Bug 113667
Summary: | Implement SVG2's buffered-rendering property for paths | ||
---|---|---|---|
Product: | WebKit | Reporter: | Philip Rogers <pdr> |
Component: | SVG | Assignee: | Philip Rogers <pdr> |
Status: | NEW | ||
Severity: | Normal | CC: | krit, syoichi, zimmermann |
Priority: | P2 | ||
Version: | 528+ (Nightly build) | ||
Hardware: | Unspecified | ||
OS: | Unspecified | ||
Bug Depends on: | 104207 | ||
Bug Blocks: |
Philip Rogers
Paths can benefit more from buffered-rendering than the other shape primitives (e.g., rect). Lets implement it!
Spec: https://svgwg.org/svg2-draft/single-page.html#painting-BufferedRendering
Attachments | ||
---|---|---|
Add attachment proposed patch, testcase, etc. |
Philip Rogers
Syoichi, I saw that you cc'ed yourself on this. Are you planning on using this feature? I'd love to get feedback on what users are looking to use it for.
Dirk Schulze
(In reply to comment #0)
> Paths can benefit more from buffered-rendering than the other shape primitives (e.g., rect). Lets implement it!
>
> Spec: https://svgwg.org/svg2-draft/single-page.html#painting-BufferedRendering
Lets hold back for a minute and see if we can optimize it in another way before exposing this property widely. What is the main problem on path and how can we fix it without relying on the property. It should be easier than the image problem.
Philip Rogers
(In reply to comment #2)
> (In reply to comment #0)
> > Paths can benefit more from buffered-rendering than the other shape primitives (e.g., rect). Lets implement it!
> >
> > Spec: https://svgwg.org/svg2-draft/single-page.html#painting-BufferedRendering
>
> Lets hold back for a minute and see if we can optimize it in another way before exposing this property widely. What is the main problem on path and how can we fix it without relying on the property. It should be easier than the image problem.
Consider a complex path such as the animating svg cougar (philbit.com/animatingcougar.svg) or an application that uses very complex paths for text (like Google Docs' presentation tool). In these cases I can see an advantage to not re-rasterizing the path every time.
This isn't a fundamental problem with paths, it's just a graphics reality: sometimes it's best to cache the rasterized result.
I'm happy to wait on this. Unlike image, I don't know of anyone asking for this feature at the moment.