthomas|away (chatzilla@moz-75DA4CCD.dsl.telstraclear.net) (Ping timeout)
- # [02:25] * Quits: kumar (kmcmillan@moz-E865F8D8.hsd1.il.comcast.net) (Quit: kumar)
- # [02:26] * Joins: jrmuizel (jrmuizel@moz-20EF8EAA.cpe.net.cable.rogers.com)
- # [02:26] * Joins: micahg_ (micahg@moz-60E06D1E.c3-0.arm-ubr1.chi-arm.il.cable.rcn.com)
- # [02:26] * Quits: micahg (micahg@moz-60E06D1E.c3-0.arm-ubr1.chi-arm.il.cable.rcn.com) (Connection reset by peer)
- # [02:26] * Quits: Pike (Pike@moz-E31CD2CB.mozilla.org) (Quit: ChatZilla 0.9.87 [Firefox 8.0.1/20111120135848])
- # [02:26] <espindola> and now I am curious is msvc does better by doing more at compile time (instead of link time) :-)
- # [02:27] * Quits: Boriss (FlyingToas@moz-C03D0C61.vlan426.asr1.sfo1.gblx.net) (Quit: Boriss)
- # [02:27] <espindola> s/is msvc/ if msvc 2010/
- # [02:27] <@dbaron> Waldo, so now that we've removed nsDerivedSafe again, we should probably add back a bunch of those 'final' annotations
- # [02:28] <gps> I have an untested theory that removing /OPT:NOREF from link.exe will cause libxul to link again. but, this would result in extra symbols cluttering the binary, so it is unsuitable for shipping
- # [02:28] <taras> ehsan: have people looked at mapfile to see what symbols are taking up most space on windows?
- # [02:29] * Joins: nthomas|away (chatzilla@moz-75DA4CCD.dsl.telstraclear.net)
- # [02:29] * jlebar is now known as jlebar|away
- # [02:29] <Waldo> dbaron: yeah, probably; there is that bug filed, listing them
- # [02:30] * Quits: mcote (mcote@moz-7CCC926E.mc.videotron.ca) (Quit: leaving)
- # [02:30] * Joins: mcote (mcote@moz-7CCC926E.mc.videotron.ca)
- # [02:31] <gps> of course size isn't the only consideration: it's how you use the symbols that's really important ;)
- # [02:31] * Joins: gal (gal@A2413EEC.695C1090.1F72B910.IP)
- # [02:31] <taras> espindola: where are you seeing update on removing template magic?
- # [02:32] <espindola> taras, the patch ehsan just landed
- # [02:32] <espindola> if I understand the question....
- # [02:33] * Quits: clee (clee@moz-BBE3ABD.mv.mozilla.com) (Quit: clee)
- # [02:33] <taras> espindola: thanks
- # [02:33] * Quits: damons (gnubeard@moz-BBE3ABD.mv.mozilla.com) (Quit: damons)
- # [02:33] <@bz_away> man
- # [02:33] <taras> and yeah that's sneaky
- # [02:33] <@bz_away> in WebKit, HTMLCollection.item and NodeList.item behave differently
- # [02:34] <gps> are the extra few instructions on cross-library calls actually noticed by talos if you break sizable chunks out of libxul? has anyone tried moving pieces out of libxul?
- # [02:34] * Joins: squib (squib@moz-F5CA0CFB.dhcp.mdsn.wi.charter.com)
- # [02:34] * Quits: IRCMonkey36567 (chatzilla@A8F7EE0.72086529.7880DB15.IP) (Ping timeout)
- # [02:35] <cpearce> bz: make any progress on bug 701662?
- # [02:36] <gps> I for one wouldn't mind a build option that sacrificed a monolithic libxul and slightly less performance for linking speed (when developing)
- # [02:36] * Joins: aja (chatzilla@A8F7EE0.72086529.7880DB15.IP)
- # [02:38] * joduinn-mtg is now known as joduinn
- # [02:38] * Quits: Honza (chatzilla@E07BF19C.5BB5597D.D0083327.IP) (Ping timeout)
- # [02:38] * Quits: ehsan (ehsan@F0B20A8D.8458880F.57F33CED.IP) (Input/output error)
- # [02:39] * Quits: hub (hub@21B7B9F2.B87E9213.6E712CE2.IP) (Ping timeout)
- # [02:39] * Quits: jhammel (jhammel@moz-C03D0C61.vlan426.asr1.sfo1.gblx.net) (Quit: hasta)
- # [02:39] <cpearce> oh, bz's not online... who would'a thought?
- # [02:39] * Quits: @mkaply (chatzilla@moz-92EDDD02.lightspeed.austtx.sbcglobal.net) (Ping timeout)
- # [02:42] * juanb is now known as juanb|afk
- # [02:42] * heycam is now known as heycam|away
- # [02:43] * alex|afk is now known as alex
- # [02:43] * Quits: tH (Rob@cpc16-seac19-2-0-cust234.7-2.cable.virginmedia.com) (Ping timeout)
- # [02:44] <dholbert> cpearce, you mean bz_away?
- # [02:44] * Quits: yuan (ywang@moz-C03D0C61.vlan426.asr1.sfo1.gblx.net) (Quit: yuan)
- # [02:44] <dholbert> cpearce, he's online, just away :)
- # [02:45] * Quits: @dbaron (dbaron@moz-BBE3ABD.mv.mozilla.com) (Quit: 8403864 bytes have been tenured, next gc will be global.)
- # [02:46] * Joins: jgoulie (jgoulie@moz-F3CABD3.tmodns.net)
- # [02:46] * Quits: mjschranz (mjschranz@117F2F8F.531FD64C.630E4E47.IP) (Ping timeout)
- # [02:46] * Quits: Enn (enn@moz-DB6467E3.cpe.net.cable.rogers.com) (Ping timeout)
- # [02:47] * Quits: micahg_ (micahg@moz-60E06D1E.c3-0.arm-ubr1.chi-arm.il.cable.rcn.com) (Ping timeout)
- # [02:47] * mcote is now known as mcote|afk
- # [02:47] * bnicholson|afk is now known as bnicholson
- # [02:48] * Quits: gal (gal@A2413EEC.695C1090.1F72B910.IP) (Ping timeout)
- # [02:48] <dolske> I've held conversations with both bz_away and bz_sleep in the past. :)
- # [02:49] * Quits: smooney (smooney@moz-BBE3ABD.mv.mozilla.com) (Quit: smooney)
- # [02:49] <dolske> though, technically for the latter he could be sleep-ircing. That's a little scary. :)
- # [02:50] * Quits: jgoulie (jgoulie@moz-F3CABD3.tmodns.net) (Quit: jgoulie)
- # [02:51] <mwu> and for the former, he could've scripted responses to all the questions he knew you were going to ask
- # [02:52] * Joins: micahg_ (micahg@moz-C7CC8D27.c3-0.arm-ubr1.chi-arm.il.cable.rcn.com)
- # [02:52] * Quits: micahg_ (micahg@moz-C7CC8D27.c3-0.arm-ubr1.chi-arm.il.cable.rcn.com) (Client exited)
- # [02:53] * Joins: micahg (micahg@moz-C7CC8D27.c3-0.arm-ubr1.chi-arm.il.cable.rcn.com)
- # [02:54] * Quits: sfink (chatzilla@moz-BBE3ABD.mv.mozilla.com) (Ping timeout)
- # [02:57] * nthomas|away is now known as nthomas
- # [02:57] * shorlander-away is now known as shorlander
- # [02:57] * joduinn is now known as joduinn-afk
- # [02:58] * davehunt is now known as davehunt|away
- # [02:58] * Quits: joduinn-afk (joduinn@moz-C03D0C61.vlan426.asr1.sfo1.gblx.net) (Quit: joduinn-afk)
- # [02:59] * Joins: Enn (enn@moz-DB6467E3.cpe.net.cable.rogers.com)
- # [03:01] * Quits: rniwa (rniwa@5CA6DC39.C60FE7DC.4065847B.IP) (Quit: rniwa)
- # [03:02] * Quits: diogogmt (kvirc@moz-4D628198.cpe.net.cable.rogers.com) (Ping timeout)
- # [03:03] * Quits: Mook_as (mook@moz-1FCC0032.activestate.com) (Quit: ChatZilla 0.9.87-rdmsoft [XULRunner 1.9.2.13/20101203074205])
- # [03:03] * Joins: diogogmt (kvirc@moz-4D628198.cpe.net.cable.rogers.com)
- # [03:04] * Quits: Enn (enn@moz-DB6467E3.cpe.net.cable.rogers.com) (Ping timeout)
- # [03:04] * Joins: mjschranz (mjschranz@117F2F8F.531FD64C.630E4E47.IP)
- # [03:05] * Joins: gwagner (idefix2@moz-3DFE0B51.hsd1.ca.comcast.net)
- # [03:06] * Quits: peterv (peterv@moz-715F6D16.access.telenet.be) (Ping timeout)
- # [03:06] * Quits: gmoro (guilherme@3DCBFB93.F520FC74.D159334F.IP) (Ping timeout)
- # [03:07] * Joins: Boriss (FlyingToas@moz-62AAA429.hsd1.ca.comcast.net)
- # [03:09] * Quits: jduell (jduell@moz-2D9EDA98.lightspeed.sntcca.sbcglobal.net) (Ping timeout)
- # [03:12] * timA is now known as timA|away
- # [03:12] * Quits: jet (junglecode@moz-C03D0C61.vlan426.asr1.sfo1.gblx.net) (Quit: jet)
- # [03:12] * Joins: lsumar (lsumar@538BABFE.A073F3E.97BBD552.IP)
- # [03:15] * dolske patiently wait for someone to claim a their reward...
- # [03:15] <dolske> firebot: bug 701655
- # [03:15] <firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=701655 nor, --, ---, nobody, NEW, Have some sort of reward for landing a changeset which gets a full set of green tests
- # [03:16] * Joins: dbaron (dbaron@moz-27F87443.dsl.dynamic.sonic.net)
- # [03:16] * ChanServ sets mode: +o dbaron
- # [03:16] <darktrojan> ooh shiny
- # [03:17] * Joins: sworkman-phone (sworkman@moz-BBE3ABD.mv.mozilla.com)
- # [03:17] <mcpherrin> dolske: I'm motivated by shiny trophies!
- # [03:17] <dolske> then start landing patches! :)
- # [03:18] <darktrojan> it's a bit sad that after a month we only have 3 winners :(
- # [03:18] <darktrojan> we're losing the WOO
- # [03:19] <dRdR> I should land a patch that fixes all intermittents all in that one changeset
- # [03:19] <njn> darktrojan: no war on an abstract noun has ever been won
- # [03:19] <philor> ITYM the WOA
- # [03:19] <philor> if we could just defeat Android, we could manage the rest
- # [03:19] * wg9s thinks he could do this (but only by cheating) used to make the network monitor at work go all red (to panic the people) by modifying the color map on the display
- # [03:19] <darktrojan> heh
- # [03:20] * Joins: Enn (enn@moz-DB6467E3.cpe.net.cable.rogers.com)
- # [03:20] <darktrojan> dolske, why do the trophies have stars on them? an all-green push needs no stars!
- # [03:20] * Quits: bholley (bholley@moz-FCAF9AAB.hsd1.ca.comcast.net) (Quit: bholley)
- # [03:21] * Quits: sworkman-phone (sworkman@moz-BBE3ABD.mv.mozilla.com) (Quit: Bye-bye!)
- # [03:21] * Quits: espindola (espindola@moz-10752211.dsl.teksavvy.com) (Quit: Ex-Chat)
- # [03:21] * Quits: Ami_Ty (Amie@moz-BBE3ABD.mv.mozilla.com) (Quit: Ami_Ty)
- # [03:21] * Quits: gkw (gkw@moz-BBE3ABD.mv.mozilla.com) (Quit: Instantbird 1.1)
- # [03:21] * Joins: damons (gnubeard@moz-A41E6911.hsd1.ca.comcast.net)
- # [03:21] <dolske> in liberal mountain view, trophies star you!
- # [03:21] <wg9s> Hmmm. Good point!
- # [03:21] * lsblakk is now known as lsblakk|afk
- # [03:22] * dolske crosses darktrojan off the elegibility list. :P
- # [03:22] <darktrojan> :(
- # [03:24] * Quits: CwiiisAway (cwiiis@moz-F15E698.croy.cable.virginmedia.com) (Ping timeout)
- # [03:25] * Quits: asac (asac@moz-584E28C9.pppoe.wtnet.de) (Ping timeout)
- # [03:25] * Quits: jgriffin (jgriffin@moz-4FBFA41D.hsd1.wa.comcast.net) (Quit: jgriffin)
- # [03:25] * Joins: eflores (AndChat@538BABFE.A073F3E.97BBD552.IP)
- # [03:27] * Joins: ashughes (ashughes@moz-4C407EE8.vc.shawcable.net)
- # [03:29] * Quits: ashughes (ashughes@moz-4C407EE8.vc.shawcable.net) (Quit: ashughes)
- # [03:30] * Quits: damons (gnubeard@moz-A41E6911.hsd1.ca.comcast.net) (Quit: damons)
- # [03:30] * KaIRC is now known as KaiRo
- # [03:31] * Joins: damons (gnubeard@moz-A41E6911.hsd1.ca.comcast.net)
- # [03:31] * Joins: bmoss (bmoss@moz-BBE3ABD.mv.mozilla.com)
- # [03:32] * heycam|away is now known as heycam
- # [03:34] * Quits: bmoss (bmoss@moz-BBE3ABD.mv.mozilla.com) (Quit: bmoss)
- # [03:34] * Joins: rshetty (quassel@6F1A83E2.CC63714F.C842849F.IP)
- # [03:34] * coop|afk is now known as coop
- # [03:35] * Quits: azakai (alon@moz-BBE3ABD.mv.mozilla.com) (Ping timeout)
- # [03:36] * Quits: mjschranz (mjschranz@117F2F8F.531FD64C.630E4E47.IP) (Client exited)
- # [03:36] * Quits: ejpbruel (ejpbruel@933E3A2E.83A511AD.F413BF4D.IP) (Quit: ejpbruel)
- # [03:37] * Quits: damons (gnubeard@moz-A41E6911.hsd1.ca.comcast.net) (Quit: damons)
- # [03:37] * Joins: Jake (Jake@moz-5D28CC35.phlapafg.dynamic.covad.net)
- # [03:39] * Quits: bsmith (bsmith@A2413EEC.695C1090.1F72B910.IP) (Ping timeout)
- # [03:40] * Quits: mccr8 (mccr8@moz-BBE3ABD.mv.mozilla.com) (Quit: mccr8)
- # [03:40] * Joins: ashughes (ashughes@moz-4C407EE8.vc.shawcable.net)
- # [03:41] * Quits: ashughes (ashughes@moz-4C407EE8.vc.shawcable.net) (Quit: ashughes)
- # [03:42] * Quits: chewey (chewey@moz-576C0852.dip0.t-ipconnect.de) (NickServ (GHOST command used by chewey_))
- # [03:42] * Joins: chewey (chewey@moz-BA1DC9B1.dip0.t-ipconnect.de)
- # [03:44] * Quits: nhirata (nhirata.bu@moz-BBE3ABD.mv.mozilla.com) (Ping timeout)
- # [03:44] <edmorley> what was ehsan's plan with https://hg.mozilla.org/mozilla-central/rev/221eccfa6a3f ?
- # [03:45] <edmorley> get someone to merge to inbound, trigger pgo and see if that did indeed fix things?
- # [03:46] <lurking> edmorley: looks like m-c tbpl already has a pgo build spinning
- # [03:47] <lurking> no idea what happens though should pgo go 'green'
- # [03:47] <edmorley> m-c doesn't have SPDY relanded or graphite preffed back on (that only happened on inbound), so PGO is already green there
- # [03:48] <lurking> oh, crap - indeed - glad you guys are keeping score - getting very confusing :)
- # [03:48] <philor> edmorley: scrollback doesn't look like he had a (workable) plan, I vote for you merging him to inbound
- # [03:48] <edmorley> sure :-)
- # [03:48] * Joins: bsmith (bsmith@E1AD879F.CC98CBF7.C589985.IP)
- # [03:48] <philor> thanks, mate, you're a rock
- # [03:48] <lurking> wonder why he didn't just push it to inbound ?
- # [03:48] * Joins: Honza (chatzilla@E07BF19C.5BB5597D.D0083327.IP)
- # [03:50] * lurking goes to bed - hopes to wake up to an 'open tree' ...
- # [03:50] <edmorley> because he knew I was dying for another excuse to make myself late to bed again :-)
- # [03:51] * Quits: bnicholson (bnicholson@moz-BBE3ABD.mv.mozilla.com) (Quit: Leaving)
- # [03:51] * Joins: bnicholson (bnicholson@moz-D1AB96C1.hsd1.ca.comcast.net)
- # [03:52] * Joins: jprmc (jprmc@moz-7F2FF3EB.cpe.net.cable.rogers.com)
- # [03:52] <wg9s> edmorley: I was going to ask the same question in the bug, but I figured I just was not privy to the plan. Guess No one else was either though.
- # [03:53] <wg9s> It would have seemed to me to make more sense to land that on inbound and do 6 pgo builds to see if they go green.
- # [03:53] <wg9s> just my kind of agreeing with you I guess.
- # [03:54] * Quits: Honza (chatzilla@E07BF19C.5BB5597D.D0083327.IP) (Ping timeout)
- # [03:54] <edmorley> it's no biggie (once TortoiseHg stops chewing up a whole core and lets me at the repo without worrying I'm going to break something by just killing it)
- # [03:54] <wg9s> I was also confused by the can someone post a patch to back out skia because it seemed to me it all alnded in one bug one patch.
- # [03:54] <wg9s> or were there followup thinkgs i missed somehow?
- # [03:55] * Joins: RyanVM (chatzilla@moz-D04D3C77.phlapa.fios.verizon.net)
- # [03:55] <edmorley> think there were dribs and drabs after
- # [03:55] * Joins: birtles (chatzilla@moz-348F61F0.mozilla.or.jp)
- # [03:55] * philor is now known as philor|away
- # [03:55] <edmorley> should be easily prefable via the makefiles though, since there are already switches per platform
- # [03:55] <edmorley> iirc
- # [03:55] <RyanVM> fyi, second-pass linking xul.dll right now off m-c tip and a vanilla .mozconfig
- # [03:56] <RyanVM> will post results to bug 709193 when done
- # [03:56] * Joins: abwillis (abwillis@moz-87B76227.hlrn.qwest.net)
- # [03:56] <edmorley> RyanVM: thanks for your stats so far, they've been helpful
- # [03:56] <RyanVM> np
- # [03:56] <edmorley> RyanVM: unfortunately m-c doesn't have SDPY relanded, inbound does
- # [03:56] <RyanVM> oh well
- # [03:57] <edmorley> so it's really inbound (with before and after ehsan's landing, which I'm about to merge), that will be extremely interesting
- # [03:57] <RyanVM> at least I've got the --enable-optimize dropped now, so it should be a better representation of what's actually being built
- # [03:57] <edmorley> :-)
- # [03:57] <aja> did SPDY get a -disable flag before relanding?
- # [03:57] <RyanVM> (and my various --disables are gone)
- # [03:57] * aja is just curious
- # [03:57] <@bz_away> aja: in what sense?
- # [03:57] * Joins: regen (Miller@moz-626FF3F1.adsl.dynamic.seed.net.tw)
- # [03:57] * RyanVM wishes skia was easier to disable
- # [03:58] <edmorley> aja: no, the relanding was just a backout of the backout cset
- # [03:58] <aja> bz_away: as in a build-time flag
- # [03:58] <@bz_away> ah
- # [03:58] <@bz_away> why would we want one?
- # [03:58] * bz_away is now known as bz
- # [03:58] <@bz> modulo the PGO mess, of course
- # [03:58] <RyanVM> edmorley: unfortunately, it takes about 3h to do a PGO build on my system and I need to go to bed relatively soon, so I won't be able to fire another one off tonight
- # [03:59] <aja> just cuz of working the 3M line
- # [03:59] <nthomas> so we an actually disable things in aurora or later ?
- # [03:59] <RyanVM> especially since I don't know of a way to actually log the peak vbyte usage
- # [03:59] <wg9s> edmorley: because someone said that was the biggest change in size in libxul under sindows and the bug seemed to be to turn on backend code and formt he comments seemed to have no real direction under when or if this code woudl ever be used so seems a prime candidate to backout with no impact on any feature schedule.
- # [03:59] <RyanVM> I just have perfmon running
- # [03:59] * Quits: Jesse (jruderman@moz-BBE3ABD.mv.mozilla.com) (Quit: Jesse)
- # [04:00] <regen> guys, can I ask something...
- # [04:00] <@bz> nthomas: well, it's behind a pref
- # [04:00] <@bz> nthomas: so we can disable its operation; just not its compilation
- # [04:00] <regen> I received a email from Mozilla about Apps developer preview...is this a scam?
- # [04:00] <edmorley> RyanVM: even m-c will be useful to know at this point - at least we'll know for sure how much SPDY/graphite adds
- # [04:01] <edmorley> regen: no
- # [04:01] <RyanVM> edmorley: and it will help assess the impact of ehsan's backout
- # [04:01] <regen> ts
- # [04:01] * Quits: jwq (jwq@moz-8238A1D6.irl.cri.nz) (Quit: ChatZilla 0.9.87 [Firefox 8.0/20111104165243])
- # [04:02] * Joins: jdm (jdm@moz-9AEDE212.cpe.net.cable.rogers.com)
- # [04:02] <edmorley> regen: https://groups.google.com/forum/#!topic/mozilla.dev.apps.firefox/69cuVaTue2E
- # [04:02] <edmorley> :-)
- # [04:02] <regen> 'oh my god
- # [04:02] <regen> it got suspicious because it asks me not to share links with others
- # [04:02] <regen> panzi scheme? @_@
- # [04:04] <wg9s> well I am suspicious becuase it says I got it becuase I am a registered Mozillain, but if I am a registered Mozillian, how come I can not do a search for other mozillians that returns anyone other than myself?
- # [04:04] <edmorley> regen: it's a limited preview outside of mozillians, hence having a separate link to ensure mozillians don't get left out
- # [04:04] <wg9s> Either the search is defective or I am ony half registered somehow.
- # [04:04] <regen> ok
- # [04:05] <edmorley> wg9s: file a bug, have heard others say similar
- # [04:05] <regen> edmorley: why not call filing a bug "to file a ticket"?
- # [04:05] <wg9s> I somehow doubt I am the only person registerd with a first name of Bill.
- # [04:05] * Quits: Enn (enn@moz-DB6467E3.cpe.net.cable.rogers.com) (Ping timeout)
- # [04:06] <edmorley> regen: for whatever reason "bug" is used on bugzilla over "ticket", to the extent that seeing ticket is unusual
- # [04:06] <edmorley> dammit closed tree hook on inbound
- # [04:06] <regen> ok
- # [04:06] <wg9s> And also that searches for aakash desai would fail.
- # [04:07] * Joins: ctopper (craig@10F5C408.23A18F00.F9E1BAE1.IP)
- # [04:07] * Quits: jhammink (textual@moz-BBE3ABD.mv.mozilla.com) (Quit: Computer has gone to sleep.)
- # [04:08] * Joins: jhammink (textual@moz-7B0110AD.mv.mozilla.com)
- # [04:08] * Quits: Jake (Jake@moz-5D28CC35.phlapafg.dynamic.covad.net) (Quit: Leaving.)
- # [04:09] * Quits: Boriss (FlyingToas@moz-62AAA429.hsd1.ca.comcast.net) (Quit: Boriss)
- # [04:10] * edmorley facepalm
- # [04:10] <edmorley> ehsan's try run was from m-c
- # [04:11] <edmorley> guess we'll see if it works in a bit :-)
- # [04:11] <RyanVM> edmorley: peak vbyte usage so far - 2.89GB :(
- # [04:11] <RyanVM> not done yet either
- # [04:12] * Joins: peterv (peterv@moz-715F6D16.access.telenet.be)
- # [04:12] <RyanVM> so maybe a bit better (could be the lack of --enable-optimize this time too)
- # [04:12] * Quits: jhammink (textual@moz-7B0110AD.mv.mozilla.com) (Quit: Computer has gone to sleep.)
- # [04:13] * Joins: damons (gnubeard@moz-A41E6911.hsd1.ca.comcast.net)
- # [04:13] * Joins: ashughes (ashughes@moz-4C407EE8.vc.shawcable.net)
- # [04:14] * Quits: timA|away (tabraldes@moz-535753DA.hsd1.or.comcast.net) (Ping timeout)
- # [04:14] * Quits: KaiRo (robert@moz-8C499365.adsl.highway.telekom.at) (Input/output error)
- # [04:14] * Quits: sworkman (sworkman@moz-BBE3ABD.mv.mozilla.com) (Ping timeout)
- # [04:14] * mcote|afk is now known as mcote
- # [04:17] * Quits: bnicholson (bnicholson@moz-D1AB96C1.hsd1.ca.comcast.net) (Quit: Leaving)
- # [04:17] * Quits: aja (chatzilla@A8F7EE0.72086529.7880DB15.IP) (Client exited)
- # [04:18] <RyanVM> Bas++
- # [04:18] <Bas> RyanVM: Hmm?
- # [04:19] <RyanVM> Skia patch :)
- # [04:19] <RyanVM> it's like you were reading my mind :P
- # [04:19] <Bas> Oh, heh, thanks :)
- # [04:19] <Bas> Khuey asked actually :)
- # [04:19] * Quits: ctopper (craig@10F5C408.23A18F00.F9E1BAE1.IP) (Quit: ctopper)
- # [04:19] <Bas> So it was his idea.
- # [04:20] <RyanVM> I'll take some credit for asking for it in bug 709193 :-P
- # [04:20] * Quits: jorendorff (jorendorff@moz-91590D94.hsd1.tn.comcast.net) (Quit: Eaten by grue.)
- # [04:22] * jcranmer|away is now known as jcranmer
- # [04:22] <roc> what happened to tbpl?
- # [04:22] * Quits: regen (Miller@moz-626FF3F1.adsl.dynamic.seed.net.tw) (Quit: regen)
- # [04:23] * Joins: Jesse (jruderman@moz-537BCF9.hsd1.ca.comcast.net)
- # [04:23] <RyanVM> edmorley: if bug 710509 doesn't land before tomorrow, I'll push it locally on my next run
- # [04:24] <edmorley> roc: wfm ..?
- # [04:25] * Quits: mayhemer (Miranda@B3D46202.F87A741B.F23860FD.IP) (Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org)
- # [04:25] * Joins: rniwa (rniwa@11828247.C8C6FF5E.4065847B.IP)
- # [04:26] * Quits: cpearce (chatzilla@538BABFE.A073F3E.97BBD552.IP) (Ping timeout)
- # [04:28] * Quits: tfair (tfairey@moz-BBE3ABD.mv.mozilla.com) (Client exited)
- # [04:28] <roc> very much not WFM
- # [04:30] * darktrojan has forgotten his mdn password
- # [04:30] <darktrojan> :/
- # [04:31] <edmorley> try's busy
- # [04:32] <gavin> darktrojan: isn't that firefox's job?
- # [04:33] * philor|away is now known as philor
- # [04:33] <darktrojan> probably
- # [04:33] <mcpherrin> gavin: isn't the sync server in charge of forgetting passwords?
- # [04:34] * Quits: rjohnson19 (rjohnson19@moz-9148485F.hsd1.ma.comcast.net) (Quit: Leaving)
- # [04:36] <RyanVM> edmorley: 2.91GB
- # [04:36] <darktrojan> now I can't remember what I was going to edit
- # [04:36] <darktrojan> heh
- # [04:36] * Joins: mjschranz (mjschranz@117F2F8F.531FD64C.630E4E47.IP)
- # [04:37] * Quits: Mardak (Mardak@moz-BBE3ABD.mv.mozilla.com) (Quit: Mardak)
- # [04:37] <@bz> roc: flushing less is apparently Very Hard... ;)
- # [04:38] * Quits: ashughes (ashughes@moz-4C407EE8.vc.shawcable.net) (Quit: ashughes)
- # [04:38] <@bz> roc: got a sec?
- # [04:39] * Quits: wg9s (bill@moz-7A06A043.hsd1.ma.comcast.net) (Quit: ChatZilla 0.9.87-4.1450hg.fc16 [XULRunner 8.0/20111108090055])
- # [04:42] <roc> yes
- # [04:46] * Joins: ctopper (craig@moz-F2793A92.hsd1.or.comcast.net)
- # [04:47] * Joins: Mook (mook@moz-B1410066.dsl.teksavvy.com)
- # [04:47] * northAway is now known as northWind
- # [04:47] * Joins: Ventron (michael@moz-ED780ED3.dyn.iinet.net.au)
- # [04:47] * Quits: mjschranz (mjschranz@117F2F8F.531FD64C.630E4E47.IP) (Client exited)
- # [04:50] <philor> few more days of this crashing on first use of the context menu after resuming from sleep, I might have to do something about it
- # [04:51] * Joins: Jake (Jake@moz-5D28CC35.phlapafg.dynamic.covad.net)
- # [04:53] * Quits: edmorley (edmorley@moz-28ED10AB.range86-144.btcentralplus.com) (Quit: nn)
- # [04:53] * glob|away is now known as glob
- # [04:53] * Quits: ctopper (craig@moz-F2793A92.hsd1.or.comcast.net) (Quit: ctopper)
- # [04:57] <@bz> man
- # [04:57] <@bz> debugging this has so far required me to patch our browser _and_ our test harness
- # [04:57] <@bz> just to be able to get basic debugging info. :(
- # [04:58] * Joins: ashughes (ashughes@moz-4C407EE8.vc.shawcable.net)
- # [04:58] * Quits: raccettura (raccettura@moz-660B8F4B.hsd1.nj.comcast.net) (Quit: raccettura)
- # [04:59] * Quits: ashughes (ashughes@moz-4C407EE8.vc.shawcable.net) (Quit: ashughes)
- # [05:00] * Quits: not_gavin (gavin@D18F6DDB.6A393516.2321E71E.IP) (Connection reset by peer)
- # [05:00] * Joins: not_gavin (gavin@D18F6DDB.6A393516.2321E71E.IP)
- # [05:03] <darktrojan> has webkit removed BlobBuilder and canvas.toBlob?
- # [05:03] * Quits: Mook (mook@moz-B1410066.dsl.teksavvy.com) (Ping timeout)
- # [05:03] * Joins: Mook (mook@moz-B1410066.dsl.teksavvy.com)
- # [05:04] <@bz> They're not present in safari 5.1.1 and chrome dev...
- # [05:04] <@bz> as for removed...
- # [05:04] <@bz> who knows
- # [05:05] * Joins: Honza (chatzilla@E07BF19C.5BB5597D.D0083327.IP)
- # [05:05] <@bz> did they use to have them?
- # [05:05] * Quits: Mook (mook@moz-B1410066.dsl.teksavvy.com) (NickServ (GHOST command used by Mook_))
- # [05:05] <darktrojan> mdn says there was WebkitBlobBuilder
- # [05:05] * Joins: Mook (mook@moz-B1410066.dsl.teksavvy.com)
- # [05:05] <@bz> oh, prefixed
- # [05:05] <darktrojan> and I swear I read it had toBlob as well
- # [05:05] <@bz> one sec
- # [05:05] <@bz> dunno
- # [05:06] <darktrojan> I think they're just trying to prevent me from doing what I want
- # [05:06] <@bz> Ah, it's there
- # [05:06] <@bz> WebKitBlobBuilder
- # [05:06] <@bz> note capital 'K'
- # [05:06] <darktrojan> ew
- # [05:06] <darktrojan> thanks
- # [05:06] * Quits: Mook (mook@moz-B1410066.dsl.teksavvy.com) (NickServ (GHOST command used by Mook_))
- # [05:06] <@bz> dunno about toBlob
- # [05:07] * Joins: Mook (mook@moz-B1410066.dsl.teksavvy.com)
- # [05:07] <darktrojan> I might've imagined it
- # [05:07] <@bz> caniuse has no mention of toblob
- # [05:07] * @bz has no idea what supports it
- # [05:07] <darktrojan> apparently nobody
- # [05:08] * Quits: RyanVM (chatzilla@moz-D04D3C77.phlapa.fios.verizon.net) (Quit: ChatZilla 0.9.87-2011052417 [Firefox 11.0a1/20111212163758])
- # [05:08] <philor> mmm, not looking like a great day to want your results from try, is it?
- # [05:08] <bbondy> damn you khuey|away
- # [05:09] <bbondy> for removing NS_SPECIALIZE_TEMPLATE :)
- # [05:10] <@bz> philor: hrmph
- # [05:10] * @bz was going to push some stuff to try tonight and hope to get results tomorrow
- # [05:10] <@bz> will that still work? ;)
- # [05:11] <philor> sure, looks like a 7 hour wait for a test slave, max
- # [05:11] * nthomas is now known as nthomas|away
- # [05:11] <philor> so maybe not first thing, but should probably be "at some time tomorrow"
- # [05:11] * Quits: @dbaron (dbaron@moz-27F87443.dsl.dynamic.sonic.net) (Connection reset by peer)
- # [05:12] * Joins: dbaron (dbaron@moz-27F87443.dsl.dynamic.sonic.net)
- # [05:12] * ChanServ sets mode: +o dbaron
- # [05:13] <@bz> oh, man
- # [05:13] <@bz> man
- # [05:13] <@bz> ok
- # [05:13] <@bz> so this test actually fails legitimately....
- # [05:13] <@bz> hrmph
- # [05:13] * Joins: hub (hub@moz-E2FCA694.figuiere.net)
- # [05:13] * @bz wonders how to fix
- # [05:15] * Quits: bsmith (bsmith@E1AD879F.CC98CBF7.C589985.IP) (Ping timeout)
- # [05:19] <@bz> I wonder whether the nsIDocument bools should be bool:1
- # [05:20] <darktrojan> bah, even if I do manage to send a blob via formdata, I can't set the name because it's not a file
- # [05:20] * Quits: markh (markh@moz-BCB4C723.vic.bigpond.net.au) (Ping timeout)
- # [05:21] <@bz> there's a spec bug on that
- # [05:21] * Quits: Waldo (waldo@moz-534B4EF1.hsd1.mi.comcast.net) (Quit: ChatZilla 0.9.87-4.1450hg.fc15 [XULRunner 8.0/20111108090029])
- # [05:21] <@bz> to allow adding a name to blobs when adding to formdata
- # [05:21] <darktrojan> sounds like just what I want :)
- # [05:21] <@bz> screw this
- # [05:21] <@bz> what's the class we have for changing values and then changing them back?
- # [05:22] * darktrojan can't wait for the internet in 2015 when people actually implement this stuff
- # [05:22] * @bz can never remember the friggin' name
- # [05:23] * Joins: markh (markh@moz-BCB4C723.vic.bigpond.net.au)
- # [05:23] * Quits: jrmuizel (jrmuizel@moz-20EF8EAA.cpe.net.cable.rogers.com) (Input/output error)
- # [05:23] * Quits: clokep (clokep@moz-69FB3955.c3-0.smr-ubr1.sbo-smr.ma.cable.rcn.com) (Quit: Instantbird 1.2a1pre)
- # [05:27] * Quits: abwillis (abwillis@moz-87B76227.hlrn.qwest.net) (Quit: ChatZilla 0.9.87 [SeaMonkey 2.5/20111023211519])
- # [05:28] <gavin> auto something?
- # [05:28] <gavin> damn now it's bugging me too
- # [05:28] <@bz> that's what I thought
- # [05:28] * Joins: jrmuizel (jrmuizel@moz-20EF8EAA.cpe.net.cable.rogers.com)
- # [05:28] <@bz> but I can't find it
- # [05:28] <@bz> not autoset
- # [05:28] <@bz> not autotoggle
- # [05:28] <@bz> aha!
- # [05:28] <@bz> autorestore
- # [05:29] <gavin> ah
- # [05:29] * @bz wonders whether that would work on a bool:1
- # [05:29] * Quits: Elen (El@moz-EB0E359F.hsd1.ut.comcast.net) (Quit: )
- # [05:30] <@bz> apparently not
- # [05:31] * Quits: peterv (peterv@moz-715F6D16.access.telenet.be) (Ping timeout)
- # [05:33] * Quits: markh (markh@moz-BCB4C723.vic.bigpond.net.au) (Ping timeout)
- # [05:33] * mcote is now known as mcote|afk
- # [05:34] * Joins: Elen (El@moz-EB0E359F.hsd1.ut.comcast.net)
- # [05:35] * Quits: darktrojan (geoff@moz-8B34E932.dsl.telstraclear.net) (Ping timeout)
- # [05:35] * Joins: peterv (peterv@moz-715F6D16.access.telenet.be)
- # [05:37] * Joins: tfair (tfairey@moz-57178BD3.hsd1.ca.comcast.net)
- # [05:39] <philor> should have merged that to inbound a little later, it's going to green up while we're still in the downtime window
- # [05:39] * Joins: markh (markh@moz-BCB4C723.vic.bigpond.net.au)
- # [05:41] * Quits: biesi (cbiesinger@BEBE4533.89BC370D.6D4D60F7.IP) (Ping timeout)
- # [05:46] <@bz> hrm
- # [05:46] <@bz> did someone break running mochitests directly in the browser?
- # [05:46] <@bz> they seem to no longer show the results?
- # [05:47] <philor> something about tables and time to run and something something
- # [05:47] <@bz> erm
- # [05:47] <@bz> but when running it standalone?
- # [05:47] <@bz> got a bug#?
- # [05:47] * philor would probably be more helpful if he actually paid attention to these things
- # [05:47] <@bz> because this is really dumb
- # [05:47] <@bz> or at least search terms?
- # [05:48] <@bz> ok
- # [05:48] <@bz> nevermind
- # [05:48] * @bz just searches
- # [05:48] <@bz> Make mochitests not build and search the results table when run inside the harness.
- # [05:48] <@bz> But I'm not running inside the harness
- # [05:48] <@bz> and this was 2010-04
- # [05:48] <philor> that sounds familiar
- # [05:49] <@bz> this is more recent
- # [05:49] * @bz really doesn't want to have to bisect this crap....
- # [05:49] <@bz> but I might have to. :(
- # [05:49] * Quits: mbrubeck (mbrubeck@moz-755AD63.hsd1.wa.comcast.net) (Ping timeout)
- # [05:50] * Quits: Honza (chatzilla@E07BF19C.5BB5597D.D0083327.IP) (Ping timeout)
- # [05:50] * philor waits for this win pgo zip to return his connection to him
- # [05:51] <philor> bug 479352 is what I was thinking of, a mere four months ago
- # [05:51] * Quits: Jake (Jake@moz-5D28CC35.phlapafg.dynamic.covad.net) (Connection reset by peer)
- # [05:52] <@bz> (gdb) p this->mInFlush
- # [05:52] <@bz> There is no member or method named mInFlush.
- # [05:52] <philor> but I doubt you accidentally have MOZ_HIDE_RESULTS_TABLE in your env
- # [05:52] <@bz> wtf?
- # [05:55] <coop> IT downtime starting in 10 minutes
- # [05:56] <@bz> well
- # [05:56] <@bz> hmm
- # [05:56] <@bz> I run it "via the harness"
- # [05:56] <@bz> but with --test-path pointing to that one file
- # [05:56] <@bz> does that still set MOZ_HIDE_RESULTS_TABLE?
- # [05:56] * @bz wouldn't think so
- # [05:56] <@bz> 'cause that would be a dumb bug, no?
- # [05:57] <@bz> yeah
- # [05:57] * Joins: cpearce (chatzilla@moz-55A2BC7D.xdsl.xnet.co.nz)
- # [05:57] <@bz> the harness doesn't set that in any case...
- # [05:57] * Joins: mbrubeck (mbrubeck@moz-755AD63.hsd1.wa.comcast.net)
- # [05:57] <@bz> <sigh>
- # [05:57] <@bz> I guess I get to bisect
- # [05:57] <@bz> later
- # [05:57] <philor> that's about how useful my memory usually is
- # [05:57] * @bz will need to find an old test
- # [05:59] * Quits: roc (chatzilla@538BABFE.A073F3E.97BBD552.IP) (Ping timeout)
- # [06:00] * Quits: glob (glob@moz-DF237567.glob.com.au) (Quit: reboot)
- # [06:01] * Joins: cjones (cjones@2BFD4C5C.906A9261.5F90E2EC.IP)
- # [06:04] * philor is now known as philor|away
- # [06:06] * philor|away is now known as philor
- # [06:07] * Quits: Polynomial-C (Poly-C@moz-9A2E8E1C.dip.t-dialin.net) (Ping timeout)
- # [06:07] * coop changes topic to 'm-c and try CLOSED for IT downtime (bug 705823). m-i closed for bug 709193 (MSVC virtual address space limit hit) || Tree rules: https://etherpad.mozilla.org/sLWdJm0zwB || Next aurora uplift: 20th Dec || If you are new or want to help, see irc://irc.mozilla.org/#introduction'
- # [06:07] <@bz> hrm
- # [06:07] <@bz> how long is Try closed for?
- # [06:07] * @bz worries about getting to push tonight....
- # [06:07] * Joins: glob (glob@moz-DF237567.glob.com.au)
- # [06:08] * Joins: Polynomial-C (Poly-C@moz-F089909D.dip.t-dialin.net)
- # [06:08] <philor> if they use it all, the window is 3 hours
- # [06:08] <@bz> <sigh>
- # [06:08] <@bz> would pushing to try 20 mins ago have helped?
- # [06:09] * Joins: jamesr (jamesr@moz-C40B3BE3.hfc.comcastbusiness.net)
- # [06:09] <philor> yep
- # [06:09] <@bz> double-sigh
- # [06:09] * @bz should not have tried to write this test
- # [06:09] * Joins: akeybl (akeybl@moz-52D39FF6.hsd1.ca.comcast.net)
- # [06:09] <philor> but come on, you've been up past 3am every night this week haven't you?
- # [06:09] <@bz> and I certainly should not have hoped that our test harness or the js engine works
- # [06:09] <@bz> nope
- # [06:09] <@bz> well
- # [06:09] <@bz> "yes"
- # [06:09] <@bz> but also getting up at 6:30
- # [06:09] <@bz> so last night I finally managed to get to bed at 10
- # [06:09] <@bz> (and slept in till 10)
- # [06:10] <@bz> but I'd really like to not get back on the old schedule again
- # [06:10] <philor> mail me the patch? I'll be up
- # Session Close: Wed Dec 14 06:13:16 2011
- #
- # Session Start: Wed Dec 14 06:13:16 2011
- # Session Ident: #developers
- # [06:13] * Disconnected
- # [06:41] * Attempting to rejoin channel #developers
- # [06:41] * Rejoined channel #developers
- # [06:41] * Topic is 'm-c and try CLOSED for IT downtime (bug 705823). m-i closed for bug 709193 (MSVC virtual address space limit hit) || Tree rules: https://etherpad.mozilla.org/sLWdJm0zwB || Next aurora uplift: 20th Dec || If you are new or want to help, see irc://irc.mozilla.org/#introduction'
- # [06:41] * Set by coop on Wed Dec 14 06:01:15
- # [06:41] * Joins: regen (Miller@moz-20F8A928.adsl.dynamic.seed.net.tw)
- # [06:41] <@bz> what's obvious to a domain expert (see philor+logs) may not be to others... ;)
- # [06:41] <philor> yeah, thus my backpeddling
- # [06:42] * ewong is now known as ewong|away
- # [06:43] * Quits: bent (chatzilla@moz-C3562645.hsd1.ca.comcast.net) (Quit: ChatZilla 0.9.87-rdmsoft [XULRunner 1.9.2.19/20110707195905])
- # [06:44] <dRdR> I'm definitely no philor
- # [06:44] * Joins: erione (erione@C06DBC77.C8B519AE.C752B3FA.IP)
- # [06:48] * Quits: juanb|afk (jbecerra@moz-BBE3ABD.mv.mozilla.com) (Quit: juanb|afk)
- # [06:48] * Joins: bsmith (bsmith@E1AD879F.CC98CBF7.C589985.IP)
- # [06:51] * Quits: jstraus (Adium@moz-69BA0CDC.hsd1.ma.comcast.net) (Ping timeout)
- # [06:51] * Joins: roc (chatzilla@C0ACF8B.5E1E9EEA.613E47D1.IP)
- # [06:52] * Joins: smooney (smooney@moz-3C7A0050.hsd1.ca.comcast.net)
- # [06:52] * Joins: gal (gal@E1AD879F.CC98CBF7.C589985.IP)
- # [06:52] * rhung is now known as rhung|AFK
- # [06:54] <dolske> there can be only one.
- # [06:54] * Quits: dbradley (dbradley@moz-8FACAA93.fuse.net) (Quit: )
- # [06:58] <philor> nice, a (final?) memorial set of "we leak when DNS is busted" runs
- # [07:01] * Joins: Honza (chatzilla@E07BF19C.5BB5597D.D0083327.IP)
- # [07:04] * Joins: sicking (chatzilla@moz-C03D0C61.vlan426.asr1.sfo1.gblx.net)
- # [07:04] <philor> ah, good, we don't have to worry about reopening the tree
- # [07:04] * Quits: fabrice (fabrice@moz-BBE3ABD.mv.mozilla.com) (Ping timeout)
- # [07:04] * Quits: zuzelvp (zuzelvp@2112147D.C3507A2D.9A8C35B4.IP) (Ping timeout)
- # [07:04] * Quits: dRdR (dRdR@moz-BBE3ABD.mv.mozilla.com) (Ping timeout)
- # [07:04] * Quits: dvander (dvander@moz-BBE3ABD.mv.mozilla.com) (Ping timeout)
- # [07:04] * Quits: luke (andhow@moz-BBE3ABD.mv.mozilla.com) (Ping timeout)
- # [07:04] * Quits: pcwalton (pcwalton@moz-7B0110AD.mv.mozilla.com) (Ping timeout)
- # [07:04] * Quits: sfink|log (sfink@moz-BBE3ABD.mv.mozilla.com) (Ping timeout)
- # [07:04] * Quits: jst (jst@moz-BBE3ABD.mv.mozilla.com) (Ping timeout)
- # [07:05] * Quits: cdleary (cdleary@moz-BBE3ABD.mv.mozilla.com) (Ping timeout)
- # [07:09] * Joins: sfink|log (sfink@moz-BBE3ABD.mv.mozilla.com)
- # [07:09] * Joins: juanb (jbecerra@moz-F1012875.hsd1.ca.comcast.net)
- # [07:09] * Joins: dRdR (dRdR@moz-BBE3ABD.mv.mozilla.com)
- # [07:09] * Quits: juanb (jbecerra@moz-F1012875.hsd1.ca.comcast.net) (Input/output error)
- # [07:12] * Joins: jst (jst@moz-BBE3ABD.mv.mozilla.com)
- # [07:12] <mbrubeck> /o\
- # [07:13] * Joins: pcwalton (pcwalton@moz-7B0110AD.mv.mozilla.com)
- # [07:14] * Joins: pnemsak (Miranda@moz-3D67D819.rainside.sk)
- # [07:14] * shorlander is now known as shorlander-away
- # [07:15] <philor> I didn't really believe, but it was nice to try to persuade myself for a while
- # [07:15] * Quits: graydot (jeba@B6001B9.4E8FB30C.700C6EB0.IP) (Quit: graydot)
- # [07:16] * dolske hoards his trophiez
- # [07:16] <dolske> precioussss
- # [07:17] * Joins: WeirdAl (chatzilla@moz-5BDC219.hsd1.ca.comcast.net)
- # [07:17] * Joins: mozillacorporation (joduinn@moz-3244C0A7.dsl.dynamic.sonic.net)
- # [07:18] * Quits: regen (Miller@moz-20F8A928.adsl.dynamic.seed.net.tw) (Quit: regen)
- # [07:19] * Joins: dvander (dvander@moz-BBE3ABD.mv.mozilla.com)
- # [07:20] * Joins: cdleary (cdleary@moz-BBE3ABD.mv.mozilla.com)
- # [07:20] * jcranmer is now known as jcranmer|away
- # [07:20] * Joins: zuzelvp (zuzelvp@2112147D.C3507A2D.9A8C35B4.IP)
- # [07:20] * Quits: Mossop (mossop@moz-E325C2EC.dsl.dynamic.sonic.net) (Ping timeout)
- # [07:21] * Joins: luke (andhow@moz-BBE3ABD.mv.mozilla.com)
- # [07:22] * Joins: fabrice (fabrice@moz-BBE3ABD.mv.mozilla.com)
- # [07:22] * Joins: Enn (enn@moz-DB6467E3.cpe.net.cable.rogers.com)
- # [07:23] * heycam is now known as heycam|away
- # [07:28] * Joins: biesi (cbiesinger@moz-5EE692A7.ca.hfc.comcastbusiness.net)
- # [07:33] * Quits: damons (gnubeard@moz-A41E6911.hsd1.ca.comcast.net) (Quit: damons)
- # [07:33] * Joins: By-Tor` (bytor@moz-46974D0B.dyn.optonline.net)
- # [07:33] * Quits: By-Tor` (bytor@moz-46974D0B.dyn.optonline.net) (Quit: Leaving)
- # [07:34] * Quits: By-Tor (bytor@moz-46974D0B.dyn.optonline.net) (Ping timeout)
- # [07:35] * Joins: graydot (jeba@63F3AA56.9CFBD76D.30E7AE68.IP)
- # [07:36] * Joins: mdas (mdas@8CB764BC.D30B51A1.412CF160.IP)
- # [07:37] * nthomas|away is now known as nthomas
- # [07:39] * Quits: paul (paul2@moz-DB556DEE.org) (Ping timeout)
- # [07:39] * Quits: alex (alex@moz-BD8D0A09.lightspeed.sntcca.sbcglobal.net) (Quit: Leaving)
- # [07:39] * Joins: By-Tor (bytor@moz-46974D0B.dyn.optonline.net)
- # [07:39] * Quits: Enn (enn@moz-DB6467E3.cpe.net.cable.rogers.com) (Ping timeout)
- # [07:44] * Quits: WeirdAl (chatzilla@moz-5BDC219.hsd1.ca.comcast.net) (Quit: ChatZilla 0.9.87 [Firefox 8.0/20111104165243])
- # [07:44] * Joins: jstraus (Adium@moz-C7B4D9FC.hsd1.ma.comcast.net)
- # [07:45] * heycam|away is now known as heycam
- # [07:46] * Quits: mdas (mdas@8CB764BC.D30B51A1.412CF160.IP) (Quit: mdas)
- # [07:48] * Joins: By-Tor` (bytor@moz-46974D0B.dyn.optonline.net)
- # [07:48] * Quits: squib (squib@moz-F5CA0CFB.dhcp.mdsn.wi.charter.com) (Quit: Leaving)
- # [07:49] * Quits: By-Tor (bytor@moz-46974D0B.dyn.optonline.net) (NickServ (GHOST command used by By-Tor`))
- # [07:49] * By-Tor` is now known as By-Tor
- # [07:50] * Joins: Boriss_ (FlyingToas@moz-62AAA429.hsd1.ca.comcast.net)
- # [07:52] * Joins: Mnyromyr (MnyroWork@moz-E2E3FF3D.tal.de)
- # [07:55] * Quits: By-Tor (bytor@moz-46974D0B.dyn.optonline.net) (Quit: Leaving)
- # [07:55] * Joins: Yoric (Yoric@moz-920DB13B.fbx.proxad.net)
- # [07:56] * Joins: RemusPop (remuspop@601F3B17.33662590.A5830293.IP)
- # [07:56] * Joins: sewardj (sewardj@moz-933C5B3E.dip.t-dialin.net)
- # [07:57] * BenWa is now known as BenWa|sms
- # [07:58] * Quits: sewardj (sewardj@moz-933C5B3E.dip.t-dialin.net) (Ping timeout)
- # [07:58] * Quits: @dbaron (dbaron@moz-27F87443.dsl.dynamic.sonic.net) (Quit: 8403864 bytes have been tenured, next gc will be global.)
- # [07:59] * Quits: chewey (chewey@moz-BA1DC9B1.dip0.t-ipconnect.de) (Ping timeout)
- # [07:59] * Quits: Polynomial-C (Poly-C@moz-F089909D.dip.t-dialin.net) (Ping timeout)
- # [07:59] * Joins: sewardj (sewardj@moz-933C5B3E.dip.t-dialin.net)
- # [07:59] * Quits: northWind (northWind@moz-20E58859.cable.teksavvy.com) (Ping timeout)
- # [08:00] * Joins: Polynomial-C (Poly-C@moz-F089909D.dip.t-dialin.net)
- # [08:00] * Joins: chewey (chewey@moz-BA1DC9B1.dip0.t-ipconnect.de)
- # [08:00] * Joins: darktrojan (geoff@moz-8B34E932.dsl.telstraclear.net)
- # [08:02] * Quits: gal (gal@E1AD879F.CC98CBF7.C589985.IP) (Quit: gal)
- # [08:02] * Quits: larfdesk (Adam_Hinke@moz-F92153ED.longlines.com) (Quit: Leaving.)
- # [08:04] * Quits: mike5w3c (MikeS@moz-98B1BB1D.pool.e-mobile.ne.jp) (Ping timeout)
- # [08:04] * Joins: mike5w3c (MikeS@moz-ABF59F31.pool.e-mobile.ne.jp)
- # [08:04] * mozillacorporation is now known as joduinn-home
- # [08:05] <@bz> man
- # [08:05] <@bz> why did ehsan do this to me? :(
- # [08:06] * Quits: eflores (AndChat@538BABFE.A073F3E.97BBD552.IP) (Quit: Bye)
- # [08:07] * Boriss_ is now known as Boriss
- # [08:08] <philor> open-tease you, or something else?
- # [08:08] <@bz> redirect a serializer review to me
- # [08:08] <@bz> for a patch with 3 lines of context
- # [08:08] <@bz> and no function name in context
- # [08:08] <@bz> written by someone who doesn't plan to update it
- # [08:08] <@bz> with serious bugs (as expected)
- # [08:09] * Joins: jduell (jduell@moz-2D9EDA98.lightspeed.sntcca.sbcglobal.net)
- # [08:09] <philor> "r-."
- # [08:09] <@bz> but most of all, making me read serializer code
- # [08:09] <@bz> well, sure
- # [08:09] <@bz> it'll totally be r-
- # [08:09] <@bz> but...
- # [08:09] <@bz> here
- # [08:09] <@bz> lemme link you
- # [08:09] <philor> the period was supposed to be significant
- # [08:09] <@bz> lol
- # [08:09] * Quits: smooney (smooney@moz-3C7A0050.hsd1.ca.comcast.net) (Quit: smooney)
- # [08:09] <philor> though I guess if the bug is actually worrisome, you can't just do that
- # [08:10] <@bz> it's a 5-digit bug
- # [08:10] <@bz> Here
- # [08:10] <@bz> check this out
- # [08:10] <@bz> http://mxr.mozilla.org/mozilla-central/source/content/base/src/nsXMLContentSerializer.cpp#224
- # [08:10] <@bz> Through line 235
- # [08:10] <@bz> compare to http://mxr.mozilla.org/mozilla-central/source/content/base/src/nsXMLContentSerializer.cpp#251 through 262
- # [08:11] * Quits: gwagner (idefix2@moz-3DFE0B51.hsd1.ca.comcast.net) (Quit: gwagner)
- # [08:11] <@bz> And to http://mxr.mozilla.org/mozilla-central/source/content/base/src/nsXMLContentSerializer.cpp#726 through 736
- # [08:11] <@bz> er, 737
- # [08:11] * Quits: diogogmt (kvirc@moz-4D628198.cpe.net.cable.rogers.com) (Quit: KVIrc 4.1.1 Equilibrium http://www.kvirc.net/)
- # [08:11] * Joins: wolfiR (wolfiR@moz-2858382B.dip0.t-ipconnect.de)
- # [08:11] <@bz> This is what you get in more or less original Netscape code... :(
- # [08:12] <@bz> (well, in the editor-and-stuff parts of it)
- # [08:13] * @bz wishes ms2ger would look at those links
- # [08:13] <@bz> that would fix the problem!
- # [08:13] <philor> heh
- # [08:14] * coop changes topic to 'm-c approval required and m-i closed for bug 709193 (MSVC virtual address space limit hit) || Tree rules: https://etherpad.mozilla.org/sLWdJm0zwB || Next aurora uplift: 20th Dec || If you are new or want to help, see irc://irc.mozilla.org/#introduction'
- # [08:14] <coop> trees are open again. enjoy!
- # [08:14] <philor> I bet I don't want to know what AppendToStringFormatedWrapped("<![CDATA[") actually does, do I?
- # [08:14] <philor> for small values of open
- # [08:14] * Quits: rail_away (rail@261201A0.73F32077.5C565F38.IP) (Ping timeout)
- # [08:15] <philor> bz: pushing to try yourself so you get email?
- # [08:15] <philor> also, what happened to that sleep thing?
- # [08:16] <@bz> philor: you probably don't need to know that append part, but what do you think about the rest?
- # [08:16] <@bz> What happened to the sleep thing was this patc
- # [08:16] <@bz> er, patch
- # [08:16] <@bz> and me wanting to catch up on reviews a bit
- # [08:16] <@bz> big mistake. :(
- # [08:19] * Joins: past (past@moz-7A93B333.dsl.dyn.forthnet.gr)
- # [08:21] * nthomas is now known as nthomas|away
- # [08:26] <@bz> philor: yeah, pushing stuff myself. Thanks!
- # [08:26] * Quits: pcwalton (pcwalton@moz-7B0110AD.mv.mozilla.com) (Quit: nbvcx)
- # [08:27] <Jesse> valgrind is telling me my bugs are in "libxul.so" rather than giving me function names :(
- # [08:28] * Quits: biesi (cbiesinger@moz-5EE692A7.ca.hfc.comcastbusiness.net) (Ping timeout)
- # [08:28] <Jesse> oh that's because i did the wrong thing
- # [08:29] * Joins: ewong (chatzilla@moz-8A30ED8F.static.netvigator.com)
- # [08:29] <@bz> libxul.so pinpoints the location, no?
- # [08:29] <@bz> and the wrong thing you did was web stuff?
- # [08:30] * philor is now known as philor|away
- # [08:30] * bz is now known as bz_sleep
- # [08:31] <Jesse> the wrong thing i did was accidentally running a build from tinderbox (which has native symbols stripped) rather than a local build
- # [08:32] * Joins: kanru (user@moz-6A9F2FC9.dynamic.hinet.net)
- # [08:33] * Quits: jhk_ (jiggy@8E6C34C1.A3F9767A.1C37C358.IP) (Connection reset by peer)
- # [08:33] * Joins: rail_away (rail@261201A0.73F32077.5C565F38.IP)
- # [08:34] * Joins: jhk_ (jiggy@8E6C34C1.A3F9767A.1C37C358.IP)
- # [08:34] <@bz_sleep> hmm
- # [08:34] <@bz_sleep> what's the right way to mail our legal folks?
- # [08:35] <glob> bz_sleep, you could file a bug
- # [08:35] <@bz_sleep> mmm
- # [08:35] <glob> https://bugzilla.mozilla.org/form.legal
- # [08:35] <@bz_sleep> thanks
- # [08:36] * Quits: hub (hub@moz-E2FCA694.figuiere.net) (Ping timeout)
- # [08:38] * Quits: slowpoke (slowpoke@moz-AE67AB45.gigabit.perfect-privacy.com) (Ping timeout)
- # [08:39] * Parts: lsumar (lsumar@538BABFE.A073F3E.97BBD552.IP) (Leaving)
- # [08:39] * philor|away is now known as philor
- # [08:40] * Joins: glazou (glazou@moz-204094DD.disruptive-innovations.fr)
- # [08:40] <dRdR> philor: have you seen this before? TEST-UNEXPECTED-FAIL | automationutils.processLeakLog() | leaked 1 instance of CondVar with size 16 bytes
- # [08:40] <dRdR> followed by a bunch of dns stuff
- # [08:40] <glazou> bonjour
- # [08:40] * Joins: nhirata (nhirata.bu@moz-2A9C9106.hsd1.ca.comcast.net)
- # [08:41] <philor> dRdR: didn't I already star those?
- # [08:41] <dRdR> philor: oh I had another build earlier that was still running
- # [08:41] <Yoric> !seen zpao|detached
- # [08:41] <dRdR> philor: https://tbpl.mozilla.org/?tree=Try&rev=864451bfe1b9
- # [08:41] <firebot> zpao|detached was last seen 12 hours, 38 minutes and 5 seconds ago, changing nick to zpao.
- # [08:41] <dRdR> thanks for starring btw
- # [08:41] <cjones> bz_sleep, ... ping
- # [08:41] <@bz_sleep> pong
- # [08:41] * @bz_sleep will regret this
- # [08:41] <dRdR> I got 3 intermittents in a row on linux debug, not sure if it's intermittent actually
- # [08:42] <dRdR> but they're different each time so I would assume so
- # [08:42] <philor> yeah, linux32 debug m1, you're lucky to get out with just three
- # [08:42] <@bz_sleep> cjones: ^
- # [08:42] <dRdR> haha, okay
- # [08:42] <cjones> bz_sleep, do you have a minute or two?
- # [08:43] * Joins: whimboo (whimboo@moz-216F9AFA.superkabel.de)
- # [08:43] <cjones> i have a question about the refresh driver
- # [08:43] <philor> dRdR: oh, oth, lemme look
- # [08:43] <dRdR> yeah
- # [08:43] <@bz_sleep> cjones: yeah
- # [08:43] <dRdR> first two are the same, third looks like a dns thing
- # [08:44] <cjones> bz_sleep, can you explain why the timer is REPEATING_SLACK?
- # [08:44] <cjones> i would expect it to be REPEATING_PRECISE_CAN_SKIP
- # [08:44] * Joins: slowpoke (slowpoke@moz-F4A45828.pools.arcor-ip.net)
- # [08:45] <philor> dRdR: a fine rogue's gallery, makes me wonder if bug 708927 is another network-related thing, but none of those worry me as being you
- # [08:45] <dRdR> philor: k thanks
- # [08:45] <dRdR> going home
- # [08:45] * Quits: dRdR (dRdR@moz-BBE3ABD.mv.mozilla.com) (Client exited)
- # [08:46] <@bz_sleep> cjones: there was some discussion about that
- # [08:46] <@bz_sleep> cjones: initially
- # [08:46] <cjones> ok
- # [08:46] <@bz_sleep> cjones: when it was first implemented, there was no REPEATING_PRECISE_CAN_SKIP
- # [08:46] <cjones> aha
- # [08:46] <@bz_sleep> cjones: just slack and precise
- # [08:46] * Quits: slowpoke (slowpoke@moz-F4A45828.pools.arcor-ip.net) (Ping timeout)
- # [08:46] <cjones> PRECISE is definitely wrong
- # [08:46] <@bz_sleep> cjones: and precise was obviously wrong
- # [08:46] <cjones> my mental model is of vsync
- # [08:46] * Joins: slowpoke (slowpoke@moz-AE67AB45.gigabit.perfect-privacy.com)
- # [08:46] <cjones> which is somewhat like PRECISE_CAN_SKIP
- # [08:47] <@bz_sleep> cjones: when I added REPEATING_PRECISE_CAN_SKIP, I made the refresh driver use it when requestAnimationFrame was in use
- # [08:47] <cjones> ok
- # [08:47] * Quits: Honza (chatzilla@E07BF19C.5BB5597D.D0083327.IP) (Ping timeout)
- # [08:47] <@bz_sleep> cjones: but otherwise left it slack to minimize the responsiveness impact when each refresh tick has to do a bunch of work
- # [08:47] * coop is now known as coop|away
- # [08:47] <@bz_sleep> cjones: for example, if REPEATING_PRECISE_CAN_SKIP is used and each tick takes 14ms to process...
- # [08:48] <@bz_sleep> cjones: then we'll only spend 2ms per frame servicing events
- # [08:48] <@bz_sleep> cjones: iirc how CAN_SKIP works correctly
- # [08:48] * @bz_sleep checks his recollection
- # [08:48] * Joins: gwagner (idefix2@moz-3DFE0B51.hsd1.ca.comcast.net)
- # [08:48] * cjones thinks
- # [08:50] <@bz_sleep> yes
- # [08:50] <@bz_sleep> so a question
- # [08:50] <@bz_sleep> how does vsync work?
- # [08:51] <@bz_sleep> because CAN_SKIP doesn't skip the next call
- # [08:51] <@bz_sleep> it just does it at the "wrong" time
- # [08:51] * Quits: gwagner (idefix2@moz-3DFE0B51.hsd1.ca.comcast.net) (Quit: gwagner)
- # [08:51] <@bz_sleep> so if you have a CAN_SKIP timer with 16ms delay
- # [08:51] <@bz_sleep> and a callback takes 18ms
- # [08:51] * glob is now known as glob|away
- # [08:51] <@bz_sleep> then the timer will run immediately after the callback is done
- # [08:51] <cjones> approximately, you block on the signal from the display HW, then rerender
- # [08:51] <@bz_sleep> (well, after a trip to the event loop)
- # [08:52] <@bz_sleep> ok
- # [08:52] <cjones> so if your callback took 18ms, you would block until the next frame
- # [08:52] <@bz_sleep> so that's not the same thing as CAN_SKIP
- # [08:52] <cjones> next signal i mean
- # [08:52] <cjones> right
- # [08:52] <@bz_sleep> maybe it should be
- # [08:52] <cjones> some HW lets you wait for an explicit signal, without blocking
- # [08:53] <@bz_sleep> If our timers were actually precise, I might be more willing to consider such skipping
- # [08:53] <@bz_sleep> as it is, I bet they're mostly nuts
- # [08:53] <@bz_sleep> for now
- # [08:53] <@bz_sleep> anyway
- # [08:53] <@bz_sleep> so that's the reasoning for the current state of affairs
- # [08:53] <cjones> ok
- # [08:53] <dolske> hmm, hazelnut
- # [08:53] <@bz_sleep> we've considered switching the refresh driver to CAN_SKIP wholesale
- # [08:53] <@bz_sleep> "we" == "roc and I"
- # [08:54] * Joins: stransky (stransky@moz-6F10C81D.net.upcbroadband.cz)
- # [08:54] * Quits: slowpoke (slowpoke@moz-AE67AB45.gigabit.perfect-privacy.com) (Ping timeout)
- # [08:54] <cjones> an async compositor thread that blocked on vsync would make this slightly easier
- # [08:54] <cjones> since it could signal the main thread(s)
- # [08:55] * Joins: jhorak (jhorak@moz-59813FB4.cust.nbox.cz)
- # [08:55] <@bz_sleep> yeah
- # [08:55] <cjones> bz_sleep, what if we set a window in which the just-slightly-late callback had to be processed before being ignored?
- # [08:55] <cjones> so in the case above, an 18ms callback
- # [08:55] <cjones> the event fired at 16ms would be dequeued
- # [08:55] <cjones> but if we set the "processing delay" at 1ms, then we could drop it and wait for the next event
- # [08:56] * @bz_sleep is not following
- # [08:56] <cjones> your concern as i understand it is
- # [08:56] * heycam is now known as heycam|away
- # [08:56] <cjones> we process a rendering callback that takes 18ms
- # [08:56] <@bz_sleep> yes
- # [08:56] <cjones> the timer fires at 16ms
- # [08:56] <cjones> during the callback
- # [08:56] <roc> we're getting closer to an async compositor thread that blocks on vsync
- # [08:57] <cjones> the callback finishes, then we hit the event loop and process the timer event
- # [08:57] <@bz_sleep> it can't for CAN_SKIP
- # [08:57] <cjones> and run the callback again
- # [08:57] <cjones> oh ok
- # [08:57] <cjones> i misunderstood
- # [08:57] <@bz_sleep> the timer fires at 18ms
- # [08:57] <@bz_sleep> and we run the event loop and do the callback again
- # [08:57] <cjones> hmm, i would expect it to fire at 32
- # [08:57] <cjones> right
- # [08:57] <@bz_sleep> that would be a REPEATING_PRECISE timer
- # [08:57] <cjones> the effect is the same
- # [08:57] <@bz_sleep> it's not the same if the callback takes 40ms. ;)
- # [08:57] <@bz_sleep> which is why we added CAN_SKIP
- # [08:58] <cjones> anyway, we don't want the two events to be processed in quick succession
- # [08:58] <cjones> that's the concern, right?
- # [08:58] <@bz_sleep> right
- # [08:58] <@bz_sleep> yes
- # [08:58] <roc> cjones: by the way, I need to ask you why ShadowableThebesLayers work the way they do. It seems they maintain a complete copy of the ThebesLayerBuffer on the content-process side. Why aren't we just rendering the invalid region into a temporary buffer during each transaction and sending that over to the compositor process/thread?
- # [08:58] <@bz_sleep> unless we really have to because the page is doing scripted animation and hence measuring the framerate
- # [08:58] <@bz_sleep> roc: I have a question for you too, once I'm done with cjones
- # [08:59] <@bz_sleep> roc: https://bugzilla.mozilla.org/show_bug.cgi?id=598482 comment 103
- # [08:59] <cjones> bz_sleep, so my proposal is
- # [08:59] <@bz_sleep> roc: Is that a "do the other thing in this bug" or a "file a followup?"
- # [08:59] <cjones> for the REPEATING_CAN_SKIP case with a too-long callback
- # [08:59] <cjones> set a deadline that's less than the timer interval in which the callback has to be processed
- # [09:00] <@bz_sleep> cjones: so 15ms, say?
- # [09:00] <cjones> which would have the effect of guaranteeing a certain percentage of the interval is not taken up by processing the callback
- # [09:00] <cjones> or 2ms
- # [09:00] <cjones> maybe interval/2
- # [09:00] <@bz_sleep> ok
- # [09:00] <cjones> i don't know what a good deadline is
- # [09:00] <@bz_sleep> alright
- # [09:00] <@bz_sleep> and use the deadline how?
- # [09:01] <cjones> if the next timer signal isn't dequeued within the deadline, it's dropped
- # [09:01] <cjones> and no callback is invoked
- # [09:01] <cjones> i.e. skip the frame
- # [09:02] <roc> bz_sleep: In this bug, I think. I don't want to regress windowed plugins by making them lag behind scrolling more
- # [09:02] <@bz_sleep> roc: ok
- # [09:02] <@bz_sleep> roc: in that case, I need to understand the proposal better
- # [09:02] <cjones> roc, keeps the update off the critical repaint path
- # [09:02] <cjones> but TBH i considered that
- # [09:02] <cjones> and wasn't sure it would be a win
- # [09:03] * Quits: stransky (stransky@moz-6F10C81D.net.upcbroadband.cz) (Quit: Connection reset by beer)
- # [09:03] <cjones> if we end up repainting most of the frame every time, it might not be
- # [09:03] <@bz_sleep> roc: where would this timer get started? How long a timer?
- # [09:03] <cjones> if we repaint only a small part, it should win
- # [09:03] <@bz_sleep> roc: (I still don't have a good grasp of how the plugin widgetry stuff is hooked up)
- # [09:03] <cjones> we could dynamically tune i guess
- # [09:03] <@bz_sleep> cjones: roc's talking about a different issue there
- # [09:03] <@bz_sleep> cjones: sorry, crossing streams. ;)
- # [09:03] <@bz_sleep> cjones: so let me see if I understand
- # [09:04] <@bz_sleep> cjones: your proposal is to have a precise 16ms timer
- # [09:04] <@bz_sleep> cjones: which, when fired on the timer thread, posts an event to the UI thread
- # [09:04] <@bz_sleep> cjones: every 16ms
- # [09:04] <cjones> yes so far
- # [09:04] <@bz_sleep> cjones: this event, when its Run() is called, compares the current time to the time when its timer fired
- # [09:04] <glazou> so the doc says fields in the app.update.url are optional but if the buildID is omitted, a "(null)" appears nin the update history and secondary updates seem to have a few issues. hrm.
- # [09:04] * Quits: sewardj (sewardj@moz-933C5B3E.dip.t-dialin.net) (Ping timeout)
- # [09:05] <@bz_sleep> cjones: and if "now" is too much later than the timer firing time, it does nothing
- # [09:05] <@bz_sleep> cjones: otherwise it calls whatever the actual callback was
- # [09:05] <cjones> yes
- # [09:05] <cjones> it would be like CAN_SKIP, except with a smaller than timer-interval deadline
- # [09:05] <cjones> if i understand CAN_SKIP correctly
- # [09:05] <@bz_sleep> cjones: I don't think you understand CAN_SKIP correctly
- # [09:06] <@bz_sleep> cjones: the way CAN_SKIP works right now is that there is a one-shot timer
- # [09:06] <@bz_sleep> cjones: set for some time
- # [09:06] <cjones> my understanding is that if Run() is called after the interval, it's dropped
- # [09:06] <cjones> ok
- # [09:06] <@bz_sleep> cjones: when the timer fires, it sets its next firing time to now+16ms
- # [09:06] <@bz_sleep> cjones: does NOT rearm itself
- # [09:06] <@bz_sleep> cjones: and posts an event
- # [09:07] <@bz_sleep> cjones: to the main thread
- # [09:07] <@bz_sleep> cjones: when the event is processed, it calls the callback
- # [09:07] * Joins: waschtl (waschtl@moz-A4ECE553.hsi4.kabel-badenwuerttemberg.de)
- # [09:07] <@bz_sleep> cjones: then rearms the timer
- # [09:07] <cjones> right
- # [09:07] <@bz_sleep> cjones: asking it to fire at that time that was computed back before we called the callback
- # [09:07] <@bz_sleep> cjones: which may or may not now be in the past
- # [09:07] * Parts: jhk_ (jiggy@8E6C34C1.A3F9767A.1C37C358.IP)
- # [09:08] <@bz_sleep> cjones: of course times in the past are clamped to "now"
- # [09:08] <cjones> ok right right right
- # [09:08] <cjones> it's call coming back now
- # [09:08] <cjones> *all
- # [09:08] <@bz_sleep> cjones: so in practice, if your callback takes 18ms every time, then it'll get invoked every 18ms
- # [09:08] <cjones> yep
- # [09:09] <@bz_sleep> cjones: whereas in your proposal it would get invoked every 18ms for a bit, then skip forward by 17ms or something
- # [09:09] <@bz_sleep> to make this concrete...
- # [09:09] <@bz_sleep> say the callback takes 18ms and we use a threshold of 8ms for your proposal
- # [09:09] <@bz_sleep> firing times will be:
- # [09:09] <darktrojan> is there an easy way to find a program on $PATH, from js?
- # [09:09] * Joins: jhk_ (jiggy@8E6C34C1.A3F9767A.1C37C358.IP)
- # [09:09] * Quits: ashughes (ashughes@moz-4C407EE8.vc.shawcable.net) (Quit: ashughes)
- # [09:09] <darktrojan> short of using nsIEnviroment.get
- # [09:09] <@bz_sleep> 18, 36, 54, 72, 80
- # [09:09] * Joins: slowpoke (slowpoke@moz-AE67AB45.gigabit.perfect-privacy.com)
- # [09:10] <@bz_sleep> er, no
- # [09:10] <@bz_sleep> that can't happen
- # [09:10] <@bz_sleep> the 80 would get skipped
- # [09:10] <@bz_sleep> so 18, 36, 54, 72, 96
- # [09:10] <@bz_sleep> if I understand your proposal correctly
- # [09:10] <@bz_sleep> because at 90, when the 80 timer would want to fire, it's delayed by too much
- # [09:11] * Joins: stransky (stransky@moz-6F10C81D.net.upcbroadband.cz)
- # [09:11] <cjones> i have 18, 36, 54, 64, 82, ... in my notebook
- # [09:11] * Joins: jfkthame (jfkthame@DBA4BEB5.90783722.9542EC20.IP)
- # [09:11] <nigelb> bz_sleep: shouldn't you be sleeping? :)
- # [09:11] <@bz_sleep> I should
- # [09:11] <@bz_sleep> cjones: let's walk through this
- # [09:11] <cjones> yes
- # [09:11] <cjones> we're on the same page up to 54 i think
- # [09:11] <@bz_sleep> let's make sure
- # [09:11] <@bz_sleep> first callback invocation at t == 0
- # [09:11] <cjones> wait, 54 gets dropped
- # [09:11] <cjones> yes
- # [09:12] <@bz_sleep> callback runs until t == 18
- # [09:12] <@bz_sleep> timer posts an event at t == 16
- # [09:12] <cjones> yes
- # [09:12] <@bz_sleep> this event is processed at t == 18
- # [09:12] <cjones> and it's processed at t == 18
- # [09:12] <@bz_sleep> timer posts an event at t == 32
- # [09:12] <@bz_sleep> callback from t==18 finishes at t == 36, processes event
- # [09:12] <cjones> yep
- # [09:12] <@bz_sleep> timer posts event at t == 48
- # [09:12] * Quits: @dveditz (dveditz@moz-34991AF4.dhcp.cruzio.com) (Quit: dveditz)
- # [09:12] <@bz_sleep> callback finishes at t == 54, processes event
- # [09:13] <cjones> right
- # [09:13] <@bz_sleep> timer posts an event at t == 64
- # [09:13] <@bz_sleep> callback finishes at t == 72
- # [09:13] <@bz_sleep> 72 - 64 == 8
- # [09:13] <cjones> yep
- # [09:13] <cjones> dropped
- # [09:13] <@bz_sleep> depending on whether we use >= or >, this might get skipped
- # [09:13] <@bz_sleep> ok
- # [09:13] <@bz_sleep> so using >= (I was using >)
- # [09:13] <cjones> i made a mistake at t=54, disregard what i had above
- # [09:14] <@bz_sleep> ok
- # [09:14] <@bz_sleep> so at 72 we do nothing
- # [09:14] * Joins: TheLink (TheLink@moz-753C1E34.pools.arcor-ip.net)
- # [09:14] <cjones> with >=, next signal is at t=80, and we process it
- # [09:14] <@bz_sleep> then we start the callback at 80
- # [09:14] <cjones> and the cycle repeats
- # [09:14] <@bz_sleep> so the callback start times are 0, 18, 36, 54, 80
- # [09:14] <cjones> yep
- # [09:14] <@bz_sleep> ok
- # [09:15] * Quits: Ziggy|AWAY (ZiggyMaes@6B780D9D.A4A6DE76.7B12EFB3.IP) (Ping timeout)
- # [09:15] * Quits: qDot (qDot@moz-4548DE5D.members.linode.com) (Ping timeout)
- # [09:15] * Quits: johnath_home (noyb@moz-5A10C3E8.dsl.teksavvy.com) (Ping timeout)
- # [09:15] * Quits: sancus (sancus@moz-3A9162D.off.net) (Ping timeout)
- # [09:15] * Quits: mkelly (mkelly@moz-99CF8925.members.linode.com) (Ping timeout)
- # [09:15] <@bz_sleep> whereas with the current CAN_SKIP we would get multiples of 18
- # [09:15] <cjones> with 10% idle time
- # [09:15] <cjones> but grouped all at the missed deadline
- # [09:15] <cjones> right
- # [09:15] * Quits: bwinton_away (bwinton@D7D3C4B5.88764A66.72A31D6.IP) (Connection reset by peer)
- # [09:15] * Quits: philor (philor@moz-638273A8.my-nick.name) (Ping timeout)
- # [09:15] <@bz_sleep> (and no idle time)
- # [09:15] * Joins: Honza (chatzilla@E07BF19C.5BB5597D.D0083327.IP)
- # [09:15] <@bz_sleep> and with SLACK we get multiples of 34
- # [09:15] * Quits: dria (dria@moz-9E3E12EC.off.net) (Input/output error)
- # [09:15] <cjones> so we'd want to model the effect of various processing times and tolerances
- # [09:15] * Joins: dria (dria@moz-9E3E12EC.off.net)
- # [09:15] <@bz_sleep> and lots of idle time
- # [09:15] * Quits: ctyler (chris@moz-A54458EA.proximity.on.ca) (Ping timeout)
- # [09:15] * Quits: mark (mark@moz-9E3E12EC.off.net) (Ping timeout)
- # [09:15] * Quits: jhford (jhford@moz-2EB93484.info) (Ping timeout)
- # [09:16] * Quits: zuzelvp (zuzelvp@2112147D.C3507A2D.9A8C35B4.IP) (Ping timeout)
- # [09:16] * Quits: beltzner (mike-beltz@moz-9E3E12EC.off.net) (Ping timeout)
- # [09:16] * Quits: Wes (chatzilla@A1FEE3E8.E3DA2587.9A5171B3.IP) (Ping timeout)
- # [09:16] <cjones> yeah
- # [09:16] * Quits: mick_laptop (mick@moz-5972833F.com) (Ping timeout)
- # [09:16] * Joins: mark (mark@moz-9E3E12EC.off.net)
- # [09:16] * Joins: bwinton_away (bwinton@D7D3C4B5.88764A66.72A31D6.IP)
- # [09:17] * Joins: zuzelvp (zuzelvp@2112147D.C3507A2D.9A8C35B4.IP)
- # [09:17] * Joins: jhford (jhford@moz-2EB93484.info)
- # [09:17] * Joins: philor|away (philor@moz-638273A8.my-nick.name)
- # [09:17] * philor|away is now known as philor
- # [09:17] <@bz_sleep> so replacing the current CAN_SKIP usage with your proposal gives a bit more idle time, but a lower frame rate (50fps vs 55fps)
- # [09:17] * Joins: qDot (qDot@moz-4548DE5D.members.linode.com)
- # [09:17] * Joins: Wes (chatzilla@A1FEE3E8.E3DA2587.9A5171B3.IP)
- # [09:17] * Joins: mkelly (mkelly@moz-99CF8925.members.linode.com)
- # [09:18] <@bz_sleep> replacing the current SLACK usage with your proposal gives a way better framerate (50 instead of 29)
- # [09:18] * Joins: dRdR (dRdR@moz-57178BD3.hsd1.ca.comcast.net)
- # [09:18] * Joins: sancus (sancus@moz-3A9162D.off.net)
- # [09:18] <@bz_sleep> the other interesting case to consider is the callback taking 15ms
- # [09:18] <@bz_sleep> for which your proposal and CAN_SKIP just coincide and give 60Hz with very little idle time
- # [09:18] * Joins: ctyler (chris@2FC2A67C.49178EC1.F061A1E6.IP)
- # [09:19] <@bz_sleep> (all this assuming ideal timers, btw, which ours are SO NOT)
- # [09:19] <cjones> yes
- # [09:19] * Joins: Ziggy|AWAY (ZiggyMaes@6B780D9D.A4A6DE76.7B12EFB3.IP)
- # [09:19] * Ziggy|AWAY is now known as Ziggy_Maes
- # [09:19] * Joins: beltzner (mike-beltz@moz-9E3E12EC.off.net)
- # [09:19] * Joins: mick_laptop (mick@moz-5972833F.com)
- # [09:21] <@bz_sleep> ok
- # [09:21] <@bz_sleep> so what are the goals here?
- # [09:22] <@bz_sleep> I'd really like us to pin that down
- # [09:22] <@bz_sleep> possible set of goals:
- # [09:22] <@bz_sleep> 1) Responsiveness
- # [09:22] <@bz_sleep> 2) Maximal average frame rate
- # [09:22] <@bz_sleep> 3) Consistent frame rate
- # [09:22] <cjones> the #1 goal is not to use SLACK, since that's hurting perf unnecessarily
- # [09:22] <cjones> perceived perf rather
- # [09:22] <@bz_sleep> 4) Syncing with some sort of external timing source
- # [09:22] <@bz_sleep> cjones: hmm
- # [09:22] * Quits: bwinton_away (bwinton@D7D3C4B5.88764A66.72A31D6.IP) (Ping timeout)
- # [09:23] <@bz_sleep> cjones: like "what the user sees" perceived?
- # [09:23] * Quits: slowpoke (slowpoke@moz-AE67AB45.gigabit.perfect-privacy.com) (Ping timeout)
- # [09:23] <cjones> correct
- # [09:23] <@bz_sleep> cjones: interesting
- # [09:23] <cjones> the effect for callback processing that takes < 16ms is to drift away from 60hz
- # [09:23] <@bz_sleep> cjones: yes
- # [09:23] <@bz_sleep> cjones: this is effectively noticeable?
- # [09:23] <cjones> yes!! :)
- # [09:23] <@bz_sleep> cjones: I see
- # [09:23] <cjones> otherwise i wouldn't be bothering you
- # [09:23] <@bz_sleep> cjones: ok
- # [09:24] * Joins: sewardj (sewardj@moz-933C5B3E.dip.t-dialin.net)
- # [09:24] <cjones> i agree with those goals
- # [09:24] <cjones> i assume (1) includes no-starvation, right?
- # [09:24] <@bz_sleep> cjones: any other goals you can think of?
- # [09:24] <cjones> responsiveness of the UI
- # [09:24] <@bz_sleep> cjones: (1) precisely means no-starvation
- # [09:24] <cjones> ok
- # [09:24] <cjones> then i agree
- # [09:25] <@bz_sleep> (clearlt the goals are self-contradictory, as usual)
- # [09:25] * Joins: bwinton_away (bwinton@D7D3C4B5.88764A66.72A31D6.IP)
- # [09:25] <@bz_sleep> er, clearly
- # [09:25] <cjones> sure
- # [09:25] <cjones> was there a test case that convinced you and roc not to use REPEATING_PRECISE_CAN_SKIP?
- # [09:25] <@bz_sleep> so just for context...
- # [09:25] <@bz_sleep> not offhand
- # [09:26] <@bz_sleep> one sec
- # [09:26] * @bz_sleep looks up bugs
- # [09:27] <@bz_sleep> ok
- # [09:27] <dolske> as I said...
- # [09:27] <dolske> [17:43] <dolske> I've held conversations with both bz_away and bz_sleep in the past. :)
- # [09:27] <@bz_sleep> so the first relevant bug is 57,444 unique urls, all taken from the top 21,000 domains
- # [09:27] <dolske> ;)
- # [09:28] <@bz_sleep> and then https://bugzilla.mozilla.org/show_bug.cgi?id=650379
- # [09:28] <@bz_sleep> er...
- # [09:28] <@bz_sleep> https://bugzilla.mozilla.org/show_bug.cgi?id=630127 and then https://bugzilla.mozilla.org/show_bug.cgi?id=650379
- # [09:28] <@bz_sleep> effing clipboard
- # [09:28] <@bz_sleep> https://bugzilla.mozilla.org/show_bug.cgi?id=587887 has the original discussion about using precise or not, of course
- # [09:28] * glob|away is now known as glob
- # [09:29] * Quits: fzzzy (donovan@moz-2B41AF9B.lightspeed.mtvwca.sbcglobal.net) (Quit: fzzzy)
- # [09:30] * jfkthame is now known as jfkthame_afk
- # [09:30] <gcp> do we have a way to see allocated memory in crash reporter?
- # [09:31] <@bz_sleep> In 650379 we never discuss switching the default mode to CAN_SKIP
- # [09:31] <@bz_sleep> I thought we had
- # [09:31] <gcp> I'm seeing an OOM bug that is probably caused by something else entirely leaking RAM, so my module fails when it tries to allocate a few megs
- # [09:31] <@bz_sleep> So in terms of our goals above, CAN_SKIP is good for #2 and #3
- # [09:31] <@bz_sleep> not good for #4, if that's a goal
- # [09:31] <@bz_sleep> and questionable for #1
- # [09:32] <@bz_sleep> I wish we had some telemetry about how long refresh ticks take...
- # [09:32] <@bz_sleep> I know that painting the browser UI was about 8ms for me when I last measured it
- # [09:32] <cjones> we can't control that in general
- # [09:32] <@bz_sleep> well, sure
- # [09:32] <nigelb> Where would the proposed identity API for DOM live?
- # [09:32] <cjones> i think it's reasonable for (1)
- # [09:32] <@bz_sleep> ok
- # [09:32] <cjones> the scheme doesn't affect event priority
- # [09:32] <@bz_sleep> in that case, my initial proposal would just be to switch to CAN_SKIP period
- # [09:33] <@bz_sleep> for the moment
- # [09:33] <cjones> it's not good for not using 100% cpu with callbacks that take near 16ms, but that's ok
- # [09:33] * joduinn-home is now known as joduinn-afk
- # [09:33] * Quits: joduinn-afk (joduinn@moz-3244C0A7.dsl.dynamic.sonic.net) (Quit: joduinn-afk)
- # [09:33] <@bz_sleep> (which happens to entail minimal work)
- # [09:33] <cjones> sounds good to me
- # [09:33] <cjones> just a pref flip
- # [09:33] * Quits: karl (karl@moz-2342089.jetstream.xtra.co.nz) (Ping timeout)
- # [09:33] <@bz_sleep> indeed
- # [09:33] <@bz_sleep> in particular, easy to land near end of cycle, because easy to back out
- # [09:33] <@bz_sleep> want to try it?
- # [09:33] <cjones> sure! want to review? :)
- # [09:34] * cjones fears talos ...
- # [09:35] * Parts: whimboo (whimboo@moz-216F9AFA.superkabel.de)
- # [09:35] <@bz_sleep> heh
- # [09:35] * Joins: regen (Miller@moz-20F8A928.adsl.dynamic.seed.net.tw)
- # [09:35] <@bz_sleep> we'll see how talos looks
- # [09:35] <@bz_sleep> I can review
- # [09:35] <@bz_sleep> but tomorrow morning
- # [09:35] <@bz_sleep> esp. because roc is gone
- # [09:35] <cjones> this is a good opportunity to hit fencepost-y style regressions
- # [09:35] <cjones> sure sure
- # [09:35] * @bz_sleep sleeps for real
- # [09:35] <roc> I'm not gone
- # [09:35] <cjones> we're already using it for b2g
- # [09:35] <@bz_sleep> oh
- # [09:35] <@bz_sleep> roc: in that case....
- # [09:35] <glazou> bz_sleep: thanks for review, will work on that as soon as I can
- # [09:36] <@bz_sleep> roc: see questions above about plugin widgets?
- # [09:36] <hsivonen> so I have a pending patch for removing the old View Source back end
- # [09:36] <hsivonen> I planned to wait until the next release train to land it
- # [09:36] <hsivonen> to do things by the book
- # [09:37] * Joins: bholley (bholley@moz-FCAF9AAB.hsd1.ca.comcast.net)
- # [09:37] <hsivonen> should I land it now in order to make libxul slightly smaller?
- # [09:38] * Joins: slowpoke (slowpoke@moz-EE751EFC.pools.arcor-ip.net)
- # [09:38] <@bz_sleep> roc: so first question is whether the important thing is to make the UpdatePluginGeometry call
- # [09:38] <@bz_sleep> roc: or something else?
- # [09:39] * Joins: Standard8 (Standard8@B7F1AE36.48015583.54C3481B.IP)
- # [09:39] <@bz_sleep> roc: I mean...
- # [09:40] * Joins: By-Tor (bytor@moz-46974D0B.dyn.optonline.net)
- # [09:40] <@bz_sleep> roc: in the old code, RequestUpdatePluginGeometry would post an event
- # [09:40] <roc> bz_sleep: instead of queuing the SynchronousPluginGeometryUpdate event, which patch #1 removes, you could just run UpdatePluginGeometry off a timer (and cancel it if UpdatePluginGeometry runs on its own before the timer expires)
- # [09:40] * Quits: slowpoke (slowpoke@moz-EE751EFC.pools.arcor-ip.net) (Ping timeout)
- # [09:40] * Joins: Ms2ger (Ms2ger@B3B463D7.44E88D0C.37724B0D.IP)
- # [09:40] * Joins: slowpoke (slowpoke@moz-AE67AB45.gigabit.perfect-privacy.com)
- # [09:41] <@bz_sleep> roc: why is that better than doing it off the refresh driver?
- # [09:41] <glazou> any one here with the app.update code?
- # [09:41] <glazou> any one here familiar with the app.update code, sorry?
- # [09:41] <@bz_sleep> roc: so if I understand the pieces right...
- # [09:42] <@bz_sleep> roc: UpdatePluginGeometry moves around the plugin widgets, right?
- # [09:42] <@bz_sleep> roc: which immediately changes where the plugins are on the screen
- # [09:42] <@bz_sleep> roc: RequestUpdatePluginGeometry gets called during reflow
- # [09:42] <roc> bz_sleep: the goal is to have UpdatePluginGeometry just before painting updates the window
- # [09:42] <@bz_sleep> roc: ok
- # [09:42] <roc> your understanding is correct
- # [09:42] <roc> so the ideal place to run UpdatePluginGeometry, right now, is in PresShell::WillPaint
- # [09:42] <@bz_sleep> roc: so right now we would tick the refresh driver
- # [09:42] <@bz_sleep> ok
- # [09:43] <@bz_sleep> that makes sense
- # [09:43] * Joins: joduinn (joduinn@moz-3244C0A7.dsl.dynamic.sonic.net)
- # [09:43] <@bz_sleep> and it does that
- # [09:43] <@bz_sleep> (well, modulo the didpaint bit)
- # [09:43] <roc> maybe DidPaint would be better but that's not at issue right now
- # [09:43] <@bz_sleep> if (!aWillSendDidPaint && rootPresContext == mPresContext) {
- # [09:43] <@bz_sleep> rootPresContext->UpdatePluginGeometry();
- # [09:43] <@bz_sleep> }
- # [09:43] <@bz_sleep> is what WillPaint does
- # [09:43] * glob is now known as glob|away
- # [09:43] <@bz_sleep> and DidPaint does:
- # [09:44] <roc> oh yeah
- # [09:44] <@bz_sleep> if (rootPresContext == mPresContext) {
- # [09:44] <roc> right
- # [09:44] <@bz_sleep> rootPresContext->UpdatePluginGeometry();
- # [09:44] <@bz_sleep> }
- # [09:44] <roc> good
- # [09:44] * joduinn is now known as joduinn-zzz
- # [09:44] * Quits: njn (chatzilla@moz-A6B647BB.dyn.iinet.net.au) (Quit: ChatZilla 0.9.87 [Firefox 11.0a1/20111212152829])
- # [09:44] <@bz_sleep> so this part is as it should be
- # [09:44] <@bz_sleep> fine
- # [09:44] <roc> anyway
- # [09:44] <roc> so the problem with putting it in the refresh driver is that means we could update plugin geometry some time before the actual paint occurs
- # [09:44] <@bz_sleep> well, hold on
- # [09:44] <@bz_sleep> before we get to that
- # [09:44] <@bz_sleep> why did we have that runnable to start with?
- # [09:45] * Joins: dao (dao@moz-7C5E3898.superkabel.de)
- # [09:45] <@bz_sleep> to handle cases where willpaint won't get called at all?
- # [09:45] <@bz_sleep> because it's getting optimized away?
- # [09:45] <roc> that's comment #47
- # [09:45] <@bz_sleep> but plugin geometry really did change, so we really do need to update it sometime?
- # [09:45] <roc> imagine, for example, that the content window is covered bya plugin
- # [09:45] <roc> a windowed plugin
- # [09:45] <@bz_sleep> yes
- # [09:45] * @bz_sleep read the comment
- # [09:46] <@bz_sleep> so in that case there is no WillPaint
- # [09:46] * Joins: kaze (kaze@moz-A2A3C9E6.vlan402.asr1.cdg3.gblx.net)
- # [09:46] <@bz_sleep> because we optimize the invalidate away
- # [09:46] <roc> if we invalidate our content area due to a layout change making the plugin invisible, the paint event won't reach the presshell, so no WillPaint, right
- # [09:46] <@bz_sleep> ok
- # [09:46] <@bz_sleep> great
- # [09:46] <roc> it's actually Windows that will make the invalidate go away
- # [09:46] <@bz_sleep> ok
- # [09:46] <@bz_sleep> so what we need to do there is to make sure it happens "sometime"
- # [09:46] <@bz_sleep> I see
- # [09:47] <roc> I think that's pretty rare, which is why I think a last-ditch timer should be OK
- # [09:47] <@bz_sleep> right
- # [09:47] <@bz_sleep> I get it now
- # [09:47] <@bz_sleep> alright
- # [09:47] <@bz_sleep> I'll post a patch tomorrow
- # [09:47] <Ms2ger> Hmm, bz's got patches scrolling before his eyes?
- # [09:47] <@bz_sleep> ms2ger: oh, I will
- # [09:47] <@bz_sleep> roc: 16ms timer sound ok?
- # [09:47] <roc> yeah
- # [09:47] <roc> sounds perfect
- # [09:47] <@bz_sleep> roc: great
- # [09:47] <@bz_sleep> roc: thanks for explaining
- # [09:48] <@bz_sleep> roc: with any luck, I can post patch and tryserver tomorrow
- # [09:48] <@bz_sleep> roc: the question is, should I land on Thurs?
- # [09:48] <@bz_sleep> roc: or wait till after the branchpoint?
- # [09:48] <@bz_sleep> roc: this is not in the "easy-to-turn-off" category
- # [09:48] * Quits: mwu (mwu@moz-59435430.hsd1.nj.comcast.net) (Quit: This computer has gone to sleep)
- # [09:49] <roc> you think the tree will be open on Thursday?
- # [09:50] <@bz_sleep> well, that's a separate issue
- # [09:50] * @bz_sleep actually forgot about that little tree closure thing
- # [09:50] <@bz_sleep> the question is really whether it's ok to land before branchpoint if the tree reopens then
- # [09:50] <philor> hrm, is the color of the jetpack tests on try actually random?
- # [09:50] * Quits: Boriss (FlyingToas@moz-62AAA429.hsd1.ca.comcast.net) (Quit: Boriss)
- # [09:51] <roc> bz_sleep: I would say, let's not
- # [09:51] <@bz_sleep> philor: they took lessons from bikeshed.com?
- # [09:52] <philor> that's an awesome feature: you'll fail 4 tests every single time, but it'll be green, or orange, or red
- # [09:52] <@bz_sleep> roc: that's my instinct too, yeah
- # [09:52] <roc> the tree is going to be pretty congested
- # [09:52] <@bz_sleep> roc: sounds like we have a plan
- # [09:52] <@bz_sleep> roc: btw, the no-op flush thing, I _will_ try to get landed
- # [09:52] <roc> I bet there's a lot of stuff that people will really want in 10
- # [09:52] <roc> er 11
- # [09:52] <@bz_sleep> roc: assuming I make it pass tests. ;)
- # [09:52] <roc> yeah
- # [09:52] <roc> sure
- # [09:52] <roc> now sleep
- # [09:52] <@bz_sleep> indeed
- # [09:52] <@bz_sleep> g'night all
- # [09:52] * Quits: rail_away (rail@261201A0.73F32077.5C565F38.IP) (Ping timeout)
- # [09:53] <bholley> hsivonen: I'd like to push something that was green on try to mozilla-aurora and then go to bed. If you see it doing anything insane while checking in on your recent push, can you handle it? This isn't something I can imagine happening since it just went green, but it'd be nice to make sure
- # [09:53] * Quits: m_kato (makoto@moz-8CEE9107.tokynt01.ap.so-net.ne.jp) (Quit: Instantbird 0.3a1)
- # [09:54] * Quits: Honza (chatzilla@E07BF19C.5BB5597D.D0083327.IP) (Ping timeout)
- # [09:54] * Joins: merinui (merinui@moz-61C7235E.osk2.eonet.ne.jp)
- # [09:54] * Joins: Goldorak (chatzilla@8109DDDE.7C10D095.187A1082.IP)
- # [09:54] * Joins: sworkman (sworkman@moz-825EC923.hsd1.ca.comcast.net)
- # [09:55] * Quits: sworkman (sworkman@moz-825EC923.hsd1.ca.comcast.net) (Quit: sworkman)
- # [09:55] * Joins: rail_away (rail@261201A0.73F32077.5C565F38.IP)
- # [09:55] <Jesse> dRdR: https://bugzilla.mozilla.org/show_bug.cgi?id=102229 ?
- # [09:56] * Joins: Honza (chatzilla@E07BF19C.5BB5597D.D0083327.IP)
- # [09:56] * Quits: birtles (chatzilla@moz-348F61F0.mozilla.or.jp) (Quit: ChatZilla 0.9.87-rdmsoft [XULRunner 1.9.0.17/2009122204])
- # [09:58] * Quits: joduinn-zzz (joduinn@moz-3244C0A7.dsl.dynamic.sonic.net) (Quit: joduinn-zzz)
- # [10:00] * Joins: joduinn (joduinn@moz-3244C0A7.dsl.dynamic.sonic.net)
- # [10:02] * Joins: pascalc (chatzilla@moz-5CB23D49.rev.sfr.net)
- # [10:03] * bholley will just push tomorrow
- # [10:03] <Yoric> Trying to understand a fine point of our build system.
- # [10:04] <Yoric> Can something in dom/ make use of something in toolkit/?
- # [10:04] * Quits: merinui (merinui@moz-61C7235E.osk2.eonet.ne.jp) (Ping timeout)
- # [10:07] <dolske> hmmmmmmmmm yes? I think we basically require toolkit in any mozilla-based app now.
- # [10:07] <dolske> (poor camino?)
- # [10:07] <nthomas|away> looks like it's in the same tier
- # [10:07] * Quits: RST-J (RST-J@moz-343E9EC3.net) (Ping timeout)
- # [10:07] * Quits: smvv (smvv@moz-ADA081D0.org) (Ping timeout)
- # [10:08] * Quits: ted2 (u2847@moz-160C58C6.com) (Ping timeout)
- # [10:08] * Quits: felipe (u2734@moz-160C58C6.com) (Ping timeout)
- # [10:08] * Quits: Unfocused (u2622@moz-160C58C6.com) (Ping timeout)
- # [10:08] * Quits: ttaubert (u2620@moz-160C58C6.com) (Ping timeout)
- # [10:08] * Quits: sid0 (u2934@moz-160C58C6.com) (Ping timeout)
- # [10:08] * Quits: jaws (u2871@moz-160C58C6.com) (Ping timeout)
- # [10:08] * Quits: Asa (u3626@moz-160C58C6.com) (Ping timeout)
- # [10:08] * Joins: gerv (gerv@moz-8E68CF56.in-addr.arpa)
- # [10:08] * Joins: ted2 (u2847@moz-160C58C6.com)
- # [10:08] * Quits: darktrojan (geoff@moz-8B34E932.dsl.telstraclear.net) (Ping timeout)
- # [10:08] * Joins: sid0 (u2934@moz-160C58C6.com)
- # [10:09] * Joins: RST-J (RST-J@moz-343E9EC3.net)
- # [10:09] * Joins: smvv (smvv@moz-ADA081D0.org)
- # [10:10] * Joins: Unfocused (u2622@moz-160C58C6.com)
- # [10:11] * Joins: jaws (u2871@moz-160C58C6.com)
- # [10:12] * Quits: joduinn (joduinn@moz-3244C0A7.dsl.dynamic.sonic.net) (Quit: joduinn)
- # [10:12] * Quits: tfair (tfairey@moz-57178BD3.hsd1.ca.comcast.net) (Client exited)
- # [10:12] * Joins: ttaubert (u2620@moz-160C58C6.com)
- # [10:12] * Joins: tH (Rob@cpc16-seac19-2-0-cust234.7-2.cable.virginmedia.com)
- # [10:13] * Joins: robhawkes (robhawkes@moz-33A339B7.dsl.cnl.uk.net)
- # [10:14] * Joins: Asa (u3626@moz-160C58C6.com)
- # [10:16] * Joins: felipe (u2734@moz-160C58C6.com)
- # [10:18] * Joins: asac (asac@moz-1C3692A7.pppoe.wtnet.de)
- # [10:18] * Quits: slowpoke (slowpoke@moz-AE67AB45.gigabit.perfect-privacy.com) (Ping timeout)
- # [10:18] * Joins: merinui (merinui@moz-61C7235E.osk2.eonet.ne.jp)
- # [10:19] <Yoric> Thanks.
- # [10:19] * Joins: msucan (msucan-@FA9E8863.56E67207.699550A1.IP)
- # [10:20] * Quits: brendan (brendaneic@moz-A286C218.hsd1.ca.comcast.net) (Quit: brendan)
- # [10:21] * Joins: grubshka (grubshka@moz-D9EA9AB9.w109-212.abo.wanadoo.fr)
- # [10:22] * Joins: bjarne (bjarne@moz-46A85847.nextgentel.com)
- # [10:24] * Joins: jacek (jacek@moz-5D707D3B.psi.wroc.pl)
- # [10:25] <glandium> Yoric: define "use"
- # [10:25] <Yoric> include and link with
- # [10:25] <glandium> link with, yes, everything is linked together
- # [10:25] <philor> dolske: not really, their current release does Places history and other bits of toolkit, and there's no reason to believe they'll ever make it past 1.9.2 where they are now
- # [10:26] <glandium> include, yes, as long as you add LOCAL_INCLUDES. that is, you can't rely on the headers being in $(DIST)/include
- # [10:26] <Ms2ger> But adding stuff to LOCAL_INCLUDES makes kittens sad
- # [10:26] <glandium> (because the build system goes into toolkit after going into dom)
- # [10:26] <Yoric> glandium: ah, thanks, I hadn't thought about LOCAL_INCLUDES
- # [10:27] <glandium> Ms2ger: why?
- # [10:28] <Ms2ger> Because touching other modules' privates isn't nice
- # [10:29] * Quits: mgoodwin (mgoodwin@moz-4BB9E5E9.cable.virginmedia.com) (Ping timeout)
- # [10:29] <glandium> Ms2ger: when you want to include something public from a module that is not yet built when you are building, you don't have much choice
- # [10:29] <Ms2ger> :(
- # [10:30] <Ms2ger> Did I mention I hate build systems? :)
- # [10:30] * Quits: dao (dao@moz-7C5E3898.superkabel.de) (Connection reset by peer)
- # [10:30] * Joins: gabor (gabor@moz-3B57BCD1.catv.pool.telekom.hu)
- # [10:31] * Joins: mgoodwin (mgoodwin@moz-4BB9E5E9.cable.virginmedia.com)
- # [10:31] <Yoric> Wasn't khuey|away supposed to design us a new one?
- # [10:31] <Yoric> (or maybe someone else, I'm not sure)
- # [10:32] <glandium> not a new one, a flat one.
- # [10:32] <glandium> and that was joey iirc
- # [10:32] * Joins: dao (dao@moz-7C5E3898.superkabel.de)
- # [10:32] * Quits: a-865 (fmcz@moz-A5D13CA.cable.mindspring.com) (Ping timeout)
- # [10:33] * Joins: a-865 (fmcz@moz-A5D13CA.cable.mindspring.com)
- # [10:33] <Yoric> What is the difference between "flat one" and "a new one"?
- # [10:33] * Joins: slowpoke (slowpoke@moz-AE67AB45.gigabit.perfect-privacy.com)
- # [10:33] <philor> about five years
- # [10:34] <Yoric> :)
- # [10:34] <Yoric> Seriously, what does this mean?
- # [10:34] * Quits: mgoodwin (mgoodwin@moz-4BB9E5E9.cable.virginmedia.com) (Ping timeout)
- # [10:34] <Unfocused> history suggests "five years" is a conservative estimate
- # [10:34] * Joins: mgoodwin_ (mgoodwin@EA3FF8A3.3B0FFCAD.302434AB.IP)
- # [10:35] * Quits: a-865 (fmcz@moz-A5D13CA.cable.mindspring.com) (Ping timeout)
- # [10:35] * Joins: darktrojan (geoff@moz-8B34E932.dsl.telstraclear.net)
- # [10:35] <glandium> personally, i don't like the idea. it will make some things that take a long time take less time, but also things that now take no time take more time
- # [10:35] <glazou> hi Unfocused
- # [10:35] * Joins: edmorley (edmorley@moz-28ED10AB.range86-144.btcentralplus.com)
- # [10:35] <glazou> Unfocused: have you worked on the app-Update code ?
- # [10:35] <Unfocused> glazou: bonjour
- # [10:35] <glazou> :)
- # [10:35] <Unfocused> i have not, and that makes me feel lucky
- # [10:35] <glazou> eheh
- # [10:35] * Quits: jhk_ (jiggy@8E6C34C1.A3F9767A.1C37C358.IP) (Connection reset by peer)
- # [10:35] <glazou> ok
- # [10:36] * Joins: a-865 (fmcz@moz-A5D13CA.cable.mindspring.com)
- # [10:36] * Yoric still has no clue what "a flat one" is all about.
- # [10:36] <Unfocused> trouble making mar files?
- # [10:36] <glazou> no
- # [10:36] <glazou> trouble understanding where the automatic look for an app update is triggered from
- # [10:36] <Unfocused> ah
- # [10:36] * Joins: mgoodwin (mgoodwin@moz-4BB9E5E9.cable.virginmedia.com)
- # [10:36] <Unfocused> no idea, sorry
- # [10:36] <glazou> np
- # [10:37] <Unfocused> rob strong would know
- # [10:37] <glandium> Yoric: 623617
- # [10:37] * Quits: mgoodwin_ (mgoodwin@EA3FF8A3.3B0FFCAD.302434AB.IP) (Ping timeout)
- # [10:37] <glazou> thanks
- # [10:37] <Yoric> Thanks
- # [10:39] * Quits: mgoodwin (mgoodwin@moz-4BB9E5E9.cable.virginmedia.com) (Input/output error)
- # [10:40] * Joins: protz (jonathan@moz-7F6750F6.xulforum.org)
- # [10:40] * Joins: jhk_ (jiggy@8E6C34C1.A3F9767A.1C37C358.IP)
- # [10:42] * Joins: daim (David_Mart@779E3E00.1773D26C.C0FF2207.IP)
- # [10:43] * philor is now known as philor|away
- # [10:45] * Quits: bholley (bholley@moz-FCAF9AAB.hsd1.ca.comcast.net) (Quit: bholley)
- # [10:46] <darktrojan> !seen boriss
- # [10:47] <firebot> boriss was last seen 7 hours, 15 minutes and 33 seconds ago, saying 'i told her to be careful around the airlock!' in #foxymonkies.
- # [10:47] * Parts: nhirata (nhirata.bu@moz-2A9C9106.hsd1.ca.comcast.net)
- # [10:48] * Quits: cpearce (chatzilla@moz-55A2BC7D.xdsl.xnet.co.nz) (Ping timeout)
- # [10:51] * Joins: dao1 (dao@moz-7C5E3898.superkabel.de)
- # [10:51] * Quits: tH (Rob@cpc16-seac19-2-0-cust234.7-2.cable.virginmedia.com) (Ping timeout)
- # [10:51] * Quits: dao (dao@moz-7C5E3898.superkabel.de) (Ping timeout)
- # [10:52] * Joins: tH (Rob@cpc16-seac19-2-0-cust234.7-2.cable.virginmedia.com)
- # [10:54] <darktrojan> dao, can I get a review on bug 469434 in the next few days?
- # [10:54] * Yoric has read project utopia and considers this definitely a _new_ build system.
- # [10:55] * Quits: gabor (gabor@moz-3B57BCD1.catv.pool.telekom.hu) (Ping timeout)
- # [10:58] * Quits: dao1 (dao@moz-7C5E3898.superkabel.de) (Ping timeout)
- # [10:58] * Joins: TheOne (TheOne@moz-D58488C3.dfki.uni-kl.de)
- # [10:58] * Joins: dao (dao@moz-7C5E3898.superkabel.de)
- # [11:01] * Joins: tabb0t (tabb0t@3836CC8C.D48495E9.B84D4DA2.IP)
- # [11:03] * mattwoodrow is now known as mattwoodrow|away
- # [11:08] * Quits: GPHemsley (GPHemsley@moz-A2DF0FC4.hfc.comcastbusiness.net) (Ping timeout)
- # [11:09] * Quits: stransky (stransky@moz-6F10C81D.net.upcbroadband.cz) (Quit: Connection reset by beer)
- # [11:10] * Quits: Polynomial-C (Poly-C@moz-F089909D.dip.t-dialin.net) (Ping timeout)
- # [11:12] <Yoric> Ok, I am progressing.
- # [11:12] <Yoric> Now, I have a problem linking my code (in toolkit/) to some of the code of jstypedarray.
- # [11:14] * Joins: magsout (magsout@moz-E559D13.fbx.proxad.net)
- # [11:14] * Quits: jwatt (roslea@jwatt.irc.users.mozilla.org) (Ping timeout)
- # [11:14] <Yoric> http://pastebin.mozilla.org/1404986
- # [11:14] * Joins: johnath_home (noyb@443AFEC5.C275C16F.F061A1E6.IP)
- # [11:14] <Yoric> Could anyone give me a hand?
- # [11:14] * Joins: jwatt (roslea@jwatt.irc.users.mozilla.org)
- # [11:15] * Joins: Cwiiis (cwiiis@moz-F15E698.croy.cable.virginmedia.com)
- # [11:16] * Quits: graydot (jeba@63F3AA56.9CFBD76D.30E7AE68.IP) (Ping timeout)
- # [11:17] * Joins: GPHemsley (GPHemsley@moz-A2DF0FC4.hfc.comcastbusiness.net)
- # [11:19] * Quits: glob|away (glob@moz-DF237567.glob.com.au) (Connection reset by peer)
- # [11:22] * Joins: florian (florian@moz-87C33FDA.kimsufi.com)
- # [11:23] <Ms2ger> Is malloc infallible already?
- # [11:25] * Parts: janv (varga@moz-C1261AFF.flarion.as5628.telecom.sk) (Leaving)
- # [11:25] * Quits: mcote|afk (mcote@moz-7CCC926E.mc.videotron.ca) (Ping timeout)
- # [11:25] * Quits: MattN (MattN@2F83A011.9167AF22.2125B4DC.IP) (Ping timeout)
- # [11:25] * Joins: mcote|afk (mcote@moz-7CCC926E.mc.videotron.ca)
- # [11:25] * Quits: dao (dao@moz-7C5E3898.superkabel.de) (Quit: Leaving.)
- # [11:26] * Joins: MattN (MattN@2F83A011.9167AF22.2125B4DC.IP)
- # [11:27] <darktrojan> can anyone show me how to get hardware accelerated graphics working on linux?
- # [11:27] * Quits: rniwa (rniwa@11828247.C8C6FF5E.4065847B.IP) (Quit: rniwa)
- # [11:28] * darktrojan is trying to play angry birds but the frame rate is poo :(
- # [11:29] * Joins: janv (varga@moz-C1261AFF.flarion.as5628.telecom.sk)
- # [11:30] * Quits: kaze (kaze@moz-A2A3C9E6.vlan402.asr1.cdg3.gblx.net) (Ping timeout)
- # [11:30] * Quits: florian (florian@moz-87C33FDA.kimsufi.com) (Quit: Instantbird 1.2a1pre)
- # [11:30] * Joins: florian (florian@moz-87C33FDA.kimsufi.com)
- # [11:31] <mounir> who should give approval for landing in m-c?
- # [11:31] * Joins: graememcc (chatzilla@moz-1850EC93.range86-162.btcentralplus.com)
- # [11:32] <edmorley> what are you landing?
- # [11:33] <Ms2ger> b2g?
- # [11:35] <mounir> edmorley: 2 patches for b2g
- # [11:35] * Quits: sicking (chatzilla@moz-C03D0C61.vlan426.asr1.sfo1.gblx.net) (Ping timeout)
- # [11:35] <mounir> one for linux
- # [11:35] <mounir> nothing related to windows
- # [11:35] <mounir> I guess I can't use m-i :(
- # [11:35] <Ms2ger> a=me
- # [11:36] <mounir> now I should find a way to push my patches applied to my m-i repo :)
- # [11:37] * Joins: dao (dao@moz-7C5E3898.superkabel.de)
- # [11:38] * Quits: jduell (jduell@moz-2D9EDA98.lightspeed.sntcca.sbcglobal.net) (Ping timeout)
- # [11:38] <Yoric> Ah, got it.
- # [11:39] <Ms2ger> mounir, hg qpop -a && hg pull -u m-c && hg qpu -a && hg qfin -a && hg push m-c
- # [11:39] * Quits: kanru (user@moz-6A9F2FC9.dynamic.hinet.net) (Ping timeout)
- # [11:40] <firebot> Check-in: http://hg.mozilla.org/mozilla-central/rev/70dea177b9d9 - Ms2ger - Bug 709516 - Remove unused _tostring from nsNPAPIPlugin.cpp; r=josh a=removal
- # [11:40] <Ms2ger> edmorley, thanks :)
- # [11:40] <edmorley> np
- # [11:41] <edmorley> there are extra csets in inbound, so don't just hg push all
- # [11:41] <edmorley> :-)
- # [11:41] <Ms2ger> Well, I hope he didn't pull the spdy relanding
- # [11:43] * Parts: dao (dao@moz-7C5E3898.superkabel.de)
- # [11:43] * Joins: dao (dao@moz-7C5E3898.superkabel.de)
- # [11:43] * Quits: pascalc (chatzilla@moz-5CB23D49.rev.sfr.net) (Ping timeout)
- # [11:44] * Joins: kaze (kaze@moz-A2A3C9E6.vlan402.asr1.cdg3.gblx.net)
- # [11:44] <mounir> Ms2ger: can you r+a this patch http://mounir.pastebin.mozilla.org/1405001
- # [11:44] * Joins: Pike (Pike@moz-E31CD2CB.mozilla.org)
- # [11:44] <mounir> so I can add the a= on it's commit ;)
- # [11:45] * Quits: MattN (MattN@2F83A011.9167AF22.2125B4DC.IP) (Ping timeout)
- # [11:46] <Ms2ger> rs+a=me
- # [11:46] * Joins: smaug (chatzilla@moz-1268A34B.elisa-mobile.fi)
- # [11:46] <mounir> Ms2ger: thanks :)
- # [11:47] <mounir> oh... now I have a remote branch
- # [11:47] * mounir tries to find the rope
- # [11:47] <firebot> Check-in: http://hg.mozilla.org/mozilla-central/rev/82187424f051 - Marek Stepien - Bug 590283 - Bookmark folder item count should use correct plural forms. r=mak77 f=pike
- # [11:47] * Joins: MattN (MattN@2F83A011.9167AF22.2125B4DC.IP)
- # [11:47] <firebot> http://hg.mozilla.org/mozilla-central/rev/6ba6336681e0 - Joshua M - Bug 710341 - Bookmarks menu button has improper toolbar-icon size when placed in the Personal Toolbar. r=dao a=css-only
- # [11:47] <firebot> http://hg.mozilla.org/mozilla-central/rev/596e3eca4196 - Chris Coulson - Bug 709259 - Try creating a named cursor before a bitmap cursor. r=karlt
- # [11:48] <mounir> oh
- # [11:49] <firebot> Check-in: http://hg.mozilla.org/mozilla-central/rev/7c621c7bfea3 - Vivien Nicolas - Bug 709010 - Add a empty components directory to b2g/. r=cjones
- # [11:49] <firebot> http://hg.mozilla.org/mozilla-central/rev/fd6ab19f312c - Mounir Lamouri - no bug - Create batteryInfo on the stack. rs=Ms2ger a=Ms2ger
- # [11:49] <firebot> http://hg.mozilla.org/mozilla-central/rev/ffea93b21d4d - Jean-Alexandre Anglès d'Auriac - Bug 562506 - Use gtk_icon_theme_add_builtin_icon and gtk_window_set_icon_name for nsWindow::SetIcon with the GTK widget. r=karlt
- # [11:49] <firebot> http://hg.mozilla.org/mozilla-central/rev/c7ffe3549389 - Vivien Nicolas - Bug 709013 - Handle the MozApplicationManifest event in b2g/. r=cjones
- # [11:50] * Quits: janv (varga@moz-C1261AFF.flarion.as5628.telecom.sk) (Ping timeout)
- # [11:56] * Quits: kaze (kaze@moz-A2A3C9E6.vlan402.asr1.cdg3.gblx.net) (Quit: WeeChat 0.3.5)
- # [11:56] * Quits: rail_away (rail@261201A0.73F32077.5C565F38.IP) (Ping timeout)
- # [11:57] * Joins: kaze (kaze@moz-A2A3C9E6.vlan402.asr1.cdg3.gblx.net)
- # [11:59] * Quits: dao (dao@moz-7C5E3898.superkabel.de) (Client exited)
- # [11:59] * Joins: rail_away (rail@261201A0.73F32077.5C565F38.IP)
- # [12:00] * Joins: dao (dao@moz-7C5E3898.superkabel.de)
- # [12:02] * Quits: peterv (peterv@moz-715F6D16.access.telenet.be) (Ping timeout)
- # [12:03] * davehunt|away is now known as davehunt
- # [12:04] * Joins: peterv (peterv@moz-715F6D16.access.telenet.be)
- # [12:05] * cjones is now known as cjones-dinner
- # [12:06] * Quits: graememcc (chatzilla@moz-1850EC93.range86-162.btcentralplus.com) (Ping timeout)
- # [12:08] * Quits: decoder (quassel@moz-216446B9.own-hero.net) (Ping timeout)
- # [12:09] * Joins: decoder (quassel@moz-216446B9.own-hero.net)
- # [12:09] * Quits: Honza (chatzilla@E07BF19C.5BB5597D.D0083327.IP) (Ping timeout)
- # [12:09] * Joins: stransky (stransky@moz-107AD163.redhat.com)
- # [12:10] * Quits: bsmith (bsmith@E1AD879F.CC98CBF7.C589985.IP) (Quit: Leaving)
- # [12:17] * Joins: bsmith (bsmith@E1AD879F.CC98CBF7.C589985.IP)
- # [12:17] * Quits: tabb0t (tabb0t@3836CC8C.D48495E9.B84D4DA2.IP) (Ping timeout)
- # [12:20] * Quits: dao (dao@moz-7C5E3898.superkabel.de) (Quit: Leaving.)
- # [12:21] * Joins: clokep (clokep@moz-69FB3955.c3-0.smr-ubr1.sbo-smr.ma.cable.rcn.com)
- # [12:22] * Quits: magsout (magsout@moz-E559D13.fbx.proxad.net) (Ping timeout)
- # [12:22] * Joins: pascalc (chatzilla@moz-A2A3C9E6.vlan402.asr1.cdg3.gblx.net)
- # [12:23] * Joins: bbondy (bbondy@moz-28CF6D1C.home.cgocable.net)
- # [12:23] * Quits: kdc (kdc@moz-2ACC6B38.pk.shawcable.net) (Connection reset by peer)
- # [12:23] * jfkthame_afk is now known as jfkthame
- # [12:24] * Quits: Yoric (Yoric@moz-920DB13B.fbx.proxad.net) (Input/output error)
- # [12:24] * Joins: Honza (chatzilla@E07BF19C.5BB5597D.D0083327.IP)
- # [12:25] * Quits: bbondy (bbondy@moz-28CF6D1C.home.cgocable.net) (Ping timeout)
- # [12:26] * Joins: magsout (magsout@moz-E559D13.fbx.proxad.net)
- # [12:28] * Joins: dao (dao@moz-7C5E3898.superkabel.de)
- # [12:30] * Joins: gmoro (guilherme@3DCBFB93.F520FC74.D159334F.IP)
- # [12:30] * Quits: bsmith (bsmith@E1AD879F.CC98CBF7.C589985.IP) (Ping timeout)
- # [12:30] * Quits: erione (erione@C06DBC77.C8B519AE.C752B3FA.IP) (Quit: Leaving)
- # [12:32] * khuey|away is now known as khuey
- # [12:33] <khuey> so, what's going on with the tree?
- # [12:33] * Quits: protz (jonathan@moz-7F6750F6.xulforum.org) (Ping timeout)
- # [12:35] <Pike> last I picked up from the newsgroup was that inbound didn't revive from the de-template stuff
- # [12:36] <khuey> ah
- # [12:36] <khuey> that's consistent with my local measurements
- # [12:36] * Quits: gerv (gerv@moz-8E68CF56.in-addr.arpa) (Ping timeout)
- # [12:37] * Joins: gerv (gerv@moz-8E68CF56.in-addr.arpa)
- # [12:38] * cjones-dinner is now known as cjones
- # [12:42] * Quits: jhk_ (jiggy@8E6C34C1.A3F9767A.1C37C358.IP) (Connection reset by peer)
- # [12:42] * Quits: regen (Miller@moz-20F8A928.adsl.dynamic.seed.net.tw) (Quit: regen)
- # [12:43] * Ms2ger curses sdwilsh
- # [12:43] <Ms2ger> Not reviewing patches for the crash I'm hitting on windows
- # [12:45] * Joins: Boriss (FlyingToas@moz-62AAA429.hsd1.ca.comcast.net)
- # [12:45] * Quits: dao (dao@moz-7C5E3898.superkabel.de) (Connection reset by peer)
- # [12:46] * Joins: dao (dao@moz-7C5E3898.superkabel.de)
- # [12:47] * Quits: Boriss (FlyingToas@moz-62AAA429.hsd1.ca.comcast.net) (Quit: Boriss)
- # [12:49] <smaug> mounir: just curious, what was the reason to split the sms patch to ~30 pieces ?
- # [12:50] * khuey watches the linker consume all his memory
- # [12:51] * Quits: klugefoo (kahr@moz-F88DFDF5.lightspeed.sntcca.sbcglobal.net) (Ping timeout)
- # [12:53] <mounir> smaug: make review easier (though, you seem to disagree) and help getting only one reviewer per patch
- # [12:53] <mounir> cjones seems to prefer small patches
- # [12:53] <mounir> you seem to disagree
- # [12:54] * Joins: jhk_ (jiggy@8E6C34C1.A3F9767A.1C37C358.IP)
- # [12:55] * Joins: KaiRo (robert@moz-8C499365.adsl.highway.telekom.at)
- # [12:55] * Joins: klugefoo (kahr@moz-F88DFDF5.lightspeed.sntcca.sbcglobal.net)
- # [12:56] * Joins: graememcc (chatzilla@moz-40DE3BF7.range86-148.btcentralplus.com)
- # [12:57] <Ms2ger> Well, there's small and too small :)
- # [12:57] * Joins: gabor (gabor@moz-3B57BCD1.catv.pool.telekom.hu)
- # [12:58] <edmorley> khuey: ehsan's original try run was from m-c, merge to inbound was still red 6 out of 8, so inbound still closed, m-c still approval
- # [12:58] <smaug> mounir: Ms2ger: yeah, there is small, and too small
- # [12:58] <smaug> or size isn't the problem
- # [12:59] <mounir> I don't think those patches were extremely small
- # [12:59] <khuey> edmorley: yeah, I've been catching up on my email
- # [12:59] <mounir> most of them are >10kb
- # [12:59] * Ms2ger hasn't actually looked
- # [12:59] <khuey> there's far too much of it :-/
- # [12:59] <smaug> but having the patch split to so many pieces that I have no idea what the whole thing is doing
- # [13:00] <darktrojan> bah, using moz-icon for gtk stock icons only works for 16px or 24px
- # [13:00] * darktrojan wants 32px
- # [13:01] * Joins: tabb0t (tabb0t@555E2CEA.3A97632F.FFE8C4AF.IP)
- # [13:01] <glandium> khuey: did you check how much memory building with the patches for media and webgl shaves off ?
- # [13:02] <khuey> glandium: nope, that's next
- # [13:03] <khuey> right now I'm doing beta, so we know how much the memory goes up per cycle
- # [13:04] <edmorley> although wasn't this cycle slightly atypical?
- # [13:04] <khuey> in what sense?
- # [13:04] <edmorley> given skia+graphite+spdy?
- # [13:04] <edmorley> and all the rest
- # [13:04] <khuey> maybe
- # [13:05] <khuey> looks like the linker consumes 51 MB more on aurora than on beta
- # [13:05] <khuey> and 64 MB more on tip than aurora
- # [13:05] <khuey> (note that's 64 MB more without SPDY and graphite)
- # [13:05] * Joins: regen (Miller@moz-4F377C4E.adsl.dynamic.seed.net.tw)
- # [13:05] <khuey> glandium: point me to the patches?
- # [13:07] <glandium> khuey: 709721 and 709914
- # [13:08] <khuey> glandium: I'm building with --disable-angle :-/
- # [13:08] <khuey> because 2005 chokes on the debug symbols in the directx sdk
- # [13:08] <gabor> edmorley: about https://bugzilla.mozilla.org/show_bug.cgi?id=697955 and https://bugzilla.mozilla.org/show_bug.cgi?id=678465 I don't have access to the try servers, so no they have not passed yet... and not very familiar with check-in process in general... is there something I can/should do?
- # [13:08] * khuey can test the media one though
- # [13:08] * Quits: dao (dao@moz-7C5E3898.superkabel.de) (Ping timeout)
- # [13:09] <Ms2ger> gabor, I can push to try
- # [13:09] <khuey> gabor: well, you should get try server access :-)
- # [13:09] <Ms2ger> Or get you try access
- # [13:09] * Joins: dao (dao@moz-7C5E3898.superkabel.de)
- # [13:09] * Joins: paul (paul2@moz-DB556DEE.org)
- # [13:09] <glandium> khuey: fwiw, --disable-angle actually builds angle. it skips the GL libraries
- # [13:09] <edmorley> gabor: I tend to push checkin needed patches through try before landing them, so I'll do it for you if they haven't been sent by the time the pgo issues are resolved
- # [13:09] * Quits: paul (paul2@moz-DB556DEE.org) (Quit: leaving)
- # [13:09] <gabor> Ms2ger: that would be great thanks! khuey: yes I should indeed, how can I get one?
- # [13:09] * Joins: paul (paul2@moz-DB556DEE.org)
- # [13:09] <edmorley> gabor: or if you file a commit level 1 bug, I'll vouch
- # [13:10] * Joins: clokep1 (clokep@moz-69FB3955.c3-0.smr-ubr1.sbo-smr.ma.cable.rcn.com)
- # [13:10] <edmorley> gabor: https://www.mozilla.org/hacking/commit-access-policy/
- # [13:10] <khuey> glandium: oh, ok
- # [13:10] <edmorley> gabor: and https://www.mozilla.org/hacking/committer/
- # [13:10] * khuey mumbles something about poorly named configure options
- # [13:10] * Quits: clokep1 (clokep@moz-69FB3955.c3-0.smr-ubr1.sbo-smr.ma.cable.rcn.com) (Quit: Instantbird 1.2a1pre)
- # [13:10] <glandium> khuey: (that got me by surprise)
- # [13:11] <khuey> yeah, me too
- # [13:11] * Quits: clokep (clokep@moz-69FB3955.c3-0.smr-ubr1.sbo-smr.ma.cable.rcn.com) (Ping timeout)
- # [13:11] <gabor> edmorley: thanks, I'll read it and do what I should do then to get one
- # [13:12] <edmorley> gabor: if you have any problems, feel free to ping me :-)
- # [13:13] <gabor> edmorley: ok, brb in a bit
- # [13:14] * Joins: janv (varga@moz-C1261AFF.flarion.as5628.telecom.sk)
- # [13:14] <Ms2ger> gabor, https://tbpl.mozilla.org/?tree=Try&rev=4a33ea8a7a0a
- # [13:14] * Quits: peterv (peterv@moz-715F6D16.access.telenet.be) (Connection reset by peer)
- # [13:15] * Quits: Honza (chatzilla@E07BF19C.5BB5597D.D0083327.IP) (Ping timeout)
- # [13:15] <gabor> Ms2ger++
- # [13:17] <smaug> mounir: why does smsfilter.next() throw ?
- # [13:19] * Joins: peterv (peterv@moz-715F6D16.access.telenet.be)
- # [13:26] * Joins: peregrino (peregrino@moz-1CB238F5.telecom.net.ar)
- # [13:26] <NeilAway> khuey: you need to apply the hotfix for 2005
- # [13:27] <NeilAway> sadly there's no hotfix available for the vista sdk compiler :s
- # [13:27] <khuey> mmm
- # [13:27] <khuey> good to know for the future
- # [13:29] * Quits: gmoro (guilherme@3DCBFB93.F520FC74.D159334F.IP) (Ping timeout)
- # [13:37] * Quits: darktrojan (geoff@moz-8B34E932.dsl.telstraclear.net) (Quit: darktrojan)
- # [13:39] * Quits: ewong (chatzilla@moz-8A30ED8F.static.netvigator.com) (Quit: ChangeNetworks)
- # [13:48] * Joins: Honza (chatzilla@E07BF19C.5BB5597D.D0083327.IP)
- # [13:49] * Joins: evilpie (chatzilla@moz-83A50905.pools.arcor-ip.net)
- # [13:49] * Quits: zuzelvp (zuzelvp@2112147D.C3507A2D.9A8C35B4.IP) (Ping timeout)
- # [13:53] <smaug> Ms2ger: ping
- # [13:53] <Ms2ger> Hi
- # [13:53] <smaug> Ms2ger: in the nsIDOMNSElement removal, why nsGenericElement::GetOnmouseenter ?
- # [13:54] <Ms2ger> Hmm, there was something with that
- # [13:55] <Ms2ger> Hrm
- # [13:55] * Quits: Honza (chatzilla@E07BF19C.5BB5597D.D0083327.IP) (Ping timeout)
- # [13:57] * Joins: jrmuizel (jrmuizel@moz-20EF8EAA.cpe.net.cable.rogers.com)
- # [13:57] * Joins: ejpbruel (ejpbruel@moz-5EE20326.adsl2.static.versatel.nl)
- # [13:58] * Joins: erione (erione@DF34D92D.B2DA192F.C752B3FA.IP)
- # [13:58] * Quits: By-Tor (bytor@moz-46974D0B.dyn.optonline.net) (Connection reset by peer)
- # [13:59] * Joins: Suresh (chatzilla@21EF890.D0998290.EB06F97B.IP)
- # [14:00] * Joins: romeo (romeo@moz-A9325888.k459.webspeed.dk)
- # [14:00] <mounir> smaug: it throws if there is no next()
- # [14:00] * Joins: mak (chatzilla@moz-D2B42BA6.retail.telecomitalia.it)
- # [14:00] <smaug> why?
- # [14:00] <smaug> mounir: I comment in the bug
- # [14:00] <smaug> commented
- # [14:00] * Parts: jhk_ (jiggy@8E6C34C1.A3F9767A.1C37C358.IP)
- # [14:00] <khuey> bug #?
- # [14:01] <edmorley> firebot: uuid
- # [14:01] <firebot> 93abae10-7024-49eb-8e05-1931343b0ebb (/msg firebot cid for CID form)
- # [14:01] * Quits: dRdR (dRdR@moz-57178BD3.hsd1.ca.comcast.net) (Client exited)
- # [14:01] <smaug> khuey: what bug #?
- # [14:02] <khuey> that you commented in?
- # [14:02] * Quits: kennyluck (kennyluck@moz-EAB1F472.dynamic.hinet.net) (Ping timeout)
- # [14:02] <smaug> bug 674725
- # [14:02] <khuey> ty
- # [14:03] <smaug> in/on whatever random article
- # [14:03] * smaug never knows which article to use, and uses some random one :)
- # [14:03] * Joins: jhk_ (jiggy@8E6C34C1.A3F9767A.1C37C358.IP)
- # [14:04] * Quits: mike5w3c (MikeS@moz-ABF59F31.pool.e-mobile.ne.jp) (Ping timeout)
- # [14:04] <Ms2ger> smaug, replied in bug 707576
- # [14:04] * Joins: yevgeni (yevgeni@moz-6EAFFC6F.hsd1.fl.comcast.net)
- # [14:04] * Joins: mike5w3c (MikeS@moz-811E91F4.pool.e-mobile.ne.jp)
- # [14:05] * Joins: Boriss (FlyingToas@moz-62AAA429.hsd1.ca.comcast.net)
- # [14:05] <khuey> glandium: http://khuey.pastebin.mozilla.org/1405071
- # [14:05] * Quits: yevgeni (yevgeni@moz-6EAFFC6F.hsd1.fl.comcast.net) (Quit: leaving)
- # [14:05] <smaug> Ms2ger: ok, makes sense
- # [14:12] <firebot> Check-in: http://hg.mozilla.org/mozilla-central/rev/9df183d280ba - Kyle Huey - Backout Bug 709657 because it doesn't help our problem.
- # [14:12] <firebot> http://hg.mozilla.org/mozilla-central/rev/0db4673fa6b5 - Kyle Huey - Bug 709193: Disable Skia on Windows. r=me a=me CLOSED TREE
- # [14:12] <firebot> http://hg.mozilla.org/mozilla-central/rev/245c74c2f4dd - Kyle Huey - Add the DLL from Bug 709657 to removed-files.in
- # [14:12] <firebot> http://hg.mozilla.org/mozilla-central/rev/0838277b7bd9 - Bas Schouten - Bug 710509: Allow Skia to be easily switched off. r=khuey
- # [14:13] * Joins: aja (chatzilla@A8F7EE0.72086529.7880DB15.IP)
- # [14:14] * Joins: andreasn (andreasn@moz-3CDFCC6A.a336.corp.bahnhof.se)
- # [14:15] * Joins: gmoro (guilherme@CF73CAA7.DC041672.32C814EB.IP)
- # [14:16] <mounir> smaug: is |if (iterator.message) iterator.next();| so weird?
- # [14:17] <smaug> mounir: I thought someone could use .next() in for ()
- # [14:17] <smaug> or while
- # [14:17] <khuey> I would expect people to do
- # [14:17] <khuey> try {
- # [14:17] <smaug> do ( something ) while(it.next())
- # [14:17] <khuey> while (iterator.next())
- # [14:17] <khuey> }
- # [14:17] <khuey> and exceptions for control flow are awful :-)
- # [14:18] <smaug> yeah
- # [14:18] <smaug> I don't recall any other iterator to throw exceptions
- # [14:18] <mounir> smaug: no, they can't do that
- # [14:18] <mounir> this api is async
- # [14:18] <smaug> ah
- # [14:18] <mounir> if you try to call next() twice in a row, you get an exception
- # [14:18] * Joins: ewong (chatzilla@moz-8A30ED8F.static.netvigator.com)
- # [14:19] <mounir> you need to wait for the success event being sent to the request object
- # [14:19] <smaug> (again, I can't see the big picture, since nothing in that patch which introduces next() says that the API is async)
- # [14:20] <smaug> mounir: how is exception better than boolean return value?
- # [14:20] <mounir> smaug: here is a usage exemple http://mounir.pastebin.mozilla.org/1405079
- # [14:20] <hsivonen> khuey: how's the libxul size situation? Should I land the old View Source back end removal without waiting for the next train?
- # [14:20] * Quits: jrmuizel (jrmuizel@moz-20EF8EAA.cpe.net.cable.rogers.com) (Input/output error)
- # [14:21] <mounir> smaug: look at function allmessages()
- # [14:22] <khuey> hsivonen: is it ready to go?
- # [14:22] <khuey> hsivonen: do you have a diff I can test the memory usage of?
- # [14:22] <mounir> smaug: why the exception instead of returning a boolean, I guess it's to prevent wrong usage like while(iterator.next())
- # [14:22] <mounir> but that behavior is copying indexeddb so we should probably ask bent and sicking
- # [14:22] * Quits: gmoro (guilherme@CF73CAA7.DC041672.32C814EB.IP) (Ping timeout)
- # [14:23] <smaug> mounir: ok. In that case the method name should be renamed
- # [14:23] <smaug> and perhaps even interface name
- # [14:23] <smaug> since it is far from common iterator
- # [14:23] * Joins: protz (protz@moz-E29D2A15.inria.fr)
- # [14:24] <mounir> smaug: next() -> continue() and SmsIterator -> SmsCursor would be fine?
- # [14:24] <smaug> continue is reserved word in JS
- # [14:24] <mounir> oh... that is what IDBCursor is using
- # [14:25] <smaug> yeah, and that is ugly
- # [14:25] <smaug> IDB uses also delete, IIRC
- # [14:25] <Ms2ger> foo.continue() works
- # [14:25] <khuey> doesn't work in IE 9 :-P
- # [14:25] <mounir> smaug: if it's SmsCursor, would next() be less wrong to you?
- # [14:25] <hsivonen> khuey: there's a reviewed patch in https://bugzilla.mozilla.org/show_bug.cgi?id=648252 Dunno how much bitrot rebasing it needs.
- # [14:26] <smaug> mounir: but, ok, if IDB uses continue(), perhaps this sms thing should too
- # [14:26] <hsivonen> khuey: I'm checking bitrot now
- # [14:26] <mounir> smaug: can we make a deal with you r+ the patch and me filing a follow-up bug?
- # [14:26] <mounir> I hardly want to fix that bug and push those patches
- # [14:26] * Joins: Yoric (Yoric@moz-A2A3C9E6.vlan402.asr1.cdg3.gblx.net)
- # [14:26] <khuey> hsivonen: just a couple chunks failed
- # [14:27] <hsivonen> khuey: I seem to have an already trivially rebased version in my queue
- # [14:27] <khuey> nice
- # [14:27] <hsivonen> khuey: so should I land it now?
- # [14:28] <gabor> edmorley: what kind of ssh key do I need to generate? dsa or rsa based?
- # [14:28] <smaug> mounir: please no. I don't want to land patches where is it know that file renames are needed immediately . Or at least land the followup in same time as rest of this stuff
- # [14:28] <khuey> hsivonen: if it's ready to go I don't see how it could hurt anything
- # [14:28] <khuey> so lets try it
- # [14:28] <gabor> edmorley: for the commit-1 access
- # [14:28] <mounir> smaug: I will just change the patch then
- # [14:28] <hsivonen> khuey: I take that as an a=khuey
- # [14:28] <khuey> hsivonen: indeed
- # [14:29] <khuey> gabor: idk if it matters, but I'm using rsa
- # [14:30] <gabor> khuey: ok thanks
- # [14:30] * Quits: kaze (kaze@moz-A2A3C9E6.vlan402.asr1.cdg3.gblx.net) (Ping timeout)
- # [14:30] * Quits: NeilAway (neil@moz-32AA0D01.in-addr.btopenworld.com) (Connection reset by peer)
- # [14:31] * bhearsum|afk is now known as bhearsum
- # [14:32] * Joins: espindola (espindola@4295960D.E84B0A4A.6F478678.IP)
- # [14:32] <smaug> mounir: sorry
- # [14:32] <mounir> smaug: btw, why 'builtin' ?
- # [14:33] * Joins: bjacob (bjacob@moz-ADCA75DC.cpe.net.cable.rogers.com)
- # [14:33] * Quits: aja (chatzilla@A8F7EE0.72086529.7880DB15.IP) (Client exited)
- # [14:33] <khuey> it stops xpconnect from allowing js objects to implement that iface
- # [14:33] <smaug> yeah.
- # [14:34] <smaug> I wonder if builtin should be the default
- # [14:34] <firebot> Check-in: http://hg.mozilla.org/mozilla-central/rev/efab5016434f - Henri Sivonen - Bug 648252 - Remove nsViewSourceHTML from the tree. r=mrbkap, a=khuey.
- # [14:34] <mounir> then why should I use that on SmsFilter but not on SmsManager or SmsMessage?
- # [14:34] <smaug> and there should be script-implementable
- # [14:34] * Joins: NeilAway (neil@moz-32AA0D01.in-addr.btopenworld.com)
- # [14:34] <smaug> mounir: some methods take smsfilter as parameter
- # [14:35] <smaug> I don't think any method takes smsmanager or smsmessage as parameter
- # [14:35] <smaug> but yes, you could make smsmanager/message builtin too
- # [14:35] * Joins: kaze (kaze@moz-A2A3C9E6.vlan402.asr1.cdg3.gblx.net)
- # [14:37] * Joins: gmoro (guilherme@3DCBFB93.F520FC74.D159334F.IP)
- # [14:37] * Joins: graydot (jeba@63F3AA56.9CFBD76D.30E7AE68.IP)
- # [14:37] * Quits: dao (dao@moz-7C5E3898.superkabel.de) (Ping timeout)
- # [14:38] * Joins: jorendorff (jorendorff@moz-91590D94.hsd1.tn.comcast.net)
- # [14:39] * Joins: joey (chatzilla@moz-EFCB4CBF.princetowncable.com)
- # [14:40] * jlebar|away is now known as jlebar
- # [14:40] <jlebar> glandium, does C promote the array index to size_t? What's the type of arr[i*2] when i is uint32_t?
- # [14:40] <glandium> khuey: not so bad
- # [14:41] <jlebar> glandium, if C promotes i to size_t before doing i*2, then the compilers' behavior makes sense.
- # [14:41] * Joins: dao (dao@moz-7C5E3898.superkabel.de)
- # [14:43] <evilpie> need some wizard
- # [14:44] <glandium> jlebar: no idea, thus my question
- # [14:44] * Joins: glob (glob@moz-DF237567.glob.com.au)
- # [14:44] * jlebar is trying now.
- # [14:44] * Quits: dao (dao@moz-7C5E3898.superkabel.de) (Quit: Leaving.)
- # [14:47] <espindola> !seen dolske
- # [14:47] <firebot> dolske was last seen 4 hours, 39 minutes and 31 seconds ago, saying '(poor camino?)' in #developers.
- # [14:47] * Joins: kaie (kaie@moz-D215FCA3.dip.t-dialin.net)
- # [14:48] * Quits: merinui (merinui@moz-61C7235E.osk2.eonet.ne.jp) (Quit: Leaving...)
- # [14:49] <gabor> khuey: edmorley: Ms2ger: any of you would be so kind to vouche it for me: https://bugzilla.mozilla.org/show_bug.cgi?id=710590
- # [14:50] <khuey> done
- # [14:50] <khuey> lol
- # [14:50] <espindola> dolske, ping
- # [14:50] <khuey> edmorley and I posted at the exact same time
- # [14:50] <khuey> edmorley: did you get a mid-air warning?
- # [14:50] <edmorley> looks like we broke the mid-air
- # [14:50] <edmorley> no
- # [14:50] <NeilAway> jlebar: I would expect it to i*2 in an int
- # [14:50] <khuey> yeah me neither
- # [14:50] <khuey> glob: ^
- # [14:50] * davehunt is now known as davehunt|away\
- # [14:50] * davehunt|away\ is now known as davehunt|away
- # [14:50] * Quits: jhk_ (jiggy@8E6C34C1.A3F9767A.1C37C358.IP) (Connection reset by peer)
- # [14:51] <glob> O_o
- # [14:51] <jlebar> glandium, okay. So i * 2 is done as a uint32_t. I dunno, then. I can put an e-mail into the clang list.
- # [14:52] <gabor> khuey++ edmorley++ :)
- # [14:52] <glandium> jlebar: with gcc, the loop is actually better with int than with size_t, here
- # [14:52] <jlebar> glandium, can you pastebin some code?
- # [14:52] <khuey> edmorley: you shouldn't move the bugs over
- # [14:52] <khuey> gavin and reed will hunt you down
- # [14:52] * Quits: bc (bc@454BDC37.1D5E8529.AC69809B.IP) (Quit: leaving)
- # [14:52] * Joins: dbradley (dbradley@moz-8FACAA93.fuse.net)
- # [14:52] * Joins: wg9s (bill@moz-7A06A043.hsd1.ma.comcast.net)
- # [14:53] <edmorley> khuey: that's the component they are supposed to start in according to https://www.mozilla.org/hacking/committer/
- # [14:53] <tbsaunde> khuey: I'd think two comments at the same time is fine so long as you don't flip other bits, so maybe its intentional that's allowd
- # [14:54] <khuey> edmorley: oh, I misread
- # [14:54] <khuey> edmorley: I thought you moved it the other way :-)
- # [14:54] <glandium> jlebar: http://pastebin.mozilla.org/1405102 (only pasted the loop itself, not the setup)
- # [14:54] <khuey> edmorley: carry on
- # [14:54] <glob> tbsaunde, it should have triggered a midair :|
- # [14:54] * Quits: jorendorff (jorendorff@moz-91590D94.hsd1.tn.comcast.net) (Quit: jorendorff)
- # [14:54] <edmorley> khuey, glob: weird, as well as no mid-air, I got bugmail listing my change as well as khuey's (my bugmail is set to defaults of not mailing own changes)
- # [14:54] <glob> edmorley, were both changes in the same email?
- # [14:55] <edmorley> yes
- # [14:55] <khuey> I don't think I got any bugmail
- # [14:55] * glob guesses they happened at the *exact* same time
- # [14:55] <evilpie> if ((int)(int8) i == i) {} is this valid for checking if a value fits into int8 ?
- # [14:55] <evilpie> or would this always be i == i
- # [14:56] * Joins: mkaply (chatzilla@moz-92EDDD02.lightspeed.austtx.sbcglobal.net)
- # [14:56] * ChanServ sets mode: +o mkaply
- # [14:56] * jlebar cannot read AT&T asm syntax.
- # [14:57] * Joins: sheppy (sheppy@moz-4992DE6D.static.kgpt.tn.charter.com)
- # [14:57] <evilpie> jlebar++ AT&T syntax is horrible
- # [14:57] * Joins: Enn (enn@F2D29657.F60B0462.67AC9B1.IP)
- # [14:59] <jlebar> glandium, the difference there seems to be re whether it's int or unsigned int.
- # [15:00] * Joins: AaronMT (AaronMT@moz-DEDA283.cpe.net.cable.rogers.com)
- # [15:00] * Joins: jhk_ (jiggy@8E6C34C1.A3F9767A.1C37C358.IP)
- # [15:00] <glandium> jlebar: yeah, i get something different with unsigned int: http://pastebin.mozilla.org/1405110
- # [15:01] <edmorley> dougt: ping
- # [15:02] * Quits: klugefoo (kahr@moz-F88DFDF5.lightspeed.sntcca.sbcglobal.net) (Ping timeout)
- # [15:02] * Quits: romeo (romeo@moz-A9325888.k459.webspeed.dk) (Quit: Leaving)
- # [15:02] <smaug> no, this must be wrong. I thought I haven't reviewed much this month but my bugzilla query says patches in 35 bugs
- # [15:04] <NeilAway> heh, MSVC uses mov eax, eax to zero-extend eax to 64 bits
- # [15:04] <glob> smaug, what's your bmo login?
- # [15:04] <jlebar> NeilAway, inside the loop?
- # [15:04] <derf> Don't use unsigned integers as loop indices.
- # [15:04] <NeilAway> jlebar: this is at /O0, but yes
- # [15:05] * NeilAway thwaps derf
- # [15:05] <jlebar> oh, okay.
- # [15:05] <NeilAway> jlebar: at /On it depends on the size of the element
- # [15:05] <jlebar> derf, ...because when they overflow, you won't notice?
- # [15:05] * Quits: regen (Miller@moz-4F377C4E.adsl.dynamic.seed.net.tw) (Quit: regen)
- # [15:05] <derf> jlebar: It's worse than that.
- # [15:05] <smaug> glob: bugs@pettay.fi
- # [15:05] * Joins: klugefoo (kahr@moz-F88DFDF5.lightspeed.sntcca.sbcglobal.net)
- # [15:05] <derf> The _compiler_ doesn't know they won't overflow.
- # [15:05] <derf> And it isn't allowed to assume they won't.
- # [15:05] <smaug> glob: my query isn't quite accurate, I know that
- # [15:06] <derf> So this disables a host of loop optimizations.
- # [15:06] <jlebar> derf, how does the compiler know that a signed int won't overflow?
- # [15:06] <derf> jlebar: It doesn't, but signed overflow is undefined behavior.
- # [15:06] <derf> So it can do whatever it wants if they do.
- # [15:06] * Quits: @mkaply (chatzilla@moz-92EDDD02.lightspeed.austtx.sbcglobal.net) (Ping timeout)
- # [15:06] <jlebar> ah!
- # [15:06] <NeilAway> jlebar: for large enough elements it rewrites the loop as (for p = arr, i = N; i; p++, i--)
- # [15:06] * Joins: Honza (chatzilla@E07BF19C.5BB5597D.D0083327.IP)
- # [15:06] <jlebar> NeilAway, that's sane...
- # [15:07] <NeilAway> jlebar: for smaller elements it uses [r+r*sizeof(T)] addressing
- # [15:07] <glob> smaug, https://bugzilla.mozilla.org/page.cgi?id=user_activity.html&action=run&who=bugs%40pettay.fi&from=2011-12-01&to=2011-12-31
- # [15:07] <jlebar> NeilAway, So you don't see a difference depending on which type you use?
- # [15:07] * KaiRo is now known as KaiRo_away
- # [15:07] <glob> smaug, lots of reviews on the same bug, probably affecting your count
- # [15:08] <NeilAway> jlebar: wait, this is madness
- # [15:08] <glandium> this discussion reminds me of http://blog.llvm.org/2011/05/what-every-c-programmer-should-know.html
- # [15:09] <NeilAway> jlebar: ok, so for unsigned, for small types, it rewrites the loop with a pointer
- # [15:09] <NeilAway> jlebar: but for size_t and signed int, it uses r+r*sizeof(T) addressing
- # [15:09] <jlebar> glandium, indeed, there's the example...
- # [15:09] <smaug> glob: well, my query returns number of bugs in which I've changed review? to + or -
- # [15:10] * khuey chuckles at http://news.ycombinator.com/item?id=3351521
- # [15:11] * Joins: mgoodwin (mgoodwin@4B81DE46.233DB7FC.2CFBC9B.IP)
- # [15:11] * Joins: mconley (mconley@F2D29657.F60B0462.67AC9B1.IP)
- # [15:11] * catlee-away is now known as catlee
- # [15:12] * rail_away is now known as rail
- # [15:12] * Quits: peregrino (peregrino@moz-1CB238F5.telecom.net.ar) (Input/output error)
- # [15:12] <glandium> jlebar: interestingly, gcc generates the same code for size_t and ssize_t
- # [15:12] * Joins: peregrino (peregrino@moz-1CB238F5.telecom.net.ar)
- # [15:12] <jlebar> indeed.
- # [15:15] <jlebar> loops are hard, let's go shopping.
- # [15:15] * Ms2ger gives jlebar a tiara
- # [15:16] <jlebar> wow, we really shouldn't use unsigned types, it looks like.
- # [15:17] <khuey> don't we pretty much only use unsigned types for loop iterators?
- # [15:17] <jlebar> Yeah.
- # [15:17] <Ms2ger> Just fix clang
- # [15:17] <jlebar> Fix C is more like it.
- # [15:18] <Ms2ger> Go for it
- # [15:18] <Ms2ger> Just make everything undefined, then we can be fast and wrong
- # [15:18] <khuey> fast and wrong is the best kind of fast
- # [15:18] <Ms2ger> twss
- # [15:20] * Quits: Honza (chatzilla@E07BF19C.5BB5597D.D0083327.IP) (Ping timeout)
- # [15:21] * Quits: srinivas (chatzilla@moz-2790C4FB.hsd1.ca.comcast.net) (Ping timeout)
- # [15:22] * Quits: gabor (gabor@moz-3B57BCD1.catv.pool.telekom.hu) (Ping timeout)
- # [15:22] <espindola> Ms2ger, what should I fix? :-)
- # [15:23] * jhopkins|away is now known as jhopkins
- # [15:23] <glandium> jlebar: especially on tight loops
- # [15:23] <Ms2ger> Loops
- # [15:23] <espindola> glandium, size_t is register sized, no?
- # [15:23] <glandium> espindola: yes
- # [15:23] <espindola> there is no overflow "problem" from the compiler point of view then
- # [15:24] <espindola> Ms2ger, sorry, what is the testcase?
- # [15:24] * glob curses mysql
- # [15:25] <Ms2ger> Apparently loops over arrays are too slow for some people
- # [15:25] * Joins: kanru (user@moz-1D154EA4.dynamic.hinet.net)
- # [15:25] <espindola> if using an unsigned index that is smaller than the register size, that can be a prolbem
- # [15:26] <espindola> problem
- # [15:26] * Quits: tabb0t (tabb0t@555E2CEA.3A97632F.FFE8C4AF.IP) (Ping timeout)
- # [15:26] <glandium> espindola: bjacob was saying the using size_t was better than using PRUint32 for array indices. It turns out in loops, it's better to use int than to use either
- # [15:26] * Joins: ajuma (ajuma@F2D29657.F60B0462.67AC9B1.IP)
- # [15:26] * coop|away is now known as coop|buildduty
- # [15:27] * Joins: srinivas (chatzilla@moz-2790C4FB.hsd1.ca.comcast.net)
- # [15:27] <espindola> PRUint32 being bad I understand, I am a bit surprised about size_t being bad
- # [15:27] <espindola> both gcc and clang?
- # [15:27] <jlebar> espindola, I don't understand why PRUint32 should be any worse than size_t.
- # [15:27] <espindola> if the register is 64 bits
- # [15:28] <jlebar> espindola, yes
- # [15:28] <espindola> the compiler has to extend the value
- # [15:28] <jlebar> espindola, because when you write to the register, x86-64 zeroes out the top 32 bits.
- # [15:28] <espindola> so that we get a wrap around
- # [15:28] <jlebar> right?
- # [15:28] <jlebar> oh, so that we overflow properly.
- # [15:28] <jlebar> I see.
- # [15:28] <jlebar> Wait...no. You do all your operations on the 32-bit flavor of the register.
- # [15:28] <khuey> hsivonen: that patch was good for another 3 MB of memory usage
- # [15:28] <jlebar> espindola, So overflow happens just like the register were actually 32 bits wide.
- # [15:29] <espindola> if you can get those, yes
- # [15:29] <jlebar> espindola, what do you mean?
- # [15:29] <espindola> but then you have a problem when computing 64 bit pointers
- # [15:29] <jlebar> espindola, just switch to the 64-bit flavor. If you've written to the register ever, you're guaranteed that the top 32-bits are zero.
- # [15:29] * Quits: smaug (chatzilla@moz-1268A34B.elisa-mobile.fi) (Ping timeout)
- # [15:30] <jlebar> no?
- # [15:30] <Ms2ger> khuey, \o/
- # [15:30] <espindola> with a 32 bit signed value you can be "sure" overflow never happens, and avoid it any extensions
- # [15:30] <khuey> Ms2ger: 3MB doesn't get us very far
- # [15:30] <khuey> but I'll take what I can get
- # [15:30] * Joins: davidb (davidb@F2D29657.F60B0462.67AC9B1.IP)
- # [15:30] <Ms2ger> And bug 709575?
- # [15:30] <jlebar> espindola, right. But just comparing uint32_t and uint64_t.
- # [15:31] <jlebar> why is uint32_t worse on a 64-bit platform?
- # [15:31] <hsivonen> khuey: \o/
- # [15:31] <jlebar> or, derf^ ?
- # [15:31] <bjacob> glandium: jlebar: i just replied to the bug. I actually use ptrdiff_t in my matrix library
- # [15:31] * Joins: mkaply (chatzilla@moz-92EDDD02.lightspeed.austtx.sbcglobal.net)
- # [15:31] * ChanServ sets mode: +o mkaply
- # [15:31] <jlebar> bjacob, aha, that's signed!
- # [15:32] * Quits: dbradley (dbradley@moz-8FACAA93.fuse.net) (Quit: )
- # [15:32] * Joins: zuzelvp (zuzelvp@F2D29657.F60B0462.67AC9B1.IP)
- # [15:32] <espindola> jlebar, I think you would still have problems with intermediate values
- # [15:33] <bjacob> jlebar: yeah, see my email. i wanted to discuss only one thing at a time, here the sizeof issue, so i took size_t for the discussion. no question that signed is better in most cases
- # [15:33] <espindola> say the array index is a+b
- # [15:33] <bjacob> again, see my email, there are other (imo better) reasons than optimization, for using signed types
- # [15:33] <espindola> if we can assume no overflow, we can compute the 64 bit value directly
- # [15:33] <espindola> bjacob, what is the email subject?
- # [15:33] <jlebar> espindola, by loading 0xdeadbeef[rax + rbx]
- # [15:34] <jlebar> espindola, but if a + b needs to be restricted to 32 bits, then we need to compute it separately.
- # [15:34] * Quits: joey (chatzilla@moz-EFCB4CBF.princetowncable.com) (Input/output error)
- # [15:34] <jlebar> That's a good point!
- # [15:34] <bjacob> espindola: see dev-platform about <stdint.h> or C types for situations where guaranteed width is not required
- # [15:34] * Joins: kdc (kdc@moz-2ACC6B38.pk.shawcable.net)
- # [15:34] <espindola> thanks
- # [15:34] <derf> jlebar: It's worse than that.
- # [15:35] <derf> If a is a pointer and b is an int, b may very well fit in 32 bits, but, say, 16*b might not.
- # [15:35] <firebot> Check-in: http://hg.mozilla.org/mozilla-central/rev/ff164dca09bc - Mihai Sucan - Bug 701790 - Remove unused group.* strings in the Style Inspector; r=rcampbell
- # [15:35] <derf> (if that's the size of your type)
- # [15:35] <glandium> bjacob: in a tight loop with arr[2*i], i get the same assembly for both size_t and ptrdiff_t. int is the best option.
- # [15:35] <firebot> http://hg.mozilla.org/mozilla-central/rev/7cf966da2e93 - Mihai Sucan - Bug 687160 - Source Editor should provide a line-based API; r=rcampbell
- # [15:35] <firebot> http://hg.mozilla.org/mozilla-central/rev/8e9120a2a121 - Paul O’Shannessy - Bug 709005 - Sync the browser.sessionstore.restore_on_demand pref [r=rnewman]
- # [15:35] <firebot> http://hg.mozilla.org/mozilla-central/rev/5bd81a413dea - Mihai Sucan - Bug 695032 - [GTK/X11] selecting text in scratchpad doesn't place it on the X primary selection; r=rcampbell
- # [15:35] <firebot> http://hg.mozilla.org/mozilla-central/rev/9366658ba029 - Brijesh Patel - Bug 704295 - Autocomplete with a variable name that is equal to a prefix of a global variable makes it impossible to use the variable; r=msucan
- # [15:35] <bjacob> glandium: int gives you better asm than ptrdiff_t ??
- # [15:35] * Joins: mayhemer (Miranda@B3D46202.F87A741B.F23860FD.IP)
- # [15:35] <firebot> http://hg.mozilla.org/mozilla-central/rev/80bb6b209ea7 - Rob Campbell - merge fx-team to m-c; a=me
- # [15:35] <firebot> http://hg.mozilla.org/mozilla-central/rev/847e7bead2db - Mihai Sucan - Bug 702331 - Update Orion from upstream; r=rcampbell
- # [15:35] <firebot> http://hg.mozilla.org/mozilla-central/rev/b30dc8e42907 - Paul O’Shannessy - Bug 625329 - Pinned tabs are used when setting homepage from current tabs [r=gavin]
- # [15:35] <bjacob> firebot: you such
- # [15:36] <bjacob> suck
- # [15:36] <firebot> bjacob: Sorry, I've no idea what 'you such' might be.
- # [15:36] * Joins: joey (chatzilla@moz-EFCB4CBF.princetowncable.com)
- # [15:36] <Ms2ger> firebot, botsnack
- # [15:36] * firebot smiles
- # [15:36] <glandium> bjacob: with gcc, yes
- # [15:36] <khuey> firebot: diaf
- # [15:36] <firebot> khuey: Die in a fire, kthx.
- # [15:36] <glandium> bjacob: with clang, it's the other way around
- # [15:36] <bjacob> glazou: amazing! how is the ptrdiff_t variant not optimal with gcc? is it not doing the i->2i variable change?
- # [15:36] <glandium> so here's the conclusion, there's no right answer, it depends on the compiler
- # [15:36] <bjacob> oops
- # [15:36] <bjacob> glandium: ^
- # [15:37] <bjacob> glazou: sorry
- # [15:37] <jlebar> derf, What's the corresponding C code to your example?
- # [15:37] * Joins: dao (dao@moz-7C5E3898.superkabel.de)
- # [15:37] <bjacob> glandium: ok
- # [15:37] <glandium> bjacob: see http://pastebin.mozilla.org/1405102
- # [15:37] <jlebar> Okay, you're saying that 16 * b always has to happen in 64-bit space? And that limits what the compiler can do...
- # [15:38] <derf> jlebar: double *a; int b; for(b=0;b<INT_MAX;b++)printf("%G\n",*(a+b));
- # [15:38] <jlebar> Because the compiler then can't, e.g. transform b --> 16 * b and keep it in a 32-bit reg.
- # [15:38] <derf> Right.
- # [15:38] <jlebar> I see.
- # [15:38] <jlebar> So now the only remaining mystery is, why does glandium's code compile better with int than with ptrdiff_t?
- # [15:38] <bjacob> glandium: i dont see a ptrdiff_t version in your pastebin
- # [15:39] <derf> Because gcc.
- # [15:39] <glandium> bjacob: same as size_t
- # [15:39] <derf> Don't expect gcc to be rational. It isn't.
- # [15:39] <bjacob> glandium: ok
- # [15:39] <glandium> bjacob: clang also generates the same code with ptrdiff_t and size_t, but not the same as gcc
- # [15:39] <derf> But 99% of all loops use int, so I'd expect that usage to be very well optimized.
- # [15:39] <jlebar> clang too, though?
- # [15:40] <glandium> clang does better with size_t than with int on my example
- # [15:40] * Joins: dbradley (dbradley@moz-8FACAA93.fuse.net)
- # [15:40] <jlebar> :sigh
- # [15:41] <glandium> with clang http://pastebin.mozilla.org/1405132
- # [15:41] * Quits: joey (chatzilla@moz-EFCB4CBF.princetowncable.com) (Ping timeout)
- # [15:41] <bjacob> derf: i dont know, e.g. all of the standard c++ library (and, i hope, the standard c library) uses size_t or ptrdiff_t, so that must be well optimized too
- # [15:41] <jlebar> bjacob, s/must/should
- # [15:42] <bjacob> jlebar: really?
- # [15:42] <jlebar> bjacob, don't we have some evidence to the contrary?
- # [15:42] <bjacob> jlebar: yes, i've looked at gnu STL code; also, int-based impl wouldn't be able to handle data structures > 2G
- # [15:42] <NeilAway> glandium: iirc msvc does the first loop with int and size_t and the second with unsigned
- # [15:42] <Ms2ger> We've got evidence all ways, afaict, so yes
- # [15:43] * Joins: joey (chatzilla@moz-781B227D.nys.biz.rr.com)
- # [15:43] * Joins: gandalf (zbraniecki@moz-9F56DED7.neoplus.adsl.tpnet.pl)
- # [15:43] * Quits: janv (varga@moz-C1261AFF.flarion.as5628.telecom.sk) (Quit: This computer has gone to sleep)
- # [15:45] <Wes_> bz_sleep: re. our css transition chat the other day: would you believe me if I told you that window.getComputedStyle(el) is observably different than el.offsetHeight = parseInt(el.offsetHeight) ? The former does not cause the transition start "bookmarking" whereas the latter does
- # [15:47] <bjacob> jlebar: actually, the fact that C/C++ std libs are often used with > 2G structures on 64bit, means that they had to be fully size_t ready as soon as they started supporting 64bit archs
- # [15:47] <jlebar> bjacob, Clearly they have to use size_t. The question is whether gcc "has" to optimize it well.
- # [15:47] * Joins: rwaldron (rwaldron@moz-BDCCF091.hfc.comcastbusiness.net)
- # [15:48] <glandium> jlebar: one would expect gcc to optimize it well. But we all know that the difference between theory and practice is that in theory, there's no difference
- # [15:48] <bjacob> jlebar: well, derf's argument was that compilers probably optimize well what is often used; the c/c++ std libs are often used
- # [15:49] * Joins: janv (varga@moz-C1261AFF.flarion.as5628.telecom.sk)
- # [15:49] * Joins: armenzg (armenzg@moz-A6FE435.build.sjc1.mozilla.com)
- # [15:49] * Quits: tH (Rob@cpc16-seac19-2-0-cust234.7-2.cable.virginmedia.com) (Ping timeout)
- # [15:49] <bjacob> jlebar: http://gcc.gnu.org/viewcvs/trunk/libstdc%2B%2B-v3/include/bits/stl_vector.h?view=markup#l230
- # [15:49] <bjacob> (and from there on, these typedefs are what they use everywhere)
- # [15:50] * Joins: tH (Rob@cpc16-seac19-2-0-cust234.7-2.cable.virginmedia.com)
- # [15:50] <jlebar> bjacob, I'm just saying, we saw gcc optimize |int| better than |size_t| in a testcase.
- # [15:50] * davehunt|away is now known as davehunt
- # [15:51] * Quits: mgoodwin (mgoodwin@4B81DE46.233DB7FC.2CFBC9B.IP) (Input/output error)
- # [15:51] <bjacob> jlebar: true... this is dumbfounding. glandium, it would be great if you filed a bug about your finding
- # [15:51] <bjacob> (gcc)
- # [15:51] * Joins: ehsan (ehsan@F2D29657.F60B0462.67AC9B1.IP)
- # [15:51] * Joins: jorendorff (jorendorff@moz-91590D94.hsd1.tn.comcast.net)
- # [15:51] <zuzelvp> https://etherpad.mozilla.org/ChallengesHackathon
- # [15:52] * Quits: erione (erione@DF34D92D.B2DA192F.C752B3FA.IP) (Quit: Leaving)
- # [15:53] <glandium> bjacob: as a matter of fact, it's not even really a case of 32-bits vs. 64-bits, it's a case of gcc using a separate register for the loop counter in one case, and just using the computed array index for the loop counter in the other. in both cases it uses a 64 bits register
- # [15:53] * Quits: graydot (jeba@63F3AA56.9CFBD76D.30E7AE68.IP) (Quit: graydot)
- # [15:54] <Wes_> in UA string, Gecko/20100101 -- what does the date mean? Is that the best way to check gecko version when working around bugs? Do non-firefox gecko browsers generally change the "firefox" part of the UA string?
- # [15:54] <dao> Wes_: it has no meaning
- # [15:54] <dao> it won't be updated
- # [15:55] <Wes_> dao: Hm. engine-bug-workarounds can be very tricky to future-proof, can't they?
- # [15:55] <dao> Wes_: you probably want to check the rv: token (sigh)
- # [15:55] <bjacob> glandium: that is specific to this benchmark though; with the merge_sort example i sent to dev-platform a while ago (i can resend to you) it was using a 32bit reg so it had to add an extra instruction to make it a 64bit value
- # [15:56] <Wes_> dao: Oh! I didn't realize *that* one had meaning. :)
- # [16:01] * Quits: dao (dao@moz-7C5E3898.superkabel.de) (Ping timeout)
- # [16:02] * Joins: Joeh (joe@5A3923AA.BC22908.C7CEC4ED.IP)
- # [16:02] * Quits: Mnyromyr (MnyroWork@moz-E2E3FF3D.tal.de) (Input/output error)
- # [16:02] * Quits: dbradley (dbradley@moz-8FACAA93.fuse.net) (Quit: )
- # [16:02] * Joins: dao (dao@moz-7C5E3898.superkabel.de)
- # [16:03] <bjacob> glandium: (jlebar has my other example; commuting now)
- # [16:03] <jlebar> bjacob, it's kind of an edge case, but for sure that's one reason to use larger registers.
- # [16:03] * Quits: bjacob (bjacob@moz-ADCA75DC.cpe.net.cable.rogers.com) (Quit: Konversation terminated!)
- # [16:03] <jlebar> bjacob, if I really care about my loop's performance, I'm going to inline, and then the problem goes away.
- # [16:05] * Joins: nerovengene (nerovengen@CB845A0D.F873E372.D5D59AD9.IP)
- # [16:06] * Joins: regen (Miller@moz-4F377C4E.adsl.dynamic.seed.net.tw)
- # [16:08] * jwir3|away is now known as jwir3
- # [16:08] <khuey> I wish the PGO linker would use the memory faster
- # [16:08] <jlebar> lol
- # [16:08] <khuey> takes so long :-P
- # [16:08] * Parts: nerovengene (nerovengen@CB845A0D.F873E372.D5D59AD9.IP)
- # [16:09] <jlebar> khuey, ramdrive?
- # [16:10] <khuey> no, the CPU is the bottleneck
- # [16:10] * Quits: wg9s (bill@moz-7A06A043.hsd1.ma.comcast.net) (Quit: ChatZilla 0.9.87-4.1450hg.fc16 [XULRunner 8.0/20111108090055])
- # [16:10] * Joins: hipokrit (hipokrit@moz-502C3BF.rackspace.net)
- # [16:10] <jlebar> khuey, actually, what you really need is http://www.youtube.com/watch?v=zUc6znC848o
- # [16:10] * Joins: hub (hub@moz-E2FCA694.figuiere.net)
- # [16:10] <khuey> jlebar: exactly
- # [16:10] * Joins: ahal (ahal@F2D29657.F60B0462.67AC9B1.IP)
- # [16:10] <khuey> I wonder if I can expense that
- # [16:11] <Ms2ger> If one can get a tiara expended...
- # [16:11] <Ms2ger> With an s, I guess
- # [16:11] * Joins: Honza (chatzilla@E07BF19C.5BB5597D.D0083327.IP)
- # [16:11] * Joins: madhava (madhava@F2D29657.F60B0462.67AC9B1.IP)
- # [16:13] * Joins: jimm (jmathies@moz-7F164CA1.pn.at.cox.net)
- # [16:13] <khuey> I wonder how quickly it goes through LN2
- # [16:15] * Quits: kaze (kaze@moz-A2A3C9E6.vlan402.asr1.cdg3.gblx.net) (Ping timeout)
- # [16:16] * Joins: tabb0t (tabb0t@4A77860B.E4AC1459.FFE8C4AF.IP)
- # [16:18] * Quits: gmoro (guilherme@3DCBFB93.F520FC74.D159334F.IP) (Ping timeout)
- # [16:18] <glandium> khuey: do you at least have ncpu times 3GB of RAM ?
- # [16:19] <glandium> since the linkage is single threaded, you could launch one per cpu
- # [16:19] * Joins: By-Tor (bytor@moz-46974D0B.dyn.optonline.net)
- # [16:19] <jprmc> khuey: so the template hack failed?
- # [16:19] <khuey> glandium: nope
- # [16:19] <espindola> glandium, http://llvm.org/bugs/show_bug.cgi?id=11567
- # [16:19] * Joins: mgoodwin (mgoodwin@E328467E.3B0FFCAD.302434AB.IP)
- # [16:19] <espindola> thanks!
- # [16:19] <khuey> glandium: my laptop is not that awesome
- # [16:19] <Ms2ger> nsTextFormatter?
- # [16:19] <khuey> jprmc: yeah, but we're making progress
- # [16:20] <jprmc> ok
- # [16:20] <glandium> espindola: you forgot to mention what T was
- # [16:20] <espindola> oops
- # [16:20] <khuey> jprmc: https://bugzilla.mozilla.org/show_bug.cgi?id=709193#c60
- # [16:21] * Joins: graydot (jeba@B6001B9.4E8FB30C.700C6EB0.IP)
- # [16:21] * Quits: bjarne (bjarne@moz-46A85847.nextgentel.com) (Quit: Leaving.)
- # [16:22] * Parts: Ventron (michael@moz-ED780ED3.dyn.iinet.net.au)
- # [16:22] * armenzg is now known as armenzg_brb
- # [16:24] <catlee> khuey: swapping?
- # [16:24] * Quits: dao (dao@moz-7C5E3898.superkabel.de) (Ping timeout)
- # [16:25] <khuey> catlee: I also don't really need to do that many measurements
- # [16:25] * Quits: mgoodwin (mgoodwin@E328467E.3B0FFCAD.302434AB.IP) (Ping timeout)
- # [16:26] * Joins: jrmuizel (jrmuizel@F2D29657.F60B0462.67AC9B1.IP)
- # [16:27] * Quits: gandalf (zbraniecki@moz-9F56DED7.neoplus.adsl.tpnet.pl) (Quit: Textual IRC Client: http://www.textualapp.com/)
- # [16:27] <khuey> glandium: so, how are those test failures coming? :-)
- # [16:27] * Quits: tonymec (tonymec@C7A040F2.516A6E14.277517C1.IP) (Input/output error)
- # [16:28] <glandium> khuey: cpearce was looking at the video ones
- # [16:29] * Joins: kaze (kaze@moz-A2A3C9E6.vlan402.asr1.cdg3.gblx.net)
- # [16:30] <khuey> so SPDY and Graphite together add 13 MB of memory usage
- # [16:30] <khuey> I expect that's mostly graphite
- # [16:31] * khuey will turn that back off
- # [16:31] * Quits: zuzelvp (zuzelvp@F2D29657.F60B0462.67AC9B1.IP) (Ping timeout)
- # [16:32] * Joins: mgoodwin (mgoodwin@E328467E.3B0FFCAD.302434AB.IP)
- # [16:34] * Quits: tonymec__ (tonymec@C7A040F2.516A6E14.277517C1.IP) (Input/output error)
- # [16:37] * Quits: mgoodwin (mgoodwin@E328467E.3B0FFCAD.302434AB.IP) (Ping timeout)
- # [16:37] <catlee> bz_sleep: would memory savings from https://bugzilla.mozilla.org/show_bug.cgi?id=700914 apply to just blank tabs, or any about:blank frames in your session?
- # [16:38] * Joins: ashughes (ashughes@moz-4C407EE8.vc.shawcable.net)
- # [16:38] * Quits: cjones (cjones@2BFD4C5C.906A9261.5F90E2EC.IP) (Quit: Leaving)
- # [16:39] * khuey is now known as khuey|away
- # [16:40] * bear-afk is now known as bear
- # [16:40] * Quits: regen (Miller@moz-4F377C4E.adsl.dynamic.seed.net.tw) (Quit: regen)
- # [16:40] <ted> khuey|away: that's interesting, i wonder what it is about graphite
- # [16:40] * Joins: erione (erione@630B0104.C5AA6A92.C752B3FA.IP)
- # [16:41] <andreasn> stransky, ping
- # [16:41] <stransky> andreasn, hi
- # [16:41] * Joins: dveditz (dveditz@moz-34991AF4.dhcp.cruzio.com)
- # [16:41] * ChanServ sets mode: +o dveditz
- # [16:42] * Joins: smooney (smooney@moz-3C7A0050.hsd1.ca.comcast.net)
- # [16:42] <andreasn> stransky, I'm looking at the gtk2 to gtk3 bug. Are any parts of it checked in or do I need to apply all patches in order to compile it?
- # [16:42] <stransky> andreasn, yes, some patches are in trunk
- # [16:42] * Quits: RemusPop (remuspop@601F3B17.33662590.A5830293.IP) (Client exited)
- # [16:43] <stransky> andreasn, and someone are waiting for review
- # [16:43] <ted> i wonder if that P4 at 5Ghz is actually faster than a single core of a Core i7
- # [16:43] * bwinton_away is now known as bwinton
- # [16:43] * gregglind_away is now known as gregglind
- # [16:43] <stransky> andreasn, I think it's better to take the complete patch and apply it to an old trunk
- # [16:44] <andreasn> stransky, so I could add "mk_add_options --enable-default-toolkit=cairo-gtk3" to my .mozconfig and get something running?
- # [16:44] <andreasn> stransky, ah, I see
- # [16:44] <NeilAway> lol @ bholley's post
- # [16:44] <ehsan> khuey|away: ping
- # [16:44] <stransky> andreasn, with the old trunk, yes
- # [16:45] * Joins: vingtetun (vingtetun@moz-A2A3C9E6.vlan402.asr1.cdg3.gblx.net)
- # [16:45] <andreasn> stransky, how old old trunk?
- # [16:45] <stransky> andreasn, like the complete patch :)
- # [16:45] <andreasn> right, thanks!
- # [16:45] <stransky> np
- # [16:46] * khuey|away is now known as khuey
- # [16:46] <khuey> ehsan: pong
- # [16:46] <khuey> ted: lots of code?
- # [16:46] <khuey> lots of templates?
- # [16:46] <khuey> idk
- # [16:46] <ehsan> khuey: any progress on the webgl test failures?
- # [16:46] * Joins: zuzelvp (zuzelvp@F2D29657.F60B0462.67AC9B1.IP)
- # [16:46] <khuey> ted: for comparison, skia is 25 mb
- # [16:46] <khuey> ehsan: I haven't been looking at those ...
- # [16:46] * Joins: kumar (kmcmillan@moz-F2D05B8.c3-0.stk-ubr1.chi-stk.il.cable.rcn.com)
- # [16:47] <ehsan> khuey: have you told bjacob about them?
- # [16:47] * Joins: biesi (cbiesinger@moz-5EE692A7.ca.hfc.comcastbusiness.net)
- # [16:47] <ehsan> maybe he can help?
- # [16:47] <khuey> ehsan: I haven't been doing anything with the webgl stuff
- # [16:47] <khuey> that's glandium's stuff
- # [16:47] <ehsan> oh ok
- # [16:47] * Quits: kumar (kmcmillan@moz-F2D05B8.c3-0.stk-ubr1.chi-stk.il.cable.rcn.com) (Quit: kumar)
- # [16:47] <ehsan> glandium: have you talked to bjacob about those?
- # [16:48] <khuey> ehsan: I applied science last night and did some measurements of memory usage
- # [16:48] <ehsan> ok
- # [16:48] <khuey> ehsan: turning off Skia should buy us 30ish MB
- # [16:48] <khuey> (half a release cycle at our current rate)
- # [16:48] <ehsan> wow
- # [16:48] <ehsan> let's kill it with fire!
- # [16:48] * Quits: TheOne (TheOne@moz-D58488C3.dfki.uni-kl.de) (Ping timeout)
- # [16:48] <khuey> glandium's two patches are another 41 MB
- # [16:49] <khuey> ehsan: I turned off skia on the tree and m-i is doing PGO builds now
- # [16:49] * Joins: dao (dao@moz-7C5E3898.superkabel.de)
- # [16:49] <catlee> ted/ehsan/khey: anything in particular I have to do to compile 32-bit on w64 other than using the 32-bit compilers?
- # [16:49] <ehsan> khuey: jrmuizel says that we don't build it on windows
- # [16:49] <ted> khuey: interesting
- # [16:49] <ted> catlee: nope
- # [16:49] <khuey> ehsan: jrmuizel was wrong
- # [16:49] <ehsan> khuey: or rather, he's not sure if we do
- # [16:49] <khuey> but we're not building it now! ;-)
- # [16:49] <ted> catlee: when i build on windows i just use start-msvcX.bat and build
- # [16:49] <ehsan> khuey: but he says we don't use it on windows
- # [16:49] <glandium> ehsan: jgilbert looked at the log and mentioned the webgl fails lookes like problems with video
- # [16:49] <khuey> ehsan: yep, which is why it's off now
- # [16:49] <ehsan> so not building it makes a lot of sense
- # [16:49] * Quits: smooney (smooney@moz-3C7A0050.hsd1.ca.comcast.net) (Quit: smooney)
- # [16:50] <glandium> i guess i could push a try with webgl moved out only
- # [16:50] <ehsan> glandium: yes I saw the bug comment, I'll ask bjacob to debug them locally when he comes in today
- # [16:50] <ehsan> I don't know if this is the sort of thing we can debug on the try server
- # [16:50] <ehsan> khuey++
- # [16:50] * Joins: gabor (gabor@moz-3B57BCD1.catv.pool.telekom.hu)
- # [16:50] <glandium> ehsan: we can at least see if they happen there
- # [16:50] <khuey> ted: yeah, what input leads to what results is interesting
- # [16:50] * ehsan is not sure how much code we have in webgl
- # [16:50] <ted> khuey: i recall googlers saying that heavy template usage was suspected to be their problem with building with PGO
- # [16:50] <ehsan> glandium: for sure
- # [16:50] <ehsan> that's a good idea
- # [16:51] <ted> khuey: definitely, it's the only thing we've got to work on here
- # [16:51] <khuey> ted: well removing ehsan's derived safe stuff cut 1.5 MB of memory usage :-/
- # [16:51] <khuey> ted: but it's possible that their other stuff uses templates in ways that use lots of memory
- # [16:51] <ehsan> I was hoping that it would cut out more
- # [16:51] <khuey> ehsan: me too :-)
- # [16:51] <ehsan> apparenlty the m-i c-set I pushed it over was just a bit over the edge
- # [16:51] <khuey> ehsan: but look at it this way, you can reland it after we get through the crisis ;-)
- # [16:51] <catlee> ted: ok, so assuming this works, we just need to deploy vs2005 and figure out the right PATHs, etc. like rail did for vs2010 to put in the mozconfig
- # [16:51] <ehsan> khuey: heh, I'm not in a rush for now :)
- # [16:51] * Quits: @dveditz (dveditz@moz-34991AF4.dhcp.cruzio.com) (Quit: dveditz)
- # [16:51] <ted> catlee: yeah, sounds reasonable
- # [16:52] * ehsan goes back to fight the windows loader
- # [16:52] <khuey> ehsan: so if the builds on m-i now go green I think we can reopen the tree
- # [16:52] <ted> catlee: how do we make things work out of the box, do we just set PATH etc like start-msvcN.bat does in the buildbot start scripts?
- # [16:52] <khuey> subject to metering
- # [16:52] <khuey> and people not landing crazy stuff
- # [16:52] <ehsan> khuey: heavy metering
- # [16:52] <ted> (for the default compiler)
- # [16:52] * Quits: zuzelvp (zuzelvp@F2D29657.F60B0462.67AC9B1.IP) (Ping timeout)
- # [16:52] <ehsan> khuey: or people testing their crazy stuff with PGO builds on try, with try links to prove it to the sheriff ;)
- # [16:52] <ehsan> who's dolske btw
- # [16:53] <khuey> heh
- # [16:53] <ted> it's kind of hard
- # [16:53] <ted> because any individual checkin might be ok
- # [16:53] <ted> but a combination of them could put us right back
- # [16:53] <ehsan> if he's not around or he's busy, I'd happily take over sheriffing if we reopen
- # [16:53] <khuey> anyways, I think we should keep driving on the other patches
- # [16:53] <khuey> more headroom is always better
- # [16:53] <ehsan> ted: for sure, but if someone has a 2mb patch they wanna land, that's a clear candidate for breaking things ;;)
- # [16:53] <catlee> ted: right now buildbot is started inside a shell where we've run start-msvc8.bat on win32
- # [16:54] * khuey is pretty much burned out, and has an exam tomorrow
- # [16:54] <ted> catlee: gotcha
- # [16:54] <ehsan> khuey: seconded
- # [16:54] <catlee> so all buildbot jobs inherit those settings
- # [16:54] <khuey> so I'm almost done with this for the time being
- # [16:54] <catlee> which is a pain
- # [16:54] <catlee> trying to move away from that now
- # [16:54] * Joins: gandalf (zbraniecki@moz-9F56DED7.neoplus.adsl.tpnet.pl)
- # [16:54] <ted> catlee: yeah, sucks there's not a better way to do it
- # [16:54] <ehsan> khuey: let me know if you have tasks I can help with
- # [16:54] <ted> but the hack we worked out with rail should suffice
- # [16:54] <ehsan> (not your exam ;)
- # [16:54] <catlee> ted: well, figuring out the magic environment settings is better I think
- # [16:54] <catlee> it's changable at runtime at least
- # [16:54] <ted> khuey: i'll take your exam for you if it gives you more time to do this :-P
- # [16:54] * Joins: dalsh (dalsh@26BD8114.B3D15CEC.163E2FB.IP)
- # [16:55] <khuey> ted: I think that's frowned upon
- # [16:55] <ted> good, because i don't know what it is and i haven't studied
- # [16:55] * armenzg_brb is now known as armenzg
- # [16:55] <ted> why were we building Skia on Windows?
- # [16:56] <ted> why are we building Skia at all, did i miss something?
- # [16:56] <khuey> "for shits and giggles"
- # [16:56] <khuey> as we determined in #gfx yesterday ;-)
- # [16:56] <ted> hah
- # [16:56] <ted> mm
- # [16:56] <catlee> nice
- # [16:56] <ted> catlee: i wonder if we can measure the peak vsize of the linker during the build process
- # [16:56] <catlee> let's just add random code into the tree and build it?
- # [16:56] <ted> to monitor this situation
- # [16:56] <khuey> catlee: more or less
- # [16:56] * Quits: jhk_ (jiggy@8E6C34C1.A3F9767A.1C37C358.IP) (Connection reset by peer)
- # [16:56] <ehsan> I wonder how that switch got turned on
- # [16:56] <catlee> ted: Tlinkvsize
- # [16:57] <ted> catlee: i am so seriously you guys
- # [16:57] <khuey> ehsan: oh, that part was intentional
- # [16:57] <ehsan> we should never be building code which we don't use
- # [16:57] <ehsan> khuey: building it on windows?
- # [16:57] <khuey> yes
- # [16:57] <catlee> ted: how does one do that on windows?
- # [16:57] <catlee> how is khuey measuring it?
- # [16:57] <ted> catlee: probably looking at task manager
- # [16:57] <khuey> http://hg.mozilla.org/mozilla-central/rev/215593486382
- # [16:57] <ehsan> catlee: I bet someone can write an app which uses perf counters or something
- # [16:57] * ewong is now known as ewong|Zzz
- # [16:57] <catlee> yeah, that's not gonna work
- # [16:57] <khuey> catlee: looking at private bytes in task manager
- # [16:57] * Joins: jhk_ (jiggy@8E6C34C1.A3F9767A.1C37C358.IP)
- # [16:57] <ehsan> or just the process query APIs
- # [16:57] <khuey> you can probably automate this with perfmon or something
- # [16:57] * Quits: By-Tor (bytor@moz-46974D0B.dyn.optonline.net) (Quit: Leaving)
- # [16:57] <ehsan> shouldn't be that hard
- # [16:58] <ted> i am absolutely sure you can do it
- # [16:58] * Joins: kumar (kmcmillan@moz-F2D05B8.c3-0.stk-ubr1.chi-stk.il.cable.rcn.com)
- # [16:58] <catlee> yeah
- # [16:58] <ted> programmatically
- # [16:58] <ehsan> someone should file that bug
- # [16:58] <catlee> we can submit it to graph server
- # [16:58] * ted will file it
- # [16:58] <ehsan> if it is not already
- # [16:58] <catlee> or just report it
- # [16:58] <ted> yeah
- # [16:58] <ted> as long as we have some way to track it
- # [16:58] <catlee> give us a heads up at least
- # [16:58] <ehsan> ted: thanks, cc me as well please
- # [16:58] <ted> we'll be able to detect when we're near danger
- # [16:58] <ted> will do
- # [16:58] <catlee> like number of static constructors
- # [16:58] <ted> yeah
- # [16:58] <ehsan> ted: as well as the rest of the windows usual suspects (bbondy, jimm, etc)
- # [16:58] <catlee> would be nice on other platforms too
- # [16:59] * Joins: mdas (mdas@8CB764BC.D30B51A1.412CF160.IP)
- # [16:59] <glandium> ted: cc me too
- # [16:59] * ehsan goes back to fight the windows loader for realz this time
- # [16:59] <catlee> linking xul is just painful
- # [16:59] * Joins: dbradley (dbradley@moz-8FACAA93.fuse.net)
- # [16:59] <catlee> it usually hangs my machine for a minute or two
- # [16:59] <ted> yeah
- # [16:59] <ted> it's pretty intensive
- # [16:59] <ted> chews up IO for a bit, then burns one core of your CPU for 1.5 hours+
- # [17:00] <catlee> I'm talking on linux
- # [17:00] * Joins: regen (Miller@moz-4F377C4E.adsl.dynamic.seed.net.tw)
- # [17:00] * Joins: smaug (chatzilla@moz-78A86AC.pp.htv.fi)
- # [17:00] <catlee> it basically flushes all my caches
- # [17:00] <ted> ah
- # [17:00] <ted> yeah, that doo
- # [17:00] <ted> too
- # [17:00] <ted> "buy more ram"
- # [17:00] <Ms2ger> *expense more ram
- # [17:01] <ted> srsly
- # [17:01] <ted> more ram for everyone!(*)
- # [17:01] <ted> (* ask your manager)
- # [17:01] <catlee> 4G should be enough for anyone!
- # [17:01] * Quits: dao (dao@moz-7C5E3898.superkabel.de) (Ping timeout)
- # [17:01] <Ms2ger> Also ponies!
- # [17:01] <Ms2ger> And dammit, I don't have a manager :(
- # [17:01] * Joins: Jake (Jake@moz-5D28CC35.phlapafg.dynamic.covad.net)
- # [17:01] * Joins: dao (dao@moz-7C5E3898.superkabel.de)
- # [17:02] <mounir> Ms2ger: ask jst ;)
- # [17:02] <ted> Ms2Manager
- # [17:02] * Quits: regen (Miller@moz-4F377C4E.adsl.dynamic.seed.net.tw) (Quit: regen)
- # [17:02] <mounir> Ms2ger: I'm pretty sure he can help with that
- # [17:02] <Ms2ger> Now all I need is a pony
- # [17:02] <Ms2ger> ted? :)
- # [17:03] <catlee> ponies we can do
- # [17:03] <catlee> unicorns you have to wait
- # [17:04] <ted> heh
- # [17:04] * glazou is now known as glazou_bbl
- # [17:04] * Quits: glazou_bbl (glazou@moz-204094DD.disruptive-innovations.fr) (Quit: glazou_bbl)
- # [17:04] <ted> Ms2ger: i'm pretty sure i've told you i can get you a pony, you just need to provide your name and address
- # [17:04] <Ms2ger> Didn't I tell you to ship it to the Paris office? :)
- # [17:04] * jfkthame is now known as jfkthame_afk
- # [17:05] <mounir> vingtetun: do you know Ms2ger and you are the same person btw?
- # [17:06] * Quits: dao (dao@moz-7C5E3898.superkabel.de) (Ping timeout)
- # [17:06] * Joins: bjacob (bjacob@F2D29657.F60B0462.67AC9B1.IP)
- # [17:07] * Joins: akeybl (akeybl@moz-52D39FF6.hsd1.ca.comcast.net)
- # [17:07] * Joins: dao (dao@moz-7C5E3898.superkabel.de)
- # [17:07] <vingtetun> mounir: hu?
- # [17:07] <vingtetun> i'm pretty sure i would be aware of it if that's the case
- # [17:08] * Joins: zuzelvp (zuzelvp@F2D29657.F60B0462.67AC9B1.IP)
- # [17:08] <smaug> Ms2ger has apparently quite a few personalities
- # [17:09] * Joins: smooney (smooney@moz-3C7A0050.hsd1.ca.comcast.net)
- # [17:10] <ted> btw
- # [17:10] <ted> firebot: https://bugzilla.mozilla.org/show_bug.cgi?id=710712
- # [17:10] <Ms2ger> That's what my other personality said
- # [17:10] <firebot> ted: Sorry, I've no idea what 'https://bugzilla.mozilla.org/show_bug.cgi?id=710712' might be.
- # [17:10] * philor|away is now known as philor
- # [17:10] <Ms2ger> firebot, bug 710712?
- # [17:10] <firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=710712 nor, --, ---, nobody, NEW, Measure peak virtual memory usage of link.exe process during libxul PGO link
- # [17:10] <ted> lame
- # [17:12] * Quits: smooney (smooney@moz-3C7A0050.hsd1.ca.comcast.net) (Quit: smooney)
- # [17:13] * Joins: northWind (northWind@moz-20E58859.cable.teksavvy.com)
- # [17:18] * jfkthame_afk is now known as jfkthame
- # [17:18] * Quits: srinivas (chatzilla@moz-2790C4FB.hsd1.ca.comcast.net) (Ping timeout)
- # [17:18] * coop|buildduty is now known as coop|afk
- # [17:20] <Callek> its weeks like this I _really_ wish that MS's linker could handle virtual memory better.... :-)
- # [17:20] <Callek> but then again I have _no_ idea of its internals or if its even feasable
- # [17:20] * Joins: fs (Elchi3@B9C9103E.56629902.2EC4CA51.IP)
- # [17:21] <ted> if they shipped a 64->32 cross-toolchain this would be a non-issue
- # [17:21] <Callek> ted: that too :-)
- # [17:22] * shorlander-away is now known as shorlander
- # [17:23] <espindola> ok to land https://bug696483.bugzilla.mozilla.org/attachment.cgi?id=581626
- # [17:23] <espindola> ?
- # [17:23] <espindola> r=mak and js only
- # [17:24] <robcee> why do I get the feeling the build-infra is a little stressed this morning?
- # [17:24] <robcee> pending and running jobs seem reasonable
- # [17:25] <Ms2ger> espindola, a=me
- # [17:25] <robcee> seeing a lot of oranges related to browser_auth
- # [17:25] <robcee> on fx-team and m-c
- # [17:25] <espindola> Ms2ger, thanks
- # [17:26] <ted> khuey: you said you've been measuring private bytes?
- # [17:26] <ted> that's not actually what we want, right?
- # [17:26] * Quits: @mkaply (chatzilla@moz-92EDDD02.lightspeed.austtx.sbcglobal.net) (Ping timeout)
- # [17:26] * Quits: tH (Rob@cpc16-seac19-2-0-cust234.7-2.cable.virginmedia.com) (Ping timeout)
- # [17:26] <ted> we want to measure virtual memory size
- # [17:26] <edmorley> gabor: unfortunately the try run shows M1 broken https://tbpl.mozilla.org/?tree=Try&rev=4a33ea8a7a0a
- # [17:27] * Joins: tH (Rob@cpc16-seac19-2-0-cust234.7-2.cable.virginmedia.com)
- # [17:27] <khuey> ted: yeah, but it's close enough I think
- # [17:27] <mounir> khuey: hey, I got no news from you in fast-package bugs
- # [17:27] <khuey> ted: I think the relative numbers are valid, at least
- # [17:28] <khuey> mounir: I've been spending the last 4 days trying to get the tree open, cut me some slack
- # [17:28] * Quits: jimb (user@moz-48241F6F.hsd1.or.comcast.net) (Connection reset by peer)
- # [17:28] <ted> hrm
- # [17:28] * Quits: protz (protz@moz-E29D2A15.inria.fr) (Quit: Leaving)
- # [17:28] <ted> khuey: i can get Working Set Size pretty easily
- # [17:28] <ted> as well as Private bytes
- # [17:28] <glandium> ted: don't we have vss in about:memory on windows ?
- # [17:28] <khuey> ted: what does Working Set Size actually mean?
- # [17:28] <ted> glandium: it's easy to get in-process
- # [17:29] <ted> it's harder to get out-of-process, apparently
- # [17:29] <gabor> edmorley: hmmm... it's weird since those ones are failing in all the previous patches as well
- # [17:29] * mkelly is now known as mkelly|wfh
- # [17:29] <glandium> ted: annoying os
- # [17:29] <ted> http://msdn.microsoft.com/en-us/library/windows/desktop/ms684879%28v=vs.85%29.aspx
- # [17:29] <ted> glandium: yeah, it's weird
- # [17:29] * Quits: Suresh (chatzilla@21EF890.D0998290.EB06F97B.IP) (Quit: Suresh)
- # [17:29] <ted> in-process you can just call GlobalMemoryStatusEx
- # [17:30] * Quits: stransky (stransky@moz-107AD163.redhat.com) (Quit: Connection reset by beer)
- # [17:30] <gabor> edmorley: or am I reading it completly wrong?
- # [17:30] <ted> and that gives you ullAvailVirtual / ullTotalVirtual
- # [17:30] * Joins: kdcw (kdc@moz-F7413045.pk.shawcable.net)
- # [17:30] * mdas is now known as mdas|breakfast
- # [17:30] <mounir> khuey: oups, sorry I didn't realize
- # [17:30] <edmorley> gabor: what do you mean by previous patches?
- # [17:31] * Joins: jdm (jdm@moz-9AEDE212.cpe.net.cable.rogers.com)
- # [17:31] <glandium> ted: isn't paged pool+unpages pool what we should be looking at?
- # [17:31] <glandium> s/unpaged/nonpaged/
- # [17:31] * Joins: necolas (necolas@moz-4197BE5D.bb.sky.com)
- # [17:31] <edmorley> gabor: the tbpl view shown at that link, is all the builds/tests run on the topmost changeset, not one per changeset in the push
- # [17:32] <ted> glandium: i honestly have no idea
- # [17:32] <ted> if that sounds right that's easy enough to do
- # [17:32] <gabor> edmorley: ah well... I read it completly wrong
- # [17:32] <gabor> edmorley: ok then I'll check what is going on :) thanks for the help
- # [17:33] <edmorley> gabor: no worries :-) if you click the orange 1s then use the view brief/full log bottom left, hopefully they should help
- # [17:33] * Joins: tfair (tfairey@moz-57178BD3.hsd1.ca.comcast.net)
- # [17:33] * Joins: mccr8 (mccr8@moz-BBE3ABD.mv.mozilla.com)
- # [17:34] * Joins: Ami_Ty (Amie@moz-BBE3ABD.mv.mozilla.com)
- # [17:34] <edmorley> gabor: just be aware aiui builds are only being retained for 4 days at the moment due to the storage issues, so grab the logs sooner rather than later
- # [17:34] * Joins: gwagner (idefix2@moz-3DFE0B51.hsd1.ca.comcast.net)
- # [17:35] * Quits: akeybl (akeybl@moz-52D39FF6.hsd1.ca.comcast.net) (Quit: Leaving...)
- # [17:35] * Joins: Radioactiveman (Radioactiv@moz-3C0E8DC1.dip.t-dialin.net)
- # [17:36] * Quits: gwagner (idefix2@moz-3DFE0B51.hsd1.ca.comcast.net) (Quit: gwagner)
- # [17:36] * Quits: florian (florian@moz-87C33FDA.kimsufi.com) (Quit: Instantbird 1.2a1pre)
- # [17:36] <Radioactiveman> Hi, I have noticed a different rendering with Aurora and want to know now if it is a bug.
- # [17:37] <Radioactiveman> The page is this one: http://mumble.sourceforge.net/
- # [17:37] * Quits: mayhemer (Miranda@B3D46202.F87A741B.F23860FD.IP) (Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org)
- # [17:37] <Radioactiveman> With Aurora "navigation", "search" and the "toolbox" are at the bottom.
- # [17:37] <Radioactiveman> FF8 displays it correctly.
- # [17:38] * Joins: joduinn-mtg (joduinn@moz-3244C0A7.dsl.dynamic.sonic.net)
- # [17:39] * mdas|breakfast is now known as mdas|afk
- # [17:39] <Radioactiveman> Do you get the same behaviour? And should I file a bugreport?
- # [17:39] * Quits: joduinn-mtg (joduinn@moz-3244C0A7.dsl.dynamic.sonic.net) (Quit: joduinn-mtg)
- # [17:39] * BenWa|sms is now known as BenWa
- # [17:40] * Joins: beaufour (beaufour@18D5CC88.C7EE4FB2.ECED8BE3.IP)
- # [17:40] * joduinn-it-helpdesk is now known as joduinn-coffee
- # [17:40] <bsmedberg> Radioactiveman: I see that behavior, but I don't know if it's a bug (yet)
- # [17:42] * Joins: lmandel (lmandel@moz-113D7D7C.cable.teksavvy.com)
- # [17:44] * Joins: nhirata (nhirata.bu@moz-BBE3ABD.mv.mozilla.com)
- # [17:44] <bjacob> glandium: re bug 709914 (split angle out) what is the rationale to fold ANGLE with video stuff? is it that fewer bigger libs mean better perf?
- # [17:47] * Quits: josh (josh@moz-2EE66546.nyc.res.rr.com) (Quit: josh)
- # [17:47] * coop|afk is now known as coop|buildduty
- # [17:48] <davidb> ttaubert: will the new page tab (thumbnail stuff) be keyboard accessible? (shouldn't be hard at all)
- # [17:48] <davidb> also should be really easy to expose page title or something for accessibility
- # [17:48] <bsmedberg> Radioactiveman: do you have some time to help investigate it?
- # [17:48] <glandium> bjacob: less files
- # [17:49] <bjacob> glandium: ok. also, ignore my last comment
- # [17:50] * Joins: nattokirai (nattokirai@moz-C97EE7C8.dynamic.ppp.asahi-net.or.jp)
- # [17:51] <bjacob> glandium: but please check comment 12. wouldn't it be neat?
- # [17:51] * mcote|afk is now known as mcote
- # [17:52] * Joins: jimb (user@moz-48241F6F.hsd1.or.comcast.net)
- # [17:52] <ttaubert> davidb: um, good question. I didn't focus on accessibility, yet :/
- # [17:52] <davidb> ttaubert: i could probably add it, if you ping me when you think the time is right.
- # [17:52] <philor> hmm
- # [17:52] <mounir> jlebar: ping
- # [17:53] <jlebar> mounir, hey
- # [17:53] <firebot> Check-in: http://hg.mozilla.org/mozilla-central/rev/a2928a1ffde5 - Rafael Ávila de Espíndola - Bug 696483 - Improve error handling. r=mak. a=Ms2ger.
- # [17:53] <davidb> ttaubert: normally i'd worry if you left it too late but in this case I don't imagine trouble
- # [17:53] <philor> Unfocused: do you have a patch for pref-flipping?
- # [17:53] <edmorley> Moth orange much :-(
- # [17:53] <mounir> jlebar: i wanted to speak a bit about hal
- # [17:53] <jlebar> mounir, sure
- # [17:54] <philor> yeah, I'm very very tempted to just revert the piss out of that tree
- # [17:54] <mounir> jlebar: currently, Hal.h has a weird hack that make all methods inside it being in hal:: hal_impl:: and hal_sandbox::
- # [17:54] <philor> STOP FUCKING CHECKING IN ON BUSTAGE
- # [17:54] <jlebar> mounir, right
- # [17:54] <philor> STOP STOP STOP IT
- # [17:54] <mounir> jlebar: would that make sense to you to change that hack a little
- # [17:54] * Quits: necolas (necolas@moz-4197BE5D.bb.sky.com) (Client exited)
- # [17:54] * Joins: ericjung (Mibbit@5210CFD5.1A5EA44.72B23B3D.IP)
- # [17:54] <davidb> ttaubert: I added a whiteboard thing [a11y-ping=davidb] so that it can be on the radar
- # [17:54] <edmorley> philor: want me to close?
- # [17:55] <mounir> jlebar: basically, Hal.h wouldn't #include HalImpl.h and HalSandbox.h but those headors would include Hal.h with #define HAL_NAMESPACE
- # [17:55] <vingtetun> dolske: can i land bug 710548? it only touches the b2g/ directory
- # [17:55] <ttaubert> davidb: cool, thanks
- # [17:55] * mdas|afk is now known as mdas
- # [17:55] <davidb> np
- # [17:55] <mounir> so when including HalImpl.h you will get the methods from Hal.h in the hal_impl namespace
- # [17:55] * Quits: biesi (cbiesinger@moz-5EE692A7.ca.hfc.comcastbusiness.net) (Ping timeout)
- # [17:55] <mounir> *and* we would be able to create HalInternal.h that would be used the same way by HalImpl.h and HalSandbox.h
- # [17:55] <jlebar> mounir, So Hal.h says ifndef HAL_NAMESPACE then define HAL_NAMESPACE hal
- # [17:55] <jlebar> ?
- # [17:55] <mounir> jlebar: yes
- # [17:56] <jlebar> mounir, If you're going to go that far, then why not create
- # [17:56] <mounir> jlebar: this is more or less what is currently done but in addition, it is including other files
- # [17:56] <jlebar> halcommon, which is included by everyone, then separate hal.h, halimpl.h, halsandbox.h, halinternal.h
- # [17:56] * mak is now known as mak|afk
- # [17:57] <Callek> khuey: uconvd changes didn't stick?
- # [17:57] <bjacob> glandium: your patch doesn't apply cleanly here, layout/media/Makefile.in doesn't exist
- # [17:57] <Callek> bustage/breakage or just "not worth it"?
- # [17:57] * Joins: fzzzy (donovan@moz-2B41AF9B.lightspeed.mtvwca.sbcglobal.net)
- # [17:57] <mounir> jlebar: Hal.h would be the public interface, I don't see why we should move this to another file
- # [17:57] <mounir> but if we do, I would prefer HalPublic ;)
- # [17:57] <Ms2ger> ApiHalPublic.h?
- # [17:57] <Radioactiveman> bsmedberg: no, sorry. I don't have the knowledge to fix it so I posted it here.
- # [17:57] * bz_sleep is now known as bz
- # [17:57] * Quits: espindola (espindola@4295960D.E84B0A4A.6F478678.IP) (Ping timeout)
- # [17:57] <jlebar> mounir, You're right, no need for HalCommon.
- # [17:58] <philor> edmorley: please, you'll do a politer message than me
- # [17:58] <jlebar> That change sgtm, but you should ask cjones, because he didn't like when I tried to do something similar.
- # [17:58] <jlebar> although your approach is cleaner.
- # [17:58] <philor> khuey: you have orange, on mozilla-central, I bet you know what that means
- # [17:58] <mccr8> Callek: factoring out the string conversion didn't really affect the PGO memory usage, if that's what you mean, so I think khuey decided to not bother.
- # [17:58] <glandium> bjacob: because you need the patch from the other bug (thus the dependency)
- # [17:58] <bsmedberg> Radioactiveman: not to fix it, but to find the regression range on nightly builds
- # [17:58] <philor> hsivonen: you have orange
- # [17:58] <philor> robcee: you have orange
- # [17:58] <edmorley> philor: already done (when I saw a "can I land xyz directed to someone else")
- # [17:58] <mounir> jlebar: hope he will agree :)
- # [17:58] <robcee> wonderful
- # [17:59] * philor doesn't bother telling the people who didn't star their Android
- # [17:59] <bsmedberg> using DOM inspector I don't see an obvious reason for that block to be down at the bottom, although it definitely looks like it is clearing the main content like it were a float
- # [17:59] <@bz> bsmedberg: which testcase are we looking at?
- # [17:59] * Joins: espindola (espindola@A0CFA52B.EFBEEE74.491DEE85.IP)
- # [18:00] <bsmedberg> bz: http://mumble.sourceforge.net/
- # [18:00] <robcee> philor: if I could get tbpl to actually give me a summary, I might be able to tell
- # [18:00] <philor> wow, is all of that orange Unfocused's fault?
- # [18:00] <bsmedberg> the "navigation" box is at the bottom sometime after FF8
- # [18:00] <@bz> bsmedberg: the "navigation" theng being all low?
- # [18:00] <@bz> ok
- # [18:00] <@bz> lemme look
- # [18:00] <robcee> philor: I don't know yet
- # [18:00] <philor> and this is *after* the network got fixed?
- # [18:00] <robcee> it's on the fx-team merge as well, obviously
- # [18:01] <philor> robcee: you can bypass the summary loader, which is killing itself hitting bugzilla, by just loading the brief logs
- # [18:01] <robcee> philor: yeah, lie I said elsewhere, the infra'
- # [18:01] <robcee> s feeling a little flakey this morning
- # [18:01] <robcee> or maybe that was here
- # [18:01] <lurking> looks like it started with khuey 's push/backout etc
- # [18:01] <philor> wherein you'll see lots of "*** LOG addons.repository: Requesting https://services.addons.mozilla.org..."
- # [18:02] <@bz> ok
- # [18:02] <@bz> so basic page structure....
- # [18:02] <smaug> jlebar: changing the method name sounds ok
- # [18:02] <@bz> you have a left column
- # [18:02] <@bz> which is just a div
- # [18:02] <@bz> you have a right column which is a 100% div right float
- # [18:02] <jlebar> smaug, sounds good. I'll try to think of something...
- # [18:02] * Quits: daim (David_Mart@779E3E00.1773D26C.C0FF2207.IP) (Client exited)
- # [18:02] <jlebar> smaug, URLIsChromeOrInPref. :)
- # [18:02] <khuey> hmm, did I break something?
- # [18:03] <robcee> not sure if it's you or infrastructure-related
- # [18:03] * Joins: mgoodwin (mgoodwin@1FA05860.3B0FFCAD.302434AB.IP)
- # [18:03] <robcee> could be me or hsivonen too!
- # [18:03] <philor> could be Unfocused
- # [18:03] <@bz> (on its face, the 100% seems bogus)
- # [18:03] * Quits: magsout (magsout@moz-E559D13.fbx.proxad.net) (Ping timeout)
- # [18:03] <robcee> seeing a bunch of browser_auth, browser_select_update, ...
- # [18:03] * Joins: victorporof (victorporo@D0477DB.DA9DCE7D.79933D60.IP)
- # [18:03] <smaug> jlebar: ChromeOrPrivilegedURL ?
- # [18:04] <jlebar> smaug, that could work...
- # [18:05] <@bz> anyway
- # [18:05] * Joins: loadbang (loadbang@moz-2BA0C010.range86-168.btcentralplus.com)
- # [18:05] <@bz> it's a 100% width float, so the in-flow content should end up below it, afaict
- # [18:05] <@bz> it's odd that it didn't
- # [18:05] <dao> espindola: mozilla-central isn't inbound. you're supposed to check the health of the tree before pushing. and if it doesn't look good, don't push.
- # [18:06] <bjacob> glandium: thanks; building locally so i can try helping
- # [18:06] <espindola> dao, sorry, should I revert?
- # [18:06] <philor> don't suppose there's anything obviously wrong in https://bug679588.bugzilla.mozilla.org/attachment.cgi?id=579043 is there?
- # [18:07] <@bz> and that it doesn't in other browsers. Hmm....
- # [18:07] * Joins: magsout (magsout@moz-E559D13.fbx.proxad.net)
- # [18:07] <philor> maybe it was just after that that we started hitting really flaky network, but the stuff in bug 709531 seems to have picked up steam right after it landed
- # [18:09] * Quits: waschtl (waschtl@moz-A4ECE553.hsi4.kabel-badenwuerttemberg.de) (Client exited)
- # [18:09] <bsmedberg> bz: would a regression range be helpful? Radioactiveman says he can help narrow it down if that would be helpful.
- # [18:09] <dao> espindola: that would trigger another set of builds, possibly delaying a hopefully-upcoming push to deal with the orange. probably not helpful at this point.
- # [18:09] <@bz> bsmedberg: I'll have one in a minute
- # [18:09] * Quits: kanru (user@moz-1D154EA4.dynamic.hinet.net) (Ping timeout)
- # [18:10] <@bz> bsmedberg: 2 builds left to test
- # [18:10] <robcee> philor: don't see anything obvious no
- # [18:11] * glob is now known as glob|away
- # [18:11] <@bz> bsmedberg, Radioactiveman: http://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=6dc468c41136&tochange=f41df039db03
- # [18:11] <philor> starting to think the place to fix this is in http://mxr.mozilla.org/mozilla-central/source/toolkit/mozapps/extensions/test/Makefile.in#55
- # [18:12] * Quits: pnemsak (Miranda@moz-3D67D819.rainside.sk) (Quit: pnemsak)
- # [18:13] * davehunt is now known as davehunt|away
- # [18:13] <robcee> Unfocused: I don't suppose you're awake are you?
- # [18:14] <bsmedberg> "implement css3 text-overflow"
- # [18:14] * Joins: caillon (caillon@moz-359E39FC.lightspeed.sntcca.sbcglobal.net)
- # [18:14] <@bz> yes
- # [18:14] <@bz> doesn't seem to be relevant here, though
- # [18:14] * @bz tries to understand why the negative margin matters
- # [18:15] <mounir> jlebar: is WindowIdentifier currently used in m-c outside of hal/?
- # [18:15] * Joins: Sander (chatzilla@moz-B871F4D3.direct-adsl.nl)
- # [18:15] <jdm> bz: probably one of bug 680505 or 677582, I suspect
- # [18:15] <@bz> because it does
- # [18:15] * Quits: magsout (magsout@moz-E559D13.fbx.proxad.net) (Client exited)
- # [18:15] <jlebar> mounir, I hope not. It's used when DOM calls into Hal, but only as an implicit conversion.
- # [18:15] <@bz> possible
- # [18:15] <@bz> So this site has a left-margin on the float
- # [18:16] <@bz> of -13.6em
- # [18:16] * Quits: joduinn-coffee (joduinn@moz-B77DEAEB.mozilla.org) (Quit: leaving)
- # [18:16] <@bz> oho
- # [18:16] <@bz> OHO
- # [18:16] * Joins: joduinn (joduinn@moz-B77DEAEB.mozilla.org)
- # [18:17] <mounir> jlebar: but methods like CancelVibrate() takes a WindowIdentifier as a param...
- # [18:17] <@bz> _Maybe_ it has that margin
- # [18:17] <jlebar> mounir, But WindowIdentifier has an implicit conversion from nsIDOMWindow*.
- # [18:17] <@bz> it loads two stylesheets
- # [18:17] <jlebar> It's horrible, precisely because the public API is a lie.
- # [18:17] <mounir> jlebar: oh, ok
- # [18:17] <@bz> main.css which sets the negative margin
- # [18:17] <@bz> and KHTMLFixes.css which sets the margin back to 0
- # [18:17] <@bz> at least in my trunk build
- # [18:17] <jlebar> mounir, The public API is "takes a nsIDOMWindow*", but the C++ says "takes a WindowIdentifier."
- # [18:17] <jlebar> But I lost that battle.
- # [18:18] * Quits: ashughes (ashughes@moz-4C407EE8.vc.shawcable.net) (Quit: office bound)
- # [18:18] <@bz> if I delete the KHTMLFixes rule, the layout is correct
- # [18:18] <@bz> (which is odd, because in my testing WebKit matches us on the behavior here)
- # [18:18] <khuey> maybe they really really meant khtml
- # [18:18] <@bz> could be
- # [18:18] <khuey> and not webkit
- # [18:18] * Quits: Honza (chatzilla@E07BF19C.5BB5597D.D0083327.IP) (Ping timeout)
- # [18:18] * @bz is checking
- # [18:19] * Joins: squib (squib-@moz-3F6F2A9C.ep.wisc.edu)
- # [18:19] <khuey> mmm sweet beautiful green
- # [18:19] <@bz> } else if (is_khtml) {
- # [18:19] <@bz> importStylesheetURI(stylepath+'/'+skin+'/KHTMLFixes.css');
- # [18:19] <@bz> var is_khtml = navigator.vendor == 'KDE' ||
- # [18:19] <@bz> ( document.childNodes && !document.all && !navigator.taintEnabled );
- # [18:19] <khuey> heh
- # [18:19] <@bz> we removed taintEnabled
- # [18:19] <@bz> but....
- # [18:20] <@bz> nobody else has it either, right?
- # [18:20] <khuey> right
- # [18:20] <khuey> that was gecko only, I think
- # [18:20] * Quits: hub (hub@moz-E2FCA694.figuiere.net) (Ping timeout)
- # [18:20] <@bz> so for us this is fallout from bug 679971
- # [18:20] * @bz is digging more
- # [18:20] <@bz> in Chrome is_khtml is true
- # [18:21] <khuey> cause they don't have it either
- # [18:21] <espindola> llvm might be getting a json parser :-)
- # [18:21] * Quits: gandalf (zbraniecki@moz-9F56DED7.neoplus.adsl.tpnet.pl) (Ping timeout)
- # [18:21] <@bz> right
- # [18:21] * @bz is not sure whether that sheet gets loaded in Chrome
- # [18:21] <@bz> checking now
- # [18:22] * Joins: gandalf (zbraniecki@moz-9F56DED7.neoplus.adsl.tpnet.pl)
- # [18:22] <khuey> http://news.ycombinator.com/item?id=3351962
- # [18:22] <@bz> yes, it is
- # [18:23] <armenzg> is DONTBUILD the correct keyword to avoid building?
- # [18:23] * Joins: damons (gnubeard@moz-A41E6911.hsd1.ca.comcast.net)
- # [18:23] <mounir> armenzg: yes
- # [18:23] * khuey hopes you're not landing onto the bright orange tree
- # [18:23] <@bz> opera possibly gets another codepath here
- # [18:24] <armenzg> thank you
- # [18:24] <@bz> bsmedberg, Radioactiveman: what would help most here now is a smaller testcase
- # [18:24] * Quits: grubshka (grubshka@moz-D9EA9AB9.w109-212.abo.wanadoo.fr) (Ping timeout)
- # [18:24] <@bz> bsmedberg, Radioactiveman: the behavior changed because they started sniffing us as webkit and sending us different styles
- # [18:24] <@bz> bsmedberg, Radioactiveman: but in my testing we handle these styles the same way as WebKit does....
- # [18:25] * joduinn is now known as joduinn-commute
- # [18:25] <philor> sigh, bugzilla has suffered an internal error
- # [18:25] <@bz> bsmedberg, Radioactiveman: So a testcase that doesn't browser-sniff (hence uses the same styles for all browsers) but shows the behavior difference would be a good start
- # [18:25] <lurking> philor: saw that earlier - but it cleared right up on a reload
- # [18:26] <jlebar> philor, you weren't trying to get any work done anyway.
- # [18:26] <jlebar> sigh, bugzilla.
- # [18:26] * bear is now known as bear-afk
- # [18:26] <@bz> yeah
- # [18:26] * robcee raises hand
- # [18:26] <@bz> Opera is NOT getting the same styles we are on this page
- # [18:27] <robcee> I saw it too (the bugzilla error)
- # [18:27] * Joins: bmoss (bmoss@moz-BBE3ABD.mv.mozilla.com)
- # [18:27] <jlebar> s/saw/see.
- # [18:28] <jlebar> and we're back.
- # [18:28] * Quits: jhk_ (jiggy@8E6C34C1.A3F9767A.1C37C358.IP) (Ping timeout)
- # [18:28] <mounir> jlebar: justin.lebar@gmail.com is your committer email address?
- # [18:28] <jlebar> mounir, yes.
- # [18:29] * Joins: cpeterson (cpeterson@moz-BBE3ABD.mv.mozilla.com)
- # [18:29] * Quits: @bz (bzbarsky@moz-69B5879F.bstnma.fios.verizon.net) (Quit: Leaving)
- # [18:29] * Joins: bz (bzbarsky@moz-69B5879F.bstnma.fios.verizon.net)
- # [18:29] * ChanServ sets mode: +o bz
- # [18:30] <philor> coop|buildduty: ping
- # [18:30] <@bz> bsmedberg, Radioactiveman: my current working hypothesis is that the site is somehow relying on a webkit bug and sending the same bug-tageted css to us
- # [18:30] <@bz> bsmedberg, radioactiveman: but it's hard to say
- # [18:31] <khuey> ehsan: based on the results on inbound I think we can reopen the tree subject to careful metering
- # [18:31] <dao> webkit still can't handle my homepage because of a combination of floats and margins, I think
- # [18:31] <khuey> ehsan: once the current orange spell passes, of course
- # [18:31] * Joins: timA (tabraldes@moz-535753DA.hsd1.or.comcast.net)
- # [18:31] <khuey> dolske: ping?
- # [18:31] <@bz> spell duration should be in the handbook
- # [18:31] <khuey> dolske: you're sheriff today
- # [18:31] <khuey> congratulations
- # [18:32] * Quits: cpeterson (cpeterson@moz-BBE3ABD.mv.mozilla.com) (Input/output error)
- # [18:32] <coop|buildduty> philor: pong
- # [18:32] * Quits: Jake (Jake@moz-5D28CC35.phlapafg.dynamic.covad.net) (Quit: Leaving.)
- # [18:33] * Quits: tfair (tfairey@moz-57178BD3.hsd1.ca.comcast.net) (Client exited)
- # [18:33] <philor> coop|buildduty: we're in an unenviable state where the health of the tree is dependent on AMO working, can you tell us when IT tells you (through however many people that goes) that it's unbusted again?
- # [18:34] <coop|buildduty> philor: will do. there are problems in the PHX datacenter right now...that's all I know at present
- # [18:34] <khuey> can we nuke whatever tests are having hte problems?
- # [18:34] <khuey> tests really shouldn't be depending on AMO ...
- # [18:34] <philor> also, could you see about someone expiditing that bug to renew my Valium prescription?
- # [18:34] <philor> khuey: sure, I pointed right at it up there, every single addonmgr browser-chrome test
- # [18:35] <khuey> ugh
- # [18:35] <@bz> hrm
- # [18:35] * @bz vaguely recalls this now
- # [18:35] <@bz> silly addonmgr. :(
- # [18:35] <philor> yeah
- # [18:35] <lurking> philor: you also should file a bug on recovering your wasted bandwidth
- # [18:36] * Quits: timA (tabraldes@moz-535753DA.hsd1.or.comcast.net) (Quit: Instantbird 1.2a1pre)
- # [18:36] * Joins: bnicholson (bnicholson@moz-BBE3ABD.mv.mozilla.com)
- # [18:36] * khuey debates whether or not to disable all their tests
- # [18:36] * Joins: necolas (necolas@moz-4197BE5D.bb.sky.com)
- # [18:37] * Quits: fzzzy (donovan@moz-2B41AF9B.lightspeed.mtvwca.sbcglobal.net) (Quit: fzzzy)
- # [18:37] * philor takes this lag to star other people's Tier 1 Android test failures
- # [18:37] * Joins: clee (clee@moz-BBE3ABD.mv.mozilla.com)
- # [18:37] * Joins: jhk_ (jiggy@8E6C34C1.A3F9767A.1C37C358.IP)
- # [18:37] <philor> we should just give up and actually formally apply inbound's rules to every tree
- # [18:37] <bhearsum> philor: really? those tests depend on AMO?
- # [18:38] <bhearsum> where do i file that bug_
- # [18:38] <dao> well, our users depend on AMO as well
- # [18:38] <bhearsum> it should be mocked out
- # [18:38] <philor> &duplicate-of=709531
- # [18:38] <khuey> our development shouldn't
- # [18:38] <bhearsum> our users depend on the internet, too, but that doesn't mean our tests should
- # [18:38] <dao> AMO being slow or unavailable is quite painful right now
- # [18:39] <mounir> jlebar: so why is WindowIdentifier living in hal?
- # [18:39] * zpao|detached is now known as zpao
- # [18:39] <jlebar> mounir, It's a hal hack.
- # [18:39] <dao> no, our UI doesn't generally depend on the internet
- # [18:39] * khuey doesn't see why people who want to land stuff should care about AMO's availability
- # [18:39] <jlebar> mounir, If you want to implement nsGlobalWindow::GetReallyUniqueIdentifier() and have it return a WindowIdentifier object, great.
- # [18:39] <jlebar> mounir, But at the moment, it's only built up in hal.
- # [18:39] * Quits: jhorak (jhorak@moz-59813FB4.cust.nbox.cz) (Ping timeout)
- # [18:40] * Quits: jwatt (roslea@jwatt.irc.users.mozilla.org) (Ping timeout)
- # [18:40] <jlebar> mounir, But because of the crummy interface, I have to expose it to DOM.
- # [18:40] * armenzg is now known as armenzg_lunch
- # [18:40] * Quits: kaze (kaze@moz-A2A3C9E6.vlan402.asr1.cdg3.gblx.net) (Ping timeout)
- # [18:40] * Joins: jwatt (roslea@jwatt.irc.users.mozilla.org)
- # [18:40] <mounir> jlebar: I guess I will go with the double entry-point
- # [18:41] <dao> bhearsum, khuey: basically, this seems to be a design flaw in the add-ons manager, not in the tests
- # [18:41] <mounir> and the potential link issue should be fixed with my patch for public/private hal apis
- # [18:41] <jlebar> mounir, not really, right?
- # [18:41] <philor> is it just me, or did the flames in phx finally get to tbpl?
- # [18:41] <jlebar> mounir, because these live in hal.h.
- # [18:41] * Joins: cpeterson (cpeterson@moz-BBE3ABD.mv.mozilla.com)
- # [18:42] <philor> even if we didn't have a single test that depends on amo in any way, the tree being open depends on either totally winning the WOO, or on bmo
- # [18:42] * Joins: myk (myk@moz-C03D0C61.vlan426.asr1.sfo1.gblx.net)
- # [18:42] <jlebar> mounir, And everything in hal.h is included in hal_sandbox:: and hal_impl::
- # [18:42] <bhearsum> are these new tests?
- # [18:42] <mounir> jlebar: if Vibrate(WindowId) lives only in HalInternal.h it shouldn't be visible outside of hal/
- # [18:42] <jlebar> mounir, HalInternal.h is for the hal:: namespace?
- # [18:42] <mounir> jlebar: no, hal_impl and hal_sandbox
- # [18:42] <khuey> haven't we learned by now that wars on abstract concepts can't be won?
- # [18:42] <mounir> jlebar: do we need Vibrate(WindowId) for hal:: ?
- # [18:43] * lsblakk|afk is now known as lsblakk
- # [18:43] <jlebar> mounir, hal_sandbox calls hal::Vibrate(WindowId)
- # [18:43] <philor> no, some sort of harness screwup, but it doesn't matter because we can't reopen without bmo, and we can't reopen without tbpl
- # [18:43] <dao> bhearsum, khuey, philor: bug 657581 might be relevant
- # [18:43] <mounir> jlebar: sic :(
- # [18:43] <jlebar> mounir, It's very messed up. What you're moving towards is the API I proposed in the first place.
- # [18:43] <jlebar> mounir, have hal_public:: with only public methods.
- # [18:43] * Joins: juanb (jbecerra@moz-BBE3ABD.mv.mozilla.com)
- # [18:43] <bhearsum> dao: thanks
- # [18:43] <jlebar> then that calls hal_private::, which calls hal_sandbox::
- # [18:44] <jlebar> mounir, hal_sandbox calls back up into *hal_private* but not into hal::
- # [18:44] <jlebar> then hal_private calls down into hal_impl.
- # [18:44] * Joins: timA (tabraldes@moz-535753DA.hsd1.or.comcast.net)
- # [18:44] <lurking> khuey: \o/ 4 green pgo's
- # [18:44] * Quits: damons (gnubeard@moz-A41E6911.hsd1.ca.comcast.net) (Quit: damons)
- # [18:44] <blassey> what is the plan for getting the tree re-opened
- # [18:45] * Quits: past (past@moz-7A93B333.dsl.dyn.forthnet.gr) (Input/output error)
- # [18:45] <jlebar> I still don't understand why cjones didn't like this. He said that what we have is simpler.
- # [18:45] * Quits: Jesse (jruderman@moz-537BCF9.hsd1.ca.comcast.net) (Quit: Jesse)
- # [18:45] <Ms2ger> blassey, fix the orange and meter
- # [18:45] <Ms2ger> AIUI
- # [18:45] <blassey> backout the push?
- # [18:45] <blassey> so, I'd ask to get to the front of any metered queue
- # [18:46] <khuey> backout the network?
- # [18:46] <khuey> margaret wants to be first, I tink
- # [18:46] <khuey> *think
- # [18:46] <mounir> jlebar: actually, if WindowIdentifier.h isn't exposed outised of hal/, hal::Vibrate(WindowId) wouldn't be callable outside of hal/
- # [18:46] <blassey> khuey: didn't know it was a network issue
- # [18:46] <mounir> I mean, it will likely produce a compile error
- # [18:46] <blassey> khuey: the updater is broken for mobile
- # [18:46] <khuey> blassey: well, the network issue is the immediate problem
- # [18:46] <khuey> blassey: ah, that's fun
- # [18:46] <khuey> I think you can go first ;-)
- # [18:46] <mounir> jlebar: sorry for that annoying discussion :)
- # [18:46] <blassey> yea...
- # [18:46] <jlebar> mounir, That's true. You won't even be able to get an object of that type.
- # [18:47] <catlee> ehsan: ash has been dormant for a few weeks
- # [18:47] <catlee> there are no builds, and no updates
- # [18:48] <margaret> khuey: blassey also knows about what i want to land, so he could land it too
- # [18:49] <blassey> margaret: you've got green try runs for your patch right?
- # [18:49] <margaret> yeah
- # [18:49] <margaret> i have it all ready to go - so i can land it whenever i get the go-ahead
- # [18:51] <edmorley> dougt: are you the best person to ask for review on bug 704056?
- # [18:51] <blassey> edmorley: yes, he is
- # [18:52] <khuey> jdm++
- # [18:52] <jdm> ?
- # [18:52] * Quits: gandalf (zbraniecki@moz-9F56DED7.neoplus.adsl.tpnet.pl) (Ping timeout)
- # [18:52] * Joins: Mook_as (mook@moz-1FCC0032.activestate.com)
- # [18:52] * Joins: gandalf (zbraniecki@moz-9F56DED7.neoplus.adsl.tpnet.pl)
- # [18:52] <edmorley> blassey: thanks :-)
- # [18:53] <khuey> jdm: choosing joe
- # [18:53] * Joins: fzzzy (donovan@moz-BBE3ABD.mv.mozilla.com)
- # [18:54] * Ms2ger imagines joe coming out of a Pokéball
- # [18:54] <Bas> Does anyone here know the new cross-thread ipdl stuff?
- # [18:54] * Joins: AutomatedTester (AutomatedT@moz-21328E59.as13285.net)
- # [18:55] <sewardj> davidb: you're not dbaron, right?
- # [18:55] <Ms2ger> No
- # [18:55] * Joins: ashughes (ashughes@21B7B9F2.B87E9213.6E712CE2.IP)
- # [18:55] <Ms2ger> dbaron is dbaron
- # [18:55] <Callek> sewardj: david-b is david boswell :-)
- # [18:56] <sewardj> !seen dbaron
- # [18:56] <firebot> dbaron was last seen 13 hours, 41 minutes and 58 seconds ago, saying 'birtles, pong... or was it just the question roc answered already?' in #gfx.
- # [18:56] <khuey> who is distinct from davidb
- # [18:56] <smaug> firebot: dbaron
- # [18:56] <firebot> smaug: dbaron is David Baron <dbaron@mozilla.com>, bugmail <dbaron@dbaron.org>, layout developer
- # [18:56] <smaug> firebot: davidb
- # [18:56] <firebot> smaug: everyone knows that! davidb is a frood who knows where his towel is AND exposes gecko to everyone (bolterbugz@gmail.com)
- # [18:56] * Quits: Enn (enn@F2D29657.F60B0462.67AC9B1.IP) (Ping timeout)
- # [18:56] <sewardj> ok, thx all
- # [18:56] <Callek> or: see mozillians
- # [18:56] <Callek> :-P
- # [18:56] <Callek> www.mozillians.org
- # [18:57] <@bz> we have a hard time answering questions
- # [18:57] * Joins: cpearce (chatzilla@moz-55A2BC7D.xdsl.xnet.co.nz)
- # [18:57] <@bz> tend to go off on tangents
- # [18:57] <khuey> srsly
- # [18:57] * Quits: bmoss (bmoss@moz-BBE3ABD.mv.mozilla.com) (Ping timeout)
- # [18:57] <philor> we prefer teaching a man how to fish. for a Cambodian fish. which went extinct 200 years ago.
- # [18:58] * Joins: mayhemer (Miranda@B3D46202.F87A741B.F23860FD.IP)
- # [18:58] * Quits: tabb0t (tabb0t@4A77860B.E4AC1459.FFE8C4AF.IP) (Ping timeout)
- # [18:58] <jlebar> mounir, should I make a new folder in dom/ for this iframe queryable interface?
- # [18:58] <jlebar> mounir, or should it go somewhere else?
- # [18:59] <Ms2ger> For what?
- # [18:59] <mounir> jlebar: what do you have in mindL
- # [18:59] <mounir> ?
- # [18:59] <glandium> ah it was about time, firefox bloat finally made it to slashdot
- # [19:00] * KaiRo_away is now known as KaiRo
- # [19:00] <jlebar> glandium, google should be playing ./ for all this advertising.
- # [19:00] <jlebar> mounir, for the interface that nsHTMLFrameElement and nsHTMLIFrameElement will be implementing.
- # [19:00] <khuey> heh
- # [19:00] <khuey> it took half a week
- # [19:00] <jlebar> mounir, which has mozQueryable and mozQueryInner
- # [19:00] * khuey is impressed
- # [19:00] * Quits: rshetty (quassel@6F1A83E2.CC63714F.C842849F.IP) (Ping timeout)
- # [19:01] * Joins: akeybl (akeybl@moz-BBE3ABD.mv.mozilla.com)
- # [19:01] <ted> glandium: someone should note that chrome can't build at all in this config
- # [19:01] <ted> for the same reasons
- # [19:01] <jlebar> ted, Do you have a citation for that?
- # [19:01] * zpao is now known as zpao|detached
- # [19:02] * Joins: tfair (tfairey@moz-BBE3ABD.mv.mozilla.com)
- # [19:02] <ted> hmm
- # [19:02] * Quits: dao (dao@moz-7C5E3898.superkabel.de) (Connection reset by peer)
- # [19:02] * Joins: bmoss (bmoss@moz-BBE3ABD.mv.mozilla.com)
- # [19:02] <khuey> well, there's http://groups.google.com/a/chromium.org/group/chromium-dev/browse_thread/thread/533e94237691e2f6
- # [19:02] <mounir> jlebar: why a new directory for that?
- # [19:02] <mounir> jlebar: you want a new interface only, right?
- # [19:02] <jlebar> mounir, where else should it go?
- # [19:02] <smaug> khuey: btw, since you're probably busy with bug 709193, who else could look at my little python script?
- # [19:02] <jlebar> mounir, I guess...
- # [19:02] <Ms2ger> And what is it for?
- # [19:02] <ted> jlebar: yeah
- # [19:03] * Joins: rshetty (quassel@D4C76247.B0D2DE49.D30E9BEF.IP)
- # [19:03] <mounir> jlebar: that could live in dom/interfaces/html/nsIDOMMozHTMlFrameElement.idl, couldn't it?
- # [19:03] <smaug> Ms2ger: the python script?
- # [19:03] <ted> http://groups.google.com/a/chromium.org/group/chromium-dev/browse_thread/thread/533e94237691e2f6
- # [19:03] <khuey> smaug: ted
- # [19:03] <jlebar> Ms2ger, https://bugzilla.mozilla.org/show_bug.cgi?id=708176
- # [19:03] <khuey> smaug: maybe jorendorff
- # [19:03] <mounir> jlebar: you have a new version of WindowIdentifier clean-up :)
- # [19:03] <khuey> I can look at it next week
- # [19:03] <Ms2ger> smaug, no, jlebar's thing
- # [19:03] <khuey> if nobody else gets to it faster
- # [19:03] * Quits: blizzard (blizzard@268BDB17.848AC7D2.16021B01.IP) (Ping timeout)
- # [19:03] <ted> jlebar: i'm pretty sure i've heard it said more specifically
- # [19:03] <ted> can't remember where
- # [19:03] <jorendorff> waht?
- # [19:04] <khuey> jorendorff: you wrote xpidl.py right?
- # [19:04] <smaug> khuey: I'm hoping to get it in soon, since I'd like to use it in few places...
- # [19:04] <jorendorff> khuey: no, i think bsmedberg wrote it
- # [19:04] <jorendorff> but i <3 it
- # [19:04] <khuey> smaug: I'm not going to be able to get to it before monday
- # [19:04] <smaug> k
- # [19:04] <khuey> jorendorff: oh I thought you wrote it for quickstubs
- # [19:04] <khuey> smaug: bsmedberg then, perhaps
- # [19:04] <@bz> what component should puppetwidget widgetry bugs go in?
- # [19:04] <jorendorff> i checked it in maybe?
- # [19:04] <khuey> yeah
- # [19:04] <khuey> I think so
- # [19:05] <@bz> bsmedberg, Radioactiveman: in any case, we should get this site to fix its sniffing....
- # [19:05] * Joins: jgriffin (jgriffin@moz-4FBFA41D.hsd1.wa.comcast.net)
- # [19:05] * Joins: anant (anant@moz-BBE3ABD.mv.mozilla.com)
- # [19:05] * Quits: clee (clee@moz-BBE3ABD.mv.mozilla.com) (Quit: clee)
- # [19:06] * Joins: azakai (alon@moz-BBE3ABD.mv.mozilla.com)
- # [19:06] * Joins: terrence (terrence@moz-BBE3ABD.mv.mozilla.com)
- # [19:07] * davehunt|away is now known as davehunt|mtg
- # [19:07] <Wes_> bz: re. our css transition chat the other day: would you believe me if I told you that window.getComputedStyle(el) is observably different than el.offsetHeight = parseInt(el.offsetHeight) ? The former does not cause the transition start "bookmarking" whereas the latter does
- # [19:07] * Quits: dalsh (dalsh@26BD8114.B3D15CEC.163E2FB.IP) (Ping timeout)
- # [19:07] * Joins: Honza (chatzilla@E07BF19C.5BB5597D.D0083327.IP)
- # [19:07] * Joins: Jesse (jruderman@moz-BBE3ABD.mv.mozilla.com)
- # [19:07] * Quits: nattokirai (nattokirai@moz-C97EE7C8.dynamic.ppp.asahi-net.or.jp) (Quit: nattokirai)
- # [19:08] <@bz> Wes_: window.getComputedStyle(el) with nothing else?
- # [19:08] <@bz> Wes_: as in, you don't actually ask the resulting object for any information?
- # [19:08] * Joins: srinivas (chatzilla@moz-B2DF94DC.hfc.comcastbusiness.net)
- # [19:08] <Wes_> bz: yes
- # [19:08] <@bz> Wes_: sure, I believe you
- # [19:08] * Joins: Jake (Jake@moz-5D28CC35.phlapafg.dynamic.covad.net)
- # [19:09] <@bz> Wes_: since it doesn't compute any styles until you ask for them
- # [19:09] <@bz> Wes_: since it's supposed to represent the computed style as of when you query it, not as of when you create the object
- # [19:09] * Quits: Jesse (jruderman@moz-BBE3ABD.mv.mozilla.com) (Quit: Jesse)
- # [19:09] <Wes_> bz: AH. Interesting. So I should be looking to force a compute specifically of the style attribute I'm trying to transition?
- # [19:10] <Wes_> (man I need to write a small test case!)
- # [19:10] * Joins: josh (josh@moz-2EE66546.nyc.res.rr.com)
- # [19:10] <@bz> Wes_: yes
- # [19:10] <decoder> Ms2ger: regarding the i+1, the original code already has i+1. not my call
- # [19:11] <Ms2ger> Yeah, it's wrong, and you're touching the code
- # [19:11] * khuey is now known as khuey|away
- # [19:11] <Wes_> bz: Thanks!
- # [19:11] <decoder> Ms2ger: im not touching the code because I broke it
- # [19:11] <@bz> Wes_: no problem
- # [19:11] * Joins: Jesse (jruderman@moz-BBE3ABD.mv.mozilla.com)
- # [19:11] * sheppy is now known as sheppy-lunch
- # [19:11] <@bz> hold on
- # [19:11] * Joins: rillian (rilllian@moz-F023CC1E.thaumas.net)
- # [19:11] <@bz> what's up with i+1?
- # [19:11] <Ms2ger> Spaces
- # [19:11] <@bz> meh
- # [19:11] <decoder> i changed this patch twice now because of the comment
- # [19:11] <sstangl> khuey|away: slashdot c.c
- # [19:12] <@bz> the comment is really important
- # [19:12] <rillian> josh: we doing the adaptive bandwidth meeting on vidyo?
- # [19:12] <decoder> bz: I agree with that
- # [19:12] <@bz> because it makes it clear why the random stuff is being done...
- # [19:12] * shorlander is now known as shorlander-away
- # [19:12] * @bz really wishes we just didn't have this code. :(
- # [19:12] * Quits: Yoric (Yoric@moz-A2A3C9E6.vlan402.asr1.cdg3.gblx.net) (Input/output error)
- # [19:12] * Joins: jhammink (textual@moz-7B0110AD.mv.mozilla.com)
- # [19:12] * Joins: dveditz (dveditz@moz-BBE3ABD.mv.mozilla.com)
- # [19:13] * ChanServ sets mode: +o dveditz
- # [19:14] <Ms2ger> bz, so, when does Peter get back? :)
- # [19:14] * Joins: gkw (gkw@moz-BBE3ABD.mv.mozilla.com)
- # [19:16] * shorlander-away is now known as shorlander
- # [19:16] * Quits: cpeterson (cpeterson@moz-BBE3ABD.mv.mozilla.com) (Input/output error)
- # [19:17] * AaronMT is now known as AaronMT|mtg
- # [19:17] * Parts: Radioactiveman (Radioactiv@moz-3C0E8DC1.dip.t-dialin.net)
- # [19:17] <bjacob> glandium: i ran locally a the webgl mochitest with your 2 patches, no failure. on what platform did you see a test failure?
- # [19:17] <@bz> Ms2ger: 3 days ago, afaik
- # [19:17] * Joins: blizzard (blizzard@268BDB17.848AC7D2.16021B01.IP)
- # [19:17] <@bz> Ms2ger: double-checking
- # [19:17] <bjacob> glandium: oh, i see the tbpl link now
- # [19:17] * Joins: protz (jonathan@moz-7F6750F6.xulforum.org)
- # [19:18] <bjacob> glandium: so, windows-only?
- # [19:18] * Quits: mayhemer (Miranda@B3D46202.F87A741B.F23860FD.IP) (Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org)
- # [19:20] * Joins: cpeterson (cpeterson@moz-BBE3ABD.mv.mozilla.com)
- # [19:20] * nhirata is now known as nhirata|mtg
- # [19:21] * Joins: mjschranz (mjschranz@moz-57B23647.senecac.on.ca)
- # [19:21] * Joins: damons (gnubeard@moz-BBE3ABD.mv.mozilla.com)
- # [19:21] * Joins: sworkman (sworkman@moz-BBE3ABD.mv.mozilla.com)
- # [19:22] * catlee is now known as catlee-lunch
- # [19:23] * Joins: stevee (Miranda@moz-BEBDF855.cable.virginmedia.com)
- # [19:23] * Joins: hub (hub@21B7B9F2.B87E9213.6E712CE2.IP)
- # [19:24] * armenzg_lunch is now known as armenzg
- # [19:24] * Quits: Boriss (FlyingToas@moz-62AAA429.hsd1.ca.comcast.net) (Quit: Boriss)
- # [19:25] <@bz> Ms2ger: nope
- # [19:25] <@bz> Ms2ger: now it's Friday, looks like
- # [19:25] * Joins: karl (karl@moz-F2842753.jetstream.xtra.co.nz)
- # [19:26] <Ms2ger> Good
- # [19:26] * khuey|away is now known as khuey
- # [19:26] <khuey> sstangl: yeah I've seen it
- # [19:27] * Joins: dbaron (dbaron@moz-27F87443.dsl.dynamic.sonic.net)
- # [19:27] * ChanServ sets mode: +o dbaron
- # [19:28] * Joins: yuan (ywang@moz-C03D0C61.vlan426.asr1.sfo1.gblx.net)
- # [19:28] * Joins: jduell (jduell@moz-2D9EDA98.lightspeed.sntcca.sbcglobal.net)
- # [19:28] * Quits: bmoss (bmoss@moz-BBE3ABD.mv.mozilla.com) (Ping timeout)
- # [19:28] * Quits: joey (chatzilla@moz-781B227D.nys.biz.rr.com) (Input/output error)
- # [19:28] * Joins: Enn (enn@F2D29657.F60B0462.67AC9B1.IP)
- # [19:29] * Joins: biesi (cbiesinger@BEBE4533.89BC370D.6D4D60F7.IP)
- # [19:29] * Quits: ddahl (ddahl@moz-6971CF66.hsd1.il.comcast.net) (Ping timeout)
- # [19:29] * davidb reads back scroll, wonders about Cambodian fish
- # [19:30] <davidb> sewardj: davidb is the dude you like hanging out with at all hands parties.
- # [19:30] * Quits: jfkthame (jfkthame@DBA4BEB5.90783722.9542EC20.IP) (Quit: jfkthame)
- # [19:30] <ehsan> catlee-lunch: I thought that vladg was using an older build
- # [19:31] <ehsan> cause that's what I asked him to do!
- # [19:31] <ehsan> btw
- # [19:31] <glandium> bjacob: the patch modified linking on windows only
- # [19:31] <ehsan> you guys might be interested to know that my new machine can do a clobber PGO build in 60 minutes!
- # [19:31] <bsmedberg> that's good
- # [19:31] <bjacob> glandium: oh ok :)
- # [19:31] <bsmedberg> we should get lots of those
- # [19:31] <khuey> ehsan: mmm
- # [19:31] <khuey> ehsan: specs?
- # [19:31] <ted> ehsan: wow
- # [19:31] <ehsan> as builders!!!
- # [19:32] <khuey> beefy CPU presumably
- # [19:32] <ted> ehsan: that is crazy fast
- # [19:32] <ehsan> it's like the standard IT spec
- # [19:32] <ted> are you sure you're doing a real PGO build? :)
- # [19:32] <Ms2ger> I propose we build all our pgos on ehsan's desktop
- # [19:32] <ehsan> Xeon E3 1225
- # [19:32] <khuey> yeah that doesn't sound right :-)
- # [19:32] <ehsan> SSD
- # [19:32] <ehsan> 16g RAM
- # [19:32] * ted forgets if he's done one on his core i7 recently
- # [19:32] <ted> hmm
- # [19:32] <ehsan> pymake with -j8
- # [19:32] <ted> xeon
- # [19:32] <bsmedberg> only j8?
- # [19:32] <ted> maybe it has a lot more l2/l3 cache?
- # [19:32] <ehsan> cause I only have 4 cores
- # [19:32] <ehsan> perhaps
- # [19:33] <khuey> PGO is pretty much entirely CPU bound
- # [19:33] <ted> i could see that making things faster if it's cpu-bound
- # [19:33] <ehsan> but that's pretty awesome
- # [19:33] * bsmedberg wonders why we're giving people desktops with only 4 cores
- # [19:33] <ted> well, a 4-core xeon
- # [19:33] <ted> is probably pretty pricey
- # [19:33] <ehsan> bsmedberg: I was quite surprised to find out that I only had 4 cores!
- # [19:33] <gps> PGO on windows only uses 2 cores
- # [19:33] <khuey> it really only uses 1
- # [19:33] <ted> gps: only really one during the final link
- # [19:33] <gps> oh, right, it has the 2nd thread for I/O only
- # [19:33] * Joins: joey (chatzilla@moz-781B227D.nys.biz.rr.com)
- # [19:33] <ehsan> oh
- # [19:33] <ted> but more parallelism helps for the rest of the build
- # [19:33] <ehsan> also vs2010, if that makes a difference
- # [19:33] <ted> ehsan: might be slightly faster, but i woudlnt' expect it to be huge
- # [19:34] <ehsan> also I had like a vm running, etc
- # [19:34] * ted finally figured out how to measure virtual size
- # [19:34] <ted> stupid undocumented NtQueryInformationProcess :-P
- # [19:34] <ehsan> ted: awesome!
- # [19:34] <ehsan> hehe
- # [19:34] * Joins: jamesr (jamesr@FB285460.D6CCE4AE.77834EAA.IP)
- # [19:34] <ehsan> ted: codesearch is your friend for that stuff
- # [19:34] <ehsan> so is wine
- # [19:34] <ted> conveniently it gives you PeakVirtualSize
- # [19:34] <ted> yeah, i found it via a wine patch
- # [19:34] <ehsan> (the software, not the drink)
- # [19:34] * ted could use some of the drink
- # [19:35] <ehsan> well yeah, maybe the drink too ;)
- # [19:35] <Ms2ger> ted, if you're using stupid undocumented stuff, I believe the normal approach is not explaining anything in the bug and not mentioning the bug in the commit message :)
- # [19:35] <ted> true
- # [19:35] <sewardj> davidb: yeah, i realised that later. Duh/Sorry.
- # [19:35] <ted> "fix some stuff r=firebot"
- # [19:35] <davidb> sewardj: np :P
- # [19:35] <ted> okay
- # [19:35] <ted> now i guess i just need to wire this in somewhere
- # [19:35] * Quits: mgoodwin (mgoodwin@1FA05860.3B0FFCAD.302434AB.IP) (Ping timeout)
- # [19:36] <ehsan> ted: please don't do it inside firefox ;)
- # [19:36] <ehsan> hmm
- # [19:36] <Ms2ger> Oh, did people see our new logo?
- # [19:36] <Ms2ger> http://img716.imageshack.us/img716/8493/fatfox.png
- # [19:36] * mcote is now known as mcote|lunch
- # [19:36] <ehsan> robcee: what's up with the tree btw?
- # [19:36] <bsmedberg> awesome
- # [19:36] <Wes_> lol
- # [19:36] <khuey> it's broken
- # [19:36] <ehsan> Ms2ger: don't see any difference
- # [19:37] <Ms2ger> ehsan, well played
- # [19:37] <ehsan> why has nobody backed out the merge?
- # [19:37] <ehsan> should I? ;)
- # [19:37] <khuey> because the network is broken
- # [19:37] <khuey> and not the code
- # [19:37] * Joins: WeirdAl (chatzilla@moz-D461843.ask.info)
- # [19:37] * Quits: cpeterson (cpeterson@moz-BBE3ABD.mv.mozilla.com) (Input/output error)
- # [19:37] <ehsan> wut?
- # [19:37] * bear-afk is now known as bear
- # [19:37] * Joins: bmoss (bmoss@moz-BBE3ABD.mv.mozilla.com)
- # [19:38] * ehsan looks at the log
- # [19:38] * joduinn-commute is now known as joduinn-mtg
- # [19:38] <ted> ehsan: hah, no
- # [19:38] <ted> in a python script
- # [19:38] <ted> monitoring link.exe somewhere
- # [19:38] <sewardj> dbaron: ping
- # [19:39] <ehsan> khuey: why is this a network problem?
- # [19:39] * ehsan is puzzled
- # [19:39] <khuey> ehsan: apparently the addom manager tests depend on AMO
- # [19:39] <ehsan> oh wow
- # [19:39] <khuey> yeah :-(
- # [19:39] <ehsan> man
- # [19:39] <ehsan> they should stop :(
- # [19:39] <Ms2ger> Yes
- # [19:39] <khuey> indeed
- # [19:39] * Joins: sheppy (sheppy@moz-FD5E5D00.static.kgpt.tn.charter.com)
- # [19:39] <khuey> I'm tempted to disable them all
- # [19:39] <Ms2ger> a=me
- # [19:39] <ehsan> khuey: please!
- # [19:39] <ehsan> r=me
- # [19:39] <khuey> really?
- # [19:39] <khuey> heh, ok
- # [19:39] <ehsan> yes!!!!!
- # [19:39] <Ms2ger> And r=philor, I do think
- # [19:40] * Joins: jfkthame (jfkthame@DBA4BEB5.90783722.9542EC20.IP)
- # [19:40] <ehsan> khuey: and please file a bug about fixing them and re-enabling them
- # [19:40] * Quits: jfkthame (jfkthame@DBA4BEB5.90783722.9542EC20.IP) (Quit: jfkthame)
- # [19:40] <khuey> there's already a bug
- # [19:41] <ehsan> ok good
- # [19:41] * ehsan goes back to the vs2010 project
- # [19:41] * Quits: tH (Rob@cpc16-seac19-2-0-cust234.7-2.cable.virginmedia.com) (Ping timeout)
- # [19:42] * Joins: tH (Rob@cpc16-seac19-2-0-cust234.7-2.cable.virginmedia.com)
- # [19:42] * Joins: waynenguyen (HP@6EB57D2.92EE8F45.B425DC1D.IP)
- # [19:43] <ehsan> khuey: quick question for you
- # [19:43] <ehsan> let's say I'm in a makefile
- # [19:43] <ehsan> and I have a var, like _MCS_VER
- # [19:43] <ehsan> and I wanna do something only if it's greater than a number
- # [19:43] <ehsan> is that something that I can do?
- # [19:43] <Ms2ger> Yes
- # [19:43] <Ms2ger> Call a python script
- # [19:44] <khuey> 302 ted
- # [19:44] <mccr8> Should I star all of my Moth failures in toolkit/mozapps with bug 709531?
- # [19:44] * armenzg is now known as armenzg_mtg
- # [19:45] <ehsan> Ms2ger: I can't tell whether you were serious or not!
- # [19:45] <ted> ehsan: it's pretty much impossible to test numeric stuff in makefiles
- # [19:45] <ted> ehsan: i recommend you check it in configure and subst a new var
- # [19:45] * Joins: Mossop (mossop@moz-C03D0C61.vlan426.asr1.sfo1.gblx.net)
- # [19:45] * catlee-lunch is now known as catlee
- # [19:45] <ted> http://mxr.mozilla.org/mozilla-central/source/configure.in#751
- # [19:45] <ted> right around there should work
- # [19:45] <ehsan> ted: ok... can I do something so that I don't need to do another full rebuild by touching the configure script?
- # [19:45] <Ms2ger> No
- # [19:46] <catlee> ehsan: yeah, but old builds get deleted
- # [19:46] <catlee> old updates anyway
- # [19:46] <Ms2ger> Except what it says when you make in your objdir
- # [19:46] <ehsan> old builds?
- # [19:46] <ehsan> I'm talking about my local objdir
- # [19:46] * Joins: pcwalton (pcwalton@moz-7B0110AD.mv.mozilla.com)
- # [19:46] <catlee> ehsan: re: ash
- # [19:46] <ehsan> catlee: ah, ok
- # [19:46] * Quits: Honza (chatzilla@E07BF19C.5BB5597D.D0083327.IP) (Ping timeout)
- # [19:46] <ehsan> :((
- # [19:46] * ehsan pushes to ash
- # [19:46] <catlee> trigger a new nightly?
- # [19:47] <ehsan> catlee: would that work?
- # [19:47] <catlee> so my win64 build using 32-bit compiler blew up
- # [19:47] <catlee> ehsan: yeah, it should
- # [19:47] <ehsan> catlee: blew up how?
- # [19:47] <ehsan> catlee: ok
- # [19:47] * Quits: smaug (chatzilla@moz-78A86AC.pp.htv.fi) (Ping timeout)
- # [19:47] <catlee> dxguid.lib(d3dxguid.obj) : fatal error LNK1103: debugging information corrupt; recompile module
- # [19:48] <khuey> yeah
- # [19:48] <jdm> jhk_: ping
- # [19:48] <khuey> you need the hotfix
- # [19:48] <khuey> apparently
- # [19:48] * khuey just --disable-angle'd
- # [19:48] * jwir3 is now known as jwir3|lunch
- # [19:48] <jhk_> jdm:pong
- # [19:48] * Joins: Boriss (FlyingToas@moz-C03D0C61.vlan426.asr1.sfo1.gblx.net)
- # [19:49] * Joins: mwu (mwu@moz-59435430.hsd1.nj.comcast.net)
- # [19:49] <catlee> ah, hmm...
- # [19:49] <jdm> jhk_: so, I think you might want to step through http://mxr.mozilla.org/mozilla-central/source/content/html/content/src/nsFormSubmission.cpp#475 and see what happens when you submit the form
- # [19:49] <catlee> khuey: which hotfix?
- # [19:50] <khuey> catlee: idk, NeilAway was telling me about it
- # [19:50] <catlee> https://connect.microsoft.com/VisualStudio/Downloads/DownloadDetails.aspx?DownloadID=18623
- # [19:50] <catlee> let's see
- # [19:50] <firebot> Check-in: http://hg.mozilla.org/mozilla-central/rev/ce998a59ee1d - Kyle Huey - Bug 709531: Turn off Addon Manager tests that depend on AMO. r=ehsan a=me CLOSED TREE
- # [19:51] <catlee> hmm, I wonder if I want the x64 version or the x86 version
- # [19:51] <bhearsum> khuey++
- # [19:51] * philor is now known as philor|away
- # [19:51] <khuey> ehsan: so ... in theory ... we're good to go
- # [19:51] <jhk_> jdm: oh okay! thanks!:)
- # [19:51] * Joins: bbondy (bbondy@moz-28CF6D1C.home.cgocable.net)
- # [19:52] <khuey> hmm
- # [19:52] <ehsan> khuey: once we get that hotfix on all 64-bit builders, right?
- # [19:52] <khuey> do we need to disable the xpinstall tests too?
- # [19:52] <khuey> ehsan: I was talking about the tree ;-)
- # [19:52] <ehsan> khuey: ah haha
- # [19:52] <jdm> jhk_: I suspect that we want to find out that the file is missing and just submit a blank value for the form field, but we should check with sicking about that
- # [19:52] <khuey> ehsan: hmm, might need to disable more tests
- # [19:53] * Joins: dao (dao@moz-7C5E3898.superkabel.de)
- # [19:53] <ehsan> khuey: if you suspect any others, I'd suggest to go ahead and disable them right now
- # [19:53] <ehsan> proactively
- # [19:53] <khuey> heh, ok
- # [19:54] <jhk_> jdm : other way is to show popup if file is missing.
- # [19:54] <Mossop> Disable ALL the tests!
- # [19:54] <jhk_> jdm : but not sure about it.
- # [19:54] <khuey> pretty mcu
- # [19:54] <khuey> *much
- # [19:54] <ehsan> catlee: so we need to have that hotfix on all x64 builders, right?
- # [19:55] <firebot> Check-in: http://hg.mozilla.org/mozilla-central/rev/a07143afa2d0 - Kyle Huey - Bug 709531: Turn off more Addon Manager tests that depend on AMO. r=ehsan a=me CLOSED TREE
- # [19:55] <jdm> jhk_: that's possible, but probably too ugly. sicking or jst or someone will make a decision about that, I expect.
- # [19:55] <khuey> ehsan: ok, now the tree should be good, in theory
- # [19:55] <catlee> ehsan: yes, along with vs2005 and some other stuff
- # [19:55] * Joins: jet (junglecode@moz-BBE3ABD.mv.mozilla.com)
- # [19:55] <ehsan> khuey: thanks
- # [19:55] <ehsan> catlee: how long is that gonna take?
- # [19:55] * Quits: Cwiiis (cwiiis@moz-F15E698.croy.cable.virginmedia.com) (Ping timeout)
- # [19:55] <catlee> ehsan: weeks
- # [19:55] <ehsan> sadness
- # [19:55] * Quits: gabor (gabor@moz-3B57BCD1.catv.pool.telekom.hu) (Ping timeout)
- # [19:56] <jhk_> jdm : me too.
- # [19:56] <catlee> indeed
- # [19:56] * Joins: bholley (bholley@moz-FCAF9AAB.hsd1.ca.comcast.net)
- # [19:58] <khuey> !seen dolske
- # [19:58] <firebot> dolske was last seen 40 minutes and 46 seconds ago, saying 'nigelb: we only use Code Bournon for emergencies' in #foxymonkies.
- # [19:58] <khuey> dolske: ping
- # [19:58] * Quits: bmoss (bmoss@moz-BBE3ABD.mv.mozilla.com) (Ping timeout)
- # [19:58] * Quits: mjschranz (mjschranz@moz-57B23647.senecac.on.ca) (Client exited)
- # [19:59] * Joins: Cwiiis (cwiiis@moz-F15E698.croy.cable.virginmedia.com)
- # [19:59] <ehsan> khuey: let's say that I propose that we just start to link against the crt statically on windows
- # [19:59] <ehsan> how sad would you be?
- # [20:00] <khuey> can we do that?
- # [20:00] <khuey> bsmedberg: ^?
- # [20:00] * Quits: WeirdAl (chatzilla@moz-D461843.ask.info) (Quit: ChatZilla 0.9.87 [Firefox 8.0/20111104165243])
- # [20:00] * Joins: smaug (chatzilla@moz-B4AE3C70.elisa-mobile.fi)
- # [20:00] <ehsan> I think we could if we wanted to!
- # [20:00] <khuey> it won't break addons or anything?
- # [20:01] <khuey> also, what would we statically link it to?
- # [20:01] <khuey> not everything, I hope
- # [20:02] <ehsan> khuey: well, linking everything would sure be a lot easier for me ;)
- # [20:02] <khuey> well linking the CRT into libxul will probably make things worse, not better!
- # [20:03] <bsmedberg> ehsan: last time we tried we had allocation mismatch problems
- # [20:03] <bsmedberg> but that may have been pre-jemalloc
- # [20:03] * Joins: brendan (brendaneic@moz-A286C218.hsd1.ca.comcast.net)
- # [20:03] * Joins: clee (clee@moz-BBE3ABD.mv.mozilla.com)
- # [20:03] <ehsan> bsmedberg: so would it be a good idea for somebody (me) to give that a shot?
- # [20:03] <ehsan> khuey: not if we're using vs2010 on win64
- # [20:03] * zpao|detached is now known as zpao
- # [20:03] <khuey> ehsan: uh ...
- # [20:04] * Quits: mike5w3c (MikeS@moz-811E91F4.pool.e-mobile.ne.jp) (Ping timeout)
- # [20:04] <ehsan> bjacob: ping
- # [20:04] <khuey> ehsan: it's still going to use more memory
- # [20:04] <catlee> are we using vs2010 on win64?
- # [20:04] <ehsan> catlee: not yet
- # [20:04] <nemo> oh god
- # [20:04] <nemo> http://developers.slashdot.org/story/11/12/14/1725205/firefox-too-big-to-link-on-32-bit-windows
- # [20:04] * Joins: mike5w3c (MikeS@moz-85E9E491.pool.e-mobile.ne.jp)
- # [20:04] <khuey> getting an additional GB and burning 400MB on linking the CRT in statically is less desirable than getting an additional GB and not burning any of it
- # [20:04] <catlee> it my experiment with vs2005 on win64 worthwhile?
- # [20:04] <nemo> queue stupid comments...
- # [20:04] <khuey> yes
- # [20:04] <khuey> don't stop experimenting
- # [20:04] <jprmc> nemo: i already sent that to press@ with some follow up
- # [20:05] <bsmedberg> ehsan: probably worth a shot, but we'd have to watch for allocator crashes, which are kinda hard to see
- # [20:05] <bsmedberg> would require coordination with crashkill
- # [20:05] <nemo> jprmc: this one that got modded funny was probably most insightful
- # [20:05] <nemo> jprmc: http://developers.slashdot.org/comments.pl?sid=2574298&cid=38372354
- # [20:05] <nemo> as well as the followups
- # [20:05] <nemo> http://developers.slashdot.org/comments.pl?sid=2574298&cid=38372978
- # [20:06] * Quits: dao (dao@moz-7C5E3898.superkabel.de) (Quit: Leaving.)
- # [20:06] <jprmc> nemo: does it mention crhome turned off PGO for similar reasons? :-)
- # [20:06] <ehsan> bsmedberg: so linking firefox.exe statically and the rest dynamically is only gonna make things trickier, isn't it?
- # [20:06] <nemo> jprmc: https://code.google.com/p/chromium/issues/detail?id=21932#c10
- # [20:06] <nemo> jprmc: (GB = 9GB
- # [20:06] <nemo> as pointed out in /. comments
- # [20:07] <nemo> jprmc: the problem for you guys is a narrative. the narrative was established before memory usage was brought under control, and is now a meme
- # [20:07] <bsmedberg> ehsan: wait, we're doing this for VC2010, right?
- # [20:07] <bsmedberg> wouldn't that mean doing it for everything?
- # [20:07] * Joins: jet_ (junglecode@moz-BBE3ABD.mv.mozilla.com)
- # [20:07] * Quits: jet (junglecode@moz-BBE3ABD.mv.mozilla.com) (Ping timeout)
- # [20:07] * jet_ is now known as jet
- # [20:07] <catlee> why do I still read /.
- # [20:07] * ashughes is now known as ashughes|meeting
- # [20:07] <catlee> it depresses me
- # [20:08] <dolske> khuey: uhm, sheriff? we still do that?!
- # [20:08] <ehsan> bsmedberg: yes to your first question
- # [20:08] <ehsan> bsmedberg: I was under the impression that we don't wanna do it for everything
- # [20:08] <sheppy> catlee: Masochism?
- # [20:08] <dolske> khuey: i can't really do it today.
- # [20:08] <ehsan> but if that's an option worth investigating, I'm willing to experiment
- # [20:09] <khuey> jprmc: so you know how you asked me if you could do something to help? ;-)
- # [20:09] <@bz> hmm
- # [20:09] * @bz can't find document.all in the html5 spec
- # [20:09] <jprmc> khuey: yup, i'm here
- # [20:09] <Ms2ger> bz, obsolete features
- # [20:09] <khuey> jprmc: we could use a sheriff
- # [20:09] <bjacob> ehsan: pong
- # [20:10] <Ms2ger> bz, right at the end of the page
- # [20:10] <khuey> jprmc: I think we can reopen with heavy metering
- # [20:10] <nemo> Ms2ger: what?? document.all was formalised??
- # [20:10] <ehsan> khuey: I could help with that
- # [20:10] * Joins: kaze (kaze@moz-79F78BFC.w90-46.abo.wanadoo.fr)
- # [20:10] <ehsan> I'm pretty strict ;)
- # [20:10] <Ms2ger> nemo, yes, ages ago
- # [20:10] <khuey> ehsan: nah you should do your crazy experiments
- # [20:10] <jprmc> khuey: assuming the AMO thing goes?
- # [20:10] <ehsan> bjacob: any luck on the webgl tests?
- # [20:10] <nemo> Ms2ger: lol
- # [20:10] <khuey> jprmc: we disabled the problematic tests
- # [20:10] <jprmc> ok
- # [20:10] * philor|away is now known as philor
- # [20:10] <ehsan> khuey: I could do both at the same time, cause most of my time is being spent thinking and failing ;)
- # [20:10] * bear is now known as bear-afk
- # [20:10] <khuey> jprmc: and the PGO situation is managable as long as we don't land anything really big
- # [20:10] <nemo> Ms2ger: ever googled for document.all? #1 hit "don't use document.all" #2) five worst pieces of code that people use with Javascript
- # [20:11] <@bz> Ms2ger: gah
- # [20:11] <khuey> jprmc: though it looks like ehsan is volunteering
- # [20:11] <@bz> Ms2ger: that use of partial interface is suck
- # [20:11] <@bz> Ms2fger: imo
- # [20:11] * ehsan is now known as ehsan|sheriff
- # [20:11] * Joins: jfkthame (jfkthame@DBA4BEB5.90783722.9542EC20.IP)
- # [20:11] <nemo> Ms2ger: wow. why would they bother doing that? It was something people were avoiding already for a long time
- # [20:11] <khuey> ehsan: if you want to do it by all means
- # [20:11] <jprmc> i think bsmedberg is available to help as well
- # [20:11] <khuey> jlebar++
- # [20:11] <ejpbruel> sicking: ping
- # [20:11] <bjacob> ehsan: i cant use my windows partition at the moment... from the logs, my only theory is that doing SurfaceFromElement on a video element is broken
- # [20:11] * ehsan|sheriff changes topic to 'm-c metered, corrdinate with sheriff. m-i closed for bug 709193 (MSVC virtual address space limit hit) || Tree rules: https://etherpad.mozilla.org/sLWdJm0zwB || Next aurora uplift: 20th Dec || If you are new or want to help, see irc://irc.mozilla.org/#introduction'
- # [20:12] <khuey> ehsan|sheriff: you'll want to merge m-i to m-c first, of course
- # [20:12] <ehsan|sheriff> bjacob: I can lend you a windows machine :)
- # [20:12] <ehsan|sheriff> khuey: will do
- # [20:12] <bjacob> ehsan|sheriff: coming
- # [20:12] <Ms2ger> nemo, because nobody is going to remove support
- # [20:12] <jprmc> why will m-i still be closed?
- # [20:12] <catlee> bz: does https://bugzilla.mozilla.org/show_bug.cgi?id=700914 reduce memory usage from just about:blank tabs, or any about:blank frame in your session?
- # [20:12] <ehsan|sheriff> jprmc: cause we need to meter
- # [20:12] <khuey> jprmc: we should clear the backlog first
- # [20:12] * Quits: Jesse (jruderman@moz-BBE3ABD.mv.mozilla.com) (Quit: Jesse)
- # [20:12] <jprmc> ok
- # [20:12] <nemo> Ms2ger: m'k - I just thought that support had never been added to anything but IE - I'm sure a lot of others thought that too
- # [20:12] <khuey> then we can reopen m-i once it's not going to get 40 pushes in an hour
- # [20:12] <nemo> Ms2ger: so all you had to do was keep not adding the support :) oh well...
- # [20:12] <khuey> ehsan|sheriff: blassey is up first btw
- # [20:13] <catlee> FYI pgo builds should now be every 3 hours
- # [20:13] * Quits: @dveditz (dveditz@moz-BBE3ABD.mv.mozilla.com) (Quit: dveditz)
- # [20:13] <catlee> and won't merge
- # [20:13] <Ms2ger> nemo, all browsers support it, and only we restrict it to quirks more
- # [20:13] <Ms2ger> *mode
- # [20:14] <@dbaron> sewardj, pong
- # [20:14] * jgriffin is now known as jgriffin-afk
- # [20:14] <@dbaron> (hard to find the one highlighted line in 6 pages of scrollback)
- # [20:14] <ehsan|sheriff> ok blassey, I'll give you the green light shortly
- # [20:14] <edmorley> khuey: want me to merge?
- # [20:14] <ehsan|sheriff> edmorley: I'm on it
- # [20:14] <khuey> edmorley: coordinate with ehsan|sheriff
- # [20:14] <edmorley> ehsan|sheriff: cool, give a shout when you need a break from hair pulling :-)
- # [20:15] <ehsan|sheriff> edmorley: it would be awesome if you can mark the bugs when the merge is done though :)
- # [20:15] <edmorley> will do
- # [20:15] <Ms2ger> <sewardj> davidb: you're not dbaron, right? <sewardj> !seen dbaron
- # [20:15] * Quits: gandalf (zbraniecki@moz-9F56DED7.neoplus.adsl.tpnet.pl) (Quit: Computer has gone to sleep.)
- # [20:15] <ehsan|sheriff> thanks
- # [20:15] * Joins: wlach (wlach@moz-C5B44C21.vif.net)
- # [20:16] * Joins: dveditz (dveditz@moz-BBE3ABD.mv.mozilla.com)
- # [20:16] * ChanServ sets mode: +o dveditz
- # [20:16] * Quits: Enn (enn@F2D29657.F60B0462.67AC9B1.IP) (Ping timeout)
- # [20:16] <joe> ehsan|sheriff: https://bug392867.bugzilla.mozilla.org/attachment.cgi?id=568841
- # [20:17] * Quits: jhammink (textual@moz-7B0110AD.mv.mozilla.com) (Quit: Computer has gone to sleep.)
- # [20:17] * Quits: pascalc (chatzilla@moz-A2A3C9E6.vlan402.asr1.cdg3.gblx.net) (Ping timeout)
- # [20:17] <ehsan|sheriff> edmorley: the ball is yours for bug marking :)
- # [20:18] <firebot> Check-in: http://hg.mozilla.org/mozilla-central/rev/850c7301c926 - Ehsan Akhgari - Merge mozilla-inbound and mozilla-central on a CLOSED TREE a=me
- # [20:18] <firebot> http://hg.mozilla.org/mozilla-central/rev/6c5cb98336ba - Ed Morley - Merge mozilla-central to mozilla-inbound
- # [20:18] <firebot> http://hg.mozilla.org/mozilla-central/rev/472dad831258 - Kyle Huey - Bug 709193: Turn Graphite back off on this CLOSED TREE.
- # [20:18] <ehsan|sheriff> khuey: is there anybody else in the queue?
- # [20:18] <ehsan|sheriff> blassey: you ready?
- # [20:18] <firebot> http://hg.mozilla.org/mozilla-central/rev/cf0b31ff2b6d - Patrick McManus - bug 528288 - reland spdy after libxul weightloss a=khuey CLOSED TREE
- # [20:18] <firebot> http://hg.mozilla.org/mozilla-central/rev/6a1184c51eae - Kyle Huey - Merge m-c to m-i. a=me CLOSED TREE
- # [20:18] <blassey> ehsan|sheriff: I am
- # [20:18] <firebot> http://hg.mozilla.org/mozilla-central/rev/f1abb2b731e0 - Kyle Huey - Bug 709193: Turn Graphite back on on this CLOSED TREE.
- # [20:18] <blassey> ehsan|sheriff: let me know when I should go
- # [20:19] * Joins: mjschranz (mjschranz@F1451709.44D93D66.1139E686.IP)
- # [20:19] <ehsan|sheriff> ah dammit
- # [20:19] <ehsan|sheriff> I don't have the admintree url in my history any more :(
- # [20:19] * Joins: jhammink (textual@moz-7B0110AD.mv.mozilla.com)
- # [20:19] * Quits: jhammink (textual@moz-7B0110AD.mv.mozilla.com) (Quit: jhammink)
- # [20:19] <@dbaron> sewardj, actually, I'll pong again in 20 minutes when I'm in the office...
- # [20:19] <khuey> ehsan|sheriff: margaret, unless blassey is landing margaret's stuff
- # [20:19] * Joins: Jesse (jruderman@moz-BBE3ABD.mv.mozilla.com)
- # [20:19] * Quits: @dbaron (dbaron@moz-27F87443.dsl.dynamic.sonic.net) (Quit: 8403864 bytes have been tenured, next gc will be global.)
- # [20:19] <khuey> ehsan|sheriff: after that, it's open
- # [20:19] * davehunt|mtg is now known as davehunt|away
- # [20:19] <blassey> khuey: I want to land margarets stuff as a seperate push
- # [20:19] <khuey> ok
- # [20:19] <blassey> there's a plan, which I haven't articulated well
- # [20:20] <blassey> but we need seperate pushes for two nightly respins
- # [20:20] <blassey> inorder to test this
- # [20:20] <ehsan|sheriff> blassey: that's fine by me
- # [20:20] <blassey> ehsan|sheriff: so, am I supposed to be pushing now?
- # [20:20] <ehsan|sheriff> blassey: alright, you're good to go
- # [20:20] <ehsan|sheriff> yep :)
- # [20:21] <ehsan|sheriff> joe: you can be in line after blassey pushes his two pushed
- # [20:21] <ehsan|sheriff> pushes, evem
- # [20:21] <ehsan|sheriff> even, even!
- # [20:21] <joe> PUSH ZEE PUSHES
- # [20:21] <ehsan|sheriff> I suck at spelling
- # [20:21] <joe> itym speling
- # [20:21] <khuey> jprmc: I sent a summary of where we are to .platform
- # [20:21] * Joins: rniwa (rniwa@moz-E171DA5.sfba.hfc.comcastbusiness.net)
- # [20:21] <froydnj> push all the changesets!
- # [20:21] * Joins: jhammink (textual@moz-7B0110AD.mv.mozilla.com)
- # [20:21] <jprmc> khuey: ok - go study for your exam :-)
- # [20:21] <khuey> yeah that's exactly why we're metered
- # [20:21] <khuey> jprmc: yep, I'm out
- # [20:22] * khuey is now known as khuey|away
- # [20:22] <ehsan|sheriff> joe: and by push, I mean only pushing the patch you showed me :P
- # [20:22] <Ms2ger> Bye, khuey|away
- # [20:22] <blassey> ehsan|sheriff: ok, I've pushed my backouts
- # [20:22] <ehsan|sheriff> thanks
- # [20:22] <blassey> next is margaret
- # [20:23] <firebot> Check-in: http://hg.mozilla.org/mozilla-central/rev/02781c0867f5 - Brad Lassey - bug 710751 - Updater is broken, backout aebdec71790e r=mfinle a=java-only
- # [20:23] <firebot> http://hg.mozilla.org/mozilla-central/rev/15a7bbbb9323 - Brad Lassey - bug 710751 - Updater is broken, backout b0165c0d85e1 r=mfinle a=java-only
- # [20:23] * armenzg_mtg is now known as armenzg_brb
- # [20:23] <blassey> aki: I need nightlies spun off of what I just pushed
- # [20:23] <blassey> or coop|buildduty I guess
- # [20:23] <ehsan|sheriff> joe: please hold on for a few minutes
- # [20:23] * Joins: billm (billm@moz-BBE3ABD.mv.mozilla.com)
- # [20:23] * Ms2ger gavinbots bholley
- # [20:23] <catlee> so we're only 10's of MB away from another linker failure?
- # [20:23] <bholley> Ms2ger ?
- # [20:24] <Ms2ger> I didn't
- # [20:24] * Joins: bmoss (bmoss@moz-BBE3ABD.mv.mozilla.com)
- # [20:24] <Ms2ger> I tried to cc on your buffer sharing bug, and failed
- # [20:24] <ehsan|sheriff> catlee: yes, which is an awesome place to be, compared to a few days ago!
- # [20:24] <catlee> I guess!
- # [20:24] <JonathanS> libxml needs to be on diet?
- # [20:24] <JonathanS> libxul
- # [20:24] <catlee> seems a bit...close for comfort
- # [20:25] * Quits: sheppy (sheppy@moz-FD5E5D00.static.kgpt.tn.charter.com) (Quit: Leaving)
- # [20:25] <bholley> Ms2ger: I don't understand what you meant
- # [20:25] * Quits: AutomatedTester (AutomatedT@moz-21328E59.as13285.net) (Quit: AutomatedTester)
- # [20:25] <Ms2ger> gavin is/was known for ccing on bugs rather fast
- # [20:25] <edmorley> Matheus is the new Gavinbot
- # [20:25] <espindola> !seen dolske
- # [20:26] <Ms2ger> True
- # [20:26] <firebot> dolske was last seen 17 minutes and 13 seconds ago, saying 'khuey: i can't really do it today.' in #developers.
- # [20:26] <ehsan|sheriff> espindola: hello :)
- # [20:26] <espindola> ehsan|sheriff, hello
- # [20:26] <ehsan|sheriff> looking for the poor sheriff?
- # [20:26] <espindola> no :-)
- # [20:26] <ehsan|sheriff> oh, shut my mouth then :)
- # [20:26] <Ms2ger> ehsan|thenewdolske?
- # [20:27] * philor loads tinderbox, gets an alert
- # [20:27] <Ms2ger> Tinderbox? What's that?
- # [20:27] <blassey> margaret: are you ready to push?
- # [20:28] <coop|buildduty> blassey: so revision 02781c0867f5, native android only for the new nightly?
- # [20:29] <blassey> coop|buildduty: yes please
- # [20:29] <espindola> dolske, do you have some time to discuss 702848?
- # [20:29] <jimm> ehsan|sheriff: has the 'no windows libxul patches' restriction been lifted?
- # [20:29] <blassey> and heads up, we're going to want another nightly spun off of what margaret pushes
- # [20:29] <blassey> so we can test that the updater is fixed
- # [20:29] * nhirata|mtg is now known as nhirata
- # [20:29] <blassey> coop|buildduty: ^
- # [20:29] <joe> jimm: ehsan|sheriff is metering changes
- # [20:30] <jimm> tree rules still state no windows code allowed, which is why I asked.
- # [20:30] <ehsan|sheriff> jimm: yes, but I'd appreciate if people would still coordinate
- # [20:30] <ehsan|sheriff> jimm: we're trying to manage the backlog, and being cautious about the big stuff
- # [20:31] <@bz> man
- # [20:31] <@bz> no useful stacks on try?
- # [20:31] <ehsan|sheriff> edmorley: thanks
- # [20:31] <jimm> ehsan|sheriff: do we have a queue set up? I'm interested in landing bug 661991.
- # [20:31] <jlebar> bz, Welcome to the world of debug plus optimizations.
- # [20:31] <@bz> jlebar: ugh
- # [20:31] <Ms2ger> ehsan|sheriff, does "metered" have a tree hook?
- # [20:31] <@bz> jlebar: this totally sucks
- # [20:31] <ehsan|sheriff> jimm: the queue currently lives in my head, but I'll go ahead and add it to the approval rules :)
- # [20:31] <@bz> jlebar: I thought we had symbols on try for opt and debug... :(
- # [20:31] <ehsan|sheriff> Ms2ger: no
- # [20:31] <JonathanS> would LLVM make debugging easier?
- # [20:31] <edmorley> ehsan|sheriff: do we want to use the etherpad for the queue?
- # [20:31] <jlebar> bz, You should have symbols, just not line numbers...
- # [20:32] * Ms2ger accidentally pushes
- # [20:32] <blassey> ehsan|sheriff: so, if margaret doesn't come back when her turn comes up, I can push her stuff for her
- # [20:32] <joe> Ms2ger: i will cut you
- # [20:32] <@bz> jlebar: nope
- # [20:32] <jlebar> :-/
- # [20:32] <@bz> Thread 0 (crashed)
- # [20:32] <@bz> 0 libxul.so + 0x909809
- # [20:32] <@bz> rbx = 0x06fbc230 r12 = 0x15e63b2f r13 = 0x00000000 r14 = 0x00000002
- # [20:32] <@bz> r15 = 0x25d53d4f rip = 0x15e63809 rsp = 0x25d53a80 rbp = 0x25d53ba0
- # [20:32] <@bz> Found by: given as instruction pointer in contex
- # [20:32] <ehsan|sheriff> blassey: oh, can you do that please?
- # [20:32] <@bz> And it goes on like that
- # [20:32] * AaronMT|mtg is now known as AaronMT
- # [20:32] <ehsan|sheriff> edmorley: yes
- # [20:32] <blassey> ehsan|sheriff: no problem
- # [20:32] * Quits: erione (erione@630B0104.C5AA6A92.C752B3FA.IP) (Quit: Leaving)
- # [20:33] <ehsan|sheriff> edmorley: if someone messes with it, they're gonna have a problem with me :P
- # [20:33] * Joins: smooney (smooney@moz-BBE3ABD.mv.mozilla.com)
- # [20:33] <coop|buildduty> blassey: nightly triggered
- # [20:33] * mcote|lunch is now known as mcote
- # [20:33] <ehsan|sheriff> jimm: hmm, that's the nsFilePicker rewrite... _shouldn't_ be that scary
- # [20:33] <bholley> Ms2ger: oh, that's the bug you're referring to
- # [20:33] <ehsan|sheriff> jimm: do you happen to have PGO builds on try?
- # [20:33] <Ms2ger> Yes
- # [20:33] <jimm> ehsan|sheriff: hope not, but it does add some new code.
- # [20:34] * Joins: abwillis (abwillis@9877934.9DD4DBBF.6A7A197.IP)
- # [20:34] <jimm> ehsan|sheriff: no, but I can fire some off
- # [20:34] <ehsan|sheriff> yeah..
- # [20:34] <ehsan|sheriff> jimm: please do
- # [20:34] <ehsan|sheriff> I'll take you off the queue for now
- # [20:34] <bholley> Ms2ger: yeah, shouldn't be too hard. the key is just to do it in a clean and sane fashion
- # [20:34] * Quits: roc (chatzilla@C0ACF8B.5E1E9EEA.613E47D1.IP) (Ping timeout)
- # [20:34] * Quits: nhirata (nhirata.bu@moz-BBE3ABD.mv.mozilla.com) (Quit: nhirata)
- # [20:34] <blassey> ehsan|sheriff: ok, margaret's patches are pushed
- # [20:34] <ehsan|sheriff> jimm: sorry but we really need to make sure that everything doesn't blow up
- # [20:34] <ehsan|sheriff> blassey: thanks
- # [20:34] <edmorley> ehsan|sheriff: I'll sort out the tbpl messages to point to the queue etc
- # [20:34] <ehsan|sheriff> jimm: make sure to use the m-c tip for your try push
- # [20:34] <blassey> coop|buildduty: so, we'll need an android-native only nightly spun off that as well
- # [20:35] <ehsan|sheriff> edmorley: awesome
- # [20:35] <firebot> Check-in: http://hg.mozilla.org/mozilla-central/rev/bc84b3376e14 - Margaret Leibovic - Bug 707886 - Platform support for non-e10s click-to-play plugins. r=jst,blassey
- # [20:35] <jimm> ehsan|sheriff: well, all I want to do is make sure those get in before the 20th. so if we think we'll have all this resolved by then, I can wait. sounds like it might not be though.
- # [20:35] <coop|buildduty> blassey: isn't that what i just triggered? did you mean XUL?
- # [20:36] <ehsan|sheriff> jimm: nobody knows, if I were you, I would try to land ASAP
- # [20:36] <blassey> coop|buildduty: no, native, just different changeset
- # [20:36] <blassey> coop|buildduty: bc84b3376e14
- # [20:36] <ehsan|sheriff> joe: you're up next in 5
- # [20:36] <jimm> ehsan|sheriff: ok cool.
- # [20:36] <blassey> coop|buildduty: I need to test the updater, so I need two working nightlies to do it
- # [20:36] <jlebar> khuey|away, what'd I do?
- # [20:36] <coop|buildduty> blassey: ah, k. i'll trigger that too
- # [20:37] <blassey> coop|buildduty: many thanks
- # [20:37] * Joins: nhirata (nhirata.bu@moz-BBE3ABD.mv.mozilla.com)
- # [20:37] <ehsan|sheriff> ted: ping
- # [20:38] <ted> ehsan|sheriff: pong
- # [20:38] <edmorley> ehsan|sheriff: "METERED" isn't recognised by the tree hook, would you prefer approval required?
- # [20:38] <ehsan|sheriff> ted: do you know where we set the crt /MD compiler flag for msvc?
- # [20:38] <ehsan|sheriff> mxr searches for that tend to be disappointing
- # [20:38] * Quits: smooney (smooney@moz-BBE3ABD.mv.mozilla.com) (Quit: smooney)
- # [20:38] * Quits: jhk_ (jiggy@8E6C34C1.A3F9767A.1C37C358.IP) (Connection reset by peer)
- # [20:38] <ted> http://mxr.mozilla.org/mozilla-central/source/config/config.mk#513
- # [20:39] * Quits: cpearce (chatzilla@moz-55A2BC7D.xdsl.xnet.co.nz) (Ping timeout)
- # [20:40] <ehsan|sheriff> edmorley: yeah, why not?
- # [20:40] * Joins: smooney (smooney@moz-BBE3ABD.mv.mozilla.com)
- # [20:40] <ehsan|sheriff> edmorley: say METERED(APPROVAL REQUIRED) maybe?
- # [20:40] <ehsan|sheriff> ted: thanks!
- # [20:40] <ted> np
- # [20:40] * Quits: jhopkins (jhopkins@moz-41E1D586.tb.shawcable.net) (Ping timeout)
- # [20:41] * Joins: bent (chatzilla@moz-C3562645.hsd1.ca.comcast.net)
- # [20:41] * Joins: jhopkins (jhopkins@moz-41E1D586.tb.shawcable.net)
- # [20:41] * Joins: dbaron (dbaron@moz-BBE3ABD.mv.mozilla.com)
- # [20:41] * ChanServ sets mode: +o dbaron
- # [20:43] * Quits: nhirata (nhirata.bu@moz-BBE3ABD.mv.mozilla.com) (Quit: nhirata)
- # [20:43] * khuey|away is now known as khuey
- # [20:43] * mdas is now known as mdas|afk
- # [20:43] <khuey> jlebar: bringing sanity to /.
- # [20:43] * khuey is now known as khuey|quite-busy
- # [20:43] <jlebar> khuey, oh. That's a bit more than I can do. :)
- # [20:44] * ashughes|meeting is now known as ashughes
- # [20:45] * Quits: mjschranz (mjschranz@F1451709.44D93D66.1139E686.IP) (Client exited)
- # [20:45] <catlee> that's a thing you can do?
- # [20:45] <bjacob> ehsan|sheriff: can i land this on central please? it will reduce libxul size by a whopping 3 bytes https://bugzilla.mozilla.org/attachment.cgi?id=581036&action=diff
- # [20:46] * edmorley changes topic to 'm-c metered, add your name to landing queue: https://etherpad.mozilla.org/sLWdJm0zwB || Tree rules: https://etherpad.mozilla.org/sLWdJm0zwB || Next aurora uplift: 20th Dec || If you are new or want to help, see irc://irc.mozilla.org/#introduction'
- # [20:46] <khuey|quite-busy> bjacob: that's the best you could do?
- # [20:46] <khuey|quite-busy> :-P
- # [20:46] * edmorley changes topic to 'm-c metered, add your name to landing queue: https://etherpad.mozilla.org/sLWdJm0zwB || Next aurora uplift: 20th Dec || If you are new or want to help, see irc://irc.mozilla.org/#introduction'
- # [20:46] * bear-afk is now known as bear
- # [20:47] <vingtetun> ehsan|sheriff: is that ok to land bug 710548 and bug 710804?
- # [20:47] <bjacob> ehsan|sheriff: actually, this is the exact patch, https://bugzilla.mozilla.org/attachment.cgi?id=581075&action=diff
- # [20:47] <ehsan|sheriff> bjacob: yes! ask joe to carry it along for you please?
- # [20:47] <bjacob> ok
- # [20:48] <ehsan|sheriff> thanks
- # [20:48] * Joins: pnemsak (Miranda@moz-7409BD53.orange.sk)
- # [20:48] <ehsan|sheriff> bjacob: he's up next in the queue
- # [20:48] <edmorley> ehsan|sheriff: tbpl x a few updated, should be all set
- # [20:48] <ehsan|sheriff> vingtetun: looking
- # [20:48] <vingtetun> that's all b2g/ only
- # [20:48] <firebot> Check-in: http://hg.mozilla.org/mozilla-central/rev/3e972d3efc11 - Chris Lord - Bug 708307 - Decouple texture size from tile size. r=pcwalton a=android-only
- # [20:49] <catlee> bz: does https://bugzilla.mozilla.org/show_bug.cgi?id=700914 reduce memory usage from just about:blank tabs, or any about:blank frame in your session?
- # [20:49] <joe> Cwiiis: YOU BASTARD
- # [20:49] <@dbaron> bugzilla is molasses-like again...
- # [20:49] <ehsan|sheriff> vingtetun: so, not built on windows?
- # [20:49] <Cwiiis> joe: did I beat you to a push?
- # [20:49] <joe> i'm in queue!
- # [20:49] <ehsan|sheriff> Cwiiis: what did you land?!!!?!
- # [20:49] <Cwiiis> oh balls
- # [20:49] * Cwiiis reads topic
- # [20:49] <joe> yes, welcome to today
- # [20:49] <ehsan|sheriff> and the TBPL status?
- # [20:50] * Joins: tonymec__ (tonymec@7923ADCD.7C3BF363.277517C1.IP)
- # [20:50] <Cwiiis> ugh....
- # [20:50] <ehsan|sheriff> Cwiiis: good thing it was not windows code ;)
- # [20:50] <joe> and the phone call we made to you?!?!?!?!?!
- # [20:50] * tonymec__ is now known as tonymec|away
- # [20:50] <Cwiiis> So, what do I need to do now?
- # [20:50] <vingtetun> ehsan|sheriff: in fact it's not build at all
- # [20:50] <joe> nothing :)
- # [20:50] * juanb is now known as juanb|afk
- # [20:50] <ehsan|sheriff> Cwiiis: nothing, I'll just reserve the right to back out one of your patches some time in the future ;)
- # [20:50] * Joins: cpeterson (cpeterson@moz-BBE3ABD.mv.mozilla.com)
- # [20:50] <Cwiiis> hah, ok :)
- # [20:50] <joe> bjacob: which patch do you want landed
- # [20:50] <joe> there are so many
- # [20:50] <joe> TWOOOOO
- # [20:50] <ehsan|sheriff> vingtetun: so it would be awesome if you can land it with DONTBUILD
- # [20:51] <ehsan|sheriff> after joe
- # [20:51] <Cwiiis> Glad you yelled quickly, I have more things to land...
- # [20:51] <bjacob> joe: if only bugzilla would reply now...
- # [20:51] * Cwiiis adds name to queue
- # [20:51] * Quits: gkw (gkw@moz-BBE3ABD.mv.mozilla.com) (Quit: Instantbird 1.1)
- # [20:51] * Joins: gkw (gkw@moz-BBE3ABD.mv.mozilla.com)
- # [20:51] <bjacob> c'mon bugzilla, you can make it. you've done it before. reply to a http request
- # [20:51] <Cwiiis> ehsan|sheriff, wait, do I add my own name or do I have to ask you?
- # [20:52] <joe> bjacob: pastebin it maybe
- # [20:52] <bjacob> joe: https://bugzilla.mozilla.org/show_bug.cgi?id=709947 works for you?
- # [20:52] <bjacob> https://bugzilla.mozilla.org/attachment.cgi?id=581075
- # [20:52] <taras> dolske: i made https://bugzilla.mozilla.org/show_bug.cgi?id=58724 a snappy:p1 since it's a pretty serious perf problem
- # [20:52] <vingtetun> ehsan|sheriff: a=DONTBUILD?
- # [20:52] * mkelly|wfh is now known as mkelly|lunch|gettingacat
- # [20:52] <margaret> blassey: thanks!
- # [20:52] <jlebar> Are there Aurora win-64 builds?
- # [20:52] <ehsan|sheriff> Cwiiis: see the part about coordinate with sheriff? ;)
- # [20:52] <bjacob> is bugzilla down?!?!
- # [20:52] <margaret> sorry, getting to the office took longer than expected
- # [20:52] <blassey> margaret: np
- # [20:53] <dolske> espindola: not until later, sorry, booked up atm.
- # [20:53] * jwir3|lunch is now known as jwir3
- # [20:53] * Quits: brendan (brendaneic@moz-A286C218.hsd1.ca.comcast.net) (Quit: brendan)
- # [20:53] <Cwiiis> ehsan|sheriff, ah, I'm too tired, it's getting late over here - I'll land the rest tomorrow morning
- # [20:53] <ehsan|sheriff> Cwiiis: alright :)
- # [20:53] <dolske> taras: I find it hard to believe that's among the most serious of our perf issues.
- # [20:53] <joe> bjacob: I have that bug loaded but that's it
- # [20:53] <@dbaron> bjacob, slower and slower, at least
- # [20:53] <bjacob> joe: http://pastebin.mozilla.org/1405673
- # [20:53] <joe> first or second attachment?
- # [20:53] <bjacob> joe: second, despite name
- # [20:54] <bjacob> ^ pastebin
- # [20:54] * Joins: pascalc (chatzilla@moz-BCFF62C7.rev.sfr.net)
- # [20:54] <joe> ok very good
- # [20:54] <bjacob> oops, comment is wrong
- # [20:54] <espindola> dolske, ok, ping we when you have time
- # [20:54] <taras> dolske: i find it hard to believe that it isn't
- # [20:54] <bjacob> joe: 1 minute please
- # [20:54] * Joins: igor (igor@169CEE78.E37E53F7.1DAC7E2F.IP)
- # [20:54] * pascalc is now known as IRCMonkey50727
- # [20:54] * Quits: davidb (davidb@F2D29657.F60B0462.67AC9B1.IP) (Quit: davidb)
- # [20:55] * Joins: mjschranz (mjschranz@moz-57B23647.senecac.on.ca)
- # [20:55] <joe> bjacob: just send it to me here
- # [20:55] <bjacob> joe: sent to ur email
- # [20:56] * Joins: jhk_ (jiggy@8E6C34C1.A3F9767A.1C37C358.IP)
- # [20:56] * armenzg_brb is now known as armenzg
- # [20:57] * Joins: Honza (chatzilla@E07BF19C.5BB5597D.D0083327.IP)
- # [20:57] * mattwoodrow|away is now known as mattwoodrow
- # [20:58] <joe> pushed
- # [20:58] <firebot> Check-in: http://hg.mozilla.org/mozilla-central/rev/6ee5c4c7beb6 - Cameron Kaiser - Bug 705516 - Always ask images loaded as OS X menu icons to decode. r=smichaud a=mac-only
- # [20:58] <joe> and forever hold your peace
- # [20:58] <firebot> http://hg.mozilla.org/mozilla-central/rev/11eb2aecc7c3 - Glenn Randers-Pehrson - Bug 392867 - libpng requires error callbacks to not return. Separate out nsPNGDecoder's ErrorCallback and WarningCallback, and call longjmp from ErrorCallback. r=joe a=ehsan
- # [20:58] <firebot> http://hg.mozilla.org/mozilla-central/rev/dc12262285a4 - Joe Drew - Bug 709947 - Don't use the ESSL backend of ANGLE - r=jgilbert a=ehsan
- # [20:58] <joe> bjacob: damn it
- # [20:58] <bjacob> joe: what?
- # [20:58] <joe> I'm listed as the patch author
- # [20:58] <taras> so what's the story with landing c++ changes
- # [20:58] <bjacob> joe: good, you'll get the flak :)
- # [20:58] <ehsan|sheriff> joe: how does it feel, plagiarizing other people's work?
- # [20:59] <taras> are we allowed to land any?
- # [20:59] <taras> or does everything have to be JS?
- # [20:59] <joe> taras: see topic
- # [20:59] <ehsan|sheriff> taras: yes, after coordinating with sheriff
- # [20:59] <jdm> joe: have you forgotten everything about academic dishonesty since leaving waterloo?
- # [20:59] <joe> jdm: FINALLY I CAN GET CREDIT FOR WORK I DIDN'T DO
- # [20:59] * Quits: tfair (tfairey@moz-BBE3ABD.mv.mozilla.com) (Client exited)
- # [21:00] <ehsan|sheriff> jdm: can I have your prof's email address please?
- # [21:00] <margaret> ehsan|sheriff: i also have a mobile-only patch to land. can i get added to the queue?
- # [21:01] <margaret> it's not super important, so more important patches can go ahead of me
- # [21:01] * khuey|quite-busy thinks ehsan is a saint for signing up for this
- # [21:01] <margaret> or can we use mozilla-inbound again?
- # [21:01] * Quits: pcwalton (pcwalton@moz-7B0110AD.mv.mozilla.com) (Quit: pcwalton)
- # [21:02] <ehsan|sheriff> margaret: vingtetun is about to land a b2g patch, can you please ask him to take you as a ride-along?
- # [21:02] * Quits: ajuma (ajuma@F2D29657.F60B0462.67AC9B1.IP) (Ping timeout)
- # [21:02] <smaug> khuey|quite-busy: you're going to graduate ?
- # [21:02] * Joins: tfair (tfairey@moz-BBE3ABD.mv.mozilla.com)
- # [21:02] <smaug> khuey|quite-busy: btw, can I land some patches
- # [21:02] <ehsan|sheriff> margaret: m-i is closed for the moment :)
- # [21:02] <ehsan|sheriff> smaug: which patches?
- # [21:02] <smaug> event ctor
- # [21:02] <vingtetun> margaret: which bug?
- # [21:03] <ehsan|sheriff> vingtetun: you're good to go btw, but see if you can land margaret's patches as well :)
- # [21:03] <khuey|quite-busy> smaug: well assuming I don't fail this exam ;-)
- # [21:03] <khuey|quite-busy> smaug: and talk to ehsan
- # [21:03] <ehsan|sheriff> smaug: bug # please?
- # [21:03] <smaug> Bug 675884, Bug 708701, Bug 709127
- # [21:03] * Quits: bmoss (bmoss@moz-BBE3ABD.mv.mozilla.com) (Ping timeout)
- # [21:03] * ehsan|sheriff looks
- # [21:03] <vingtetun> ehsan|sheriff: thanks, waiting for margaret patch and i will land
- # [21:03] <espindola> would https://bug709406.bugzilla.mozilla.org/attachment.cgi?id=581452
- # [21:03] <espindola> be ok under current rules?
- # [21:04] <margaret> vingtetun: ok one sec, i'll upload the un-bitrotted patch to the bug
- # [21:04] <joe> bjacob: have you landed bug 709947 in aurora?
- # [21:04] <ehsan|sheriff> espindola: yes, you wanna be in the queue?
- # [21:04] * sheppy-lunch is now known as sheppy
- # [21:04] <bjacob> joe: i just requested approval and replied to the release-drivers thread
- # [21:04] * joduinn-mtg is now known as joduinn-lunch
- # [21:04] <joe> nod
- # [21:04] <espindola> ehsan|sheriff, yes, please
- # [21:04] * Joins: ajuma (ajuma@F2D29657.F60B0462.67AC9B1.IP)
- # [21:04] <ehsan|sheriff> espindola: ok, smaug is ahead of you
- # [21:04] * Joins: pcwalton (pcwalton@moz-7B0110AD.mv.mozilla.com)
- # [21:04] <espindola> ok
- # [21:05] <bjacob> joe: somehow, the patch in my MQ (what I sent you) doesn't have author info, which is why it used your identity
- # [21:05] <ehsan|sheriff> smaug: hmm, have you done a try PGO build?
- # [21:05] <smaug> um
- # [21:05] <smaug> does that happen by default?
- # [21:06] <margaret> vingtetun: https://bug695444.bugzilla.mozilla.org/attachment.cgi?id=581737
- # [21:06] <smaug> with try: -a
- # [21:06] <ehsan|sheriff> no
- # [21:07] <smaug> then no
- # [21:07] <ehsan|sheriff> smaug: I mean, your patches seem to be strictly adding code, and I'm worried that they might tip us over to linker crashes...
- # [21:07] <smaug> I have no idea how to do try pgo
- # [21:07] <ehsan|sheriff> smaug: mk_add_options MOZ_PGO=1 in the mozcongig
- # [21:07] <smaug> and have only that thing in .mozconfig?
- # [21:08] <khuey|quite-busy> ehsan|sheriff: fwiw, we should have a good bit of headroom
- # [21:08] <ehsan|sheriff> smaug: add it to browser/config/mozconfigs/nightly
- # [21:08] <ehsan|sheriff> khuey|quite-busy: oh really?
- # [21:08] <khuey|quite-busy> I was able to chuck in all of graphite on top of what's in the tree with no problems
- # [21:08] <ehsan|sheriff> ok I didn't know that
- # [21:08] * Quits: cpeterson (cpeterson@moz-BBE3ABD.mv.mozilla.com) (Input/output error)
- # [21:08] <ehsan|sheriff> smaug: in that case you can just land
- # [21:08] <ehsan|sheriff> jimm: so can you
- # [21:08] <smaug> k
- # [21:08] <khuey|quite-busy> ehsan|sheriff: sorry, I thought I had mentioned that
- # [21:09] <smaug> waiting for vingtetun
- # [21:09] <khuey|quite-busy> ehsan|sheriff: as long as people aren't adding gobs of code we should be ok, I think
- # [21:09] <jwir3> so win7 isn't on the try chooser syntax builder. anyone know what the switch is for this platform?
- # [21:09] <smaug> and waiting for my local build to compile
- # [21:09] * Joins: tonymec (tonymec@7923ADCD.7C3BF363.277517C1.IP)
- # [21:09] <ehsan|sheriff> khuey|quite-busy: awesome!
- # [21:09] <ehsan|sheriff> smaug: cool
- # [21:09] * Quits: gkw (gkw@moz-BBE3ABD.mv.mozilla.com) (Ping timeout)
- # [21:09] <ehsan|sheriff> smaug: if jimm shows up before you, let him land first
- # [21:09] * mdas|afk is now known as mdas
- # [21:09] <ehsan|sheriff> cause he asked earlier
- # [21:10] <smaug> k
- # [21:10] * Joins: gkw (gkw@moz-BBE3ABD.mv.mozilla.com)
- # [21:10] <ehsan|sheriff> thanks :)
- # [21:10] <ehsan|sheriff> espindola: you're also in the queue
- # [21:10] <vingtetun> smaug: done
- # [21:10] <jimm> I'm in the middle of a build, hold on a sec..
- # [21:10] <espindola> ehsan|sheriff, thanks
- # [21:10] <firebot> Check-in: http://hg.mozilla.org/mozilla-central/rev/fe5bd75f1553 - Vivien Nicolas - Bug 710548 - Add offline-app and webapp-manage permissions to pre-installed apps r=cjones DONTBUILD since this is b2g/ only
- # [21:10] <firebot> http://hg.mozilla.org/mozilla-central/rev/b7bafe0967a9 - Sriram Ramasubramanian - Bug 695444 - Form history autocomplete. r=mfinkle a=android-only
- # [21:10] <firebot> http://hg.mozilla.org/mozilla-central/rev/89cdde42ded6 - Vivien Nicolas - Bug 710804 - Use postMessage instead of an event for asking the homescreen to close an app r=fabrice DONTBUILD since this is b2g/ only
- # [21:10] <ehsan|sheriff> jimm: no rush
- # [21:10] * Joins: cpeterson (cpeterson@moz-BBE3ABD.mv.mozilla.com)
- # [21:10] * nthomas|away is now known as nthomas
- # [21:13] <margaret> vingtetun: thanks!
- # [21:13] * Joins: Polynomial-C (Poly-C@moz-F089909D.dip.t-dialin.net)
- # [21:14] * IRCMonkey50727 is now known as pascalc
- # [21:14] <ehsan|sheriff> bjacob: your build failed
- # [21:14] * khuey|quite-busy is now known as khuey|away
- # [21:15] <bjacob> ehsan|sheriff: coming
- # [21:17] * Joins: qwebirc (blah@A3C15257.C8511E92.79496794.IP)
- # [21:18] * Joins: TheOne (TheOne@moz-DFAC5DAC.dip.t-dialin.net)
- # [21:18] * Joins: priya (priya@moz-BBE3ABD.mv.mozilla.com)
- # [21:18] <qwebirc> hi, there's a bug with Firefox 11 and qwebirc 0.85, http://webchat.irchighway.net try logging in and it doesn't work, works in Firefox 8.0.1, I haven't tested with 9.0
- # [21:18] * Joins: bmoss (bmoss@moz-BBE3ABD.mv.mozilla.com)
- # [21:19] <bjacob> qwebirc: file a bug?
- # [21:19] * Quits: dcamp (dave@moz-8EBEC133.hsd1.wa.comcast.net) (Connection reset by peer)
- # [21:19] <qwebirc> Firefox 11 does work with qwebirc 0.91 though, I tested on their irc server here, http://webchat.quakenet.org/
- # [21:19] <qwebirc> very odd
- # [21:19] * Joins: dcamp (dave@moz-8EBEC133.hsd1.wa.comcast.net)
- # [21:20] <catlee> link.exe is at 2.8GB and climbing...
- # [21:20] * Joins: wolfiR_ (wolfiR@moz-2858382B.dip0.t-ipconnect.de)
- # [21:20] <smaug> qwebirc: do you get any warnings in error console when using 0.85?
- # [21:20] <@bz> erm
- # [21:20] <qwebirc> no :(
- # [21:20] <@bz> layout/generic/test/test_plugin_focus.html gives me:
- # [21:20] <@bz> todo | Platform not supported
- # [21:20] <@bz> why?
- # [21:20] * @bz is on Linux
- # [21:20] <@bz> and this is a 32-bit build...
- # [21:21] * Joins: sicking (chatzilla@moz-1FF05400.hsd1.ca.comcast.net)
- # [21:21] <philor> hmm, turning off the addonmgr xpinstall tests broke a couple of browser/base/content/ tests?
- # [21:21] * Joins: ddahl (ddahl@moz-CD6D5794.hfc.comcastbusiness.net)
- # [21:21] <@bz> erm
- # [21:21] <@bz> that test doesn't claim to run on Linux
- # [21:21] <@bz> so..
- # [21:21] <@bz> wait
- # [21:21] <philor> those aren't actually socks we're wearing inside our clownshoes, are they?
- # [21:21] * Joins: alexp (alexp@6BFC7CE9.2CABD6FE.22C20A9F.IP)
- # [21:21] <@bz> why the hell is it crashing on Linux, then?
- # [21:22] <qwebirc> oh wait, I do see some errors in the error console
- # [21:22] <qwebirc> Error: arguments.callee.caller is null
- # [21:22] <qwebirc> Source File:
- # [21:22] <qwebirc> http://webchat.irchighway.net/js/qui-867f7629c339.js
- # [21:22] <qwebirc> Line: 44
- # [21:22] * Quits: liuche (liuche@moz-B77DEAEB.mozilla.org) (Quit: Lost terminal)
- # [21:23] * Quits: mjschranz (mjschranz@moz-57B23647.senecac.on.ca) (Client exited)
- # [21:23] <espindola> what is the order in the queue? first one is at the top?
- # [21:23] <ehsan|sheriff> espindola: yes
- # [21:23] <nemo> qwebirc: wow. that's an old version of mootools
- # [21:23] <espindola> ok, thanks
- # [21:23] <smaug> qwebirc: you might want to ask #jsapi if there has been some changes to arguments.callee.caller handling
- # [21:23] <ehsan|sheriff> jimm: are you ready to land?
- # [21:23] * Quits: kdcw (kdc@moz-F7413045.pk.shawcable.net) (Quit: HydraIRC -> http://www.hydrairc.com <- s0 d4Mn l33t |t'z 5c4rY!)
- # [21:23] * @bz has seen a bug about that go by
- # [21:24] <@bz> sometime recently
- # [21:24] <jimm> ehsan|sheriff: no, I had to do a pull/config/build with these patches first. I'll need 20 minutes or so.
- # [21:24] * Quits: TheOne (TheOne@moz-DFAC5DAC.dip.t-dialin.net) (Ping timeout)
- # [21:24] * Joins: liuche (liuche@moz-B77DEAEB.mozilla.org)
- # [21:25] <qwebirc> not really inclined to report this as a bug since 0.91 works fine with Firefox 11
- # [21:25] <ehsan|sheriff> jimm: ok
- # [21:25] <ehsan|sheriff> smaug: you wanna land in the mean time?
- # [21:25] * Quits: bmoss (bmoss@moz-BBE3ABD.mv.mozilla.com) (Ping timeout)
- # [21:26] <smaug> my build should be ready real soon
- # [21:26] * Quits: pcwalton (pcwalton@moz-7B0110AD.mv.mozilla.com) (Quit: pcwalton)
- # [21:27] <nemo> qwebirc: huh. mootools still did useragent sniffing back then.
- # [21:27] * timA is now known as timA|lunch
- # [21:27] <qwebirc> from the about page, I see 0.85 using mootools 1.2 and 0.91 uses mootools 1.2.5
- # [21:27] * Quits: pnemsak (Miranda@moz-7409BD53.orange.sk) (Ping timeout)
- # [21:28] * Joins: pnemsak (Miranda@moz-7409BD53.orange.sk)
- # [21:28] <nemo> qwebirc: heh. was 0.91 release two years ago or something? :)
- # [21:28] <qwebirc> no idea, thats whats running on the qwebirc site
- # [21:29] <qwebirc> Copyright © 2008-2011 Chris Porter and the qwebirc project.
- # [21:29] <ehsan|sheriff> smaug: cool, if you finished before jimm, please land
- # [21:29] <nemo> oh. actually 1.2.5 is 1¼ years old :)
- # [21:29] <nemo> http://mootools.net/blog/2010/09/15/mootools-core-1-2-5/
- # [21:29] <qwebirc> qwebirc seems to be 2011
- # [21:29] <smaug> ehsan|sheriff: --rebasing...
- # [21:29] <qwebirc> for 0.91
- # [21:29] * Quits: wolfiR_ (wolfiR@moz-2858382B.dip0.t-ipconnect.de) (Ping timeout)
- # [21:29] <nemo> contains fixes for Firefox 4 beta! :D
- # [21:29] * Quits: tH (Rob@cpc16-seac19-2-0-cust234.7-2.cable.virginmedia.com) (Ping timeout)
- # [21:29] <ehsan|sheriff> bjacob: your build is failing again
- # [21:29] <ehsan|sheriff> no idea why
- # [21:29] <bjacob> coming
- # [21:30] <philor> ehsan|sheriff: did you notice that you have permaorange?
- # [21:30] <nemo> qwebirc: http://stackoverflow.com/questions/103598/why-was-the-arguments-callee-caller-property-deprecated-in-javascript
- # [21:30] * Joins: tH (Rob@cpc16-seac19-2-0-cust234.7-2.cable.virginmedia.com)
- # [21:30] <smaug> some unstar'ed oranges
- # [21:30] <ehsan|sheriff> philor: oh noez
- # [21:30] <nemo> qwebirc: deprecated years ago eh. wow. that site there is cutting edge ;)
- # [21:31] <qwebirc> nemo shrug, I guess I can only hope irchighway would update their qwebirc to 0.91 which works with Firefox 11 and probably 12 next week
- # [21:31] <nemo> https://developer.mozilla.org/en/JavaScript/Reference/Functions_and_function_scope/arguments/callee
- # [21:31] * armenzg is now known as armenzg_mtg
- # [21:32] <philor> ehsan|sheriff: I haven't looked, but I bet those tests use something installed by the toolkit/.../xpinstall/ tests
- # [21:32] <ehsan|sheriff> ah ok
- # [21:32] <philor> given where it starts
- # [21:32] <nemo> qwebirc: looks like they might want to bump mootools up a bit while they are at it :)
- # [21:32] <ehsan|sheriff> philor: those are fixed I believe
- # [21:33] * Joins: mjschranz (mjschranz@moz-57B23647.senecac.on.ca)
- # [21:33] <nemo> oh. wow. hg pull -u worked w/o failing despite fact I hadn't updated m-c in a couple of months
- # [21:33] <philor> thus my fear that rather than socks inside our clownshoes, we actually have another pair of clownshoes in there, all the way down
- # [21:33] <smaug> ehsan|sheriff: I can't get the summary for the browser-chrome failures
- # [21:33] * timA|lunch is now known as timA
- # [21:33] <nemo> that's never happened before :)
- # [21:34] <nemo> bjacob: heeey, what was that config option you said I should enable for profiling?
- # [21:34] * Quits: jfkthame (jfkthame@DBA4BEB5.90783722.9542EC20.IP) (Quit: jfkthame)
- # [21:34] <timA> firebot: uuid
- # [21:34] <firebot> 4b26b116-39c0-46d4-aa62-b601b60d1f4a (/msg firebot cid for CID form)
- # [21:34] <nemo> bjacob: I thought I'd try that 2D sprite thing again, since as noted we were worried about that webgl port
- # [21:34] <bjacob> nemo: --enable-profiling, it implites -fno-omit-frame-pointers
- # [21:34] <ehsan|sheriff> Mossop: you around?
- # [21:34] <bjacob> implies
- # [21:35] <philor> ehsan|sheriff: fixed where? khuey disabled a directory on which tests in another directory depend, and nothing put it back
- # [21:35] <ehsan|sheriff> ah dammit
- # [21:35] <ehsan|sheriff> right
- # [21:35] <ehsan|sheriff> ok
- # [21:35] <ehsan|sheriff> I'll fix it
- # [21:35] <philor> thx
- # [21:35] <nemo> bjacob: gracias
- # [21:36] <jlebar> taras, So about:telemetry shows me slow SQL statements, but only the total amount of time I spent in them. I kind of want a histogram.
- # [21:36] <db48x> philor: meta-clownshoes?
- # [21:37] * Quits: anant (anant@moz-BBE3ABD.mv.mozilla.com) (Connection reset by peer)
- # [21:37] * Joins: anant (anant@moz-BBE3ABD.mv.mozilla.com)
- # [21:38] * Joins: bjarne (bjarne@moz-B436FB98.nextgentel.com)
- # [21:38] <smaug> ehsan|sheriff: so, are you fixing something, or is it ok to land?
- # [21:38] <ehsan|sheriff> smaug: ok, you can land now
- # [21:38] * @bz thinks we should rename "sheriff" to "liposuctor" or something
- # [21:38] <@bz> obviously
- # [21:39] <ehsan|sheriff> haha
- # [21:39] * Quits: tfair (tfairey@moz-BBE3ABD.mv.mozilla.com) (Client exited)
- # [21:39] * Joins: DGMurdockIII (dgmurdocki@moz-E933B63.hsd1.il.comcast.net)
- # [21:39] <firebot> Check-in: http://hg.mozilla.org/mozilla-central/rev/4d2921df8f34 - Ehsan Akhgari - Bug 709531 - Remove two more tests which depend on the xpinstall directory; a=me DONTBUILD
- # [21:39] * mcote is now known as mcote|bbiab
- # [21:39] <nemo> bjacob: this is my first firefox build since I bumped this machine up to 16GiB - I'm looking forward to linking :)
- # [21:39] <philor> "ABORT: PresArena: poison overwritten" sound sort of like it isn't joking around
- # [21:40] * Quits: graydot (jeba@B6001B9.4E8FB30C.700C6EB0.IP) (Quit: graydot)
- # [21:40] <bjacob> nemo: use gold
- # [21:41] <mcpherrin> bjacob: how do I change linkers in our build system?
- # [21:41] <smaug> what, do I need a=
- # [21:41] <smaug> bah
- # [21:41] <jlebar> mcpherrin, export LD=ld.gold
- # [21:41] <bjacob> mcpherrin: just install gold, your distro should make it the default linker
- # [21:41] <jlebar> or that.
- # [21:42] <jlebar> apt-get install binutils-gold
- # [21:42] <qwebirc> ok, it's not the version of qwebirc, http://webchat.freenode.net/ uses 0.90 with mootools 1.2 and it doesn't work with Firefox 11 either
- # [21:42] <qwebirc> time to bug these people to update their mootools :(
- # [21:42] * Quits: waynenguyen (HP@6EB57D2.92EE8F45.B425DC1D.IP) (Connection reset by peer)
- # [21:42] <nemo> qwebirc: not surprising if they are using something deprecated 3 or 4 years ago :)
- # [21:42] <nemo> but then again, deprecating is a tricky thing on the web
- # [21:42] <mcpherrin> bah, arch linux doesn't ship gold
- # [21:43] * armenzg_mtg is now known as armenzg
- # [21:43] <nemo> bjacob: I'm not switching immediately. dammit. I want to enjoy my 16GiB :-p
- # [21:43] <smaug> ehsan|sheriff: done
- # [21:43] <jlebar> So what happens if you take the 64-bit MSVC linker and pass /MACHINE:X86?
- # [21:43] <jlebar> Does it invoke a 32-bit process?
- # [21:43] <jlebar> ted ^?
- # [21:43] <firebot> Check-in: http://hg.mozilla.org/mozilla-central/rev/141ae3a6ccd2 - Olli Pettay - Bug 708701 - Implement HTML event ctors, r=jst
- # [21:43] <firebot> http://hg.mozilla.org/mozilla-central/rev/01a26239462a - Olli Pettay - Bug 675884 - Implement Event constructors, part1, r=jst
- # [21:43] * Joins: WeirdAl (chatzilla@moz-D461843.ask.info)
- # [21:43] <firebot> http://hg.mozilla.org/mozilla-central/rev/0507a68e994b - Olli Pettay - Bug 709127 - Implement MouseEvent and UIEvent ctors, r=jst, a=sheriff
- # [21:43] <nemo> jlebar: this /. person claims linking is always 32bit? http://developers.slashdot.org/comments.pl?sid=2574298&cid=38372996
- # [21:43] <firebot> http://hg.mozilla.org/mozilla-central/rev/1ab5ed4f93bf - Olli Pettay - Bug 675884 - Implement Event constructors, part2, r=jst
- # [21:44] <nemo> is that /. people just making stuff up?
- # [21:44] <nemo> he's got a +5 insightful, he must be right!
- # [21:44] <jlebar> nemo, he's saying that the 64-bit linker is a 64-bit program?
- # [21:44] <jlebar> I did not think that.
- # [21:45] <nemo> "cl.exe is a 32-bit process. Always."
- # [21:45] <jlebar> But someone else is saying that the 64-bit linker can link 32-bit programs with /MACHINE:X86.
- # [21:45] <ehsan|sheriff> smaug: cool, thanks!
- # [21:45] * Quits: tonymec (tonymec@7923ADCD.7C3BF363.277517C1.IP) (Input/output error)
- # [21:45] <ted> jlebar: i have no idea
- # [21:45] <jlebar> I have to admit, that MSDN page sure looks like he could be right.
- # [21:45] <nemo> jlebar: yeah, but you were asking 'bout 32bit process, and seemed like he was saying it always was
- # [21:45] <ehsan|sheriff> jimm: when will you be ready?
- # [21:45] <nemo> jlebar: I guess that means the other half of his post is right too :-p
- # [21:45] <jlebar> :D
- # [21:45] <nemo> "incredibly deep-rooted problems within the architecture of Firefox"
- # [21:46] <bjacob> nemo: if you have a big CPU cache, you can go crazy with make -j20
- # [21:46] <jimm> ehsan|sheriff: getting closer, maybe ten minutes
- # [21:46] <bjacob> this way you'll enjoy your memoryu
- # [21:46] <ehsan|sheriff> jimm: 10 mins is awesome
- # [21:46] <ted> nemo: i love hearing people talk about things they have no idea about
- # [21:46] * Joins: tonymec (tonymec@7923ADCD.7C3BF363.277517C1.IP)
- # [21:46] <ted> would they make the same statement about chrome?
- # [21:46] <Mossop> ehsan|sheriff: Whut?
- # [21:46] <nemo> ted: chrome didn't attempt pgo, according to that bug it required too much memory
- # [21:47] <ehsan|sheriff> Mossop: nothing, disabled two more tests, see the link in the bug
- # [21:47] <nemo> ted: someone did point it out in the /. post
- # [21:47] <@bz> nemo: they used to do pgo
- # [21:47] <qwebirc> Chrome needs more than 3GB to link?
- # [21:47] <@bz> nemo: they stopped when they started running out of memory
- # [21:47] <nemo> qwebirc: supposedly 9GiB
- # [21:47] <@bz> qwebirc: yes, even without pgo
- # [21:47] <@bz> qwebirc: they build on 64-bit machines
- # [21:47] <qwebirc> oh
- # [21:47] * Quits: tH (Rob@cpc16-seac19-2-0-cust234.7-2.cable.virginmedia.com) (Ping timeout)
- # [21:47] <@bz> nemo: that's unlikely
- # [21:48] <@bz> nemo: since the linker can't do that
- # [21:48] <nemo> bz: http://developers.slashdot.org/comments.pl?sid=2574298&cid=38373000
- # [21:48] <nemo> bz: right. that was someone's guess
- # [21:48] <nemo> that it couldn't be done
- # [21:48] <@bz> nemo: oh, with pgo
- # [21:48] <ehsan|sheriff> yeah, if they needed 9gigs, they couldn't ship on windows ;)
- # [21:48] * Joins: wolfiR_ (wolfiR@moz-2858382B.dip0.t-ipconnect.de)
- # [21:48] <nemo> bz: yep
- # [21:48] <ted> nemo: i know, that was the point i made
- # [21:48] <nemo> ehsan|sheriff: https://code.google.com/p/chromium/issues/detail?id=21932#c10 as someone pointed out (GB is probably 9gb w/ a shift :)
- # [21:49] <ted> chrome can't even build as pgo for tis same reason
- # [21:49] <ehsan|sheriff> nope
- # [21:49] <ehsan|sheriff> wrong conclusion :)
- # [21:50] * Joins: Enn (enn@A974B2E7.1BC119D8.8C65A0C1.IP)
- # [21:50] <ted> there are people claming that you can make the 64-bit toolchain produce 32-bit binaries
- # [21:50] <ted> i would be interested to see
- # [21:50] <Mossop> ehsan|sheriff: ftr I think browser_bug557956.js is the faulty test that isn't resetting state properly. I am stuck in meetings now though so can't fix it
- # [21:50] <jlebar> ted, metoo.
- # [21:50] <jduell> bz / biesi : ping
- # [21:50] * Quits: KWierso (chatzilla@moz-608EC7C5.desm.qwest.net) (Ping timeout)
- # [21:50] * Joins: roc (chatzilla@538BABFE.A073F3E.97BBD552.IP)
- # [21:51] <ehsan|sheriff> Mossop: well, the directory that the test depends on won't exist any more...
- # [21:51] <@bz> roc: hey
- # [21:51] <Mossop> I mean that is why we disabled all the tests, because that one test isn't resetting extensions.getAddons.cache.enabled
- # [21:51] <@bz> roc: so asked you for a review, but that patch is not quite right
- # [21:51] <@bz> roc: so wait an hour or so, please
- # [21:51] <roc> sigh. The Slashdot thread about linker OOM is an explosion of ignorance and error. At least the modded-up comments were mostly good ones.
- # [21:51] * Joins: KWierso (chatzilla@moz-4E7BE0E1.desm.qwest.net)
- # [21:51] <@bz> roc: yeah
- # [21:51] <@bz> roc: verily
- # [21:51] * Quits: clee (clee@moz-BBE3ABD.mv.mozilla.com) (Connection reset by peer)
- # [21:52] * Joins: clee (clee@moz-BBE3ABD.mv.mozilla.com)
- # [21:52] <jduell> bz: tracking down some loadgroup issues with websockets. I noticed that WS creates an HTTP channel and calls SetLoadGroup(loadGroup) on it, but doesn't calls loudGroup->AddRequest(httpChan).
- # [21:52] <jlebar> but seriously...I wonder what happens when you pass /MACHINE:X86 to 64-bit link.exe.
- # [21:53] <@bz> jduell: does it set the loadgroup before asyncopen?
- # [21:53] * Quits: Mook_as (mook@moz-1FCC0032.activestate.com) (Quit: ChatZilla 0.9.87-rdmsoft [XULRunner 1.9.2.13/20101203074205])
- # [21:53] <bsmedberg> where do you get a 64-bit link.exe?
- # [21:53] <bwinton> bsmedberg: You just concatenate two 32-bit link.exes…
- # [21:53] * Quits: Honza (chatzilla@E07BF19C.5BB5597D.D0083327.IP) (Ping timeout)
- # [21:53] * Joins: Waldo (waldo@moz-534B4EF1.hsd1.mi.comcast.net)
- # [21:53] <roc> from the 64-bit toolchain
- # [21:53] * Joins: Mook_as (mook@moz-1FCC0032.activestate.com)
- # [21:53] * Joins: cpearce (chatzilla@538BABFE.A073F3E.97BBD552.IP)
- # [21:53] <bsmedberg> I have VC10 installed on my win7-64bit
- # [21:54] <bsmedberg> and it's all in Program Files (x86)
- # [21:54] <jduell> bz: yes, before AsyncOpen
- # [21:54] <@bz> jduell: then that's fine; asyncopen will add to the loadgroup
- # [21:54] <roc> bsmedberg: yes, but the tools under bin/amd64 are 64-bit
- # [21:54] <jduell> bz: right--thanks!
- # [21:54] <@bz> jduell: no problem
- # [21:55] <biesi> jduell, pong
- # [21:55] * Quits: Enn (enn@A974B2E7.1BC119D8.8C65A0C1.IP) (Ping timeout)
- # [21:55] <biesi> jduell, what bz said
- # [21:55] <jduell> biesi: right :)
- # [21:55] <roc> $ dumpbin -headers /c/Program\ Files\ \(x86\)/Microsoft\ Visual\ Studio\ 10.0/VC/bin/amd64/link.exe
- # [21:55] <biesi> :)
- # [21:55] <roc> FILE HEADER VALUES
- # [21:55] <roc> 8664 machine (x64)
- # [21:55] <@bz> amd64
- # [21:55] <@bz> nice
- # [21:56] <@bz> program installation conventions?
- # [21:56] <roc> and if you disasm, you can see it using 64-bit pointers
- # [21:56] <@bz> What are those?
- # [21:56] * Joins: davidb_ (davidb@moz-976CD2.dsl.bell.ca)
- # [21:56] <philor> I know this will come as a shock to one and all, but
- # [21:56] <philor> **************
- # [21:56] <philor> someone needs to actually look at Android failures
- # [21:56] <philor> *********************
- # [21:56] <bsmedberg> shocking
- # [21:57] <ted> jlebar: doesn't seem to work
- # [21:57] <sid0> The whole separate installation dirs for 32 and 64 bit idea is just so fundamentally broken
- # [21:57] <catlee> I, for one, am shocked
- # [21:57] <Waldo> philor: your winnings, sir
- # [21:57] <ted> roc: huh
- # [21:57] <jlebar> ted, what happens?
- # [21:58] * Quits: Waldo (waldo@moz-534B4EF1.hsd1.mi.comcast.net) (Quit: system updates)
- # [21:58] <ted> http://pastebin.mozilla.org/1405749
- # [21:58] <ted> test.obj was produced with the 32-bit cl.ee
- # [21:58] <ted> exe
- # [21:58] <ted> both vc2008
- # [21:58] <jlebar> aha
- # [21:59] <jlebar> well, it was a nice idea.
- # [21:59] <ted> nice, but wrong
- # [21:59] <ted> like all things on slashdot
- # [21:59] <philor> Cwiiis: you need to back out
- # [21:59] <ehsan|sheriff> we can binary patch the linker to jump over that compare instruction ;)
- # [21:59] <philor> world: you need to not pile on
- # [21:59] * Quits: karl (karl@moz-F2842753.jetstream.xtra.co.nz) (Ping timeout)
- # [22:00] <ehsan|sheriff> philor: I will happily take care of that :)
- # [22:00] <nemo> bjacob: sorry. bad memory etc - you recommended "prof" for profiling?
- # [22:01] <@bz> ehsan|sheriff: we could binary patch it to also produce the right output
- # [22:01] <ehsan|sheriff> philor: Cwiiis: backed out
- # [22:01] <@bz> ehsan|sheriff: this may or may not be more work than writing our own linker. ;)
- # [22:01] <taras> jlebar: patches welcome
- # [22:01] <ehsan|sheriff> bz: we can always write our own linker in C++ and compile it!
- # [22:01] <bjacob> nemo: perf, if you're on linux
- # [22:01] * Joins: mgoodwin (mgoodwin@moz-4BB9E5E9.cable.virginmedia.com)
- # [22:01] <jlebar> taras, Even displaying the max would be more useful than the total.
- # [22:02] * bjacob is now known as bjacob_meeting
- # [22:02] <firebot> Check-in: http://hg.mozilla.org/mozilla-central/rev/d508455660d3 - Ehsan Akhgari - Backout changeset 3e972d3efc11 (bug 708307) because it breaks Android tests; a=me
- # [22:02] <jlebar> taras, We're not going to send the total to telemetry, right?
- # [22:02] <taras> jlebar: note these are a sum of queries over 100ms
- # [22:02] <taras> jlebar: total is
- # [22:02] <taras> what we will send
- # [22:03] <taras> jlebar: the problem these are solving is helping us prioritize queries to move off main thread in near term
- # [22:03] * Joins: Waldo (waldo@moz-534B4EF1.hsd1.mi.comcast.net)
- # [22:03] <jlebar> taras, But how can you distinguish between something which usually runs at 100ms but occasionally runs at 1000ms, and something which always runs at 150ms?
- # [22:03] <jlebar> taras, One of those is a bigger priority.
- # [22:03] <@bz> ehsan|sheriff: precisely
- # [22:04] * Joins: Enn (enn@A974B2E7.1BC119D8.8C65A0C1.IP)
- # [22:04] <taras> jlebar: it's true more precision will be good
- # [22:04] <jlebar> taras, More precision is *essential*!
- # [22:04] <taras> i disagree
- # [22:04] <taras> having data is required
- # [22:04] <jlebar> This data is misleading.
- # [22:04] <taras> once you have data, every improvement is incremental
- # [22:04] * Joins: tH (Rob@cpc16-seac19-2-0-cust234.7-2.cable.virginmedia.com)
- # [22:04] <taras> ie now we dont even know which db to blame for io
- # [22:05] * Quits: jrmuizel (jrmuizel@F2D29657.F60B0462.67AC9B1.IP) (Input/output error)
- # [22:05] * Quits: peregrino (peregrino@moz-1CB238F5.telecom.net.ar) (Quit: peregrino)
- # [22:05] * Joins: dRdR (dRdR@moz-57178BD3.hsd1.ca.comcast.net)
- # [22:07] * Quits: vingtetun (vingtetun@moz-A2A3C9E6.vlan402.asr1.cdg3.gblx.net) (Ping timeout)
- # [22:07] <jlebar> taras, I'm very worried when potentially misleading data is collected. An important purpose is to catch things like that bad sqlite hang we had a few versions ago. In order to catch that, I think we need better data than this.
- # [22:08] <jlebar> taras, But my point was just to make sure that improving the data is on your radar.
- # [22:08] * Joins: vladan (vladan@F2D29657.F60B0462.67AC9B1.IP)
- # [22:08] <taras> jlebar: it is, i want to see existing data and go from there
- # [22:08] <@bz> oh, heycam....
- # [22:08] * Quits: mjschranz (mjschranz@moz-57B23647.senecac.on.ca) (Client exited)
- # [22:08] <taras> there is a lot of data and limited dev resources, thus baby steps
- # [22:09] * Joins: jrmuizel (jrmuizel@F2D29657.F60B0462.67AC9B1.IP)
- # [22:09] * jgriffin-afk is now known as jgriffin
- # [22:09] <jlebar> taras, Understood. But in the future, I submit that "total time" is not an appropriate metric for snappy.
- # [22:09] <jlebar> taras, Total time spent GC'ing would not be useful.
- # [22:10] * Quits: Enn (enn@A974B2E7.1BC119D8.8C65A0C1.IP) (Ping timeout)
- # [22:10] <ted> heh, my linker is using more than 3GB of VM already
- # [22:10] <jlebar> Total time spent on the network would not be useful.
- # [22:10] * Joins: Boriss_ (FlyingToas@moz-C03D0C61.vlan426.asr1.sfo1.gblx.net)
- # [22:11] <Cwiiis> ehsan|sheriff, ah, what was the problem?
- # [22:11] * Quits: Boriss (FlyingToas@moz-C03D0C61.vlan426.asr1.sfo1.gblx.net) (Ping timeout)
- # [22:11] * Boriss_ is now known as Boriss
- # [22:12] <Cwiiis> ehsan|sheriff, I assume the test failures were due to something underneath it? I'll see about getting it in tomorrow
- # [22:12] * Joins: alice0775 (Mibbit@moz-A36EF071.osaka.ocn.ne.jp)
- # [22:13] * Quits: Jake (Jake@moz-5D28CC35.phlapafg.dynamic.covad.net) (Quit: Leaving.)
- # [22:14] <NeilAway> ted: you can use the 64-bit linker but only without pgo of course ;-)
- # [22:14] * khuey|away is now known as khuey|quite-busy
- # [22:14] * Joins: jhorak (jhorak@moz-59813FB4.cust.nbox.cz)
- # [22:16] <ted> NeilAway: ah
- # [22:17] <philor> Cwiiis: since every single test on your push and above crashes, and every single test on the one below it did not, I don't think that's a very good bet
- # [22:17] * Joins: mcsmurf (mcsmurf@moz-31272968.dip.t-dialin.net)
- # [22:19] <Unfocused> ugh, sorry for the network-related oranges :( filed the bug the other day to fix that, and promptly fell asleep. burn-out ftl
- # [22:19] <jimm> ehsan|sheriff: ok to land?
- # [22:19] * Joins: Wevah (Wevah@moz-97AD33CE.stcd.qwest.net)
- # [22:20] <philor> is every orange and red on the tree starred?
- # [22:21] * Joins: mjschranz (mjschranz@moz-57B23647.senecac.on.ca)
- # [22:21] * philor should probably treat his blood pressure by just changing the tree rules to reflect reality
- # [22:21] <ehsan|sheriff> Cwiiis: sorry, no idea
- # [22:21] <froydnj> philor: embrace borkenness and anarchy!
- # [22:22] * Quits: fs (Elchi3@B9C9103E.56629902.2EC4CA51.IP) (Client exited)
- # [22:22] * Quits: Hendikins (wolfox@moz-9A361C95.static.internode.on.net) (Ping timeout)
- # [22:22] * Waldo hates mailing lists and reply-to
- # [22:22] * Joins: Hendikins (wolfox@moz-9A361C95.static.internode.on.net)
- # [22:23] * Joins: mkaply (chatzilla@moz-92EDDD02.lightspeed.austtx.sbcglobal.net)
- # [22:23] * ChanServ sets mode: +o mkaply
- # [22:25] * Quits: Pike (Pike@moz-E31CD2CB.mozilla.org) (Input/output error)
- # [22:25] <ehsan|sheriff> jimm: I'm cleaning up the oranges...
- # [22:26] <mfinkle> ehsan|sheriff, can I get added to the queue?
- # [22:26] <ehsan|sheriff> mfinkle: of course
- # [22:26] <mfinkle> android-only landing
- # [22:26] * Joins: rs (rs@moz-217F02CE.lightspeed.sntcca.sbcglobal.net)
- # [22:26] <mcsmurf> is the queue for important checkins only? I have a Windows-only installer patch, which I wanted to check in for weeks now ;)
- # [22:26] <ehsan|sheriff> mfinkle: added you
- # [22:26] <ehsan|sheriff> mcsmurf: no, for all
- # [22:27] <ehsan|sheriff> mcsmurf: you wanna be added?
- # [22:27] <mcsmurf> yeah
- # [22:27] <ehsan|sheriff> done
- # [22:27] * Waldo resists the urge to troll now, seeing as he really doesn't have anything that important that he must land immediately
- # [22:27] <mcsmurf> :P
- # [22:28] * Quits: WeirdAl (chatzilla@moz-D461843.ask.info) (Quit: ChatZilla 0.9.87 [Firefox 8.0/20111104165243])
- # [22:28] * Quits: jrmuizel (jrmuizel@F2D29657.F60B0462.67AC9B1.IP) (Input/output error)
- # [22:28] <ehsan|sheriff> blassey: wanna file https://tbpl.mozilla.org/php/getParsedLog.php?id=7934499&tree=Firefox&full=1?
- # [22:28] <ehsan|sheriff> jimm: green light
- # [22:29] * mcote|bbiab is now known as mcote
- # [22:29] * Joins: lsumar (lsumar@538BABFE.A073F3E.97BBD552.IP)
- # [22:29] * Joins: Pike (Pike@moz-E31CD2CB.mozilla.org)
- # [22:30] * Quits: zuzelvp (zuzelvp@F2D29657.F60B0462.67AC9B1.IP) (Ping timeout)
- # [22:30] <firebot> Check-in: http://hg.mozilla.org/mozilla-central/rev/7f5a60c89420 - Jim Mathies - Bug 661991 - close file pickers when the parent window closes. r=ehsan
- # [22:30] <firebot> http://hg.mozilla.org/mozilla-central/rev/df958b75de87 - Jim Mathies - Bug 661991 - Add query interface support, com dialog events, and split up xp/vista+ show calls. r=ehsan
- # [22:30] <firebot> http://hg.mozilla.org/mozilla-central/rev/5131c0b1982f - Jim Mathies - Bug 661991 - tests. r=bbondy, a=ehsan
- # [22:30] <firebot> http://hg.mozilla.org/mozilla-central/rev/46f1d7608fb4 - Jim Mathies - Bug 661991 - add support for new folder pickers. r=ehsan
- # [22:30] <firebot> http://hg.mozilla.org/mozilla-central/rev/487e81eb1b6b - Jim Mathies - Bug 661991 - add support for new file pickers. r=ehsan
- # [22:30] <firebot> http://hg.mozilla.org/mozilla-central/rev/77bcf3671260 - Jim Mathies - Bug 661991 - dynamically load entry point to SHCreateItemFromParsingName and share this with jump list code. r=ehsan.
- # [22:30] <firebot> http://hg.mozilla.org/mozilla-central/rev/4f2b48a62b75 - Jim Mathies - Bug 661991 - Cleanup win widget nsFilePicker. r=neil, a=ehsan
- # [22:30] <@bz> do we produce localized nightlies?
- # [22:30] <khuey|quite-busy> yes
- # [22:31] <@bz> ok
- # [22:31] <@bz> and we have about 90 locales
- # [22:31] <@bz> right?
- # [22:31] <ehsan|sheriff> khuey|quite-busy: go and study!
- # [22:31] <ehsan|sheriff> yes
- # [22:31] * Joins: zuzelvp (zuzelvp@F2D29657.F60B0462.67AC9B1.IP)
- # [22:31] <mcsmurf> ehsan|sheriff: so many checkins, will that mean that I'll have to wait quite a bit until I can check-in?
- # [22:31] <ehsan|sheriff> iirc we don't generate localized builds for all of them...
- # [22:31] * Quits: wolfiR_ (wolfiR@moz-2858382B.dip0.t-ipconnect.de) (Quit: Leaving)
- # [22:31] <ehsan|sheriff> Pike: ^
- # [22:31] <ehsan|sheriff> mcsmurf: no, hopefully ~30 mins
- # [22:31] <blassey> ehsan|sheriff: is that the Ts from my push?
- # [22:31] <ehsan|sheriff> or less
- # [22:32] <mcsmurf> ah ok, that's fine
- # [22:32] <Pike> bz: what's the question?
- # [22:32] <ehsan|sheriff> blassey: no, osx10.7 m-4
- # [22:32] * Joins: kats (kats@moz-DE13FC16.compute-1.amazonaws.com)
- # [22:32] <@bz> Pike: "how many nightly builds do we produce?"
- # [22:32] <ehsan|sheriff> blassey: I triggered another build and it was green
- # [22:32] <Pike> bz: thousands, probably
- # [22:32] <@bz> Pike: well, on Windows
- # [22:32] <ehsan|sheriff> espindola: you ready?
- # [22:32] * Quits: sheppy (sheppy@moz-4992DE6D.static.kgpt.tn.charter.com) (Quit: sheppy)
- # [22:32] <@bz> Pike: I assueme 4*(numlocales)
- # [22:32] <@bz> Pike: (32/64, opt/debug)
- # [22:33] <espindola> ehsan|sheriff, ready
- # [22:33] <blassey> ehsan|sheriff: sure thing then
- # [22:33] <Pike> bz: we only do repacks for l10n, so not real builds
- # [22:33] <@bz> Pike: ah, ok
- # [22:33] <Pike> bz: download the en-US, unpack, rip out en-US, package in l10n
- # [22:33] <@bz> Pike: so 4 real builds, then?
- # [22:33] * Parts: kats (kats@moz-DE13FC16.compute-1.amazonaws.com)
- # [22:33] * Quits: evilpie (chatzilla@moz-83A50905.pools.arcor-ip.net) (Quit: ChatZilla 0.9.87 [Firefox 11.0a1/20111213031157])
- # [22:33] <ehsan|sheriff> espindola: please land
- # [22:34] <Pike> bz: yeah, we only compile stuff for en-US
- # [22:34] * bear is now known as bear-afk
- # [22:34] * bjacob_meeting is now known as bjacob
- # [22:34] <Pike> bz: and we only repack 32bit opt, so one windows build (and two for linux, and one for mac)
- # [22:34] <espindola> ehsan|sheriff, done, thanks
- # [22:34] <Pike> bz: the repack itself for windwos takes about 3 mins per locale
- # [22:35] <firebot> Check-in: http://hg.mozilla.org/mozilla-central/rev/99dedca7064e - Rafael Ávila de Espíndola - Bug 709406: Fix up slow SQL reporting. r=taras. a=ehsan.
- # [22:35] * davehunt|away is now known as davehunt
- # [22:36] * Quits: cpeterson (cpeterson@moz-BBE3ABD.mv.mozilla.com) (Input/output error)
- # [22:36] <Callek> khuey|quite-busy, ehsan|sheriff: shall I presume these test disables/removes are _simply_ to keep things green during this current phx outage, and will get re-enabled soon?
- # [22:36] <ehsan|sheriff> mfinkle: so you said yours is android only?
- # [22:36] <mfinkle> ehsan|sheriff, yep
- # [22:36] <Callek> I really really would hate if this leads to us breaking the functionality tested there
- # [22:37] <ehsan|sheriff> Callek: they will get re-enabled when the tests are fixed :P
- # [22:37] <ehsan|sheriff> Callek: the bug is a P1 blocker
- # [22:37] <ehsan|sheriff> mfinkle: land in 5 mins please?
- # [22:37] <mfinkle> ehsan|sheriff, ok
- # [22:37] * Joins: bmoss (bmoss@moz-BBE3ABD.mv.mozilla.com)
- # [22:37] * Quits: ericjung (Mibbit@5210CFD5.1A5EA44.72B23B3D.IP) (Quit: http://www.mibbit.com ajax IRC Client)
- # [22:38] <ejpbruel> sicking: ping
- # [22:38] <Callek> ehsan|sheriff: P1 blocker changed by philor today, not sure if the owner of the bug prioritizes based on what non-employees set things as :-P
- # [22:38] <Unfocused> Callek / ehsan|sheriff: i'm working on fixes right now
- # [22:38] <@bz> how many win builders do we actually have?
- # [22:38] <Callek> Unfocused: :-)
- # [22:38] <sicking> ejpbruel: pong
- # [22:38] * Quits: micahg (micahg@moz-C7CC8D27.c3-0.arm-ubr1.chi-arm.il.cable.rcn.com) (Quit: Ex-Chat)
- # [22:38] <philor> bz: http://build.mozilla.org/builds/last-job-per-slave.txt
- # [22:39] <Unfocused> (bare with me - i'm exhausted)
- # [22:39] <ehsan|sheriff> Unfocused: thanks a million!
- # [22:39] <Callek> Unfocused: I just get concerned when I see big blocks of tests disabled due to "bustage" even if it is test issues, history (as far as I see) shows that it usually ends up with the tests not getting fixed... so as long as it is, I won't argue
- # [22:39] <ehsan|sheriff> bz: 5, probably ;)
- # [22:39] <ejpbruel> sicking: could you give me some feedback on bug 701591?
- # [22:39] <ehsan|sheriff> bz: to make things worse, I've got one of them out of the pool
- # [22:40] <ehsan|sheriff> it's an x64 builder though
- # [22:40] * Joins: micahg (micahg@moz-C7CC8D27.c3-0.arm-ubr1.chi-arm.il.cable.rcn.com)
- # [22:40] * Quits: qwebirc (blah@A3C15257.C8511E92.79496794.IP) (Quit: )
- # [22:40] <@bz> philor: hmm. Do we use the same pool for build and test?
- # [22:40] <Callek> bz: no
- # [22:40] <ejpbruel> sicking: specifically, is the behavior described intended, an unfortunate but unfixable side-effect, or is it fixable?
- # [22:40] <Callek> bz: different physical machine pools
- # [22:40] <philor> build is way down the line, 37 of them, 30 active
- # [22:40] * Quits: azakai (alon@moz-BBE3ABD.mv.mozilla.com) (Quit: Ex-Chat)
- # [22:40] <@bz> ok
- # [22:40] <@bz> thanks
- # [22:41] <philor> cmd+f, mw32-ix-slave02 will get you down there
- # [22:41] <timA> firebot: uuid
- # [22:41] <firebot> c3d1f00a-8146-43b4-a044-cb246233358e (/msg firebot cid for CID form)
- # [22:41] <philor> 29-4 for try
- # [22:41] <ehsan|sheriff> mfinkle: please land whenever you're ready
- # [22:41] <Ms2ger> I sure hope nobody is going to push addon manager code
- # [22:42] * Joins: cpeterson (cpeterson@moz-BBE3ABD.mv.mozilla.com)
- # [22:42] <sicking> ejpbruel: I'm not super familiar with the code sending out status notifications for pageloads
- # [22:42] <sicking> ejpbruel: i suspect it's a docshell problem
- # [22:42] <sicking> ejpbruel: i don't think it's unfixable, few things are :)
- # [22:42] * davehunt is now known as davehunt|mtg
- # [22:43] <ejpbruel> sicking: right, thats all i really needed :)
- # [22:43] * Joins: karl (karl@538BABFE.A073F3E.97BBD552.IP)
- # [22:43] * Quits: msucan (msucan-@FA9E8863.56E67207.699550A1.IP) (Quit: .)
- # [22:43] <firebot> Check-in: http://hg.mozilla.org/mozilla-central/rev/b0213baf2905 - Mark Finkle - Bug 710697 - LinkPreference causes a null preference to be sent to Gecko [r=mbrubeck a=android-only]
- # [22:43] <firebot> http://hg.mozilla.org/mozilla-central/rev/5ef118dcf1a1 - Matt Brubeck - Bug 710578 - Pages without contentDocument cause JSON exception in handleLocationChange [r=mfinkle a=javascript]
- # [22:43] <mfinkle> ehsan|sheriff, done
- # [22:43] <ehsan|sheriff> ty
- # [22:44] <ehsan|sheriff> mcsmurf: you can land whenever you're ready
- # [22:44] <mcsmurf> ok
- # [22:44] <Ms2ger> sicking, did you just claim docshell is not unfixable? :)
- # [22:44] <Cwiiis> philor, oh, really? :/ My patch doesn't touch anything outside of android though, that seems surprising to me...
- # [22:44] <khuey|quite-busy> it's fixable
- # [22:45] <khuey|quite-busy> it's just a question of how many years of bz's life you're willing to sacrifice at that alter
- # [22:45] <khuey|quite-busy> *altar
- # [22:45] <ehsan|sheriff> Cwiiis: iirc it broke every test in android opt
- # [22:45] <Cwiiis> ehsan|sheriff, ah, ok, that makes more sense
- # [22:45] <Cwiiis> damn, ok
- # [22:45] <ehsan|sheriff> khuey|quite-busy: would 3 do it?
- # [22:45] <Cwiiis> ehsan|sheriff, will sort out tomorrow
- # [22:45] <khuey|quite-busy> probably
- # [22:45] <ehsan|sheriff> ok
- # [22:45] <firebot> Check-in: http://hg.mozilla.org/mozilla-central/rev/7e74a12601a4 - Frank Wein - Bug 688615 - Wrong reg key used in Installer shell icon handler code, r=rob_strong, a=ehsan
- # [22:45] <mcsmurf> ehsan|sheriff: done
- # [22:45] * Cwiiis is now known as CwiiisAway
- # [22:45] <ehsan|sheriff> mcsmurf: thanks
- # [22:45] <ehsan|sheriff> with that, our queue is empty
- # [22:46] <@bz> docshell is totally fixable
- # [22:46] <@bz> with a sane spec (a big if)
- # [22:46] * juanb|afk is now known as juanb
- # [22:46] * Quits: jacek (jacek@moz-5D707D3B.psi.wroc.pl) (Client exited)
- # [22:46] <@bz> as in, a spec that actually matches reality
- # [22:47] <ehsan|sheriff> bz: how long would that take to write?
- # [22:47] * Joins: njn (chatzilla@moz-A6B647BB.dyn.iinet.net.au)
- # [22:48] <@bz> ehsan|sheriff: I don't know. :(
- # [22:48] * Quits: fzzzy (donovan@moz-BBE3ABD.mv.mozilla.com) (Quit: fzzzy)
- # [22:48] <ehsan|sheriff> sadface
- # [22:48] <nemo> bjacob: http://m8y.org/tmp/temp.txt - does this say anything interesting to you? that was on the site from https://bugzilla.mozilla.org/show_bug.cgi?id=708641 where Firefox was like 4 times slower than Chrome
- # [22:48] <Ms2ger> khuey|quite-busy, why are you on IRC? :)
- # [22:49] <nemo> Ms2ger: so you can see his status and not stop by his office?
- # [22:49] <bjacob> nemo: clearly you're spending most of the time in the JS engine here, so file a Core -> JS Engine bug
- # [22:49] <Ms2ger> Well, he's got an exam tomorrow
- # [22:49] * Quits: northWind (northWind@moz-20E58859.cable.teksavvy.com) (Quit: )
- # [22:49] * Joins: nhirata (nhirata.bu@moz-BBE3ABD.mv.mozilla.com)
- # [22:49] <nemo> bjacob: aight
- # [22:49] <ted> there are a lot of angry people on slashdot
- # [22:49] <ted> i think this is one of the reasons i stopped reading it
- # [22:50] <@bz> nemo: I was going to profile this...
- # [22:50] * Quits: Mook_as (mook@moz-1FCC0032.activestate.com) (Quit: ChatZilla 0.9.87-rdmsoft [XULRunner 1.9.2.13/20101203074205])
- # [22:50] <@bz> nemo: just hadn't gotten to it yet
- # [22:50] <nemo> bz: oh. um. do you have a bug already?
- # [22:50] <ted> like, even the ones that have a clue are so full of rage that you don't even care what their point is
- # [22:50] <nemo> bz: I have no idea how to easily install kernel symbols under ubuntu
- # [22:50] <@bz> nemo: well, 708641
- # [22:50] <@bz> nemo: as you pointed out
- # [22:50] <nemo> bz: ah. so. no point in making a 2nd bug?
- # [22:50] <nemo> should I attach that temp file there?
- # [22:50] <@bz> nemo: I don't think so
- # [22:50] <@bz> nemo: yeah
- # [22:50] * Quits: clee (clee@moz-BBE3ABD.mv.mozilla.com) (Quit: clee)
- # [22:50] <@bz> nemo: I'll do a profile in a few mins
- # [22:51] * @bz needs to rebuild his opt build
- # [22:51] <@bz> nemo: the other thing that might help is a regression range
- # [22:51] * Joins: clee (clee@moz-BBE3ABD.mv.mozilla.com)
- # [22:51] <ehsan|sheriff> ted: popular forums with open comments tend to attract angry people :/
- # [22:51] * Joins: jwq (jwq@moz-8238A1D6.irl.cri.nz)
- # [22:52] <@bz> witness bugzilla!
- # [22:52] * Quits: CwiiisAway (cwiiis@moz-F15E698.croy.cable.virginmedia.com) (Quit: Ex-Chat)
- # [22:52] <ted> true
- # [22:52] <nemo> bz: welp. now that my build works, I guess I could do that :)
- # [22:52] <nemo> hg bisect isn't too horribly tedious...
- # [22:52] * Quits: sewardj (sewardj@moz-933C5B3E.dip.t-dialin.net) (Ping timeout)
- # [22:53] <@bz> nemo: can't you just do it on nightlies?
- # [22:53] <ted> heh, i figured i looked like some kind of old-timer on /. with my ~100k UID
- # [22:53] <njn> Hacker News has a less angry vibe
- # [22:53] <ted> but BZ's is like 40k
- # [22:53] <@bz> nemo: I mean.... an hg bisect would be even better, but probably slower!
- # [22:53] <@bz> ted: yeah, I was young and in college... ;)
- # [22:53] * Quits: mjschranz (mjschranz@moz-57B23647.senecac.on.ca) (Client exited)
- # [22:53] <ted> me too
- # [22:53] <ted> heh
- # [22:53] <nemo> ted: mine is low 4 digit :) I put off registering for a while. seemed like a fad
- # [22:53] <ted> nice
- # [22:53] <ted> i used to have a ICQ UIN that was < 250k
- # [22:54] <ted> but i forgot the password and now nobody cares
- # [22:54] <ehsan|sheriff> fwiw, I'm very happy that we use irc which not everybody knows how to use
- # [22:54] * Joins: eflores (AndChat@538BABFE.A073F3E.97BBD552.IP)
- # [22:54] <@bz> yeah
- # [22:54] <nemo> ICQ UIN - I used to know that better than my phone number
- # [22:54] * @bz no longer recalls his ICQ login....
- # [22:54] <ehsan|sheriff> as opposed to, let's say, a web accessible chat group!
- # [22:54] <Ms2ger> nemo, better than your parents' phone number? ;)
- # [22:54] <@bz> ehsan|sheriff: there are irc webclients!
- # [22:54] <nemo> Ms2ger: I still don't know that one...
- # [22:54] <smaug> what...
- # [22:54] * Quits: klugefoo (kahr@moz-F88DFDF5.lightspeed.sntcca.sbcglobal.net) (Ping timeout)
- # [22:54] <smaug> I got first "update Nightly" dialog
- # [22:54] <@bz> ehsan|sheriff: we have one right on our video streaming page!
- # [22:55] <smaug> and shortly after that "Update Aurora"
- # [22:55] <smaug> I'm using an Aurora build
- # [22:55] <ehsan|sheriff> bz: I know, but people don't know about them :)
- # [22:55] <smaug> (this is actually Aurora 9)
- # [22:55] <hub> what's ICQ? ;-)
- # [22:55] <Ms2ger> Pah
- # [22:55] <Ms2ger> Even I know ICQ
- # [22:55] <ehsan|sheriff> hub: you're young!
- # [22:55] <hub> ehsan|sheriff: nah. Just ignored all that stuff for a while
- # [22:56] <mcsmurf> the cool kids use Facebook chat these days or what ;)
- # [22:56] <ehsan|sheriff> heh
- # [22:56] <nemo> bz: hm. WRT bisecting. I guess since 9 was the least-bad. I should start testing around the time 9 was branched off?
- # [22:56] <hub> mcsmurf: I must be old. I don't have Facebook
- # [22:56] * Quits: Goldorak (chatzilla@8109DDDE.7C10D095.187A1082.IP) (Quit: ChatZilla 0.9.87 [Firefox 8.0/20111104165243])
- # [22:57] * smaug has used Facebook chat once. That was enough.
- # [22:57] * Quits: jhk_ (jiggy@8E6C34C1.A3F9767A.1C37C358.IP) (Connection reset by peer)
- # [22:57] <mcsmurf> heh :P
- # [22:57] <ted> roc: so uh, if the VC2010 32-bit linker is 64-bit
- # [22:57] * Joins: jhk_ (jiggy@8E6C34C1.A3F9767A.1C37C358.IP)
- # [22:57] <ted> then we should really be making that change
- # [22:57] <roc> it's not
- # [22:57] <ted> oh
- # [22:57] <ted> er
- # [22:57] <ted> was that a mistake?
- # [22:57] <mcsmurf> I think FB chat is the reason some study claims "40% of online time spent on Facebook" (or something like that)
- # [22:58] * Joins: klugefoo (kahr@moz-F88DFDF5.lightspeed.sntcca.sbcglobal.net)
- # [22:58] <ted> oh, right
- # [22:58] <ted> that was the 64-bit linker
- # [22:58] <mbrubeck> ehsan|sheriff: Can I get into the landing queue?
- # [22:58] <ted> just installed in (x86)
- # [22:58] <ted> hah
- # [22:58] * Quits: davidb_ (davidb@moz-976CD2.dsl.bell.ca) (Quit: davidb_)
- # [22:58] <ehsan|sheriff> mbrubeck: since you're now the only person in it, you can just land :)
- # [22:58] <mbrubeck> thanks!
- # [22:58] <mbrubeck> updating my tree, this'll take just a minute
- # [22:59] <@bz> nemo: yeah
- # [22:59] <ehsan|sheriff> ted: yeah, their x64 toolchain doesn't know a thing about x86
- # [22:59] <@bz> nemo: that would make the most sense to me
- # [22:59] * Quits: wolfiR (wolfiR@moz-2858382B.dip0.t-ipconnect.de) (Quit: Leaving)
- # [22:59] <ted> ehsan|sheriff: it does have /MACHINE, but that doesn't seem to work with PGO anyway
- # [22:59] * Joins: northWind (northWind@2F50C7BC.10A7D831.6816E6B7.IP)
- # [22:59] <ted> probably because of it having to do the actual compiling
- # [22:59] <ted> it is kind of weird that they don't make a combined toolchain
- # [23:00] <ehsan|sheriff> ted: indeed
- # [23:00] <ted> i have to imagine most of the code is shared anyway
- # [23:00] <ehsan|sheriff> ted: I hear their compiler code base is horrible
- # [23:00] <ehsan|sheriff> so it's not exactly 64-bit safe ;)
- # [23:00] <ehsan|sheriff> ted: the code gen is probably not shared
- # [23:00] <ted> ah
- # [23:00] <ted> crazy
- # [23:00] <ehsan|sheriff> which is what link /ltcg mostly runs
- # [23:00] <ted> right
- # [23:00] <ted> insane
- # [23:01] <ted> although i guess they probably have code dating from VC 1.0 still in there
- # [23:01] <ted> :-P
- # [23:01] * Quits: Joeh (joe@5A3923AA.BC22908.C7CEC4ED.IP) (Ping timeout)
- # [23:01] <ted> vc++ 1.0 came out in 1993
- # [23:01] * Joins: kats (kats@moz-DE13FC16.compute-1.amazonaws.com)
- # [23:01] <ehsan|sheriff> as if we don't have code form the phoenix days around...
- # [23:01] <ted> i wonder if any code from microsoft C 1.0 is still in there
- # [23:01] <ehsan|sheriff> :P
- # [23:01] <ted> which came out in 1983
- # [23:01] * Joins: davidillsley (chatzilla@moz-D3C5EBB2.range86-144.btcentralplus.com)
- # [23:02] <firebot> Check-in: http://hg.mozilla.org/mozilla-central/rev/759a688577d5 - Matt Brubeck - Bug 708772 - (3/3) Use honeycomb theme and tablet layout on ICS tablets [r=mfinkle, a=ehsan]
- # [23:02] <kats> ehsan|sheriff: i have a patch for m-c; android java-only change
- # [23:02] <firebot> http://hg.mozilla.org/mozilla-central/rev/ae066b87a525 - Matt Brubeck - Bug 708772 - (2/3) Add tablet flag to jar manifest parser and nsSystemInfo [r=bsmedberg]
- # [23:02] <firebot> http://hg.mozilla.org/mozilla-central/rev/7cb2752fce12 - Matt Brubeck - Bug 708772 - (1/3) Add IsTablet method to AndroidBridge [r=dougt]
- # [23:02] * Quits: yuan (ywang@moz-C03D0C61.vlan426.asr1.sfo1.gblx.net) (Quit: yuan)
- # [23:02] <ehsan|sheriff> kats: land away :)
- # [23:02] <kats> ehsan|sheriff: thanks :)
- # [23:03] * Joins: davidb_ (davidb@moz-976CD2.dsl.bell.ca)
- # [23:03] <@bz> old code never dies
- # [23:03] <@bz> it just hides and causes bugs
- # [23:04] <Ms2ger> Hear, hear
- # [23:04] * Joins: tfair (tfairey@moz-BBE3ABD.mv.mozilla.com)
- # [23:04] * Joins: Honza (chatzilla@E07BF19C.5BB5597D.D0083327.IP)
- # [23:04] <catlee> woo! got a 32-bit firefox built on w64!
- # [23:04] <firebot> Check-in: http://hg.mozilla.org/mozilla-central/rev/41f75cbb91f2 - Kartikaya Gupta - Bug 710393 - Stop using Math.copySign which doesn't exist on Froyo. r=Cwiiis a=ehsan
- # [23:04] * Joins: yuan (ywang@moz-C03D0C61.vlan426.asr1.sfo1.gblx.net)
- # [23:04] <@bz> catlee: nice
- # [23:04] <ehsan|sheriff> cool!
- # [23:04] <catlee> ehsan|sheriff: do you have a m-c revision which was known to fail on w32 that I can try?
- # [23:04] <ehsan|sheriff> yes, just a sec
- # [23:05] <catlee> I used a2928a1
- # [23:05] <ehsan|sheriff> catlee: a64d8d8b512e
- # [23:05] * Joins: mjschranz (mjschranz@moz-57B23647.senecac.on.ca)
- # [23:05] <catlee> ehsan|sheriff: ok, I'll kick that off now
- # [23:06] <ehsan|sheriff> bjacob: there's a test failure sitting here staring at ,e
- # [23:06] <ehsan|sheriff> *me
- # [23:06] <ehsan|sheriff> catlee: thanks
- # [23:06] * Quits: davidb_ (davidb@moz-976CD2.dsl.bell.ca) (Quit: davidb_)
- # [23:06] <bjacob> ehsan|sheriff: coming soon
- # [23:06] * jmaher is now known as jmaher|afk
- # [23:06] <ted> oh good, then all we need to do is get a zillion more win64 slaves
- # [23:07] <ehsan|sheriff> ted: do we have a zillion 32-bit builders?
- # [23:07] * Joins: fzzzy (donovan@moz-2B41AF9B.lightspeed.mtvwca.sbcglobal.net)
- # [23:07] * Quits: joey (chatzilla@moz-781B227D.nys.biz.rr.com) (Connection reset by peer)
- # [23:07] <ted> i don't know offhand, but i'd guess yes
- # [23:08] <ted> maybe catlee knows
- # [23:08] <catlee> we have ~70 win32 builders
- # [23:08] <catlee> and ~30 win64
- # [23:08] <ted> only 70?
- # [23:08] <ted> wow
- # [23:08] <ted> i thought we had more
- # [23:08] <catlee> so we'd need to reimage the win32 machines with win64
- # [23:08] <catlee> or get more
- # [23:08] * Joins: joey (chatzilla@moz-EFCB4CBF.princetowncable.com)
- # [23:09] <ted> i guess if we were switching to doing all our builds on win64 then reimaging makes sense
- # [23:09] <catlee> I think it's like 70..
- # [23:09] <ted> because only old branches would need the 32-bit ones
- # [23:09] <catlee> yeah
- # [23:09] <catlee> better machine usage
- # [23:10] * Quits: Ami_Ty (Amie@moz-BBE3ABD.mv.mozilla.com) (Quit: Ami_Ty)
- # [23:10] <catlee> the next linux machines will be doing cross compiles for 32-bit too I think
- # [23:10] <@bz> hmm
- # [23:10] <@bz> so philor's link didn't show all the builders?
- # [23:10] <ehsan|sheriff> catlee: I'm assuming that the hardware on our 32-bit builders can handle x64...
- # [23:10] <catlee> yes
- # [23:11] <catlee> all the hw is identical
- # [23:11] * Quits: dRdR (dRdR@moz-57178BD3.hsd1.ca.comcast.net) (Client exited)
- # [23:11] <catlee> for our {linux,win}{32,64} machines
- # [23:11] * Joins: Mook_as (mook@moz-1FCC0032.activestate.com)
- # [23:11] * Joins: jrmuizel (jrmuizel@B31A279D.5BCEC6DB.DA78B690.IP)
- # [23:12] <ehsan|sheriff> good
- # [23:12] * Quits: mjschranz (mjschranz@moz-57B23647.senecac.on.ca) (Client exited)
- # [23:12] * Joins: pcwalton (pcwalton@moz-7B0110AD.mv.mozilla.com)
- # [23:12] <ted> at least that's something
- # [23:12] <catlee> xeon X3430 2.4GHz w/ 8GB ram
- # [23:13] * Quits: loadbang (loadbang@moz-2BA0C010.range86-168.btcentralplus.com) (Quit: loadbang)
- # [23:13] * Quits: bbondy (bbondy@moz-28CF6D1C.home.cgocable.net) (Ping timeout)
- # [23:13] * Quits: jet (junglecode@moz-BBE3ABD.mv.mozilla.com) (Quit: jet)
- # [23:13] * Joins: azakai (alon@moz-BBE3ABD.mv.mozilla.com)
- # [23:14] * Parts: kats (kats@moz-DE13FC16.compute-1.amazonaws.com)
- # [23:15] * Quits: @dbaron (dbaron@moz-BBE3ABD.mv.mozilla.com) (Quit: 8403864 bytes have been tenured, next gc will be global.)
- # [23:16] <mfinkle> ctalbert|afk, can we say bug Compute raw start is done?
- # [23:16] <mfinkle> err bug 695216
- # [23:16] * Quits: jhorak (jhorak@moz-59813FB4.cust.nbox.cz) (Quit: Leaving)
- # [23:17] <NeilAway> ted: at the very least, they probably still have the same code for /nologo ;-)
- # [23:17] * heycam|away is now known as heycam
- # [23:18] <philor> bz: ~70 as in 37 in the buildpool, and 29 in the trybuildpool, I think was what he meant
- # [23:18] <@bz> philor: ah, ok
- # [23:18] * Quits: asac (asac@moz-1C3692A7.pppoe.wtnet.de) (Quit: leaving)
- # [23:18] <@bz> philor: that's fine; we clearly care about trybuildpool
- # [23:19] <NeilAway> ted: wouldn't you want to keep some win32 machines for running tests?
- # [23:19] * Joins: mjschranz (mjschranz@moz-57B23647.senecac.on.ca)
- # [23:19] * Joins: darktrojan (geoff@moz-8B34E932.dsl.telstraclear.net)
- # [23:20] <catlee> tests run in a different pool
- # [23:20] * rhung|AFK is now known as rhung
- # [23:20] <catlee> minis!
- # [23:21] * joduinn-lunch is now known as joduinn-mtg
- # [23:21] * Quits: tH (Rob@cpc16-seac19-2-0-cust234.7-2.cable.virginmedia.com) (Ping timeout)
- # [23:21] * Quits: jrmuizel (jrmuizel@B31A279D.5BCEC6DB.DA78B690.IP) (Ping timeout)
- # [23:22] * Quits: mjschranz (mjschranz@moz-57B23647.senecac.on.ca) (Client exited)
- # [23:22] * Joins: tH (Rob@cpc16-seac19-2-0-cust234.7-2.cable.virginmedia.com)
- # [23:22] * Joins: KWierso_ (chatzilla@moz-4E7BE0E1.desm.qwest.net)
- # [23:23] * catlee is now known as catlee-mtg
- # [23:23] * Quits: Jesse (jruderman@moz-BBE3ABD.mv.mozilla.com) (Quit: Jesse)
- # [23:23] * Quits: KWierso (chatzilla@moz-4E7BE0E1.desm.qwest.net) (Ping timeout)
- # [23:23] * KWierso_ is now known as KWierso
- # [23:23] * Joins: ericjung (Mibbit@moz-87C31D97.bostonballet.org)
- # [23:25] * Joins: gwagner (idefix2@moz-C03D0C61.vlan426.asr1.sfo1.gblx.net)
- # [23:25] <philor> does pushing older branches to try still work, and would we be breaking it?
- # [23:26] * Quits: protz (jonathan@moz-7F6750F6.xulforum.org) (Quit: Leaving)
- # [23:26] * philor bets on no and yes
- # [23:26] * Joins: gmoro (guilherme@3DCBFB93.F520FC74.D159334F.IP)
- # [23:26] <darktrojan> worked for me a few weeks ago
- # [23:27] <@bz> how older?
- # [23:27] <@bz> pushing beta works
- # [23:27] <darktrojan> it's probably configured differently now, but it works
- # [23:27] <philor> yeah, one of my "older" definitions includes 1.9.2, but not very strongly
- # [23:27] <heycam> smaug++ for event constructors!
- # [23:27] <philor> since it barely even works on its own tree
- # [23:27] * Quits: mgoodwin (mgoodwin@moz-4BB9E5E9.cable.virginmedia.com) (Input/output error)
- # [23:28] <smaug> heycam: I need to get the tool to handle dictionary-like interfaces done
- # [23:28] <heycam> smaug, oh right you'd need that for those constructors
- # [23:28] <heycam> smaug, just done by hand atm?
- # [23:29] <smaug> ctors work
- # [23:29] <smaug> but right now everything is done by hand
- # [23:29] <smaug> heycam: but the tool will make value reading a lot easier
- # [23:29] * Quits: tH (Rob@cpc16-seac19-2-0-cust234.7-2.cable.virginmedia.com) (Ping timeout)
- # [23:29] <smaug> heycam: Bug 709569
- # [23:30] <@bz> heycam!
- # [23:30] <heycam> hello bz
- # [23:30] <@bz> heycam: so I filed this bug on you....
- # [23:30] <smaug> heycam: I don't know when we're getting real dictionary support, though in practice one can use interfaces
- # [23:30] * heycam figures he knows what bz will mention
- # [23:30] <heycam> yes I saw
- # [23:30] <mfinkle> ehsan|sheriff, back for another landing (android only)
- # [23:30] <heycam> but having just worken up I haven't looked at it yet
- # [23:30] * bwinton is now known as bwinton_away
- # [23:30] <@bz> heycam: I was 99% sure the test was just backwards, then I realized that would be totally wrong
- # [23:31] <@bz> heycam: ok
- # [23:31] <@bz> heycam: let me know if I can help somehow
- # [23:31] <heycam> bz, ok, I'll take a look this morning
- # [23:31] * @bz would really like his working tests back ;)
- # [23:31] * Joins: jrmuizel (jrmuizel@B31A279D.5BCEC6DB.DA78B690.IP)
- # [23:31] * Joins: Joeh (joe@5A3923AA.BC22908.C7CEC4ED.IP)
- # [23:32] * Joins: wg9s (bill@moz-7A06A043.hsd1.ma.comcast.net)
- # [23:32] <ehsan|sheriff> mfinkle: go ahead
- # [23:33] * Joins: sewardj (sewardj@moz-933C5B3E.dip.t-dialin.net)
- # [23:33] <mfinkle> ehsan|sheriff, ty
- # [23:34] <firebot> Check-in: http://hg.mozilla.org/mozilla-central/rev/83f4fe779fb1 - Mark Finkle - Bug 710783 - about:firefox won't load [r=mbrubeck a=android-only]
- # [23:34] * Quits: jwq (jwq@moz-8238A1D6.irl.cri.nz) (Ping timeout)
- # [23:35] <smaug> jst: did you ever figure out if we have some licenses for Zoom?
- # [23:35] * fabrice is now known as fabrice|afk
- # [23:35] <darktrojan> wow these images of mobile look hot
- # [23:35] * Quits: gerv (gerv@moz-8E68CF56.in-addr.arpa) (Quit: Leaving.)
- # [23:35] <smaug> (my evaluation license expired)
- # [23:35] * Joins: mayhemer (Miranda@B3D46202.F87A741B.F23860FD.IP)
- # [23:36] * armenzg is now known as armenzg_dinner
- # [23:36] <gavin> smaug: how do you like it (Zoom)?
- # [23:36] * Quits: mcsmurf (mcsmurf@moz-31272968.dip.t-dialin.net) (Quit: )
- # [23:36] <smaug> gavin: it was quite ok
- # [23:36] * jcranmer|away is now known as jcranmer
- # [23:37] <smaug> at least it was very easy to use, though I didn't quite get as much data as from Shark
- # [23:37] <smaug> but I'm not sure if I should have used some compiler flags to get better data
- # [23:37] * Joins: jwq (jwq@moz-8238A1D6.irl.cri.nz)
- # [23:38] * Quits: Sander (chatzilla@moz-B871F4D3.direct-adsl.nl) (Quit: And back he spurred like a madman, shrieking a curse to the sky.)
- # [23:39] * Quits: jamesr (jamesr@FB285460.D6CCE4AE.77834EAA.IP) (Ping timeout)
- # [23:39] * Quits: sicking (chatzilla@moz-1FF05400.hsd1.ca.comcast.net) (Ping timeout)
- # [23:39] <nemo> http://m8y.org/tmp/irc.txt - heh. (#css on freenode)
- # [23:39] * Quits: jrmuizel (jrmuizel@B31A279D.5BCEC6DB.DA78B690.IP) (Ping timeout)
- # [23:39] <ehsan|sheriff> bjacob: what was the bug #?
- # [23:40] * Quits: Honza (chatzilla@E07BF19C.5BB5597D.D0083327.IP) (Ping timeout)
- # [23:40] * Quits: MattN (MattN@2F83A011.9167AF22.2125B4DC.IP) (Ping timeout)
- # [23:40] <bjacob> ehsan|sheriff: 709914
- # [23:41] <Callek> Mossop, Unfocused: ping
- # [23:41] <Mossop> Callek: pong
- # [23:41] * Joins: Honza (chatzilla@E07BF19C.5BB5597D.D0083327.IP)
- # [23:42] <Callek> Mossop, Unfocused: Bug 709531 missing the re-adds of the tests from https://hg.mozilla.org/mozilla-central/rev/4d2921df8f34
- # [23:42] <gregglind> this might be a strange question... but is there an xpath-like (or other 'search') for looking through json?
- # [23:42] * Quits: armenzg_dinner (armenzg@moz-A6FE435.build.sjc1.mozilla.com) (Ping timeout)
- # [23:42] <Mossop> Yeah I couldn't be bothered to include them in the patch
- # [23:42] * Quits: lsumar (lsumar@538BABFE.A073F3E.97BBD552.IP) (Ping timeout)
- # [23:42] * Joins: sicking (chatzilla@moz-87306D4.dsl.snfc21.pacbell.net)
- # [23:42] * Joins: gandalf (zbraniecki@moz-9F56DED7.neoplus.adsl.tpnet.pl)
- # [23:42] <Callek> Mossop: so leaving them disabled??
- # [23:42] <Mossop> No
- # [23:42] <Ms2ger> bz, really? (re: doc.all)
- # [23:43] * Joins: jrmuizel (jrmuizel@D8C22410.5BCEC6DB.DA78B690.IP)
- # [23:43] <NeilAway> catlee: fair enough
- # [23:43] * Quits: bholley (bholley@moz-FCAF9AAB.hsd1.ca.comcast.net) (Quit: bholley)
- # [23:44] * Joins: lsumar (lsumar@E2F7352.56080C53.97BBD552.IP)
- # [23:44] <@bz> Ms2ger: what about it?
- # [23:44] <Ms2ger> item(string)
- # [23:44] * Joins: MattN (MattN@2F83A011.9167AF22.2125B4DC.IP)
- # [23:45] <@bz> Ms2ger: fwiw, WebKit and IE do that for HTMLCollection too
- # [23:45] <Ms2ger> ...
- # [23:45] * Quits: mconley (mconley@F2D29657.F60B0462.67AC9B1.IP) (Input/output error)
- # [23:45] <Callek> Mossop: as long as you are certain it won't be forgotten I guess I won't argue
- # [23:45] * Quits: clee (clee@moz-BBE3ABD.mv.mozilla.com) (Quit: clee)
- # [23:46] <blizzard> bz: did we decide that we shouldn't fix this? https://bugzilla.mozilla.org/show_bug.cgi?id=709448
- # [23:46] * Joins: tH (Rob@cpc16-seac19-2-0-cust234.7-2.cable.virginmedia.com)
- # [23:46] <blizzard> bz: based on coversations with jst?
- # [23:46] <Unfocused> Mossop: hah, just saw your patch. i was (slowly) going through all the tests and making the same changes
- # [23:46] * Joins: clee (clee@moz-BBE3ABD.mv.mozilla.com)
- # [23:46] * Joins: Josh_T (Meph@moz-E0BF24D4.dynamic.dsl.as9105.com)
- # [23:46] <@bz> blizzard: which?
- # [23:46] <@bz> Ms2ger: so what Gecko does for document.all....
- # [23:46] <Unfocused> and also doing a more extensive reset
- # [23:46] <ehsan|sheriff> bjacob: I think I figured out the problem
- # [23:46] <@bz> Ms2ger: is this:
- # [23:46] <Mossop> Unfocused: Sorry, had planned to get it up and ready before you woke up, but then meetings happened
- # [23:47] <@bz> if (id == sItem_id || id == sNamedItem_id) {
- # [23:47] <@bz> // Define the item() or namedItem() method.
- # [23:47] <@bz> JSFunction *fnc = ::JS_DefineFunctionById(cx, obj, id, CallToGetPropMapper,
- # [23:47] <@bz> 0, JSPROP_ENUMERATE);
- # [23:47] <Unfocused> np :)
- # [23:47] <@bz> Then CallToGetPropMapper does:
- # [23:47] <@bz> // Convert all types to string.
- # [23:47] <@bz> JSString *str = ::JS_ValueToString(cx, JS_ARGV(cx, vp)[0]);
- # [23:47] <@bz> // If we are called via document.all(id) instead of document.all.item(i) or
- # [23:47] <@bz> // another method, use the document.all callee object as self.
- # [23:47] <blizzard> bz: calling .click() from a mouseup
- # [23:47] <@bz> and converts the whole thing into a property get:
- # [23:47] <@bz> return ::JS_GetUCProperty(cx, self, chars, length, vp);
- # [23:47] <smaug> is bugzilla down?
- # [23:47] <@bz> blizzard: I have no idea
- # [23:47] <blizzard> bz: or maybe I'm thinking of sicking
- # [23:47] <bjacob> ehsan|sheriff: great
- # [23:48] <blizzard> sicking: ^^^
- # [23:48] <@bz> blizzard: we allow window.open() from mouseup
- # [23:48] <@bz> blizzard: I think you're thinking sicking
- # [23:48] * Quits: stevee (Miranda@moz-BEBDF855.cable.virginmedia.com) (Quit: Miranda IM - Multi protocol instant messenger @ www.miranda-im.org)
- # [23:48] <@bz> Ms2ger: so the point is, document.all is f-ed up
- # [23:48] <blizzard> bz: yeah, but we don't allow you to call .click() on a file control from mouseup
- # [23:48] <blizzard> bz: which was an issue for some of Google's properties
- # [23:48] <ehsan|sheriff> bjacob: so we have this extern variable in libvpx
- # [23:48] <Ms2ger> Heh
- # [23:48] <Unfocused> Mossop: i have changes on top of that... there are tests that reset prefs for urls, and they reset to pointing to AMO :\ (because automation.py.in uses user_pref() instead of changing the defaults)
- # [23:48] <nemo> bz: so. I decided to do bisect, just for the heck of it. (and masochism?) and I noticed something kind of weird
- # [23:48] <ehsan|sheriff> bjacob: and we try to read it in nsWebMReader.cpp
- # [23:48] <Ms2ger> bz, I think that has been established a while ago :)
- # [23:48] <ehsan|sheriff> that's gonna fail across modules!
- # [23:49] <sicking> blizzard: so i think we should change our code such that input.click() is allowed anywhere where window.open is
- # [23:49] <sicking> blizzard: apparently input.click() is a bit more restrictive currently
- # [23:49] <nemo> bz: I built the -g and the -b the same way, but the one around early Nov did the demo at 2000 sprites for 30fps while the latest did 300 sprites for 30fps
- # [23:49] <Mossop> Unfocused: Hrm really? Which URLs? If it's the AddonRepository URL then it doesn't matter so long as caching is disabled, but the others would be bad
- # [23:49] <sicking> blizzard: but i don't think we should change things beyond that based on the feedback from google
- # [23:49] <nemo> bz: these are debug builds, but... that's quite a difference
- # [23:49] <Mossop> Unfocused: Why don't you just post your patch instead then
- # [23:50] * Quits: Tobbi (Tobbi@16BAC97A.933EA279.AC7F8427.IP) (Ping timeout)
- # [23:50] * Quits: Josh_T (Meph@moz-E0BF24D4.dynamic.dsl.as9105.com) (Quit: ChatZilla 0.9.87 [Firefox 10.0a2/20111213042020])
- # [23:50] <blizzard> sicking: so we want to fix that bug or not?
- # [23:50] <Unfocused> Mossop: getAddons.get, etc is reset in a couple of places
- # [23:50] <blizzard> sicking: just trying to figure out if I should wontfix it
- # [23:50] <bjacob> ehsan|sheriff: hah
- # [23:50] <sicking> blizzard: i think we should fix the bug yes
- # [23:50] <Unfocused> will do, just finishing it off (and eating lunch)
- # [23:50] <blizzard> sicking: ok
- # [23:50] <bjacob> ehsan|sheriff: why'd we not get a linking error?
- # [23:50] <sicking> blizzard: which may or may not help google, i'm not 100% sure
- # [23:50] <Unfocused> my patch also re-enables the tests, fwiw
- # [23:50] * Quits: hipokrit (hipokrit@moz-502C3BF.rackspace.net) (Client exited)
- # [23:51] * Joins: dRdR (dRdR@moz-BBE3ABD.mv.mozilla.com)
- # [23:51] <blizzard> sicking: I know they fixed their problems another way (closure, I guess) but I was wondering if wanted to fix it anyway
- # [23:51] * davehunt|mtg is now known as davehunt
- # [23:52] <sicking> blizzard: right. I think we should fix the bug no matter what
- # [23:52] <sicking> blizzard: consistency is good
- # [23:52] <blizzard> sicking: ok, thx
- # [23:52] * blizzard keeps it on his list
- # [23:52] * Joins: Jesse (jruderman@moz-537BCF9.hsd1.ca.comcast.net)
- # [23:52] * Joins: IRCMonkey20419 (Tobbi@16BAC97A.933EA279.AC7F8427.IP)
- # [23:53] * Quits: Honza (chatzilla@E07BF19C.5BB5597D.D0083327.IP) (Ping timeout)
- # [23:53] <mounir> sicking: may I say I will have a look when you will be done with my reviews? :)
- # [23:53] <ehsan|sheriff> bjacob: do not know yet
- # [23:53] * gregglind is now known as gregglind_away
- # [23:53] * Quits: jhk_ (jiggy@8E6C34C1.A3F9767A.1C37C358.IP) (Ping timeout)
- # [23:54] * Quits: gwagner (idefix2@moz-C03D0C61.vlan426.asr1.sfo1.gblx.net) (Quit: gwagner)
- # [23:54] <blizzard> heh
- # [23:54] <sicking> mounir: sold!
- # [23:54] <blizzard> mounir: :D
- # [23:54] * Quits: tH (Rob@cpc16-seac19-2-0-cust234.7-2.cable.virginmedia.com) (Ping timeout)
- # [23:55] * Joins: tH (Rob@cpc16-seac19-2-0-cust234.7-2.cable.virginmedia.com)
- # [23:55] <Mossop> I love having to reboot windows in order to delete files from my objdir
- # [23:56] * Quits: graememcc (chatzilla@moz-40DE3BF7.range86-148.btcentralplus.com) (Ping timeout)
- # [23:56] * Quits: jhammink (textual@moz-7B0110AD.mv.mozilla.com) (Ping timeout)
- # [23:56] <khuey|quite-busy> "ur doin it wrong"
- # [23:56] * Quits: Mossop (mossop@moz-C03D0C61.vlan426.asr1.sfo1.gblx.net) (Quit: Fleeing the scene)
- # [23:56] <@bz> nemo: don't do debug builds
- # [23:56] <rs> Mossop: check for anything in Mozilla build that could keep them in use'
- # [23:56] <@bz> nemo: measuring perf in debug builds == pointless
- # [23:56] <@bz> nemo: please test opt builds!
- # [23:56] <@bz> nemo: fwiw, I just commented in the bug
- # [23:57] <Ms2ger> bz, even better, HTMLCollection.item(string) in chrome can return a nodelist :/
- # [23:57] * Joins: jamesr (jamesr@FB285460.D6CCE4AE.77834EAA.IP)
- # [23:57] * Joins: jhammink (textual@moz-7B0110AD.mv.mozilla.com)
- # [23:57] * Joins: bbondy (bbondy@moz-28CF6D1C.home.cgocable.net)
- # [23:57] * Joins: graememcc (chatzilla@moz-40DE3BF7.range86-148.btcentralplus.com)
- # [23:57] <nemo> bz: ack. bjacob didn't mention that :-/
- # [23:57] <nemo> dammit
- # [23:57] <@bz> nemo: um... it should be obvious
- # [23:57] <Callek> mossop: few things that could prevent that, a hung process, a running dll, a command shell open in the objectdir, a pop-up window hidden behind everything else (like a MS crash dialog)
- # [23:57] <@bz> nemo: since debug builds do extra checks that opt builds don't, and you don't care how fast those are
- # [23:57] <@bz> ms2ger: yep
- # [23:57] <nemo> bz: not to me. in our simple apps, bad code is still slow even unoptimised
- # [23:57] <@bz> ms2ger: it's all fucked up
- # [23:57] <Callek> mossop: also (possibly) the new windows-service-updater thing iirc
- # [23:58] <nemo> well. relatively simple
- # [23:58] <@bz> nemo: well, yes. But good code can get slow when unoptimized
- # [23:58] * Joins: Mossop (mossop@moz-C03D0C61.vlan426.asr1.sfo1.gblx.net)
- # [23:58] <rs> Callek: it hasn't even landed
- # [23:58] <Callek> rs: ooo I thought elm merged in already
- # [23:58] <@bz> nemo: keep in mind that debug is not just lack of -O, it's also addition of assertions
- # [23:58] <@bz> nemo: various js and layout code has different big-O behavior in debug vs opt, say
- # [23:58] <nemo> m'k...
- # [23:58] <nemo> welp. n/m then
- # [23:58] <@bz> nemo: because it might be doing an assert that takes O(N) time to compute on every operation....
- # [23:58] * Quits: @mkaply (chatzilla@moz-92EDDD02.lightspeed.austtx.sbcglobal.net) (Ping timeout)
- # [23:59] <nemo> well, you appear to have it under control and I was doing it all wrong anyway... *resets the bisect*
- # Session Close: Thu Dec 15 00:00:00 2011
The end :)