Separate compiled JARs into GECKOVIEW_JARS and FENNEC_JARS, and run
AnnotationProcessor separately on each set. The GeckoView bindings are
put into widget/android/GeneratedJNI* (same as before), while the
Fennec-specific bindings are put into widget/android/fennec/FennecJNI*.
This patch includes a small memory optimization of using ArrayList and
`volatile int` for storing the pending thumbnails list and pending
width, instead of using LinkedList and AtomicInteger, respectively.
The patch also fixes a possible race condition due to calling
processNextThumbnail outside of a lock. Now it must be called inside a
lock and its name is changed to reflect that.
Previous design is only to request audio focus for audible media, but now we
also request focus for non-audible media.
It's simple that the app should own the focus when users start watching media.
MozReview-Commit-ID: 3eJP26h4kh7
--HG--
extra : rebase_source : b35c4ef7d6560635f428eba445f089abd3844bda
This patch does the following.
- Removes the return value, because none of the call sites check it.
- Removes the empty implementations from several nsIWidget instances, because
they can use the nsBaseWidget one.
--HG--
extra : rebase_source : 1c42bb32a662f2659c934a245ecd0025045120a5
This patch does the following.
- Removes the return value, because none of the call sites check it.
- Puts an empty implementation into nsBaseWidget.
- Removes the empty implementations from several nsIWidget instances, because
they can use the nsBaseWidget one.
--HG--
extra : rebase_source : 2a94a47f1e7fe986d1efc3854c72968a3e28e365
This patch does the following.
- Removes the return value, because none of the call sites check it.
- Removes the empty implementations from the android nsIWidget instance,
because it can use the nsBaseWidget one.
--HG--
extra : rebase_source : f276e1629fcd0ffccb8c8dbec7bba6639a5ebd9f
This patch does the following.
- Removes the return value, because none of the call sites check it.
- Removes the empty implementations from several nsIWidget instances, because
they can use the nsBaseWidget one.
--HG--
extra : rebase_source : 82733f682c9e254e20354cc6908955a1e7485ee7
This patch does the following.
- Removes the return value, because none of the call sites check it.
- Puts an empty implementation into nsBaseWidget.
- Removes the empty implementations from several nsIWidget instances, because
they can use the nsBaseWidget one.
--HG--
extra : rebase_source : ccf64aaa9364d096e1f060ef77be7e8455b11e1f
This patch does the following.
- Removes the return value, because none of the call sites check it.
- Removes the empty implementations from several nsIWidget instances, because
they can use the nsBaseWidget one.
--HG--
extra : rebase_source : b8a0d9a49b31929dd06af9e61fc57f484af7671d
This patch does the following.
- Removes the return value, because none of the call sites check it.
- Puts an empty implementation into nsBaseWidget.
- Removes the empty implementations from several nsIWidget instances, because
they can use the nsBaseWidget one.
--HG--
extra : rebase_source : e98b7f9970a920c39e941bb531dfc098125913d1
Make the GeckoAppShell.cameraCallbackBridge callback use the new native
JNI bindings. Also refactor the relevant code in GeckoAppShell and
CameraStreamImpl.
Get rid of LayerView.Compositor.getSurface and just pass in the surface
when creating or resuming the compositor. That also lets us get rid of
some synchronization required for getSurface.
Now that nsWindow doesn't own LayerViewSupport, we shouldn't be using
WindowEvent for LayerViewSupport calls. This patch converts the calls
that dispatch to proxy to dispatch directly to Gecko. For
SyncResumeResizeCompositor, it used a proxy to call OnResumedCompositor
on the Gecko thread; this patch makes SyncResumeResizeCompositor post an
event to call OnResumedCompositor directly, without going through the
proxy.
Right now we report Java exception stacks through two code paths:
- GeckoAppShell.reportJavaCrash for exceptions caught by Java code
- jni::HandleUncaughtException for exceptions caught by native code
This patch combines the code for reporting the Java stack trace into
ReportException, to make it easier to add new features such as reporting
OOM exceptions.
We used to only handle certain exceptions in native code, and otherwise
forward the exception to Java handling code. Handling exceptions in
native code gives us better native stack traces in crash reports, so
this patch makes us always handle uncaught Java exceptions in native
code if possible, and only forward to Java as a fallback.
The patch is generated from following command:
rgrep -l unused.h|xargs sed -i -e s,mozilla/unused.h,mozilla/Unused.h,
MozReview-Commit-ID: AtLcWApZfES
--HG--
rename : mfbt/unused.h => mfbt/Unused.h
Make the generated SDK binding files depend on the annotation processor
jar file, so that changes in the annotation processor will cause the
bindings to regenerate.
We need to set the compositor-created flag in LayerView correctly when
reattaching to a new LayerView, so that we resume the compositor instead
of trying to create the compositor again, which we already did
previously.
This patch consists of general code cleanup and a new reattach method in
LayerView.Compositor that we call in order to set the compositor-created
flag correctly.
Its return value is only checked in one low-value assertion.
The patch also does the following.
- Removes the Android and GTK overloadings of EnableDragDrop(), which are
identical to the nsBaseWidget one.
- Streamlines the Windows implementation: fixes the indentation and takes
advantage of infallible |new|.
--HG--
extra : rebase_source : d090848cf5ea2e92c0188b07559c1e1f3899829f
The only implementation of SetSizeMode() that can fail is the nsCocoaWindow
one, on an Objective C exception, which is unlikely and can be swallowed.
This allows some nsGlobalWindow functions to become infallible as well.
--HG--
extra : rebase_source : 85d54185f63d2c138ee3f3e0e0bfe3b5e805a981
The ANativeWindow methods in AndroidBridge are just wrappers around
ANativeWindow calls in the NDK. We should just use those calls instead
of going through the wrappers.
Remove the presentation methods from nsWindow/AndroidBridge now that
they are not used. This also eliminates gGeckoViewWindow from nsWindow.
NS_PRESENTATION_WINDOW and NS_PRESENTATION_SURFACE constants are added
for GetNativeData/SetNativeData to provide replacements that support
multiple GeckoViews/nsWindows.
Put presentation surface native methods in PMPM Instead of in
GeckoAppShell, and also make these methods work with multiple GeckoViews
by having them take a GeckoView parameter if necessary. This also lets
us eliminate the static nsWindow reference in nsWindow.cpp.
They don't need to be NS_IMETHOD, but they should be MOZ_MUST_USE. Adding the
latter catches a few missing checks, which the patch adds.
The patch also gives PuppetWidget an InfallibleCreate() function, which makes
the infallibility of PuppetWidget creation clear.
This patch makes GetBounds(), GetScreenBounds() and GetClientBounds() more
obviously infallible, like existing functions such as GetNaturalBounds() and
GetClientSize(). This results in clearer behaviour in nsCocoaWindow.mm if
Objective C exceptions occur. Along the way, the patch removes some useless
failure checks for these functions.
The patch also removes the NS_IMETHOD from GetRestoredBounds and makes that
function MOZ_MUST_USE.
It's possible for LayerView to be destroyed before it's finished
initializing on the Gecko thread. In that case, we should not continue
with normal destruction of NPZC. Checking isGeckoReady() in
GeckoLayerClient accomplishes that. Also, guard against a case where
LayerView.onSizeChanged is called after LayerView destruction and
results in a NPE from mCompositor being null.
This patch makes most Run() declarations in subclasses of nsIRunnable have the
same form: |NS_IMETHOD Run() override|.
As a result of these changes, I had to add |override| to a couple of other
functions to satisfy clang's -Winconsistent-missing-override warning.
--HG--
extra : rebase_source : 815d0018b0b13329bb5698c410f500dddcc3ee12
This is not just a refactoring. It ensures that the early return in the
factored-out code only skips the dispatch to observers, *not* the
additional processing by ChromeProcessController.
MozReview-Commit-ID: F7xCoORKRlG
--HG--
extra : rebase_source : f9ef57d25a54442dd083560029437fcad885149c
Combine the callObserver and removeObserver methods used for notifying
alert listeners into the native method
GeckoAppShell.notifyAlertListener. Keep track of the listener and the
alert cookie on the native side so that we don't need
GeckoAppShell.ALERT_COOKIES anymore.
Refactor the notification-related methods in GeckoAppShell. In
particular, combine showPersistentAlertNotification into
showAlertNotification, and move code out of AndroidBridge.
Add native method calls to AndroidGamepadManager to replace the gamepad
events in GeckoEvent. Implement those calls in AndroidGamepad.cpp. The
jni/Refs.h change is necessary to fix a compile error when using
jni::BooleanArray.
Add a native method in ZoomedView to request zoomed view data. The
native method is implemented in the ThumbnailHelper class on the C++
side because it shares a lot of code with the thumbnailing code.
Refactor the dispatching code to be more streamlined, and add two
optimizations to the way we dispatch a call:
* avoid a pair of unnecessary calls to add/delete the global class ref,
when dispatching a static call to the Gecko thread without a class ref
parameter.
* avoid an extra allocation when dispatching to a proxy function.
Remove uses of UsesNativeCallProxy and UsesGeckoThreadProxy, now that
they are not needed.
Remove cases where we had to invoke a call in a proxy, because the call
is now specified to be invoked directly in the WrapForJNI annotation,
without the need to go through the proxy.
For SmsManager and AlarmReceiver, we no longer need to manually dispatch
everything to the Gecko thread because that's now handled automatically.
Implement checking the calling thread of a JNI call based on the
calledFrom attribute set in WrapForJNI. Also implement automatic call
dispatching based on the dispatchTo attribute set in WrapForJNI. This
eliminates the use of UsesNativeCallProxy and UsesGeckoThreadProxy.
Update the code generator and related classes in annotation processor to
use the new WrapForJNI flags. Also add some more sanity checking to make
sure the flags are used correctly.
Replace old flags in WrapForJNI usages with new flags. The calledFrom
and dispatchTo flags are set based on whether the method is native or
non-native, and how the method is used.
Also fix testEventDipatcher to respect NativeJSObject's calledFrom =
"gekco" flag, by moving a test to Gecko thread.
We need to register the Java UI thread in native code very early in the
startup process, before libxul registers its JNI entry points. So it
makes sense to register the Java UI thread in mozglue.
There were a couple of problems when delivering tap gestures to content with
full zoom applied. One was that the ConverToGecko function converted the coords
into "CSS pixel" space by using the web content's CSS-to-LD scale, but also
applied that on the translation from the chrome area. Moving that conversion
to later in the process (after the coords got passed through TabParent::
AdjustTapToChildWidget) corrected that issue.
The other problem was that bits of code in APZEventState and APZCCallbackHelper
were using the widget->GetDefaultScale() value as the CSS-to-LD scale, but that
omitted the full zoom value. Getting the CSS-to-LD scale from the presShell and
propagating that through corrected that issue.
MozReview-Commit-ID: KdrkdEZslHo
This patch replaces all references to |nsAutoPtr| in widget/ by references
to |UniquePtr|. |nsAutoPtr| is deprecated and will go away soon.
MozReview-Commit-ID: 8xAS79wTkPC
PrefsHelper uses Maybe<> as a replacement for a ternary expression, in
order to work around jni::StringParam's lack of copy constructor.
However, we can add a move constructor to StringParam, which lets us use
it in a ternary expression, and avoid the awkwardness of Maybe<>.
This removes the unnecessary setting of c-basic-offset from all
python-mode files.
This was automatically generated using
perl -pi -e 's/; *c-basic-offset: *[0-9]+//'
... on the affected files.
The bulk of these files are moz.build files but there a few others as
well.
MozReview-Commit-ID: 2pPf3DEiZqx
--HG--
extra : rebase_source : 0a7dcac80b924174a2c429b093791148ea6ac204
This patch replaces all references to |nsAutoPtr| in widget/ by references
to |UniquePtr|. |nsAutoPtr| is deprecated and will go away soon.
MozReview-Commit-ID: 8xAS79wTkPC
PrefsHelper uses Maybe<> as a replacement for a ternary expression, in
order to work around jni::StringParam's lack of copy constructor.
However, we can add a move constructor to StringParam, which lets us use
it in a ternary expression, and avoid the awkwardness of Maybe<>.
Add a loadUri method in GeckoView that replaces the functionality of the
current LOAD_URI event in GeckoEvent, and make GeckoApp and BrowserApp
use the new call.
The implementation for loadUri differs from the previous implementation
of LOAD_URI by directly calling nsIBrowserDOMWindow::OpenURI, instead of
going through the command-line handler. This more direct approach lets
us get rid of the Fennec command-line handler entirely.
Keep a pointer to nsPIDOMWindowOuter in GeckoViewSupport so that
GeckoView native method implementations can access the DOM window. This
also lets us simplify GeckoViewSupport::Close by calling
nsPIDOMWindowOuter::ForceClose directly.
Remove SENSOR_EVENT from GeckoEvent and implement it as a native method
in GeckoAppShell that is invoked by the sensor event listener in
GeckoAppShell.
The intention is that GeckoView consumers will not use browser.xul,
but instead geckoview.xul; and that they'll provide content URLs from
their assets/ directories. This lets them do things without modifying
the omnijar and in particular without setting Gecko's prefs *before*
Gecko startup.
MozReview-Commit-ID: 2QLQQ5RphFu
--HG--
extra : rebase_source : b80eb1b15a54dac3c19ba9c7bfe4f0ac7dae78c6
Currently, all widgets request selection change notifications to IMEContentObserver. Additionally, IMEContentObserver needs to listen selection changes for caching latest selection for eQuerySelectedText. Therefore, it doesn't make sense to keep defining nsIMEUpdatePreference::NOTIFY_SELECTION_CHANGE.
If widgets didn't need selection change notifications, they could just ignore the unnecessary notifications.
Note that all widgets don't need selection change notifications if a plugin has focus and IMEContentObserver cannot observe selection changes in the plugin. Therefore, if IMEContentObserver is initialized with a plugin, it shouldn't listen selection changes (and doesn't need to notify widgets of selection changes).
MozReview-Commit-ID: FOVFFgA2nOz
--HG--
extra : rebase_source : 3e16d5023835f99f82934e754d2e7db70474f9ee
In an expression such as:
const auto& x = cond() ? AClass(...) : AClass();
the C++ standard specifies that the copy constructor of AClass is
invoked on the result of the conditional expression ([expr.cond]p6).
GCC does not honor this part of the specification, whereas clang does;
clang therefore complains about instances of code such as:
const auto& jstrVal = type == widget::PrefsHelper::PREF_STRING ?
jni::StringParam(strVal, aPrefName.Env()) :
jni::StringParam(nullptr);
as jni::StringParam is not copy-constructable.
The simplest solution that does not introduce unnecessary allocation
uses mozilla::Maybe to hold the temporary objects and to hide some of
the details of constructing objects in-place. The compiler may even be
able to optimize away some of the unnnecessary checks that Maybe
introduces (e.g. checking for whether the Maybe is a Some or None at
certain points).
The new names Create{,PreservingTransform}OrNull() better communicate that
these functions (a) do object creation, and (b) are fallible.
--HG--
extra : rebase_source : a36bd9a2bcdfae281868959403f811f2bc690ad4
clang complains that it's unable to instantiate this template because
the functions being passed as arguments are MOZ_JNICALL, while the
actual parameter to the function has no such attribute. Adding the
attribute makes everything line up.
clang complains about specializations of NativeStubImpl not being able
to see the private constructor of ProxyNativeCall. While
ProxyNativeCall includes a friend declaration for NativeStubImpl, it's
not obvious *which* NativeStubImpl is being friended, as NativeStubImpl
hasn't been forward declared and exists in a completely separate
namespace. Forward declaring NativeStubImpl and adjusting the friend
declaration makes everything more correct.
clang complains that a constexpr definition of methods[] cannot refer to
members of the incomplete Impl template parameter, and rightly so.
Making these const is sufficient for our purposes, and that enables us
to move the declaration outside of the class, where it will be
instantiated lazily (presumably at the point when |Impl| is a complete
class definition). We also need to declare the length of methods[], as
other parts of the code require knowing the length of methods[] at
compile time.
For making our code clearer by the stronger type check, we should change the anonymous enum for NS_TEXTRANGE_* to enum class whose name is "TextRangeType" and whose type is "RawTextRangeType" which is an alias of uint8_t.
Additionally, this also adds some utility methods for them.
Note that some lines which are changed by this patch become over 80 characters but it will be fixed by the following patches.
MozReview-Commit-ID: 76izA1WqTkp
--HG--
extra : rebase_source : 27cd8cc8f7f8e82055dbfe82aba94c02beda5fa4
Even though a class may not be multi-threaded, its object references can
still be passed to different threads. As long as the references are not
accessed, we should allow it, and therefore GlobalRef should use
GetEnvForThread for ref/unref, instead of using FindEnv, which can
trigger assertions on non-main threads.
And mCharCode shouldn't be compared with NS_VK_*, nsIDOMKeyEvent::DOM_VK_*. Additionally, when it's compared with a character constant, cast isn't necessary.
MozReview-Commit-ID: JMT614copjG
--HG--
extra : rebase_source : 69ee3c589e5a71c814ec9a40ac3aab39c789c11d
And also WidgetKeyboardEvent::mKeyCode should be compared with NS_VK_* rather than nsIDOMKeyEvent::DOM_VK_*.
MozReview-Commit-ID: IKjQ1nr8XYe
--HG--
extra : rebase_source : 83125cd2523f6b70759f621470aad23b00aae8ae
WindowEvent is not specific to GeckoViewSupport, and classes other than
GeckoViewSupport use it, so this patch moves it to outside of
GeckoViewSupport.
The number of recent successive crashes is now tracked wholly within Java, so we can remove the old Gecko pref and the associated reset code.
MozReview-Commit-ID: 7bR9wqJsLoi
--HG--
extra : transplant_source : %A3%9C%89%87%E9Z%9B%C6%900%23%27%C1%CF%B7%DD1D%DD%FC
Clear any JS exceptions in appropriate places in NativeJSContainer, so
the exceptions don't affect subsequent JS API calls. We don't actually
"handle" the exceptions because we throw a Java exception instead or it
is safe to ignore the JS exception.
GeckoThread.waitOnGecko blocks the current thread and waits for previous
events on the Gecko thread to finish executing before continuing. This
is implemented by synchronously running a dummy event on the Gecko
thread. This patch also lets us get rid of sendEventToGeckoSync in
GeckoAppShell.
Jim Chen
Carsten "Tomcat" Book
Kartikaya Gupta
Alastor Wu
Randall Barker
Lee Salzman
Makoto Kato
Nicholas Nethercote
Wes Kocher
Kan-Ru Chen
John Lin
Botond Ballo
James Cheng
James Willcox
Dylan Roeh
Nicolas Silva
Thomas Zimmermann
Tom Tromey
Chris Peterson
Ryan Hunt
Nick Alexander
Bill McCloskey
Masayuki Nakano
Chih-Yi Leu
Eugen Sawin
Jeff Gilbert
Nathan Froyd
SUN Haitao
Jonathan Watt
Milan Sreckovic
Benoit Girard
Ryan VanderMeulen
David Anderson
Bob Owen
Neil Deakin
Tetsuharu OHZEKI
Kyle Huey
Sebastian Hengst
Jan Henning
Jamie Nicol
Nathan Froyd
Shatur
Mark Capella
Mouaad