[Webkit-unassigned] [Bug 49799] createImageBuffer fail should be handled in the same way as other fails

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=49799





--- Comment #11 from Dirk Schulze <krit at webkit.org>  2010-11-19 07:55:12 PST ---
(In reply to comment #10)
> (In reply to comment #9)
> > (In reply to comment #7)
> > > (From update of attachment 74375 [details] [details] [details])
> > > View in context: https://bugs.webkit.org/attachment.cgi?id=74375&action=review
> > > 
> > > Maybe I  misunderstood the intention of the after reading the text in the ChangeLog.
> > > 
> > > > LayoutTests/svg/filters/filter-x.svg:7
> > > > +<filter id="filt" filterUnits="userSpaceOnUse" x="9.51" y="5" width="10" height="10">
> > > > +    <feFlood flood-color="red" />
> > > > +</filter>
> > > > +</defs>
> > > > +<rect x="0" y="0" width="10" height="10" filter="url(#filt)"/>
> > > 
> > > Why does it test the case, where width or height are < 0.5 like in the bug description? Just x has a fractional part. How is it related to the described problem?
> > Talked with Niko about the image size. Sorry the test _is_ correct for the check of the correct behavior _inside_ of SVGFilter.Nevertheless, a red reference image is not a good idea.
> > 
> > Not related to this bug: We should create a ImageBuffer for 0.49 sized images. This is why I was talking about enclosingIntRect/size.
> Really, do you think there's any gain, filtering a 0.49px image buffer? :-)
> Anyhow, however wants to work on it, be careful with the lroundfs, they're there on purpose.
> 
> Suppose you want a 199.51 px image buffer, we'd now allocate a 199px image buffer, and _scale the context_ accordingly by 199/199.51, so it perfectly fits into the buffer. Otherwhise you'll get artefacts on zooming.

But is it really a big problem (performance and allocation) to make it 1px larger and scale it up instead, but have a correct result, always?

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