=== yofel_ is now known as yofel
[20:51] <bdrung> micahg: are you involved in the firefox package?
[20:51] <micahg> bdrung: yes
[20:52] <bdrung> micahg: i tried the ff 4.0 daily from the ppa and webm didn't work.
[20:52] <micahg> bdrung: hmmm, well chrisccoulson has been doing most of the maintenance for 4.0, but I can take a look a little later
[20:53] <bdrung> micahg: is webm expected to work?
[20:53] <micahg> bdrung: I would think so
[20:53] <chrisccoulson> it works here....
[20:54] <bdrung> chrisccoulson: with which site did you test it?
[20:54] <chrisccoulson> http://devfiles.myopera.com/articles/1891/custom-controls-webm-360p.html
[20:57] <bdrung> chrisccoulson: that works. i tried youtube html5. maybe it's a youtube issue
[20:59] <chrisccoulson> have you got a link which doesn't work?
[20:59] <micahg> bdrung: Youtube HTML5 was H.264 last time I checked, did they change that?
[21:00] <bdrung> micahg: they provide both (some h.264 and some webm)
[21:00] <micahg> bdrung: ah, ok
[21:01] <bdrung> chrisccoulson: http://www.youtube.com/watch?v=8QSgNM9yNjo - you have to enable html5 first
[21:01] <bdrung> then you should see "html5 webm"
[21:02] <bdrung> chromium uses webm - on ff 4.0 i see only the flash player
[21:15] <chrisccoulson> bdrung, it's just a user agent issue, it works fine in the beta 6 release build
[21:16] <chrisccoulson> micahg - when did you get home?
[21:16] <bdrung> chrisccoulson: how can i adjust the user agent?
[21:16] <chrisccoulson> i didn't see you on friday evening ;)
[21:16] <micahg> chrisccoulson: about an hour ago
[21:16] <micahg> chrisccoulson: yeah, I went somewhere to spend the Sababth
[21:17] <micahg> *Sabbath
[21:17] <chrisccoulson> bdrung - you can re-enable the old format in about:config
[21:17] <chrisccoulson> general.useragent.compatMode.firefox
[21:18] <bdrung> thx. where can i see which user agent is used? the about window doesn't show the user agent any more
[21:18] <chrisccoulson> bdrung - http://whatsmyuseragent.com/
[21:19] <chrisccoulson> youtube just needs to fix their check, which i suspect will happen once b7 is released, and people start complaining about it
[21:21] <bdrung> you can't get the user agent directly from the browser?
[21:22] <fta> chrisccoulson, micahg: you have prism/natty/i386 stuck for days on a builder
[21:24] <chrisccoulson> fta - i guess that's the same gcc issue that affects all mozilla i386 builds atm
[21:24] <micahg> fta: can you stop the natty dailies for the moment?
[21:24] <fta> chrisccoulson, yep, xul stuck in malloc()
[21:24] <micahg> I'll ping someone for the failed build, thanks for the heads up
[21:24] <fta> easy to reproduce locally
[21:25] <chrisccoulson> fta - yeah, that's bug 663294
[21:25] <ubot2> Launchpad bug 663294 in gcc-4.5 (Ubuntu) "Firefox built with gcc-4.5 is a non-starter on i386 (affects: 2) (heat: 16)" [High,Confirmed] https://launchpad.net/bugs/663294
[21:25] <fta> yep, read it already
[21:26]  * micahg added to topic already
[21:26] <fta> i have some issues with chromium too
[21:26] <fta> but it's tougher to identify the issue
[21:27] <chrisccoulson> i talked to kees at UDS last week, we're going to switch off pie for the time being
[21:27] <chrisccoulson> once i've checked it with gcc-snapshot, i'll send the bug upstream to gcc
[21:28] <fta> for ch, it's not pie, nor the rest of the hardening, i tried
[21:28] <fta> could be a gcc 4.5 bug though but i doubt it and a gcc 4.4 ch has the bug on natty
[21:29] <fta> -and+as
[21:29] <chrisccoulson> that's not good. what's the actual issue?
[21:29] <fta> http://code.google.com/p/chromium/issues/detail?id=61311
[21:29] <chrisccoulson> i guess i should kick off a build of firefox with gcc-snapshot, whilst i'm not doing anything else
[21:29] <fta> ch already has this: http://codereview.chromium.org/download/issue4116008_1.diff
[21:31] <fta> that fixed a regression for searches in the omnibar
[21:31] <chrisccoulson> interesting
[21:31] <chrisccoulson> what does -fno-ipa-cp do? it's not in the GCC manual?
[21:32] <chrisccoulson> oops, the last sentence actually wasn't a question
[21:32]  * chrisccoulson needs more sleep
[21:32] <fta>        -fipa-cp
[21:32] <fta>            Perform interprocedural constant propagation.  This optimization
[21:32] <fta>            analyzes the program to determine when values passed to functions
[21:32] <fta>            are constants and then optimizes accordingly.  This optimization
[21:32] <fta>            can substantially increase performance if the application has
[21:32] <fta>            constants passed to functions.  This flag is enabled by default at
[21:32] <fta>            -O2, -Os and -O3.
[21:34] <fta> we should have a wiki page or something for the regressions in natty
[21:34] <fta> i'm sure there are plenty
[21:34] <fta> i have another on in emacs23 with ucb
[21:34] <fta> onE
[21:35] <chrisccoulson> yeah, and based on what i've seen in firefox, i bet there's a whole load of regressions we haven't found yet
[21:35] <chrisccoulson> i'm sure once people start upgrading to natty in larger numbers, more of these issues will appear
[21:35] <chrisccoulson> scary ;)
[21:35] <chrisccoulson> i'll probably upgrade tomorrow
[21:35] <micahg> chrisccoulson: fun :), we have less than 6 months to find them :), it'll probably go fasterc-4.5 once upstream switches to gc
[21:36] <chrisccoulson> heh :)
[21:36] <fta> i doubt they will
[21:43] <fta> funny, the chromium translations are not open yet, but i already received almost 700 strings