Previously we were using native config flags in some places that expected
Java flags, and vice-versa. All usages of config flags are now annotated
to ensure we're using the right type.
Cleans up annotations on most methods that were touched.
Bug: 21161798
Change-Id: Ifd87dfb12199fc8258915d8a510e03ddb681ca89
With this CL, we expand the target of Bug 27129703 and Bug 27348943 to
non-keyboard subtypes.
Suppose there is a handwriting IME (mode == "handwriting") that supports
the following 5 subtypes.
- en-US
- en-GB
- fr
- sr-Cyrl
- sr-Latn
Also suppose the system languages are configured as follows.
1. sr-Latn-RS
2. ja-JP
3. fr-FR
4. en-GB
5. en-US
In this case we want to enable [sr-Latn, fr, en-GB] by default when
"use system language" is checked in the subtype enabler.
See previous commits [1][2] about how we addressed those issues for
keyboard subtypes.
[1]: Iaf179d60c12b9a98b4f097e2449471c4184e049b
e985c240e3feb62ea38d5b4e386be083ca0f215b
[2]: I8fc774154f5175abff2f16e8f12a4847bf5f5b7c
072a95a3094af2ced4f009ad62c4553c28e3f830
Bug: 27560993
Change-Id: I416b5671602c550805ed6267dd210968aa1de83c
- Ensure that we start the screenshot as a foreground service to reduce
likelihood that it is killed while taking a screenshot.
- If the screenshot process times out or gets killed for any reason,
ensure that we update the notification with an appropriate error
message.
Bug: 27389179
Change-Id: I5007bda95538044bc753e4ceffd2f59a069c857b
Since DRAG_END is now being delivered only to views that
returned true from DRAG_STARTED, make sure to clear
mCurrentDragStartEvent when returning false from DRAG_STARTED,
to avoid incorrectly sending DRAG_STARTED events to children
views on visibility changes.
Bug: 27595763
Change-Id: Ic18628b48b474351e2433e871bf545657a6ebf52
If we are pinning an activity whose task should return home, then move
the home stack forward so that the home activity is visible behind it
instead of the next task in the fullscreen stack.
Bug: 26697763
Change-Id: Iadb7f5453d4351bc4200bcc2feed0ab9fe41e530
When there are not enough number of suggestions, previously used
suggestions was continue to be used. This fixes it and simplifies code
by always allocating SuggestionInfo array.
Bug: 27602619
Change-Id: I9378315b8c810f126f65fade96ce9f6ff5271957
This CL addresses a regression introduced by a recent CL [1] that
non-keyboard subtypes are no longer implicitly enabled based on the
system language (a.k.a. "use system language" in the subtype enabler)
due to a type mismatch in comparison.
Here is the original logic:
if (language.equals(systemLanguage) && systemLocale.startsWith(locale))
And here is the logic replaced by [1]:
if (locale != null && locale.equals(systemLanguage)) {
The new logic is simply broken, because locale is a Locale object while
systemLanguage is a String object. It never matches.
With this CL we will compare the system language with the locale
language again, with several test cases that should have been included
in [1], as a temporary solution until we start relying on
LocaleUtils#filterByLanguage() for non-keyboard subtypes.
[1]: Iaf179d60c12b9a98b4f097e2449471c4184e049b
e985c240e3feb62ea38d5b4e386be083ca0f215b
Bug: 27560993
Change-Id: If2d1710174853180465832e6ecbbb91235b76210
This was somehow missed in the previous commit [1] that added
isAsciiCapable attribute to InputMethodSubtype.
For devices from API level 14 to 23, IME developers might be able to
work around this issue by specifying "AsciiCapable" [2] in
imeSubtypeExtraValue, which is not coverted by CTS though.
[1]: Ic3ace4b6e0432d56696bcbc0be336aec1dc744a5
dc8abf6cee0bcf44e2cad8155f0c151105d46471
[2]: I0bc9954f163a3ec38d08b9ba842a8a31176eb6a6
8e303cc5dd4860b6050d5725ce60ca7e6fb00c7b
Bug: 27603986
Change-Id: Ifb6bc83e782ac05df180dfe0d689897d07441a0c
Update docs to reflect that values in fields store sizes available for
application (not full screen size).
Bug: 27449789
Change-Id: I92f7552a92fa1b66c5eba4cf28693528cbff21d3