Before, after using the Browser, memory-hungry apps could
become very sluggish. This was because the search dialog in the
system process had the BrowserProvider open, which in turn had
EnhancedGoogleSearch open. Since EhancedGoogleSearch runs in acore,
the system would keep both the Browser process and acore to stay
around forever.
The cause (or at least one common cause) for this was that
if the user types quickly, and clicks on a suggestion before
the displayed suggestions have caught up, some suggestion cursors
are not be closed.
This change solves this problem by adding a close() method to
SuggestionsAdapter. SuggestionsAdapter now closes any cursors
that are passed to it after close() is called.
Fixes http://b/issue?id=2078226
"global search holding reference to browser: system -> browser -> acore = :("
- @hides relevant APIs
- removes relevant javadoc
- enforces that only system apps can participate
note: general support is still there, will be easy to reenable when we are ready.
Includes:
- an overview section on including search suggestions in quick search box
- additional table of searchable metadata attributes related to quick search box
- additional column descriptions that we've added for quick search box.
- description of how to expose your content provider's search suggestions to
the search infrastructure if you have permissions guarding your content provider.
close itself directly because it may not happen correctly for some cursors
currently. This fixes http://b/2036290, which is being caused by
http://b/2015069 which we are not fixing for Donut, so this is a hack around
that for the time being.
search. Without this, because of the recent changes to how the search
dialog hides and resumes, we end up re-showing the search dialog after
returning to the app from voice search, obscuring the new voice search
results behind.
Fixes http://b/2025364
When the suggstion provider is in a different package from the searchable
activity, the icons stored for shortcuts used the package of the
activity instead of the provider.
Fixes http://b/issue?id=2023993
to the left of the search field. Also, because this removes context
about whether you're in browser search or global search, we make
sure to clear any entered text if you jump out to global search from
within browser search.
This is a really ugly hack, but was required by the UI team. We will
find a better way to reconcile in Eclair.
Re-arrange various things to ensure that the search dialog is told about system windows being
closed before it is told about the navigation back to home.
* changes:
Fix resource code and version attributes Create a new package setting object for updated system apps rather than moving around the same setting. This updates the resource, code and version correctly. For updating system packages, disable the package first which removes the entry from internal structures, create a new package setting, add it to list of user id's then rest of installation steps, kill the process if needed via ActivityManager then add this setting if everything was successful. This also fixes issues with updating values prematurely.
Create a new package setting object for updated system apps rather than moving
around the same setting. This updates the resource, code and version correctly.
For updating system packages, disable the package first which removes the entry
from internal structures, create a new package setting, add it to list of user id's
then rest of installation steps, kill the process if needed via ActivityManager
then add this setting if everything was successful. This also fixes issues with
updating values prematurely.
When a new version of system package is available via OTA, just physically remove
entries for pkg. Note that the component and other info will be eventually updated
later on when scanning the package.
Also move certificate verification slightly ahead before scanning packages.
Some null checks
New api's in ActivityManager to kill an application pkg before finishing installation
This adds a hidden method AutoCompleteTextView.isImeHidden(),
and uses that in SearchDialog to cancel the search dialog
when BACK is pressed, if there is no previous search component
to return to.
mlebeau says:
If we fill the whole screen then it makes the issue of the back
button a little more important. Specifically, right now if you have
the list expanded and you press back, the keyboard hides but it's not
really showing any more anyway so it seems like pressing the button
does nothing. We rationalized this by saying "part of the keyboard
will be showing so it won't be completely non-obvious that it was
hidden". But since really the right UX is to fill the screen, as part
of this we should probably also add logic to the back button such
that if it is pressed when the list is obscuring the keyboard
(i.e. softInputMode on the PopupWindow is INPUT_METHOD_NOT_NEEDED)
then we should hide the dialog entirely rather than closing the
keyboard.
This is part of the fix for http://b/issue?id=2014450
We now distinguish between in-app search (pivoted in from GlobalSearch)
and real in-app search mode. Only the latter dismisses the search
dialog when a suggestion is clicked. Also, the drop-down
is now always visible except in real in-app search mode.
Fixes http://b/issue?id=2014626
* AccessibilityService -- document onBind() to not be implemented.
* GestureLibrary.getLearner() -- needs to be hidden.
* IntentSender -- remove protected constructors, document that it is retrieved from a PendingIntent.
* Hide permissions: SHUTDOWN, STOP_APP_SWITCHES.
* Context -- hide BACKUP_SERVICE.
* ContextWrapper -- hide getSharedPrefs bla h blah
* Intent.parseUri() -- fix docs.
* ApplicationInfo.FLAG_TEST_ONLY?!?
* Hide MockContext.getSharedPrefs blah blah
An issue with the density API is that bitmaps assumed the old default density,
so new programs would have to explicitly set the correct density for every bitmap
they create.
This is an attempt to fix that situation, by define the default density of bitmaps
to be the main screen's density, except for old apps where it is the original default
density.
Actually implementing this is not so great, though, because the Bitmap constructors
can't really know anything about who is calling them to know which density to use.
So at this level the compatibility mode is defined per-process -- meaning the initial
package loaded into a process defines the default bitmap density, and everyone else
loaded in later on has to live with that.
In practice this shouldn't be much of a problem, there shouldn't be much mixing of
old vs. new apps in a process. It does mean that, going forward, if a developer is
going to use shared user IDs for this, they will need to make sure either that all of
their apps are in the same compatibility mode, or that their code explicitly sets the
density of bitmaps it receives. This isn't all that great, but I think it is worth
the benefit of allowing people who write modern apps to not have to deal with bitmap
densities.
This change also does some cleanup of the density management (making sure to always
copy over bitmap densities, etc) and adds java docs to explain the various ways
density is set and used by the system.
This change allows us to use drawables that match the current screen
density even when being loaded in compatibility mode. In this case,
the bitmap is loaded in the screen density, and the bitmap and
nine-patch drawables take care of accounting for the density difference.
This should be safe for existing applications, for the most part, since
they shouldn't really be pulling the bitmap out of the drawable. For
the small rare chance of them breaking, it worth getting the correct
graphics. Also this will only happen when there is actually a resource
of the matching density, and no existing apps should have resources for
anything besides the default density (though of course all of the
framework resources will be available in the native density).
As part of this, the bitmap density API has been changed to a single
integer provider the DPI unit density.
use by SearchDialog. SearchDialog wants to be able to say a dropdown should
not always be visible but still let it ignore outside touch when it does
show.
Fixes http://b/1997170
Sometimes when searching, some of the suggestions had no left padding.
The left-hand side icons were flush with the left edge of the screen.
The problems was that setting a StateListDrawable as a background
will always set the padding of a View, because of a problem in
DrawableContainer.
DrawableContainer.DrawableContainerState.getConstantPadding()
will always return a Rect if mVariablePadding is false, which
makes DrawableContainer.getPadding() return true, which
causes View to change the padding.
As a workaround, we use setVariablePadding(true) on the background
that we create.
Fixes http://b/editIssue?id=1984813
the search manager at that point.
Still works:
- hitting search when a managed dialog of an app is showing will dismiss
the dialog and start in-app search (if it supports it), falling back
on global search
- hitting search when a dialog of an app (not managed by the activity) will
dismiss the dialog and start global search
- hitting search when a system dialog is showing will just dismiss the dialog.
