summaryrefslogtreecommitdiff
path: root/tools/aapt2/java/JavaClassGenerator_test.cpp
diff options
context:
space:
mode:
authorRobert Carr <racarr@google.com>2017-06-13 12:41:53 -0700
committerRobert Carr <racarr@google.com>2017-06-14 14:02:12 -0700
commit6837519a9181891c55ad286dd24aa211a536842d (patch)
treef742a1f034723f55dfe5287837bb45cbc8f7a68b /tools/aapt2/java/JavaClassGenerator_test.cpp
parent7c41ddb9750c74541d4df9134064f418dd347fe8 (diff)
Correct errant entry in to DRAG_RESIZE_FREEFORM state.
Observing the logic in computeDragResizing(), we see that whenever an AppWindowToken has frozen bounds we will enter the drag resizing state. Furthermore, in setDragResizing() we see that if the docked divider is not resizing, then we set ourselves as freeform resizing. Also observe that AppWindowToken#startRelaunching unconditionally freezes the bounds for tasks not in the freeform stack. This means we are almost always entering the freeform resize state while relaunching. Within computeDragResizing, the StackId check and the MATCH_PARENT check have prevented this from being an issue in too many places, espescially in pinned. Once we are in this state, calculateSurfaceBounds() will force us to 0,0 as it thinks we are using the big surface approach. There's no need to freeze the bounds if we weren't drag resizing at the time that the app is relaunched, and this CL enforces that. Bug: 62430780 Test: Manual repro from bug. To verify no regressions, create app with long delay in onResume (>1s) verify no flickers when releasing the docked divider and triggering a resize. go/wm-smoke Change-Id: I1d9bdfbe815ff48c884f933acd65612429d633ef
Diffstat (limited to 'tools/aapt2/java/JavaClassGenerator_test.cpp')
0 files changed, 0 insertions, 0 deletions