[Webkit-unassigned] [Bug 204575] New: Make CSSPropertyParserHelpers::consumeColor safe to use within a worker

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Mon Nov 25 06:35:42 PST 2019


https://bugs.webkit.org/show_bug.cgi?id=204575

            Bug ID: 204575
           Summary: Make CSSPropertyParserHelpers::consumeColor safe to
                    use within a worker
           Product: WebKit
           Version: WebKit Nightly Build
          Hardware: Unspecified
                OS: Unspecified
            Status: NEW
          Severity: Normal
          Priority: P2
         Component: CSS
          Assignee: webkit-unassigned at lists.webkit.org
          Reporter: clord at igalia.com
            Blocks: 183720

After bug 182686, CSSParser::parseColorWorkerSafe was introduced, which uses CSSParserFastPaths:parseColor with a given CSSValuePool that is safe to use within the worker and thus allows CSS colour parsing within a worker.

parseColorWorkerSafe skips using the full CSS parser, however, as for that to be safe within a worker requires a more significant change. This bug tracks making full colour parsing work within workers.

My current WIP patch for this splits the functions in CSSParserFastPaths into worker-safe variants and the standard, non-worker-safe variants. This is done based on my inspection and some small changes - for the most (entire?) part, what makes colour-parsing not safe within a worker is using CSSValuePool::singleton() (the result of which should only be used on the main thread), so adding a CSSValuePool& parameter makes them safe to use within workers, assuming the given CSSValuePool was created on the same thread. Whether my analysis is accurate is of course within question.


Referenced Bugs:

https://bugs.webkit.org/show_bug.cgi?id=183720
[Bug 183720] Complete OffscreenCanvas implementation
-- 
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/20191125/30a851ed/attachment.htm>


More information about the webkit-unassigned mailing list