summaryrefslogtreecommitdiff
path: root/tools/aapt2/java/JavaClassGenerator_test.cpp
diff options
context:
space:
mode:
authorWinson Chung <winsonc@google.com>2019-06-04 21:20:41 -0700
committerWinson Chung <winsonc@google.com>2019-06-05 15:33:16 -0700
commitdec1a6665a76e1a9b6dbf514b5194458e5d7dc2e (patch)
tree4460e441b01e891391654a1c92c0b908572fb00e /tools/aapt2/java/JavaClassGenerator_test.cpp
parentf5671d390d0dfebec37f62bfff7228cd23cd1b59 (diff)
Fix regressions in PiP movement bounds handling
- When we return early when detecting when to skip applying the bounds the stored bounds and rotation weren't being updated - In landscape, the addition of the nav bar inset in the bottom pushed the movement bounds to be out of range, which cause an incorrect offset to be calculated - Simplified the logic to calculate the offset for the adjustment for IME and shelf -- keep track of the previous offset, check the previous movement+offset bounds with the next movement+offset bounds and only update if the PIP is currently between those two bounds - Tweak to landscape policy, if there isn't enough space for the ime offset, then don't apply it as a part of the shift Bug: 126199952 Test: Go to landscape + pip, open IME and dismiss IME and ensure the PIP is pushed off and restored correctly Test: In portrait, test going from app/home/ime and verify that the right offsets are applied, and then if the stack is not in the bounds of the offset change, that it is not applied Change-Id: I67eee7d321995f403137db940b06f685d0c919af
Diffstat (limited to 'tools/aapt2/java/JavaClassGenerator_test.cpp')
0 files changed, 0 insertions, 0 deletions