summaryrefslogtreecommitdiff
path: root/core/java/android/util/SparseSetArray.java
diff options
context:
space:
mode:
authorJacky Kao <jackykao@google.com>2021-11-17 15:27:42 +0800
committerJacky Kao <jackykao@google.com>2022-03-02 14:29:11 +0800
commit081a7c61e6b277efc65f292488618650ac86394c (patch)
treebf24dc0a96b7f9caffdbd01cc5ca49b883596165 /core/java/android/util/SparseSetArray.java
parent532df4fc033bbd10b1d444cec3bd666db9eac832 (diff)
Descreses the counts of windows_change_bounds events
When the display magnification is activated, the window bounds of the windows callbacked from the SF will be the real position after the magnifying. We need to use the inverse matrix of magnification spec to compute the original bounds without the magnification effect. And we found if the magnification spec changed too quickly, the computed window bounds won't be correct because the callbacked windows will get one frame behind. This symptom will make the window bounds of the same window isn't the same, and send the windows_change_bounds event out. To descreses the counts of the window_change_bounds events, we stores two consecutive magnification specs, and found out which one is the proper one closing the identity matrix for generating the inverse matrix based on below formula under window is at the stable state based on the assumption, We don't consider the transition effect because this condition wouldn't happen with changing the magnification spec simultaneously: WindowState#transform * MagnificationSpecMatrix * inputWindowHandle#transform = IdentityMatrix Bug: 191736824 Test: a11y CTS & unit tests Test: Manual testing including the A11y services Change-Id: I760b66a50a18c3d6be0cebc0a9dbc03a1bfe6d2c
Diffstat (limited to 'core/java/android/util/SparseSetArray.java')
0 files changed, 0 insertions, 0 deletions