diff options
| author | Yohei Yukawa <yukawa@google.com> | 2014-09-02 14:18:40 -0700 |
|---|---|---|
| committer | Yohei Yukawa <yukawa@google.com> | 2014-09-08 02:17:54 +0000 |
| commit | 5f183f0671dfa1d87ca6d741deb457170c432493 (patch) | |
| tree | bbe26c8efa6ac5b36856227c2f58b437321aee9f /core/tests | |
| parent | a3ca5a31a5db14e7c4597de074f01c1af6161872 (diff) | |
| download | frameworks_base-5f183f0671dfa1d87ca6d741deb457170c432493.zip frameworks_base-5f183f0671dfa1d87ca6d741deb457170c432493.tar.gz frameworks_base-5f183f0671dfa1d87ca6d741deb457170c432493.tar.bz2 | |
L API proposal: Introduce IS_RTL flag
This CL introduces CursorAnchorInfo.FLAG_IS_RTL for better
RTL support. This CL also renames *CharacterRect() with
*CharacterBounds() so that they can look more consistent
with other existing APIs.
Rationale:
CursorAnchorInfo.FLAG_IS_RTL addresses following issues.
1. There is no way to associate the RTL information with
the insertion marker.
2. Returning mirrored (right < left) RectF for RTL in
CursorAnchorInfo#getCharacterRect() is turned out
to be bug-prone. Such usage of RectF is not fully
supported. For example, RectF#isEmpty() always returns
false when right < left.
3. There is no reliable to provide the RTL information
when CursorAnchorInfo#getCharacterRect() returns an
empty (right == left) RectF. Perhaps we could use +0.0
and -0.0, but I'm afraid that it is also bug-prone.
BUG: 17365414
BUG: 17335734
Change-Id: Ic8c6fab58c01206872a34e7ee604cdda1581364d
Diffstat (limited to 'core/tests')
0 files changed, 0 insertions, 0 deletions