-
Notifications
You must be signed in to change notification settings - Fork 27
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[offscreen-canvas] Remove 2d.path.stroke.prune.* tests for open paths from interop-2023 #392
Comments
Seems ok |
See web-platform-tests/interop#392 for details. These tests relate to a longstanding open spec issue, and should not be included in interop-2023 as long as that is unresolved.
@fserb can you review for Chromium? |
lgtm |
See web-platform-tests/interop#392 for details. These tests relate to a longstanding open spec issue, and should not be included in interop-2023 as long as that is unresolved.
This misses the 4 following tests. Could we remove them from interop as well? |
Is there ambiguity about how those should behave? The others were about unclosed paths, where there may potentially be line-caps involved, but that's not an issue for closed paths. |
Hi, I am very new to the Path spec. Please help me to understand the spec for closed curves. In the spec, https://html.spec.whatwg.org/multipage/canvas.html#trace-a-path, item 5 says
since the start and the end is the same point, it basically added a subpath of length 0. Will the line caps be drawn in this case? |
I don't think it would ever make sense to draw line-caps for a closed subpath (whatever its length). Caps go on open ends of a path. A closed subpath doesn't have open ends, so it doesn't get caps. |
Caps go on each lines of the dash strokes, so you can have these even on closed pathes, e.g. when the dash lines are smaller than the path's total length. What happens for zero length closed pathes is relatively unclear when you don't prune them at the beginning. Note that SVG does render the caps in this case. |
Yes, caps would go on dash strokes. But intuitively, it seems to me that if the path length is zero, then there would be no instances of the dash, and hence no caps either. |
I'm not saying your reasoning isn't sound, but I fear that both interpretations are, and that would maintain an interop issue with SVG which I believe is against most views expressed in the issue there. |
nt1m could you review the change web-platform-tests/wpt-metadata#4775? Thank you so much. |
Test List
https://wpt.fyi/results/html/canvas/offscreen/path-objects/2d.path.stroke.prune.arc.html
https://wpt.fyi/results/html/canvas/offscreen/path-objects/2d.path.stroke.prune.arc.worker.html
https://wpt.fyi/results/html/canvas/offscreen/path-objects/2d.path.stroke.prune.curve.html
https://wpt.fyi/results/html/canvas/offscreen/path-objects/2d.path.stroke.prune.curve.worker.html
https://wpt.fyi/results/html/canvas/offscreen/path-objects/2d.path.stroke.prune.line.html
https://wpt.fyi/results/html/canvas/offscreen/path-objects/2d.path.stroke.prune.line.worker.html
Rationale
The behavior of 0-length subpaths with line-caps is subject to a longstanding canvas2d spec issue that has never been resolved, although discussion there appeared to converge on a consensus that would align behavior with SVG rendering and would not match the existing WPT tests here.
Given that the existing spec text is confusing/ambiguous, and discussion in the open spec issue favored a different result than the existing tests expect, I think we should remove these tests from the interop-2023 set until such time as there's a clear resolution of the spec issue.
The text was updated successfully, but these errors were encountered: