231014 – Move canCompositeClipPath to RenderLayer

Rob Buis

Reported 2021-09-30 02:37:13 PDT

Move canCompositeClipPath to RenderLayer.

Rob Buis

Comment 1 2021-09-30 02:39:17 PDT

Created attachment 439710 [details] Patch

Simon Fraser (smfr)

Comment 2 2021-09-30 10:05:16 PDT

Comment on attachment 439710 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=439710&action=review > Source/WebCore/rendering/RenderLayer.h:241 > + bool canCompositeClipPath() const; Moving it here has added ambiguity. It's not clear if this is "if this layer were composited, can it composite clip-path", or "is this layer composited and has a clip path and can composite it".

Rob Buis

Comment 3 2021-09-30 10:18:57 PDT

Created attachment 439751 [details] Patch

Rob Buis

Comment 4 2021-09-30 10:23:52 PDT

Comment on attachment 439710 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=439710&action=review >> Source/WebCore/rendering/RenderLayer.h:241 >> + bool canCompositeClipPath() const; > > Moving it here has added ambiguity. It's not clear if this is "if this layer were composited, can it composite clip-path", or "is this layer composited and has a clip path and can composite it". Fair enough, I did another stab. To be clear, what I do not like about current code: - unneeded layer accessing - unneeded ASSERTs Of these to me the ASSERTs are more annoying. Note that this is not high priority at all but let me know if you think we can make this code a bit nicer/more efficient this way or some other.

Rob Buis

Comment 5 2021-09-30 10:24:45 PDT

Comment on attachment 439710 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=439710&action=review >>> Source/WebCore/rendering/RenderLayer.h:241 >>> + bool canCompositeClipPath() const; >> >> Moving it here has added ambiguity. It's not clear if this is "if this layer were composited, can it composite clip-path", or "is this layer composited and has a clip path and can composite it". > > Fair enough, I did another stab. > To be clear, what I do not like about current code: > - unneeded layer accessing > - unneeded ASSERTs > Of these to me the ASSERTs are more annoying. Note that this is not high priority at all but let me know if you think we can make this code a bit nicer/more efficient this way or some other. Nm, thanks :)

Rob Buis

Comment 6 2021-09-30 10:24:45 PDT

Comment on attachment 439710 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=439710&action=review >>> Source/WebCore/rendering/RenderLayer.h:241 >>> + bool canCompositeClipPath() const; >> >> Moving it here has added ambiguity. It's not clear if this is "if this layer were composited, can it composite clip-path", or "is this layer composited and has a clip path and can composite it". > > Fair enough, I did another stab. > To be clear, what I do not like about current code: > - unneeded layer accessing > - unneeded ASSERTs > Of these to me the ASSERTs are more annoying. Note that this is not high priority at all but let me know if you think we can make this code a bit nicer/more efficient this way or some other. Nm, thanks :)

EWS

Comment 7 2021-09-30 14:20:41 PDT

Committed r283340 (242357@main): <https://commits.webkit.org/242357@main> All reviewed patches have been landed. Closing bug and clearing flags on attachment 439751 [details].

Radar WebKit Bug Importer

Comment 8 2021-09-30 14:21:14 PDT

<rdar://problem/83737585>

Attachments
Patch (5.10 KB, patch) 2021-09-30 02:39 PDT, Rob Buis	no flags	Details Formatted Diff Diff
Patch (5.19 KB, patch) 2021-09-30 10:18 PDT, Rob Buis	no flags	Details Formatted Diff Diff
Show Obsolete (1) View All Add attachment proposed patch, testcase, etc.