[webkit-qt] Minutes from the Status Meeting in #qtwebkit on Freenode IRC network

[Jocelyn Turcotte]

well :)

summary from most of what I remember:
1. For the scene graph integration.
The problem: We need a way to render inside the scene graph in a sustainable way.
Conclusion: The graphic guys know that rendering directly using OpenGL in the scene graph would be the best for us. Having a custom scene graph node that calls us when it's time to take over the GL context could however prevent them from doing optimizations in the future. So they will give it a try and get an implementation to have a better idea.

2. For the html5 canvas:
The problem: The Qt port is lagging behind other browsers which use the graphic card to blit images on the 2D canvas buffer.
Conclusion: The possibilities include:
- putting man power to morph the OpenGL paint engine to do all canvas rendering. Requires a lot of effort.
- switching to skia, which already support this, for all WebProcess rendering. Requires a lot of effort.
- waiting to see what will happen with WebGL, which might take over 2D canvas for most serious web apps out there, and focus our efforts on it instead.
For now we have plenty on our plate, but it would be worth keeping our eyes open for possible optimizations in this area if we judge them worthy.

Jocelyn


On Wed, 14 Dec 2011 17:01:02 +0100
ext Simon Hausmann <simon.hausmann at nokia.com> wrote:

> 
> /me: patch reviews, a bit more QML/V8 work, looong discussion with jocelyn, 
> gunnar, kim and samuel about scene graph integration and html5 canvas
> (I should post a summary of that separately, unless someone beats me to it ;)
> 
> 
> Simon
> 
> On Wednesday, December 14, 2011 06:30:18 AM ext Qt WebKit StatusBot wrote:
> > Updates:
> >   * jprvita status: gardening: looking for skipped layout tests to fix
> >   * elproxy status: started looking at fonts again, but might have to fiddle
> > around the mobile theme and squircles a bit longer to please kenneth... *
> > bbandix status: still working on the scaling code: resizing the page item
> > changes the coodinate system compared to the webcore page, and we end up
> > translating event positions on both ends of the process boundary, trying to
> > find a nice way to avoid back and forth conversions if possible * ahf
> > status Cleaning up. Waiting for EWS on
> > https://bugs.webkit.org/show_bug.cgi?id=74499 * kenneth status: looking
> > into testing systems for the viewport
> >   * Ossy status: gardening a little, buildfixes, review small patches,
> > experimenting kbalazs's testfont patch -> zillion test will be enabled on
> > Qt5 asap * laknudse status trying to get an overview of AC issues
> >   * jeez_ status: API work pending review, will spend half of the day with
> > interviews related stuff * darktears status : work on <ol> to upload a new
> > version of the patch * Smith status: debug my plugin patch, why dosn't work
> > it
> >   * mibrunin status Experimentending the WebPreferences, looking at touch
> > testing * kbalazs status: plugin process nTh round, gardening
> >   * azbest_hu status: try to run LayoutTests on N9; testing kbalazs'
> > testfont patch (bug 72513) on amazon ec2; fine tuning rsync scripts for N9
> > * Zoltan status: profiling memory with Valgrind in MiniBrowser on N9 *
> > zherczeg status: css tokenizer
> >   * tczene status: implement opencl to svg
> >   * cmarcelo status: removing QtWebPageProxy.
> >   * lmoura status: checking touch issues with MiniBrowser on N950
> >   * torarne status: first aid course, learned how to do cpr and shout
> > "clear" before shocking the patient * luck status: managed to get qtwebkit
> > compiling inside qt5 tree using old scripts. Now doing proper testing and
> > git scripts.
> > 
> > Missing updates from: TwistO, andris88, hugopl, igoroliveira, jturcotte,
> > kenneth_, kkristof, kling, loki04, mulvad, rafaelbrandao, reni, setanta,
> > tronical, zalan, zalbisser _______________________________________________
> > webkit-qt mailing list
> > webkit-qt at lists.webkit.org
> > http://lists.webkit.org/mailman/listinfo.cgi/webkit-qt
> _______________________________________________
> webkit-qt mailing list
> webkit-qt at lists.webkit.org
> http://lists.webkit.org/mailman/listinfo.cgi/webkit-qt