Icons reported as unused by running the script "findUnusedAssets". This
script checks for any reference to a drawable in all XML and Java files.
Some _holo_light icons are removed as they are no longer in the designer
git and should not be used.
Change-Id: Ie42d13394d6d81d35fcb48f8f76c6abd4cb09ee4
Use the new icons where appropriate. This is not a complete change as there are
still some icons missing.
Change-Id: I51d1458749343ddfc6456dc2a9853b8bc2540e26
We discovered that AsyncTask.cancel() doesn't quite perform as expected;
In particular, if you call cancel() during a particularly slow background
worker, the result is discarded and onPostExecute() is never called. If
the result is an open cursor, then we "leak" by not closing it.
For AccountFolderList, which has a multi-step doInBackground():
1. Check for isCancelled() during the long doInBackground() which will
reduce the number of discarded cursors in the first place.
2. Check for isCancelled() at the end of the long doInBackground() and
if true, close the result cursors and return null.
3. In the existing isCancelled() code in onPostExecute(), close the
cursors.
For other AsyncTasks (with simpler configurations):
1. Check for isCancelled() at the end of doInBackground() and if true,
close the just-opened cursor and return null.
See also Change-Id: Ie63a3197af563baa8bb0fe6f1ef9423e281cbf4c
Bug: 3088870
Change-Id: I51b11bb5b23f209f8b61962500a063040484fd24
* Add simple sanity checks to scanned accounts, skip over if bad
* Fix existing unit tests and add new unit test for this change
* Also fixed minor bug in EmailContent that was never triggered in
production code (only discovered it via a unit test).
* Also fixed minor bug in an existing unit test
Bug: 2937595
Change-Id: Id60bbb5d8bd923db043d46891c7f89d7debb0a11
for the selected message.
This also fixes wired background flicker when you long-press.
(Apparently you can't share StateListDrawable among multiple items...)
Change-Id: I922f50a5b2a33731919e26a55bd3537d55057af5
Now it should be close enough...
Also fixed the "selected mailbox won't get hilighted" issue.
Bug 3137965
Change-Id: Ifba83783121e62363fac6428534394d2e0a4bee9
* Sending too many mailboxes to EmailProvider in a single batch can
lead to a binder exception, so we limit them
* Move all provider updates into the commit() method
Bug: 3325331
Bug: 3314442
Change-Id: I1bd8d19a96f642b6e49e2f47a5ddcebaa55c0ec0
* Update MockTransport to allow TLS connections
* Test TLS connection in ImapStore unit tests
* The bugfix: Re-query capabilities after closing/reopening parser for TLS
(Note: Actually, this is required by the IMAP RFC 3501, 6.2.1)
Bug: 3315939
Change-Id: I51f838043e87750b5712a1bd2e4f9c821b58c808
We discovered that AsyncTask.cancel() doesn't quite perform as expected;
In particular, if you call cancel() during a particularly slow background
worker, the result is discarded and onPostExecute() is never called. If
the result is an open cursor, then we "leak" by not closing it.
For AccountFolderList, which has a multi-step doInBackground():
1. Check for isCancelled() during the long doInBackground() which will
reduce the number of discarded cursors in the first place.
2. Check for isCancelled() at the end of the long doInBackground() and
if true, close the result cursors and return null.
3. In the existing isCancelled() code in onPostExecute(), close the
cursors.
For other Activities (with simpler configurations):
1. Check for isCancelled() at the end of doInBackground() and if true,
close the just-opened cursor and return null.
Bug: 3088870
Change-Id: Ie63a3197af563baa8bb0fe6f1ef9423e281cbf4c
Specifically, added configuration settings for international Yahoo
domains (e.g. yahoo.ca, yahoo.de, yahoo.fr, etc), including partially
or wholly owned Yahoo subsidiaries; as well as partner domains from AT&T,
British Telecom, and Rogers whose email services are managed by Yahoo.
For domains supporting IMAP, changed to IMAP+SSL for mail handling to
support a better user experience on mobile devices.
Bug: 3258944
Original Author: Kristopher Giesing
Original CL: If5bdc70f70820f3f8aada6d2ee9b0cdb115432b5
Change-Id: Ib117f3755ff9250746f05debd1a6d1cda44c7855
I'm not yet 100% sure what's going on, but the old code, not stopping
the scheduled forceLoad() in onReset() is apparently wrong.
onReset() is a new callback introduced in the recent loader change, so
this can explain why it started after the change.
Also, call super.onContentChanged() instead of directly calling forceLoad(),
as Loader.onContentChanged() now does more than just calling forceLoad.
Bug 3308465
Change-Id: Ica46bf994b571a26dd532734307a5c7ec6df0268
