GitLab

Since iOS 7, the iPhone 5s, the iPad Air and the iPad Mini Retina
supports 64-bits binaries. This requires building a fat binary with
support for both armv7 and arm64 (and optionally armv7s).

In an effort to prepare building such a fat binary for Chrome on
iOS, introduce a "target_subarch" variable that control the variant
we want to build (32-bit only, 64-bit only -- mainly for debugging,
fat binary).

This require introducing way to use the clang compiler shipped with
Xcode, as the clang compiler shipped with Chromium code does not
support "arm64" architecture yet (Apple has not yet opensource the
implementation).

BUG=341453, 314720
R=mark@chromium.org

Review URL: https://codereview.chromium.org/156473005

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@251083 0039d316-1c4b-4281-b951-d872f2087c98

This partially reverts the "cleanup" of r250260. The "cleaner" version is not equivalent if the call to -[NSWindow animations] returns nil.

BUG=none
TEST=no change

Review URL: https://codereview.chromium.org/148113011

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@251082 0039d316-1c4b-4281-b951-d872f2087c98

Previously we would check for unused args on a per-build basis. In the GYP build we run the build multiple times, and not all args might be declared in all types of builds. For example, clang args would only be relevant for clang builds, which might only be in the host GYP build.

Makes unused args errors nonfatal. This was coming up a lot in benign cases. It's important to notify people about unused arguments, but it's not critical to block progress on this problem.

R=scottmg@chromium.org

Review URL: https://codereview.chromium.org/137713007

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@251080 0039d316-1c4b-4281-b951-d872f2087c98

The manager will now wait 400ms after it receives both a request and its view
before showing a bubble.  This allows permissions that are requested
quickly in sequence to be shown together, in only one bubble.  Permission
requests that occur when there is already a bubble visible are queued and not
shown until the current bubble disappears.

This simple wait period is probably simpler than the final design, but is a
good stepping stone toward a better algorithm.

BUG=None

Review URL: https://codereview.chromium.org/150103013

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@251079 0039d316-1c4b-4281-b951-d872f2087c98

BUG=NONE

Review URL: https://codereview.chromium.org/138913011

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@251078 0039d316-1c4b-4281-b951-d872f2087c98

I'm planning on reusing the name test_support for something slightly
different: a dynamic library that can thunk to an implementation
provided by the test runner (much like we do for system and gles2). This
is so we can remove public/tests's base dependency (by injecting it
instead).

R=sky@chromium.org

Review URL: https://codereview.chromium.org/132353006

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@251077 0039d316-1c4b-4281-b951-d872f2087c98

BUG=None
R=darin@chromium.org, ben

Review URL: https://codereview.chromium.org/162213002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@251076 0039d316-1c4b-4281-b951-d872f2087c98

The old default causes weird scaling on the Chromebook Pixel. Try to use
the display size to set the scale.

TEST=content_shell with embedded=1 on pixel
TBR=rjkroege

Review URL: https://codereview.chromium.org/131373007

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@251075 0039d316-1c4b-4281-b951-d872f2087c98

This means no file can be opened for write during the test run.

R=vadimsh@chromium.org
NOTRY=true
BUG=116251

Review URL: https://codereview.chromium.org/137133002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@251074 0039d316-1c4b-4281-b951-d872f2087c98

The |nestable_tasks_allowed| parameter was always being set to true from the
callsites. So get rid of the variable, and use ScopedNestableTaskAllower in
the RunWithDispatcher implementations.

BUG=none
R=sky@chromium.org

Review URL: https://codereview.chromium.org/162633002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@251072 0039d316-1c4b-4281-b951-d872f2087c98

BUG=294477
TEST=browser_tests:CaptivePortalWindowCtorDtor*

Review URL: https://codereview.chromium.org/163413002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@251071 0039d316-1c4b-4281-b951-d872f2087c98

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@251070 0039d316-1c4b-4281-b951-d872f2087c98

BUG=338338

Review URL: https://codereview.chromium.org/157053010

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@251069 0039d316-1c4b-4281-b951-d872f2087c98

Because we draw things in reverse in Cocoa, we should iterate through the
avatar items in reverse too.

BUG=341972

Review URL: https://codereview.chromium.org/149603009

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@251068 0039d316-1c4b-4281-b951-d872f2087c98

Per discussion with ananta, opening a system modal dialog from a
background thread is a bad idea, so reverting this (from long ago).
Windows doesn't preserve the order correctly when the window is not on
the main UI thread, and in the bug here, it interacts very badly with
the IME toolbar window, causing the modal print dialog to go behind the
browser window in Z order making chrome appear to be hung.

The original goal was to fix tasks not being dispatched while the dialog
was open. We can instead use the ScopedNestedTaskAllower as we've done
for other system modal dialogs which I'll do in a separate CL.

BUG=342697,180997


> Use BaseShellDialog for print dialog on Windows
> 
> This puts the print dialog on a background thread which is necessary so other
> top level windows can keep painting as Aura does the compositor swaps on the
> UI thread.
> 
> R=sky@chromium.org,vitalybuka@chromium.org
> 
> 
> Review URL: https://codereview.chromium.org/27441003

TBR=scottmg@chromium.org,sky@chromium.org,vitalybuka@chromium.org,ananta@chromium.org

Review URL: https://codereview.chromium.org/164013002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@251066 0039d316-1c4b-4281-b951-d872f2087c98

NOTRY=True

Review URL: https://codereview.chromium.org/163883002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@251065 0039d316-1c4b-4281-b951-d872f2087c98

This patch simplifies the priority math calculation and moves it
into layers space. This makes the priority to be less correct,
since we're inverse transforming the viewport into layer space
instead of forward transforming tiles into screen space in all cases.

In particular, with perspective and rotated layers, we will be marking
more tiles as visible than before. This should be OK, since the amount
of those layers in practice is limited.

BUG=339142

Review URL: https://codereview.chromium.org/140513006

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@251064 0039d316-1c4b-4281-b951-d872f2087c98

> Move Windows-specific code into NativeAppWindowViewsWin
> 
> NativeAppWindowViews has Windows-specific code in #ifdefs. Move this
> into derived class NativeAppWindowViewsWin.
> 
> BUG=None
> 
> Review URL: https://codereview.chromium.org/161753002

Broke windows build:

c:\b\build\slave\win\build\src\chrome\browser\ui\views\apps\native_app_window_views_win.cc(36) : error C2027: use of undefined type 'base::SequencedWorkerPool'
        c:\b\build\slave\win\build\src\base\threading\thread_restrictions.h(82) : see declaration of 'base::SequencedWorkerPool'
c:\b\build\slave\win\build\src\chrome\browser\ui\views\apps\native_app_window_views_win.cc(36) : error C2227: left of '->RunsTasksOnCurrentThread' must point to class/struct/union/generic type

TBR=tmdiep@chromium.org

Review URL: https://codereview.chromium.org/163963002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@251063 0039d316-1c4b-4281-b951-d872f2087c98

NOTRY=True

Review URL: https://codereview.chromium.org/163873002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@251060 0039d316-1c4b-4281-b951-d872f2087c98

* Fix FD leak when reading output file.
* Update OS X version check.

BUG=None

Review URL: https://codereview.chromium.org/163183003

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@251059 0039d316-1c4b-4281-b951-d872f2087c98

BUG=313751
TBR=glider@chromium.org
NOTRY=true

Review URL: https://codereview.chromium.org/163783003

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@251058 0039d316-1c4b-4281-b951-d872f2087c98

BUG=294008
TBR=yoz@chromium.org
NOTRY=true

Review URL: https://codereview.chromium.org/163813002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@251056 0039d316-1c4b-4281-b951-d872f2087c98

BUG=313220
NOTRY=TRUE

Review URL: https://codereview.chromium.org/163673002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@251055 0039d316-1c4b-4281-b951-d872f2087c98

This means no file can be opened for write during the test run.

Windows can't be done at the moment due to bug 342913.

R=vadimsh@chromium.org
BUG=116251
BUG=342913

Review URL: https://codereview.chromium.org/136973003

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@251054 0039d316-1c4b-4281-b951-d872f2087c98

Review URL: https://codereview.chromium.org/121223002

BUG=324535

Review URL: https://codereview.chromium.org/121223002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@251053 0039d316-1c4b-4281-b951-d872f2087c98

Argument copying and lookups are a bit more complicated than they have to be.
This cleans up the code a little bit to make the trusted plugin more compact.

BUG=
R=dmichael@chromium.org

Review URL: https://codereview.chromium.org/161403002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@251052 0039d316-1c4b-4281-b951-d872f2087c98

Cleanups:
  Explicitly handle snapped case in SetChildBounds
  Use SetMaximizedOrFullscreenBounds in OnWindowAddedToLayout

BUG=none

Review URL: https://codereview.chromium.org/161293003

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@251051 0039d316-1c4b-4281-b951-d872f2087c98

I'm still tweaking the script to accommodate specific libraries. Rebuilding the entire list of libraries with every change is less than fun.

R=glider@chromium.org
BUG=313751,343515
NOTRY=true

Review URL: https://codereview.chromium.org/161133002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@251050 0039d316-1c4b-4281-b951-d872f2087c98

WebRTC browser tests is based on Manycam virtual capture device,
but this does not work with AVFoundation, see bug.

Termporarily, disable AVFoundation for those tests, hence using
QTKit capture Framework.

BUG=327618, 288562

Review URL: https://codereview.chromium.org/163243002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@251049 0039d316-1c4b-4281-b951-d872f2087c98

BUG=343006

Review URL: https://codereview.chromium.org/132793014

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@251048 0039d316-1c4b-4281-b951-d872f2087c98

c1cb193 disable NEON for arm64 platform

BUG=314720, 342387
R=urvang@google.com
TBR=cpu@chromium.org

Review URL: https://codereview.chromium.org/153273005

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@251047 0039d316-1c4b-4281-b951-d872f2087c98

Use GCC's __builtin_expect() branch prediction hint to tell the compiler
that trace events are likely to be disabled. This lets the compiler
optimize trace event shim code embedded in functions by moving it
outside the critical path and by annotating it with hints for the CPU's
runtime static branch predictor (depending on the architecture).

The exact performance impact of this change is difficult to quantify,
but the disassembly shows improvements to code locality. For example,
cc::ThreadProxy::CompositeAndReadback, which begins with a call to
TRACE_EVENT0(), looks like this before the change (x64):

00000000004bc2e0 <_ZN2cc11ThreadProxy20CompositeAndReadbackEPvRKN3gfx4RectE>:
  [... prologue ...]
  4bc30a:	48 8d 35 e7 e7 2d 00 	lea    0x2de7e7(%rip),%rsi        # 79aaf8 <.L.str>
  4bc311:	48 8d 15 8b 60 2e 00 	lea    0x2e608b(%rip),%rdx        # 7a23a3 <.L.str7>
  4bc318:	48 8d bc 24 f0 03 00 	lea    0x3f0(%rsp),%rdi
  4bc31f:	00

  /* Initialize scoped trace memory tracker. */
  4bc320:	e8 7b 19 0a 00       	callq  55dca0 <_ZN4base5debug17ScopedTraceMemory10InitializeEPKcS3_>

  /* Have seen this trace event before? */
  4bc325:	48 8b 1d 94 78 3d 00 	mov    0x3d7894(%rip),%rbx        # 893bc0 <_ZZN2cc11ThreadProxy20CompositeAndReadbackEPvRKN3gfx4RectEE28trace_event_unique_atomic149>
  4bc32c:	48 85 db             	test   %rbx,%rbx

  /* No? Better initialize it then */
  4bc32f:	75 16                	jne    4bc347 <_ZN2cc11ThreadProxy20CompositeAndReadbackEPvRKN3gfx4RectE+0x67>
  4bc331:	48 8d 3d c0 e7 2d 00 	lea    0x2de7c0(%rip),%rdi        # 79aaf8 <.L.str>
  4bc338:	e8 63 6f 09 00       	callq  5532a0 <_ZN4base5debug8TraceLog23GetCategoryGroupEnabledEPKc>
  4bc33d:	48 89 c3             	mov    %rax,%rbx
  4bc340:	48 89 1d 79 78 3d 00 	mov    %rbx,0x3d7879(%rip)        # 893bc0 <_ZZN2cc11ThreadProxy20CompositeAndReadbackEPvRKN3gfx4RectEE28trace_event_unique_atomic149>

  /* Trace event has been initialized. Is its category enabled? (atomic load) */
  4bc347:	48 c7 84 24 d0 03 00 	movq   $0x0,0x3d0(%rsp)
  4bc34e:	00 00 00 00 00
  4bc353:	f6 03 05             	testb  $0x5,(%rbx)

  /* Trace category is disabled => jump to rest of function. */
  4bc356:	0f 84 a8 00 00 00    	je     4bc404 <_ZN2cc11ThreadProxy20CompositeAndReadbackEPvRKN3gfx4RectE+0x124>

  /*
   * Trace category is enabled => compute timestamp and arguments and
   * append the event to trace log.
   */
  4bc35c:	e8 df cc 0b 00       	callq  579040 <_ZN4base14PlatformThread9CurrentIdEv>
  4bc361:	4d 89 e6             	mov    %r12,%r14
  4bc364:	41 89 c4             	mov    %eax,%r12d
  4bc367:	e8 a4 3d 0c 00       	callq  580110 <_ZN4base9TimeTicks22NowFromSystemTraceTimeEv>
  4bc36c:	48 89 84 24 80 00 00 	mov    %rax,0x80(%rsp)
  4bc373:	00
  4bc374:	e8 47 69 09 00       	callq  552cc0 <_ZN4base5debug8TraceLog11GetInstanceEv>
  4bc379:	48 8d 8c 24 80 00 00 	lea    0x80(%rsp),%rcx
  4bc380:	00
  4bc381:	48 89 0c 24          	mov    %rcx,(%rsp)
  4bc385:	c7 44 24 30 00 00 00 	movl   $0x0,0x30(%rsp)
  4bc38c:	00
  4bc38d:	48 c7 44 24 28 00 00 	movq   $0x0,0x28(%rsp)
  4bc394:	00 00
  4bc396:	48 c7 44 24 20 00 00 	movq   $0x0,0x20(%rsp)
  4bc39d:	00 00
  4bc39f:	48 c7 44 24 18 00 00 	movq   $0x0,0x18(%rsp)
  4bc3a6:	00 00
  4bc3a8:	48 c7 44 24 10 00 00 	movq   $0x0,0x10(%rsp)
  4bc3af:	00 00
  4bc3b1:	c7 44 24 08 00 00 00 	movl   $0x0,0x8(%rsp)
  4bc3b8:	00
  4bc3b9:	48 8d 2d e3 5f 2e 00 	lea    0x2e5fe3(%rip),%rbp        # 7a23a3 <.L.str7>
  4bc3c0:	be 58 00 00 00       	mov    $0x58,%esi
  4bc3c5:	45 31 c0             	xor    %r8d,%r8d
  4bc3c8:	48 89 c7             	mov    %rax,%rdi
  4bc3cb:	48 89 da             	mov    %rbx,%rdx
  4bc3ce:	48 89 e9             	mov    %rbp,%rcx
  4bc3d1:	45 89 e1             	mov    %r12d,%r9d
  4bc3d4:	4d 89 f4             	mov    %r14,%r12
  4bc3d7:	e8 24 a4 09 00       	callq  556800 <_ZN4base5debug8TraceLog37AddTraceEventWithThreadIdAndTimestampEcPKhPKcyiRKNS_9TimeTicksEiPS5_S3_PKyPK13scoped_refptrINS0_24ConvertableToTraceFormatEEh>

  [... actual function code ...]

After the change:

00000000004bcab0 <_ZN2cc11ThreadProxy20CompositeAndReadbackEPvRKN3gfx4RectE>:
  [... prologue ...]
  4bcad8:	48 8d 35 89 f0 2d 00 	lea    0x2df089(%rip),%rsi        # 79bb68 <.L.str>
  4bcadf:	48 8d 15 2d 69 2e 00 	lea    0x2e692d(%rip),%rdx        # 7a3413 <.L.str7>
  4bcae6:	48 8d bc 24 f0 03 00 	lea    0x3f0(%rsp),%rdi
  4bcaed:	00

  /* Initialize scoped trace memory tracker. */
  4bcaee:	e8 fd 20 0a 00       	callq  55ebf0 <_ZN4base5debug17ScopedTraceMemory10InitializeEPKcS3_>

  /* Have seen this trace event before? */
  4bcaf3:	48 8b 1d c6 80 3d 00 	mov    0x3d80c6(%rip),%rbx        # 894bc0 <_ZZN2cc11ThreadProxy20CompositeAndReadbackEPvRKN3gfx4RectEE28trace_event_unique_atomic149>
  4bcafa:	48 85 db             	test   %rbx,%rbx

  /* No? Jump to the initialization code at the end of the function. */
  4bcafd:	0f 84 7e 03 00 00    	je     4bce81 <_ZN2cc11ThreadProxy20CompositeAndReadbackEPvRKN3gfx4RectE+0x3d1>

  /* Is the trace category enabled? (atomic load) */
  4bcb03:	48 c7 84 24 d0 03 00 	movq   $0x0,0x3d0(%rsp)
  4bcb0a:	00 00 00 00 00
  4bcb0f:	f6 03 05             	testb  $0x5,(%rbx)

  /* If so, jump to the event appending code at the end of the function. */
  4bcb12:	0f 85 84 03 00 00    	jne    4bce9c <_ZN2cc11ThreadProxy20CompositeAndReadbackEPvRKN3gfx4RectE+0x3ec>

  [... actual function code ...]

  [... code for initializing trace event...]

  [... code for appending trace event...]

In other words, the trace event related code has been moved outside the
main function code flow. This should improve code cache utilization.

Note that this change currently has no effect on ARM, since
__builtin_expect() is a no-op because of -Os and -mthumb. Also, despite
being a GCC extension, __builtin_expect() is also supported by Clang.

Review URL: https://codereview.chromium.org/160673002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@251046 0039d316-1c4b-4281-b951-d872f2087c98

This change introduces the prerequisites for the upcoming Android
backend tracing classes of the memory_inspector tool. In particular:
 - The mock_adb binary, which will be used by the unittests to mock
   adb responses.
 - The prebuilts_fetcher.py which pulls and syncs the prebuilts from
   GCS (in a similar way to what telemetry currently does).

BUG=340294
NOTRY=true

Review URL: https://codereview.chromium.org/158993002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@251045 0039d316-1c4b-4281-b951-d872f2087c98

When the supervised user changes his avatar, the change is
synced to his other devices, as well as to the settings of
the custodian who sees the updated avatar when opening the
supervised user import overlay.

BUG=339060
TEST=unit_tests

Review URL: https://codereview.chromium.org/147083016

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@251044 0039d316-1c4b-4281-b951-d872f2087c98

Didn't finish for M34, hiding the flag for the cut.

BUG=341556

Review URL: https://codereview.chromium.org/162833002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@251043 0039d316-1c4b-4281-b951-d872f2087c98

This is a non-perfect, temporary fix for the bug. It should be revisited after the M34 cut. There's e.g. discussions about tying this to getUserMedia. Bug for this: https://code.google.com/p/chromium/issues/detail?id=342770

BUG=342689

Review URL: https://codereview.chromium.org/152923005

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@251041 0039d316-1c4b-4281-b951-d872f2087c98

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@251040 0039d316-1c4b-4281-b951-d872f2087c98

This is the client side part, the server config is in CL 61376672.

BUG=338260,338289

Review URL: https://codereview.chromium.org/149503006

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@251038 0039d316-1c4b-4281-b951-d872f2087c98

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@251037 0039d316-1c4b-4281-b951-d872f2087c98

TBR=dsinclair

Review URL: https://codereview.chromium.org/162883003

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@251036 0039d316-1c4b-4281-b951-d872f2087c98