This fixes the MOZ_TRACE_* crap in Console.app (it just disables it entirely except for debug builds). Also, we did try aligning JaegerStubVeneer to 16-bytes (for the interested, it's .align 4) and it makes things worse on G4 and G5, and I don't understand these processors sometimes. 12 will come out sometime after Firefox 12b3. Once again, release notes and versions (this version is still 11.0, but has a later build ID):
So I bring up 11 r2 & the page greeting says:
ReplyDeleteYou're using the unstable version and should be a smart enough chap/dame/whatsit to check if you're up to date yourself.
I am Shiva destroyer of worlds!(my comment)
So you *are* smart enough! :P
DeleteMy bad..less unstable than the previous unstable version.
ReplyDeleteSmart enough to test launching before releasing!
ReplyDeleteDon't make me comment again...I'm getting a headache from the kepthas...they just keep getting harder
ReplyDelete[rant]
ReplyDeleteToo sad, the 3D inspector requires WebGL… On the other hand, I wonder why a tool that 99% of people will never use must be in the browser core. The same functionality has been in the Web Developer Toolbar add-on for years. Do Mozilla folks ask themselves "How can we make it use even more RAM" all day? They have time to unnecessarily duplicate functionality, but they don't have time to support PPC…
[/rant]
I'm really glad you keep cranking out these builds--but I've been wondering lately why the latest Camino browser *feels* so much faster than the TFF browser?
ReplyDeleteThe 'numbers' indicate otherwise ( http://peacekeeper.futuremark.com/results?key=53NW&resultId=2227194) but on something like Google Reader especially, TFF seems to "bog-down" when I'm 'nexting' through articles.
TFF has a bunch of nifty addons that I use--I suppose that could account for some slowing? I'm also willing to concede that there might be 'placebo' effects going on here.
Any thoughts?
This comment has been removed by the author.
ReplyDeleteThis comment has been removed by the author.
ReplyDeleteCamino may feel lighter due to the pure Cocoa API and built-in adblock but it will never have the lightning fast modern JIT that TenFourFox has or Mozilla's legendary extensibility.
ReplyDeleteAs commenter pointed out, the UI in Camino is written in Objective-C and leverages native Cocoa, whereas Firefox and TenFourFox simulate the native interface in XUL (and therefore must do extra work). It's no longer possible to do otherwise, in fact: Camino does its magic by embedding Gecko as its page layout system wrapped in Cocoa widgets, but embedding is no longer supported as of Mozilla 2.0 which is why Camino won't be Gecko after Camino 2.1 (Mozilla 1.9.2).
ReplyDeletePlus, the addons probably do have some role to play. Firebug, for example, has been implicated as part of the perennial complaints of Mozilla's memory overuse. I only use the QTE, AdBlock and OverbiteFF personally.
This might become interesting: H.264 support for html5 video.
ReplyDeletehttp://www.h-online.com/open/news/item/Mozilla-looks-at-supporting-H-264-video-again-1471283.html
so this update runs slower than previous versions on G4 and G5's?
ReplyDeleteThe code above was a failed experiment, so it was removed prior to final build.
ReplyDeleteNot sure what Mozilla is going to do with H.264. Hope they implement it using QT, though, if they decide to bite the bullet. I'm not sure what this means for us with licensing and that's the part I'm most concerned about.
If that comes to pass someone really must build TenFiveFox!
ReplyDeleteWhat are your concerns? Isn't playing H.264 content free (unlike encoding, which requires a payed license)? It looks like WebM is not going to make it, and H.264 support is only a matter of time. Mozilla doesn't seem to intend to ship an H.264 decoder – they want to use the one that's already installed in the OS.
ReplyDeletehttp://hacks.mozilla.org/2012/03/video-mobile-and-the-open-web/
http://blog.lizardwrangler.com/2012/03/18/video-user-experience-and-our-mission/
This comment has been removed by the author.
ReplyDeleteWell, that's just it. I don't see anywhere that says that decoders are free, and the MPEG-LA only has a grace period for a limited time.
ReplyDeleteI am hoping they do just as you say, i.e., use the system codecs, because then we just piggyback on QuickTime (assuming they are not using some API only in QT X). I would not want to distribute ffmpeg or something if we can possibly avoid it, but Mozilla may have to do something like that for Win XP.
@commenter, the major problem is the cost to run the hardware. I could probably scare up a G5 to act as a public buildbot, but at 200W, 24/7/365, plus whatever heat it adds to the A/C system in my server room, it's not cheap. The bill for the T1 and electricity to run the POWER6, the G4 file server, my personal quad and a couple ancillary servers is around $500-$600/mo currently.
ReplyDeleteDo the VLC people pay for H.264 decoding? Or anyone who uses ffmpeg?
ReplyDeleteIt's not so much "do they" as it is "should they" and "if they don't, can they be sued."
ReplyDeleteI doubt if we did that the MPEG-LA would go after a non-profit like us, but I really don't want to find out, because I would probably be personally liable if they could. I'd really like Mozilla to either negotiate something to protect their downstreams, or to use system codecs. If they required the use of ffmpeg, I'd probably sit it out if there were no clear guidance from their legal counsel. Just defending such a suit could be ruinous.
Oh, I see.
ReplyDeleteShooting from the hip here, but is it possible MPEG-LA has a non-profit clause?
ReplyDeleteMPEG-LA definitely does have a non-profit clause for CONTENT PROVIDERS who DISTRIBUTE files with H.264 content.
DeleteThey have no such clause for SOFTWARE VENDORS who SUPPORT PLAYBACK of such files.
For software vendors who want to decode H.264, the strategy I've most often seen recommended, is to look for ways to use a concept known as "patent exhaustion": If some other software already present on your computer system has come with a paid license for H.264 video decoding, and it also provides an API which allows your own software to invoke that other software's decoder, than that other software's paid H.264 license will most likely be considered to extend to cover your own software too.
Since every Mac OS X computer running QuickTime 7 (or above) automatically contains such a paid H.264 license, and also contains such an API for offloading video decoding, patent exhaustion would almost certainly apply for any software which actually relied on offloading to QT.
This comment has been removed by the author.
Delete=) Thanks a lot: FF11.0r2 still runs fine on my G4mdd1250/OSX.5.8 (about:config > tenfourfox.plugins.enabled).
ReplyDelete