Issue metadata
Sign in to add a comment
|
matrixTransform() and transformPoint() 2d code path uses m12 instead of m21
Reported by
sim...@opera.com,
Jun 7 2017
|
||||||||||||||||||||||||
Issue descriptionhttps://cs.chromium.org/chromium/src/third_party/WebKit/Source/core/geometry/DOMPointReadOnly.cpp?l=45&rcl=6a40e94018f930fd154f12a309b0a211e9e5f915 https://cs.chromium.org/chromium/src/third_party/WebKit/Source/core/geometry/DOMMatrixReadOnly.cpp?l=279&rcl=3ff6ed64df1419714cfa65f287f107a934f519f4 m12 here should be m21. Compare with the 3d code path. In the 2d code path, m12 is used for both x and y, but m21 is not used at all.
,
Jun 7 2017
Thanks! If you could also add tests under external/wpt/css/geometry-1/ for this difference, that would be A+. :-)
,
Jun 9 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/f81049d5a4f02b598815c00baa0d135b79304588 commit f81049d5a4f02b598815c00baa0d135b79304588 Author: hs1217.lee <hs1217.lee@samsung.com> Date: Fri Jun 09 03:38:04 2017 [geometry] replace m12 with m21 when transform DOMPoint from DOMMatrix. there is wrong calculate when transform DOMPoint from DOMMatrix. we should replace m12 with m21. unfortunately all of m12 and m21 values were zero in layout tests. so it could not detect problems. BUG= 730441 Review-Url: https://codereview.chromium.org/2930883002 Cr-Commit-Position: refs/heads/master@{#478185} [modify] https://crrev.com/f81049d5a4f02b598815c00baa0d135b79304588/third_party/WebKit/LayoutTests/fast/dom/geometry-interfaces-dom-matrix-transformPoint.html [modify] https://crrev.com/f81049d5a4f02b598815c00baa0d135b79304588/third_party/WebKit/LayoutTests/fast/dom/geometry-interfaces-dom-point-readonly.html [modify] https://crrev.com/f81049d5a4f02b598815c00baa0d135b79304588/third_party/WebKit/Source/core/geometry/DOMMatrixReadOnly.cpp [modify] https://crrev.com/f81049d5a4f02b598815c00baa0d135b79304588/third_party/WebKit/Source/core/geometry/DOMPointReadOnly.cpp
,
Jun 9 2017
@simonp i will upload other tests for external/wpt/css/geometry-1/ indirectly. as far as i know, external/wpt/css/geometry-1/ is synchronized from https://github.com/w3c/web-platform-tests periodically.
,
Jun 13 2017
tests are merged for external/wpt/css/geometry-1/ at https://github.com/w3c/web-platform-tests. (https://github.com/w3c/web-platform-tests/commit/29ff90825e8af4b5ced39300ce16f7c5a4b16ef6)
,
Jun 14 2017
external/wpt/css/geometry-1/ tests were synchronized. https://chromium.googlesource.com/chromium/src/+/e2d1d93c5b61ce06be4b0620b7e40429365e3582%5E%21/#F57 https://chromium.googlesource.com/chromium/src/+/e2d1d93c5b61ce06be4b0620b7e40429365e3582%5E%21/#F58
,
Jun 14 2017
\o/ This can be closed as fixed, then, right? (Since there is two-way syncing for wpt, it should be possible to add tests in a Chromium CL and it will be automatically upstreamed to web-platform-tests, but either way works of course.)
,
Jun 14 2017
\o/ This can be closed as fixed, then, right? (Since there is two-way syncing for wpt, it should be possible to add tests in a Chromium CL and it will be automatically upstreamed to web-platform-tests, but either way works of course.)
,
Jun 14 2017
yes. i will close this issue as fixed. thank you. |
|||||||||||||||||||||||||
►
Sign in to add a comment |
|||||||||||||||||||||||||
Comment 1 by hs1217....@samsung.com
, Jun 7 2017Status: Started (was: Untriaged)