* We were using the getSelectedItem() from the deletion spinner to
set the account's deletion policy, even if that spinner was
invisible (which it would be for IMAP).
* The result of this is indeterminate; sigh
* The fix is to make sure the spinner is visible before using its
value
Bug: 5216422
Change-Id: I7e44b5e8127f5277693f7e962899e8642be55239
- this is leftover from a time where we re-used a single WebView for
multiple messages. The current architecture does not rely on that, so
the clearView() is actually not useful. Worse - it's causing issues
because the operation in the WebView is asynchronous, so a clearView and
a load of the content in quick succession causes a race condition,
causing items not to load
Bug: 5205296
Change-Id: I5ee6697bdf53a033ba1c7f21170820259bf830d1
- minor changes to some icons (can't see a difference in some of the
cases?), larger changes to others (e.g. mark unread)
- xhdpi assets for a bunch that were missing
Change-Id: Id71d22aa50440cf6c83e4244bb807e3d237949c6
* We weren't checking for it in determining whether our policies
were active; because of this, we never actually SET the policy
in the DPM
Bug: 5193399
Change-Id: I276901be21be681f66891f5374ec58cf1ea7b4be
We had two separate "height" values in XML that weren't consistent - one
was used for the coordinates and one was used for the actual item i nthe
adapter. This made things off slightly when the values didn't match
(i.e. on tablet wide mode).
Also fixed an issue with paperclip icon not aligning with date on wide
mode
Bug: 5184169
Change-Id: I1d80bb9c685fdd8216fcaef26bf6ad715eb74c4c
* The existing IMAP search code is well-known to be primitive and
largely broken. In particular, it fails with any non-ASCII
character and with a variety of symbols (e.g. quotes, slashes,
etc.) Basically, it's an accident waiting to happen, returning
empty data sets even when the query might reasonably be expected
(or known) to return valid data.
* The current CL uses the IMAP literal string format to represent
the query text; this string can be sent either in ascii or in
UTF-8, and since it is sent as an octet (byte) count followed by
8-bit data, it also solves any quoting issues that might come
up. So, we kill two birds with one stone.
* The bug in question was punted to a subsequent MR; however, I
think it would be a mistake to ship the code without this
CL, which has been tested against the three common IMAP servers
that we aim to support.
Bug: 4690713
Change-Id: Iaa542bfc56737871f3cbc9c83f0e768415a7f2b6
The MessageListFragment doesn't guarantee to return a non-null Mailbox -
it can be null for virtual mailboxes. We weren't checking this properly.
Bug: 5171577
Change-Id: I272f6a71633680d3424b049d730500fb5f04935e
