div with scaleX(-1) transform rendering incorrectly with preserve-3d and hidden backface.
Reported by
syke....@gmail.com,
Mar 17 2016
|
||
Issue descriptionUserAgent: Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/49.0.2623.87 Safari/537.36 Steps to reproduce the problem: 1. Visit https://jsfiddle.net/jnu0snw3/ or attached file 2. Note red div in same z-order as green div, yet the black div renders on top of it. 3. Removing scaleX(-1) will make the div render correctly. What is the expected behavior? Red div renders on top of black div. What went wrong? Red div renders below the black div. Did this work before? N/A Chrome version: 49.0.2623.87 Channel: stable OS Version: 6.1 (Windows 7, Windows Server 2008 R2) Flash Version: Shockwave Flash 21.0 r0 Does it occur on multiple sites: N/A Is it a problem with a plugin? N/A Did this work before? Yes Version <= 48 Does this work in other browsers? Yes
,
Apr 11 2016
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/94858ade402cdf1fec143c154d600ee7b17c3557 commit 94858ade402cdf1fec143c154d600ee7b17c3557 Author: petermayo <petermayo@chromium.org> Date: Mon Apr 11 21:04:42 2016 Enhance unit tests for normals. We asserted that changing vertex winding shouldn't change a polygon's normal when it doesn't have to. We asserted that inversion on the X axis (or Y-Axis) should change the normal, but it shouldn't. We also tested the normal calculation of degenerate polygons (colinear vertices) when these don't have to work, and weren't what the test was claiming to check. This introduces DISABLED tests to fix, separates the set of normal calculations into separate tests so that failures are reported better, and uses better formed polygons for the touching and splitting test cases. BUG= 595820 TEST=self CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel Review URL: https://codereview.chromium.org/1860823003 Cr-Commit-Position: refs/heads/master@{#386470} [modify] https://crrev.com/94858ade402cdf1fec143c154d600ee7b17c3557/cc/quads/draw_polygon_unittest.cc
,
Apr 13 2016
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/59d3768dd104c22554d86f410d22ad8f714c82eb commit 59d3768dd104c22554d86f410d22ad8f714c82eb Author: petermayo <petermayo@chromium.org> Date: Wed Apr 13 14:37:35 2016 Correct normal computation for inverted polygons. Rather than rely solely the vertices, and their winding order for sign, we rely on the transform of the previous normal for the sign, and the vertices (cross product of edge vectors, similar to Newell's algorithm) for the numerics of the direction. BUG= 595820 TEST=unittest CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel Review URL: https://codereview.chromium.org/1883623002 Cr-Commit-Position: refs/heads/master@{#386984} [modify] https://crrev.com/59d3768dd104c22554d86f410d22ad8f714c82eb/cc/quads/draw_polygon.cc [modify] https://crrev.com/59d3768dd104c22554d86f410d22ad8f714c82eb/cc/quads/draw_polygon_unittest.cc [modify] https://crrev.com/59d3768dd104c22554d86f410d22ad8f714c82eb/ui/gfx/transform.cc [modify] https://crrev.com/59d3768dd104c22554d86f410d22ad8f714c82eb/ui/gfx/transform.h
,
Apr 13 2016
This is now fixed. Verification can proceed as the builds & branches make it through the normal process. |
||
►
Sign in to add a comment |
||
Comment 1 by ranjitkan@chromium.org
, Mar 18 2016Labels: -Type-Bug M-51 has-Bisect OS-Linux OS-Mac Type-Bug-Regression
Owner: petermayo@chromium.org
Status: Assigned (was: Unconfirmed)