Ideally, I would like to allow Activities in the same package to intercept mime
types but there is no way to get the current Activity information from a generic
Context object. For now, skip resolved Activities that are in the same package.
Bug: 2314904
drag. It was added for the better performance of panning
while loading. But it does make sites using touch event
look bad as we don't have the display tree until touch up.
Fix http://b/issue?id=2310295
Move the change in position/size before the call to bring into view,
so that the properly positioned field will be brought into view. It
is also now before we check to see if it is the same textfield, so
it happens in both cases.
Also remove some code which checks the selection and then sets it
unnecessarily. It was originally surrounding some code which called
setRect, which had a side-effect of changing the selection. That
code was removed (rendering the remaining code useless), and setRect
no longer has the side effect (the side effect was a result of calling
reqestFocus, which it no longer does - further, the change in selection from
requestFocus no longer affects the textfield's text).
Help fix http://b/issue?id=2150538 and http://b/issue?id=1890360
Use the <input type> information to set the InputType, so that
the IME can show the correct options. Also consolidate setup
of WebTextView into setType().
Requires a change in external/webkit.
WebCore treats */*+xml as content that it can render. This prevents Activities
from intercepting those mime types. Check the same regular expression as
DOMImplementation and then use the PackageManager to look for a matching
Activity.
Requires a change to webcore.
Bug: 2256744
Partial fix for http://b/issue?id=2232001
We were calling obtainMessage with the node pointer as an argument. What we
really need to do is set arg1 to the node pointer. Also, avoid checking for
saved form data for password fields and fields whose names are blank.
If a touch event was targeted at a ScrollView child, the content would jump after
the user moved her finger by the required threshold. This is inconsistent with
ListView's behavior.
Fixes#2296622.
Fixes#2296601.
WebView and its related APIs are keeping static objects that hold references to
Context instance. This could cause applications to leak their first Activity
and all the associated resources.
This change adds 3 new interfaces for plugins to the framework. This
change also includes extensive cleanup as we consolidate internal plugin
functions into the pluginManager. Also using the new interfaces we no
longer need to pass additional parameters in quite a few methods.
We were previously calling nativeTextInputMotionUp, which was
intended to be faster than nativeMotionUp, which does more work.
However, as discussed in http://b/issue?id=2257787 , I think we
should simply call nativeMotionUp, which does everything the old
method did, while also correcting things if necessary. This fixes
the second problem listed in http://b/issue?id=2266756 , since
nativeMotionUp will recognize that the <select> element should
take the click.
Since I have removed a native method, this requires a change to
external/webkit.
We remove the WebTextView while zooming. If the WebTextView ended up
offscreen, we remove it completely, thus hiding the IME. In some cases,
however, the WebTextView is only offscreen because the IME overlapped it.
Check to see if the IME is showing, and if so, adjust the WebTextView and
let it get scrolled into view.
Also perform the removal/change in text size inside the method, so it
happens in both places.
Lastly, do not call didUpdateTextViewBounds if there is no WebTextView.
Fix for http://b/issue?id=2266066
Also sets an explicit type for GeolocationPermissions.getOrigins.
This is a partial fix for bug http://b/issue?id=2271636
Change-Id: Ife478917a979b68e242c467798358a018455957f
the content. Limit the overview viewport width to
a level that if a site behaves badly, the WebView
still does reasonably.
Partially fix http://b/issue?id=2209659
maps.yahoo.com keeps growing due to a layout bug.
This prevents the browser to go insane.
Partially fix http://b/issue?id=2215387
caltrain.com has a super long which caused
the page super wide. This prevents the browser to
become unusable.
This change will make sites like, caltrain.com/timetable.html
not fit all in the view. You have to scroll left
and right. And it is same as iPhone.
When removing a WebView from its Window, remove the WebTextView.
Remove the old attempt to do the same, which was not being called.
When attaching a WebView to a Window, call onWindowFocusChanged so
its state of drawing the selection ring returns.
Fixes http://b/issue?id=2257787 and http://b/issue?id=2260867
Merge commit 'dc30d28bc8321171fbf4ed92a9486eb8dd382259' into eclair-mr2
* commit 'dc30d28bc8321171fbf4ed92a9486eb8dd382259':
Honor the mInitialScale set on WebView. This was
Merge commit 'bf5b632abbc2f4b96323ff428be585fd2a3d381c' into eclair-mr2
* commit 'bf5b632abbc2f4b96323ff428be585fd2a3d381c':
Use getViewHeightWithTitle() to honor the scrollbar
Merge commit 'd7439f4bdb31d005011b69112e82a9fed9be4470' into eclair-mr2
* commit 'd7439f4bdb31d005011b69112e82a9fed9be4470':
Use computeVerticalScrollRange() to find the bottom
Merge commit 'c996689109b1ac0849aa675e1987e3851ee77a95' into eclair-mr2
* commit 'c996689109b1ac0849aa675e1987e3851ee77a95':
Return an empty PluginList instead of null for the
