While Leon mulls over how to rewrite WebTextView to
handle webkit driven focus changes, these fixes make
the current code slightly more stable.
Keep track of when keyDown is sent again so that the
WebTextView won't be rebuilt in that case.
Don't clear the cursor from WebTextView if the there's
no native focus pointer.
Only set the selection to its initial state once when
passing events down to webkit.
partially fixes http://b/issue?id=2201049
Merge commit 'f8e3ac851959efbd21da930a802f8efb65ccec5b' into eclair-mr2
* commit 'f8e3ac851959efbd21da930a802f8efb65ccec5b':
In onKeyUp for DPAD_CENTER, show the IME for text input.
Remove the ro.config.sim_password_clear property; go back to enabling
this for all devices. For GSM devices, we clear the password on SIM
change as before. For CDMA devices, we only clear when the device is
provisioned with a different number; we don't clear the password while
the device is unprovisioned.
0 height to get the exact height. But if the View system
doesn't think the WebView dimension changed, even we
call requestLayout, we won't get onSizeChanged. So
we never notify the WebKit about the final view size.
If updateLayout is true, which is probable because that
we set the height to be 0 to WebKit, just call requestLayout()
even the meansured height / width is matching the
content height / width.
Fix http://b/issue?id=2162991
Without the size checks it's possible for calls to glBufferData
and glBufferSubData to read off the end of the Buffer object's
data, which can cause page faults.
Fix end-of-line characters for the "spec" files. (That's why
every line of these files is changed.)
Enhance our code emitter to properly handle bounds checks for
possibly-null pointers.
If the typed key doesn't go with the current text area, clear
the focus so the key won't return to that text area. However,
don't clear the focus if it is already off the old text area.
fixes http://b/issue?id=2201049
The ViewRoot is responsible for calculating the current view's surface's
transparent region and reporting this to the window server, which in
turn notifies Surface Flinger.
To minimize cross-process traffic, ViewRoot keeps track of the
most recent transparent region that it has reported to the window
server, and only sends updates when the region shape changes.
ViewRoot is effectively shadowing the internal state of the window
server, and usually this works fine, because ViewRoot is usually the
only thing that changes this value.
But when the surface is recreated, the window server resets the
transparent region to empty. So in order to properly update the window
server we need to reset our copy of its state to empty too.
Fixes issue #2133090: GL Gears runs at half speed in some cases
Fix for http://b/issue?id=2186188. Keep track of <optgroup> labels
separately from disabled <option> labels. Requires a change to
external/webkit.
In CheckedTextView, if the CheckMarkDrawable is set to null,
remove it.
* changes:
It turns out to be useful to allow enumeration on a per-tag basis, and it's easy to support based on the data structures we have, so add a tag parameter to getNextEvent().
WebView.mHeldMotionless should only be set to MOTIONLESS_FALSE
if some scrolling occurs, even in the case where motion is discarded
by snap-scroll mode.
Refactoring the code allows the code to be smaller, and provides
a single point where scrollBy() is called while dragging.
This doesn't change the snap-scroll logic in any way.
related to http://b/issue?id=2092446
Merge commit 'cdc8b326dd0989003e2fdd7d5a605b5755acf221' into eclair-mr2
* commit 'cdc8b326dd0989003e2fdd7d5a605b5755acf221':
Fix high density bug 2179543: use 'dip' instead of 'px'
Make movement of text selection arrow and I-beam respond to
both trackball and d-pad. Make the 'click' action work for
both trackball-up and d-pad center.
fixes http://b/issue?id=2186069
* changes:
If openOrCreateDatabase() throws an exception, delete the old db and re-do it. If it still fails, something bad happens, like the directory may have the different permission. Let it throw as WebView needs the db.
Merge commit 'd7786b41b78ba8592daea708b024b5a75e441992' into eclair-mr2
* commit 'd7786b41b78ba8592daea708b024b5a75e441992':
Power Manager support for light sensor backlight management.
Merge commit 'f37447bad3773b62176baa837908daf6edb44273' into eclair-mr2
* commit 'f37447bad3773b62176baa837908daf6edb44273':
Proper fix for zero signal strength and no_service. Fixes#2176141
Track phone service state changes and use a separate timer for out-of-service
since the hunting can timeout on some devices.
Store the timeout value in the config.xml, as it is device/network specific.
Settings App will also change to use the hunting duration to compute the cost
of zero signal.
Since references can be queued in another thread, the first entry in the pool
could have been queued after processPoolLocked. Check for null and create a new
chunk if the check fails.
The fix relaxes our processing of the content-disposition header by allowing unquoted filenames, and adds the m4v type to the MIME type map incase the content-disposition header was not specified.
Change-Id: Iaab0c5b17991a2c05eff5db01babe084444fd3b7
This change is likely incomplete and perhaps not right in other ways.
The gist of the change is that the span can return the number of lines
to which to apply the "leading margin".
Some specific things that should be looked at:
1) if the user has nested multiple
LeadingMarginSpans then they will inherit the "line count" feature.
This is wrong but I didn't want to spend time fixing it until it
was clear that this overall approach was acceptible.
2) The units for how many lines should indented is "lines" rather than
something like dips.
3) I wasn't sure what our strategy was for binary compatibility so
I didn't want to modify the methods in LeadingMarginSpan. Instead I
made another interface with extends LeadingMarginSpan that has the
extra method to return the line count.
This change will reduce the occurence rate of A2DP sink suspend resume failures observed in issues 2184627, 2181005 and possibly 2189628.
More robust suspend/resume logic.
Use only the suspend request to audio hardware to avoid having two concurent suspend resume control paths.
* Fixes#2192935
* Exchange attachment handling depends on MimeTypeMap to determine the
mime type of attachments based on their file extension
Change-Id: I44bbc36d7a5aa0800f072553274346d4547f8d4c
We will use the System.SCREEN_BRIGHTNESS_MODE Settings value instead.
Add SCREEN_BRIGHTNESS_MODE_MANUAL and SCREEN_BRIGHTNESS_MODE_AUTOMATIC constants.
Change-Id: I01935be3fcb48cf76392d2c594205cb47babc5b2
Signed-off-by: Mike Lockwood <lockwood@android.com>
