commit 4740f488f57876dbc84d9a3c54890c6b814f325f
author Fredrik Söderquist <fs@opera.com> Tue Oct 03 18:01:52 2017
committer Commit Bot <commit-bot@chromium.org> Tue Oct 03 18:01:52 2017
tree 3361179b81009b12a1cc716c28e13e72660906d9
parent 3e22e068e2eaa0323ee1cb6f7c801345e8ba1415

Revert behavior for non-scaling-stroke transform

https://codereview.chromium.org/2711503002 fixed getScreenCTM by
including the actual CTM from the element to the "screen" ("host"
coordinate space.) This of course also included the scale factor
contributed by any ancestor transforms. Including these transforms
however made evaluation of the non-scaling-stroke transform (the host
transform) unreliable, because of how that transform is "baked" into
the display list. (So if the host transform is sampled when it is
degenerate, all rasterization using that display list will produce
"weird" results - which will remain until an new display list is
recorded.)
Revert the computation of the non-scaling-stroke transform back
something akin to what it was previously until we can figure out a
better way to deal with this situation.

Bug: 747708
Change-Id: I8fdd19829c44f88d4d5e0bbbe4dec3ff2bee52a4
Reviewed-on: https://chromium-review.googlesource.com/697809
Reviewed-by: Stephen Chenney <schenney@chromium.org>
Commit-Queue: Fredrik Söderquist <fs@opera.com>
Cr-Commit-Position: refs/heads/master@{#506099}