[offscreen-canvas] Remove 2d.path.stroke.prune.* tests for open paths from interop-2023

[jfkthame]

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.

[nt1m]

Seems ok

[chrishtr]

@fserb can you review for Chromium?

[fserb]

lgtm

[yiyix]

This misses the 4 following tests. Could we remove them from interop as well?
2d.path.stroke.prune.closed.html
2d.path.stroke.prune.closed.worker.html
2d.path.stroke.prune.rect.html
2d.path.stroke.prune.rect.worker.html

[jfkthame]

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.

[yiyix]

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

Add a straight closing line to each closed subpath in path connecting the last point and the first point of that subpath; change the last point to a join (from the previously last line to the newly added closing line), and change the first point to a join (from the newly added closing line to the first line).

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?

[jfkthame]

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.