All our widgets support it with a constant true.
MozReview-Commit-ID: JMEItUsxYWq
--HG--
extra : rebase_source : e7e0a3f83001813239338bc5b3895252e1fb3ea6
We have a lot of crash reports in OnEventNeedingAckHandled() due to unexpected
state (hit MOZ_RELEASE_ASSERT). However, it's unclear what occurs and we're not
sure there are how many cases to crash because the stack trace is too short
because the method is called when TabParent receives event handled message from
the remote process. I.e., it doesn't show what happens immediately before the
crash.
This patch puts 2 sets of composition events to app notes of crash report when
it needs to crash. This *might* make damage to the performance. If so, after
fixing the crashes, we should back this out. Fortunately, we have a lot of
reports from either Nightly or Beta.
MozReview-Commit-ID: 9tDrEIf72MG
--HG--
extra : rebase_source : 523c183466740e08d6c8cc3836b6b52310c1e53a
Currently the Gecko Profiler defines a moderate amount of stuff when
MOZ_GECKO_PROFILER is undefined. It also #includes various headers, including
JS ones. This is making it difficult to separate Gecko's media stack for
inclusion in Servo.
This patch greatly simplifies how things are exposed. The starting point is:
- GeckoProfiler.h can be #included unconditionally;
- everything else from the profiler must be guarded by MOZ_GECKO_PROFILER.
In practice this introduces way too many #ifdefs, so the patch loosens it by
adding no-op macros for a number of the most common operations.
The net result is that #ifdefs and macros are used a bit more, but almost
nothing is exposed in non-MOZ_GECKO_PROFILER builds (including
ProfilerMarkerPayload.h and GeckoProfiler.h), and understanding what is exposed
is much simpler than before.
Note also that in BHR, ThreadStackHelper is now entirely absent in
non-MOZ_GECKO_PROFILER builds.
(Path is actually r=froydnj.)
Bug 1400459 devirtualized nsIAtom so that it is no longer a subclass of
nsISupports. This means that nsAtom is now a better name for it than nsIAtom.
MozReview-Commit-ID: 91U22X2NydP
--HG--
rename : xpcom/ds/nsIAtom.h => xpcom/ds/nsAtom.h
extra : rebase_source : ac3e904a21b8b48e74534fff964f1623ee937c67
Removes the XPCOM interface for nsIDOMHTMLCanvasElement, replacing it
with binding class usage.
MozReview-Commit-ID: DQJhqGlY8U6
--HG--
extra : rebase_source : a33146a22ee356a0840bb9fef82d51b2b315f6d7
When "mozAwesomebar" is set to inputmode value, that means that the Smart
Location Bar gets focus. In that case, we should notify IME of input scopes
as "URL" because on-screen keyboard for URL has some useful additional keys
but they are not hindrances even when users want to type non-URL text.
On the other hand, MS-IME for Japanese and Google Japanese Input changes their
open state to "closed" if we notify them of URL input scope. A lot of users
complain about this behavior. Therefore, we should notify only them of
"Default" input scope even when "mozAwesomebar" has focus.
MozReview-Commit-ID: DIgqpR7TXQx
All our widgets support it with a constant true.
MozReview-Commit-ID: JMEItUsxYWq
--HG--
extra : rebase_source : a2661dce1ac191fdf098e631cd7878f0215643d5
This warning is saying the operation may be undefined because the value of
->left is not guarenteed to be the same because of undefined order of operations.
Fortunately, this seems like a typo and we actually meant to assign ->bottom
MozReview-Commit-ID: H5G8fnDwIJP
--HG--
extra : rebase_source : cd6a2463a90b6675b12bb8255fe605937771bee8
When we UnregisterAudioSessionNotification in response to an OnSessionDisconnected
message, we unintentionally decrement the refcount before restoring it. This KungFu
grips us for the duration of that operation.
--HG--
extra : rebase_source : 7c7e786064811d040d23741c143d9eef04edb5e0
IMEStateManager has a static instance of InputContext which has some nsString
instances. Then, their members will be released after XPCOM shutdown and
that is detected as unexpected destruction.
Therefore, IMEStateManager should release them by itself at XPCOM shutdown.
Using nsTSubstring::SetCapacity(0) is the simplest way to release only the
string buffers.
MozReview-Commit-ID: LMrQxQF9xPn
--HG--
extra : rebase_source : 7cbaf19b76be8db083c10270d0ee436a1c061ea2
Avoid going through GeckoAppShell and move the start child process JNI
call directly to GeckoProcessManager.
MozReview-Commit-ID: KU62TiHVQJX
--HG--
extra : rebase_source : 0e8546da502257e1c59bc00b79f50c79a314f3e6
This patch merges nsAtom into nsIAtom. For the moment, both names can be used
interchangeably due to a typedef. The patch also devirtualizes nsIAtom, by
making it not inherit from nsISupports, removing NS_DECL_NSIATOM, and dropping
the use of NS_IMETHOD_. It also removes nsIAtom's IIDs.
These changes trigger knock-on changes throughout the codebase, changing the
types of lots of things as follows.
- nsCOMPtr<nsIAtom> --> RefPtr<nsIAtom>
- nsCOMArray<nsIAtom> --> nsTArray<RefPtr<nsIAtom>>
- Count() --> Length()
- ObjectAt() --> ElementAt()
- AppendObject() --> AppendElement()
- RemoveObjectAt() --> RemoveElementAt()
- ns*Hashtable<nsISupportsHashKey, ...> -->
ns*Hashtable<nsRefPtrHashKey<nsIAtom>, ...>
- nsInterfaceHashtable<T, nsIAtom> --> nsRefPtrHashtable<T, nsIAtom>
- This requires adding a Get() method to nsRefPtrHashtable that it lacks but
nsInterfaceHashtable has.
- nsCOMPtr<nsIMutableArray> --> nsTArray<RefPtr<nsIAtom>>
- nsArrayBase::Create() --> nsTArray()
- GetLength() --> Length()
- do_QueryElementAt() --> operator[]
The patch also has some changes to Rust code that manipulates nsIAtom.
MozReview-Commit-ID: DykOl8aEnUJ
--HG--
extra : rebase_source : 254404e318e94b4c93ec8d4081ff0f0fda8aa7d1
- Count open file descriptors instead of testing the existence of a file
(because the clipboard is now only reachable through file descriptors,
and not through a file path).
- Use a fixed string instead of a random string. The previous way of
generating a string was non-deterministic, and there was a very small
chance that the generated string was not large enough to trigger the
cache-to-disk-mode.
MozReview-Commit-ID: KOkYOm280Oh
--HG--
extra : rebase_source : 32f1eb0c5beb4f8ebb1728bda481ded21247e0ec
Ensure that only DataStruct::mData + mDataLen, XOR
DataStruct::mCacheFD is used.
(Previously it was possible that all of these members were populated,
which is a waste of memory.)
The effect of this change is visible when SetTransferData is called
multiple times with the same flavor, but with one below the threshold
for storing in-memory, and the other above (=store in a file).
MozReview-Commit-ID: 4UlkKAYsjf
--HG--
extra : rebase_source : fff484301dc87f12685575084993b6bf877f9a9b
The cache file is never directly exposed to consumers of DataStruct,
so it does not make sense to keep the clipboardcache file around
forever.
The only change in this commit is to switch from using a filename to
using a file descriptor. In the destructor, the FD is explicitly closed
(which releases the file data). nsAnonymousTemporaryFile takes care
of removing the file when the destructor is not called (e.g. crashes).
Previously, the clipboard cache was stored in a file called:
TmpD/clipboardcache-N
As of this commit, the clipboard cache is stored at:
TmpD/mozilla-temp-randomN (macOS and Linux)
TmpD/mozilla-temp-files/mozilla-temp-randomN (Windows)
(see xpcom/io/nsAnonymousTemporaryFile.{h,cpp} for more details)
To verify that these files are really gone:
1. Create a document with 500k+ characters, open it in Firefox.
2. Copy its content - this will trigger the clipboard cache.
3. Look for the open file descriptor of the deleted file:
( macOS and Linux: )
lsof +L1 | grep mozilla-temp
4. Copy anything (under the 500k threshold), or quit/kill Firefox.
5. Repeat step 3 and observe that the number of file descriptors
has decreased.
MozReview-Commit-ID: 85GlKQrNUl5
--HG--
extra : rebase_source : b0c6e9540e53ad5e713a9db55f3001d7f53c308a
It has no users in mozilla-central or comm-central. This also lets us remove
nsBaseWidget::GetCursor().
--HG--
extra : rebase_source : 3875ce37c78f82285c4e5dbd8133e9ae3dafa0de
Because nsAString is nicer to work with than char16_t*. The patch relatedly
changes nsIEmbeddingSiteWindow::title and nsIWindowMediator::updateWindowTitle
as well.
--HG--
extra : rebase_source : 0bf332dec3e09af6c39c676f8795b368768a6046
The NS_LITERAL_STRING macro creates a temporary nsLiteralString to encapsulate the char16_t string literal and its length, but AssignLiteral() can determine the char16_t string literal's length at compile-time without nsLiteralString.
MozReview-Commit-ID: 6vgQiU8zN3o
--HG--
extra : rebase_source : 1b536b92ef43f610db057ace6f108620e8d8b4d5
extra : source : 336e21386d5eeb16f1c9893c29377f23b67cc4b0
The problem happened only on themed widget. The problem was triggered by Bug 888870. It just added Windows 8.1 support. When high contrast mode was enabled on Win 10, background color problem happend on themed widgets. There were 2 patterns.
- After hover the menu item, its background color remained highlighted
- After hover the menu item, its background color became black
From it, the problem seemed to be caused by background color drawing of themed widgets. nsNativeThemeWin::DrawWidgetBackground() does the background color drawing. AssumeThemePartAndStateAreTransparent() controls skipping background color drawing of themed widgets. If AssumeThemePartAndStateAreTransparent() was removed, the problem was addressed. From it, how DrawThemeBackground() works seems to be changed on high contrast mode since since Win8.1. To address the problem, the patch remove the skipping on high contrast mode since since Win8.1
Add a way to append app notes to the crash report from Java, so we can
associate additional crash data with the crash.
MozReview-Commit-ID: 6I6pGOWb1sZ
In order to let necko postpone the load of favicon, we have to set request context ID to the http channel that is created to load favicon.
This patch starts with passing a request context ID to nsContentUtils::LoadImage and makes other necessary changes to set the request context ID to the channel.
For layers-full mode, we set the backface-visibility to visible because
visibility would be handled by FLB and layers.
MozReview-Commit-ID: CUbeUabfC7K
For layers-full mode, we set the backface-visibility to visible because
visibility would be handled by FLB and layers.
MozReview-Commit-ID: CUbeUabfC7K
--HG--
extra : rebase_source : ae88579f3718884bd4ddd89dfc3885bada3f15a4
dcWidth and dcHieght are used only once, we do not need to define a local
variable for any of them.
MozReview-Commit-ID: Fw5rHqtUPQN
--HG--
extra : rebase_source : 52b5381319fc6626e7f1ec4111dad70482a092cd
Currently, WindowsEMF::InitForDrawing can be used once. With the change in this
patch, we can call WindowsEMF::InitForDrawing and Playback/SaveToFile in pair
as many times as we want.
MozReview-Commit-ID: 4fbY4Q6i9v5
--HG--
extra : rebase_source : 4df3fc7d8663d7f552c96204a16d73261bfc9dd7
Englightenment closes fd 0 on child processes and so pipe() can return a zero fd.
MozReview-Commit-ID: 5d9xQXgwgfv
--HG--
extra : rebase_source : c31aa7ce731ba325993f463b79b446ae67c932dd
CopySurface doesn't respect any existing clips that are applied to the draw
target. It doesn't look like nsNativeThemeGTK uses CopySurface normally, but
when OMTP is enabled the draw target is a DrawTargetCapture which can't
LockBits() which causes us to use CopySurface. This can cause a reftest failure.
MozReview-Commit-ID: 1mYMhQB0r3M
--HG--
extra : rebase_source : d2683dee37ce32049d8daeee786cbb06ba624438
extra : amend_source : 59c48b66f733108636d52047a55889ee0c4e2746
Because UBSan complains about casting -1:
> runtime error: load of value 4294967295, which is not a valid value for type 'nsCursor'
--HG--
extra : rebase_source : 037a96700228ea0d427afa7c25c40490c701cdc4
Due to the indeterminate nature of Gecko shutdown, the ref count of the
nsWindow on Android would sometimes go to zero before the XPCOM shutdown
observer was called in nsBaseWindow which is where the compositor thread
IPC is shutdown. If nsBaseWindow::Shutdown does not get called, then the
compositor thread IPC is shutdown in the nsBaseWindow destructor.
Unfortunately while the nsWindow is being deleted, it can be accessed in
the compositor thread and cause a crash in
LayerManagerComposite::RenderToPresentationSurface.
Calling nsBaseWidget::DestroyLayerManager() in nsWindow::Destroy()
ensures it gets called before the destructor gets invoked typically in
the Destroy() call.
MozReview-Commit-ID: KCv8SCmEjnb
--HG--
extra : rebase_source : f48fe01b15ff7e2e8763628dd22a2c926d051e9d
Due to content sandbox, NSPrintJobSavingURL will throws an exception on content process. So we don't use NSPrintJobSavingURL on content process and print.print_via_parent=true.
Also, even if chrome process, SetToFileName should store file name to mToFileName. Because SetToFileName uses mToFileName.
MozReview-Commit-ID: J7vlOG0gyGP
--HG--
extra : rebase_source : dfda2f625ca9dabefe131e40424609a2b921d430
A nsPrintOptionsImpl has char16_t internal variables, and gecko will not release
these allocated variable. So we should use nsString instead of char16_t.
MozReview-Commit-ID: 4ZMtUKoVCcR
--HG--
extra : rebase_source : 1720483301b900f8135c3d5f8f32d7051191aa8f
We'll need it to determine if NV12 D3D11 surface are operationals.
MozReview-Commit-ID: FvsxGhnv7H1
--HG--
extra : rebase_source : 4580819b19aa36d7d26ee7d05934eaaddd17206a
Currently, we initialize TSF modules when there is only message window (this
started from bug 1341915). At this time, QQ Input (Simplified Chinese TIP)
fails to initialize itself.
Therefore, we should put off to initialize TSF modules after creating first
normal window. Then, initialize its IMC and input context for the window.
IMEHandler::InitInputContext() should be called when each normal window is
created. Therefore, calling Initialize() from it can guarantee there is
at least one normal window when Initialize() is called.
MozReview-Commit-ID: IfR4y3pYv6J
--HG--
extra : rebase_source : 3f36acfa65567f140eea9c7fc12d6268aadbd5ef
Additionally, do not call RegisterDragDrop for hidden windows.
MozReview-Commit-ID: Fv8j9FntGGT
--HG--
extra : rebase_source : fccbc576887d000805da8ac915d16fc4fddfeceb
This fixes improper usages of Find where an offset was actually being use for
the boolean ignore case flag. It also fixes a few instances of passing in a
literal wchar_t to our functions where a NS_LITERAL_STRING or char16_t should
be used instead.
This fixes improper usages of Find where an offset was actually being use for
the boolean ignore case flag. It also fixes a few instances of passing in a
literal wchar_t to our functions where a NS_LITERAL_STRING or char16_t should
be used instead.
--HG--
extra : rebase_source : 5de1e9335895d65e6db06c510e8887d27be3390f
extra : source : f762f605dd83fc6331161a33e1ef5d54cafbd08d
The kEnableSelectionRB flag is already set in nsPrintEngine code and the generic
nsPrintOptions::SerializeToPrintData() serializes this flags, so we don't need
to check NSPrintSelectionOnly and we don't need to override the kEnableSelectionRB
setting in nsPrintOptionsX::SerializeToPrintData().
MozReview-Commit-ID: I3iYdx5OQCp
--HG--
extra : rebase_source : e66c0bf2538b546e4778b7b2ea640642cedb3aca
GTK should be used only from the main thread, but the specific symptom before
this change was that creating a GtkEntry causes
pango_cairo_font_map_get_default() to be called. This function returns a
different font map on each thread. The font map is leaked when StyleThreads
are leaked at exit.
Font caches are usually expensive and so using the existing font map on the
main thread is preferable.
A GtkEntry already exists on the main thread, as well as style contexts for
most other system fonts, and so it is more efficient to create these on the
main thread while the style contexts exist.
Doing this also avoids the need for Gecko_nsFont_InitSystem() to hold a global
lock to avoid concurrently calling into GTK through LookAndFeel::GetColor().
MozReview-Commit-ID: DSOwEUeYmtV
--HG--
extra : rebase_source : 971bc565e465b2d2482308006f68451ecaa6e264
This old code no longer builds and is superseded by the GTK3 port.
See also bug 1278282.
MozReview-Commit-ID: 9KyrPuNAluv
--HG--
extra : rebase_source : 8d5de061743e0281bf4ed03c68286512e7b492e9
If a keydown event handler moves focus like Ctrl+PageDown handler, IM context
may be changed to DISABLED or something. In such case, native IME would stop
current composition because focus moving in Gecko causes making IME blurred.
However, IMContextWrapper::DispatchCompositionStart() always dispatches
eCompositionStart even in such case.
So, it should stop dispatching eCompositionStart if IME enabled state is
changed during dispatching the preceding keydown event.
Note that this patch moves the setter of mComposingContext from
OnStartCompositionNative() which is a signal listener of "preedit_start" to
DispatchCompositionStart() because if IME starts composition without
"preedit_start" signal, DispatchCompositionStart() will be called but
OnStartCompositionNative() isn't called. However, this fix needs
mComposingContext.
MozReview-Commit-ID: F3F6NuCOrkJ
--HG--
extra : rebase_source : 513528eba0f29eb9b6ce8c5f47e4badbde9cbdb8
Doing this avoids -Wundefined-var-template warnings with clang. This
warning is mostly harmless (clang is trying to tell you a linker error
might be awaiting you later), but being careful with this might make
using C++ modules easier somewhere down the line.
clang seems to think that `Base` is unused despite its obvious uses
inside this class. We can deal with the `Base::` qualifications for
name lookup by using `this->` instead, and the other uses of `Base::`
can simply use the explicit name of the superclass.
Prevent default on pointerdown will stop firing the subsequent mouse events. Ignore the case that preventDefault by WebExtensions to avoid breaking some websites.
MozReview-Commit-ID: 9ztW1WfEg9a
With this change, we are able to load PDFium engine from different library.
MozReview-Commit-ID: ErAZCPRzRR5
--HG--
extra : rebase_source : f0421af9c70d69564ed8f96d32870314871aabef
* OSX
Make the lock of the type kIOPMAssertionTypeNoDisplaySleep and kIOPMAssertionTypeNoIdleSleep
as a singleton. Won't need to require an extra lock.
* Windows
Add |mRequireForDisplay| to ensure the "audio-playing" won't overwrite the previous
display requirement.
* Android
Add "audio-playing" and "video-playing", and make sure the audio-lock won't be cancel
when receiving "WakeLockDelegate.STATE_LOCKED_BACKGROUND".
MozReview-Commit-ID: 97oNX7H2qij
--HG--
extra : rebase_source : 24fa8b267ad97d668fa55462d1f61ef5c92b632f
According to [1], kIOPMAssertionTypeNoDisplaySleep prevents display sleep and
kIOPMAssertionTypeNoIdleSleep prevents idle sleep.
We should use kIOPMAssertionTypeNoIdleSleep for audio playing, because it won't
need to block the display.
[1] https://developer.apple.com/library/content/qa/qa1340/_index.html
---
On Windows, ES_DISPLAY_REQUIRED forces display on, and ES_SYSTEM_REQUIRED forces
system on working state [2].
[2] https://msdn.microsoft.com/zh-tw/library/windows/desktop/aa373208(v=vs.85).aspx
MozReview-Commit-ID: Izs29PdzQOW
--HG--
extra : rebase_source : d0726131735b9cedf566f937204e585345b9d5e3
For knowing the wake lock usage more clearly, we should use more specific topic name.
In OSX, you can use "$ pmset -g assertions" to check all the wakelock.
In Windows, using "$ powser -energy" to generate the energy report.
MozReview-Commit-ID: rAXnkxTvLc
--HG--
extra : rebase_source : 42ebf204673d3c913739f64c71c24af20d37c95d
Bug 1354020 causes that IMM-IME on Windows 7 doesn't work with --no-remote. Although I think that this issue is OS or IME bug, when default window proceduce by RegisterClass is DefaultWindowProcW, Google Japanese IME doesn't work.
I am not sure why this issue occurs when lpfnWndProc is DefWidnowProcW and DDE isn't started. But for workaround, we should revert a part of bug 1354020 changes.
MozReview-Commit-ID: BkxlZnm8mIh
--HG--
extra : rebase_source : 21c1d86fb5628dbd726a15e5057a536271cfb26d
This is straightforward, with only two notable things.
- `#include "nsXPIDLString.h" is replaced with `#include "nsString.h"`
throughout, because all nsXPIDLString.h did was include nsString.h. The
exception is for files which already include nsString.h, in which case the
patch just removes the nsXPIDLString.h inclusion.
- The patch removes the |xpidl_string| gtest, but improves the |voided| test to
cover some of its ground, e.g. testing Adopt(nullptr).
--HG--
extra : rebase_source : 452cc4a08046a1adb1a8099a7e85a1917de5add8
We should not be declaring forward declarations for nsString classes directly,
instead we should use nsStringFwd.h. This will make changing the underlying
types easier.
--HG--
extra : rebase_source : b2c7554e8632f078167ff2f609392e63a136c299
Emilio Cobos Álvarez
Masayuki Nakano
Sebastian Hengst
Ryan Hunt
Nicholas Nethercote
Stephen A Pohl
Kyle Machulis
Tom Ritter
Louis Chang
John Lin
Wes Kocher
Stone Shih
James Willcox
Catalin Badea
Eugen Sawin
David Parks
Lee Salzman
Markus Stange
Milan Sreckovic
Jim Chen
Mats Palmgren
Jonathan Watt
Rob Wu
Chris Peterson
sotaro
Chris Coulson
Kershaw Chang
Morris Tseng
Brad Werth
cku
Karl Tomlinson
sebastian@ifyouwantblood.de
Randall Barker
Makoto Kato
Andrea Marchesini
Mantaroh Yoshinaga
Jean-Yves Avenard
Jim Mathies
Ryan VanderMeulen
Kirk Steuber
Eric Rahm
Will Wang
Josh Aas
Nathan Froyd
David Anderson
Nicolas Silva
Haik Aftandilian
Jan Steffens
Adam Gashlin
Sylvestre Ledru
Alastor Wu
Martin Stransky
aceman
Jessica Jong
Farmer Tseng