[chromium] Add more compositor unit tests for high DPI transforms
Created attachment 153645 [details] Patch
Created attachment 153646 [details] Patch
Created attachment 153655 [details] Fix egregious typo
Created attachment 153681 [details] More tests; fix scale for duplicateChild test
Sorry for all the patch noise. The previous patches had a bug in them that caused problems in dependent patches. This is ready for review.
Comment on attachment 153681 [details] More tests; fix scale for duplicateChild test View in context: https://bugs.webkit.org/attachment.cgi?id=153681&action=review nice tests! lgtm > Source/WebKit/chromium/tests/CCLayerTreeHostCommonTest.cpp:3722 > + EXPECT_TRANSFORMATION_MATRIX_EQ(parent->screenSpaceTransform(), expectedParentScreenSpaceTransform); nit: expected value should come first in all these
Comment on attachment 153681 [details] More tests; fix scale for duplicateChild test Nice! LGTM, too.
Seems great to me too, but I did have one question - do you think we need to be testing more explicitly something about order of operations, just in case? it looks like we are implicitly testing order of operations between the scale and the layer position translation, but do we need to also check how the contents scale is ordered with respect to the compositeLayerTransform and compositeSublayerTransform? I'm OK with the patch as-is, just wanted to throw that idea out there in case someone felt strongly.
Created attachment 153911 [details] Rebased to ToT
Created attachment 153913 [details] Reorder expect calls
Thanks for that. @kbr or @senorblanco for review?
The diff from two patches ago is just to remove the clip rect line to rebase, one typo fix, and to reorder expect calls as Dana asked.
Comment on attachment 153913 [details] Reorder expect calls rs=me based on Dana's review.
Comment on attachment 153913 [details] Reorder expect calls Clearing flags on attachment: 153913 Committed r123413: <http://trac.webkit.org/changeset/123413>
All reviewed patches have been landed. Closing bug.