[Webkit-unassigned] [Bug 27561] ImageDecoder enhancements for WINCE port

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=27561





--- Comment #42 from Adam Treat <treat at kde.org>  2009-08-14 13:19:04 PDT ---
(In reply to comment #41)
> > My object is to reduce the complexity and special-casing of shared code,
> > especially the decoders themselves.  Let's say platform X supports native image
> > storage in various formats (565, 888, 8888, etc.).  The pattern that makes
> > sense to me is:
> > 
> > Image Decoder reads the pixel values out of the source image and calls setRGBA
> > with them
> > ImageFrameSink/RGBA32Buffer/whatever passes these values along to the backend's
> > underlying NativeImage functions
> > These functions decide how to downsample to an appropriate format
> > 
> > With your existing code, the downsampling would need to be done in the decoder,
> > which would need to check whether the image frame claims to be 16-bit.  All
> > that seems to do is cause us to plumb more state up through shared code and
> > push the complexity on everyone.
> 
> I think you make a good point here.

You both are assuming that the decoder can stream pixels directly into the
native image buffer.

That simply isn't the case for some of the existing users of the built-in image
decoders.  This was covered I think by Holger.  It might not even be the case
that the native image buffer is in main memory.

At this point, I think the discussion should move away from WinCE's
ImageFrameSink and more towards the path we want to take in the future.  I
still do not believe we are all on the same page.  Perhaps a conference call or
two with all interested parties would be in order.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.