[Webkit-unassigned] [Bug 208482] New: SVG and Canvas content should factor into visually-non-empty heuristic
bugzilla-daemon at webkit.org
bugzilla-daemon at webkit.org
Mon Mar 2 15:48:54 PST 2020
https://bugs.webkit.org/show_bug.cgi?id=208482
Bug ID: 208482
Summary: SVG and Canvas content should factor into
visually-non-empty heuristic
Product: WebKit
Version: WebKit Local Build
Hardware: Unspecified
OS: Unspecified
Status: NEW
Severity: Normal
Priority: P2
Component: Layout and Rendering
Assignee: webkit-unassigned at lists.webkit.org
Reporter: nham at apple.com
CC: bfulgham at webkit.org, simon.fraser at apple.com,
zalan at apple.com
Currently, we block the initial render on our "visually non-empty" heuristic. This heuristic depends on (among other things) how much text is in the document, how many image pixels are in the document, and whether or not the main resource is fully parsed. However, SVG and canvas content is not considered by the heuristic.
This causes a delayed first contentful paint for certain websites. For instance, YouTube's first contentful paint is often delayed until far after the main HTML resource is parsed:
1. They don't ship down any text initially, so we won't trip over the visually non-empty character threshold
2. They don't ship down any actual images initially, so we won't trip over the filled-image-pixel threshold (everything that looks like an image is actually an SVG)
3. On an uncached load, we won't paint after parsing the initial HTML either, since there's usually an outstanding web font loading, and our heuristic delays the first paint until the web fonts load if there isn't enough character or image content
We could improve our FCP for YT and other similar sites by including an estimate of SVG and canvas content sizes in the visually-non-empty heuristic. This will also improve FCP as reported by https://w3c.github.io/paint-timing/#first-contentful-paint.
--
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20200302/2a37c691/attachment.htm>
More information about the webkit-unassigned
mailing list