This change updates isEmergencyNumberInternal() to always return false if
you pass in a SIP address, since the concept of "emergency numbers" is
only meaningful for calls placed over the cell network.
Previously we *did* try to compare SIP addresses against the list of known
emergency numbers, which caused bad behavior with SIP addresses that even
contained "911"/"112"/etc as a substring (since we were filtering out
non-dialable characters before doing the comparison!)
TESTED:
- Before this change, calls to "abc911def@example.com" or
"911abcdef@example.com" were incorrectly detected as emergency
numbers, and fail.
- After this change, SIP addresses like "abc911def@example.com" and
"911abcdef@example.com" work fine.
- Also, confirmed that this change doesn't break the restriction that
3rd party apps shouldn't be able to make emergency calls.
Specifically, I fired off ACTION_CALL intents (using the CallDialTest
activity) for a bunch of numbers *similar* to emergency numbers, and
confirmed that none of them actually resulted in an emergency call
being placed.
The specific ACTION_CALL intents I tested were:
"911" ==> Didn't place the call; brought up dialer instead
"tel:911" ==> Didn't place the call; brought up dialer instead
"911@foo" ==> Tried to start a SIP call (which failed)
"911%40foo" ==> Tried to start a SIP call (which failed)
"tel:911@foo" ==> Tried to start a SIP call (which failed)
"tel:911%40foo" ==> Tried to start a SIP call (which failed)
"911@example.com" ==> Tried to start a SIP call (which failed)
"sip:911" ==> Didn't place the call; brought up dialer instead
"sip:911@foo" ==> Tried to start a SIP call (which failed)
"sip:911%40foo" ==> Tried to start a SIP call (which failed)
Bug: 5515452
Change-Id: I6f9f8690b08564c53c7a76f480654477b475d94d
* commit 'e88ca718ed9af788f2dc46b5f2db494906bbf41d':
Hack to workaround the fact that the EGL context can be removed from under us by framework in low memory condition.
Copy the value of AUTO_TIME to AUTO_TIME_ZONE.
Don't upgrade screen timeout and auto brightness.
Bug: 5607851, 5217460
Also fix the default for vibrate mode. Bug: 5553900
Change-Id: I1b8d5215d55953f2b686e77bf55fd07110a08b8d
Move NTP updates outside locks to avoid blocking ShutdownThread. Add
logging around trim events that remove substantial history. Include
history start in sample events.
Bug: 5627247, 5584564
Change-Id: If9bbd93842c710efcdec94e68da87e9e42a5f961
On AudioTrack and AudioRecord stop or failed start, restore the priority
and cgroup of the caller to their previous values, rather than forcing
to NORMAL. Dependent on new thread APIs.
Also fixes bug where priority was set to AUDIO but cgroup not set.
Change-Id: Ib83893918fb4fdf57c6b87884b51038997a631d8
This class provides an API for an application to know when it is time to
destroy its EGL context when memory is being trimmed. By having this in
the framework, we can still detect whether it will be useful to destroy
any EGL contexts (because we know if doing so will destroy all of them).
Change-Id: I1eac8d640052778052926b875c7928008f752182
Updated WebView documentation to clarify this behavior.
Additionally, if target-densityDpi is specified in the meta viewport
tag, update the display density in ZoomManager.
Bug: 5477652
Bug: 5327492
Change-Id: I176e87489f76a2ee6595d76a006e302b07595de1
...through setup wizard after wipe data
Deal with finish() being called when there are no running activities
on the stack.
Also some improved debugging output.
Change-Id: Ia1d3f3f7e7b79c06ca95c738081322fc80282e0d
