GitLab

template, and if we stop always passing some files in chrome/ on every
invocation it is possible to generalize it and move it to build/util
instead (and fix some layering violation comments along the way).

Compared to its previous incarnation, process_version() no longer passes
LASTCHANGE, BRANDING and VERSION automatically to version.py. This makes
it easier for callers to specify their own |sources| with values that
may override those set in those 3 files.

A new wrapper, process_version_rc_template(), was introduced to cater
for callers (generally Windows ones) who need to process a .rc.version
file that requires the values from LASTCHANGE, BRANDING and VERSION. It
always passes those 3 files to process_version() before any optional
additional |sources|, and |template_file| defaults to
chrome_version.rc.version.
Since this template depends on files from //chrome, it lives in the
newly-added //chrome/process_version_rc_template.gni.

R=dpranke@chromium.org,brettw@chromium.org

Review-Url: https://codereview.chromium.org/2308313003
Cr-Commit-Position: refs/heads/master@{#418514}

ExceptionState is now a scope-like object, and guarantees
that an exception is automatically thrown at destruction.
ExceptionToPromiseScope similarly guarantees to return
a reject promise instead of throwing an exception.

A new Jinja2 filter "format_remove_duplicates" helps
write a template more easily without caring duplicates.
    {% if A_enabled %}
    ScriptState* scriptState = ...;
    A(scriptState);
    {% endif %}
    {% if B_enabled %}
    ScriptState* scriptState = ...;
    B(scriptState);
    {% endif %}
produces two lines of definitions of |scriptState|,
however, format_remove_duplicates removes the duplicates.

Jinja2 macro "propagate_error_with_exception_state" is no
longer necessary, because ExceptionState and
ExceptionToPromiseScope automatically throws an exception
or returns a reject promise appropriately.

This CL alone does not fix crbug.com/627309 , but the CL
makes it easy to fix the issue.

BUG=627309

Review-Url: https://codereview.chromium.org/2301993002
Cr-Commit-Position: refs/heads/master@{#418513}

|context_| has already been created in MojoChildConnection constructor
initialization list, no need to create it again in construtor body.

BUG=none

Review-Url: https://codereview.chromium.org/2338753002
Cr-Commit-Position: refs/heads/master@{#418512}

Revert of 🍵 Refactor the various locale_paks() templates to be more shared (patchset #2 id:260001 of https://codereview.chromium.org/2339783002/ )

Reason for revert:
The refactoring of the ios file is likely incorrect as it breaks the autoroller (as some files are no longer generated or have different name).

Reverting as half of the touched files are in ios/.

Original issue's description:
> Reland of Refactor the various locale_paks() templates to be more shared
>
> Reverted by:
> https://codereview.chromium.org/2340673002/
>
> Reason for reland:
> Now gn gens for enable_resource_whitelist_generation = true
>
> Main motivation for this is to more easily add per-target
> repack_whitelist.
>
> TBR=michaelbai@chromium.org,thakis@chromium.org,sdefresne@chromium.org,estevenson@chromium.org,rsesek@chromium.org,brettw@chromium.org
> BUG=645716
>
> Committed: https://crrev.com/d83512c644c7c61331a32c97ba549fcb1d8793cb
> Cr-Commit-Position: refs/heads/master@{#418450}

TBR=michaelbai@chromium.org,thakis@chromium.org,estevenson@chromium.org,rsesek@chromium.org,brettw@chromium.org,agrieve@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=645716

Review-Url: https://codereview.chromium.org/2338213002
Cr-Commit-Position: refs/heads/master@{#418511}

Current clang version supports the UUID syntax in <robuffer.h>, removing
additional declaration.

BUG=512433
R=toyoshim@chromium.org

Review-Url: https://codereview.chromium.org/2341603002
Cr-Commit-Position: refs/heads/master@{#418510}

BUG=

Review-Url: https://codereview.chromium.org/2334003004
Cr-Commit-Position: refs/heads/master@{#418509}

Previously, loadFile() is called to load the test page during performance
measurement in PerfTestRunner.measurePageLoadTime(), and thus the results of
blink_perf.svg tests were affected by the performance of loadFile() (that is not
related to SVG).
This CL makes the loadFile() to be executed before measurement and thus avoids
blink_perf.svg regressions caused by loadFile() performance regression.

The loadFile() performance regression is caused by [1] because, after [1],
synchronous XHRs are evicted from MemoryCache when Oilpan GC is executed.
[1] https://codereview.chromium.org/2296913003
This is expected and we can ignore the regression unless it causes regressions
in real use cases.

CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.perf:android_s5_perf_cq;master.tryserver.chromium.perf:linux_perf_cq;master.tryserver.chromium.perf:mac_retina_perf_cq;master.tryserver.chromium.perf:winx64_10_perf_cq
BUG=645363

Review-Url: https://codereview.chromium.org/2323313002
Cr-Commit-Position: refs/heads/master@{#418508}

For element.animate() calls of the form
"element.animate({property: [a, b, c]}, timing);" we were not accepting
non-array iterables for the value list.

This patch extends support for generic iterable objects to property indexed
keyframe values.

This behaviour corresponds to step 6.3 of
http://w3c.github.io/web-animations/#process-a-keyframe-like-object
where we cast the values as a (DOMString or sequence<DOMString>) object.

BUG=645393

Review-Url: https://codereview.chromium.org/2334123002
Cr-Commit-Position: refs/heads/master@{#418507}

BUG=610195

Review-Url: https://codereview.chromium.org/2319033002
Cr-Commit-Position: refs/heads/master@{#418506}

Summary of changes available at:
https://chromium.googlesource.com/v8/v8/+log/9d075810..96eff685

Please follow these instructions for assigning/CC'ing issues:
https://github.com/v8/v8/wiki/Triaging%20issues

Please close rolling in case of a roll revert:
https://v8-roll.appspot.com/
This only works with a Google account.

CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel;master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel

TBR=hablich@chromium.org,machenbach@chromium.org,littledan@chromium.org,vogelheim@chromium.org

Review-Url: https://codereview.chromium.org/2338163002
Cr-Commit-Position: refs/heads/master@{#418505}

A mostly-automated change to convert instances of WrapUnique(new Foo(...)) to
MakeUnique<Foo>(...). See the thread at
https://groups.google.com/a/chromium.org/forum/#!topic/chromium-dev/iQgMedVA8-k
for background.

To avoid requiring too many manual fixups, the change skips some cases that are
frequently problematic. In particular, in methods named Foo::Method() it will
not try to change WrapUnique(new Foo()) to MakeUnique<Foo>(). This is because
Foo::Method() may be accessing an internal constructor of Foo.

Cases where MakeUnique<NestedClass>(...) is called within a method of
OuterClass are common but hard to detect automatically, so have been fixed-up
manually.

The only types of manual fix ups applied are:
1) Revert MakeUnique back to WrapUnique
2) Change NULL to nullptr in argument list (MakeUnique cannot forward NULL
   correctly)
3) Add base:: namespace qualifier where missing.

WrapUnique(new Foo) has not been converted to MakeUnique<Foo>() as this might
change behaviour if Foo does not have a user-defined constructor. For example,
WrapUnique(new int) creates an unitialised integer, but MakeUnique<int>()
creates an integer initialised to 0.

git cl format has been been run over the CL. Spot-checking has uncovered no
cases of mis-formatting.

  BUG=637812

Review-Url: https://codereview.chromium.org/2260433003
Cr-Commit-Position: refs/heads/master@{#418504}

TBR=hubbe
BUG=646708

Review-Url: https://codereview.chromium.org/2337423002
Cr-Commit-Position: refs/heads/master@{#418503}

http://crrev.com/59aa6bb1162b31d2 added a parameter of BindStateBase,
and that causes 0.8% (40kB) binary size bloat on cronet, by adding
a new instruction to the constructor invocation at BindState ctor.

To reduce the binary size, this CL splits BindStateBase into two: one
for cancellable bind and another for non-cancellable one, so that
non-cancellable bind doesn't need to pass an extra argument to
BindStateBase.

This reduces the stripped binary size of chrome on Linux by 56kB.

Review-Url: https://codereview.chromium.org/2322313002
Cr-Commit-Position: refs/heads/master@{#418502}

This patch converts "editing/deleting/merge-paragraph-from-h6*.html" to use w3c
test harness to simplify test script, and consolidates into
"merge_paragraph_from_h6.html" since these test scripts use similar test sample,
for improving code health.

BUG=n/a
TEST=n/a; no behavior changes

Review-Url: https://codereview.chromium.org/2336173002
Cr-Commit-Position: refs/heads/master@{#418501}

This patch converts "editing/deleting/merge-paragraph-from-p*.html" to use w3c
test harness to simplify test script, and consolidates into
"merge_paragraph_from_p.html" since these test scripts use similar test sample,
for improving code health.

BUG=n/a
TEST=n/a; no behavior changes

Review-Url: https://codereview.chromium.org/2333223002
Cr-Commit-Position: refs/heads/master@{#418500}

AecDumpMessageFilter should not be included if enable_webrtc=false.

BUG=642210
TEST=

Review-Url: https://codereview.chromium.org/2334173002
Cr-Commit-Position: refs/heads/master@{#418499}

This patch converts "editing/execCommand/queryCommandState-01.html" to use w3c
test harness to simplify test script for improving code health.

BUG=n/a
TEST=n/a; no behavior changes

Review-Url: https://codereview.chromium.org/2342513002
Cr-Commit-Position: refs/heads/master@{#418498}

This patch converts "editing/execCommand/insertHTML.html" to use w3c test
harness to simplify test script for improving code health.

BUG=n/a
TEST=n/a; no behavior changes

Review-Url: https://codereview.chromium.org/2330353004
Cr-Commit-Position: refs/heads/master@{#418497}

TestSimpleTaskRunner::GetPendingTasks returns a full copy of the queued
pending tasks, which is less clean, and also that prevents a refactoring
that makes tasks move-only.

This CL replaces them with an equivalent NumPendingTasks to avoid the copy.

BUG=554299

Review-Url: https://codereview.chromium.org/2335153002
Cr-Commit-Position: refs/heads/master@{#418496}

V8's named interceptor never traps |0| of type number nor |"0"|
of type string.  Both of |0| and |"0"| trigger indexed interceptor.

In order for named properties to capture property keys such as 0
or "0", we have to first trap them at indexed interceptor, and
then fall through to an named property handler.

If an interface supports indexed properties, we don't need this
fallback mechanism.  If an interface only supports named
properties, let 0 or "0" fall through to an named property
handler.

BUG=229750

Review-Url: https://codereview.chromium.org/2321073002
Cr-Commit-Position: refs/heads/master@{#418495}

AppendIfNotPresent() and Insert() now take unique_ptr instead of a raw
pointer.

BUG=581865
TBR=brettw@chromium.org

Review-Url: https://codereview.chromium.org/2336863003
Cr-Commit-Position: refs/heads/master@{#418494}

Part of making PassRefPtr move only, in preparation for the removal of
PassRefPtr.

BUG=640449

Review-Url: https://codereview.chromium.org/2308353002
Cr-Commit-Position: refs/heads/master@{#418493}

Summary of changes available at:
https://chromium.googlesource.com/v8/v8/+log/0ea5d982..9d075810

Please follow these instructions for assigning/CC'ing issues:
https://github.com/v8/v8/wiki/Triaging%20issues

Please close rolling in case of a roll revert:
https://v8-roll.appspot.com/
This only works with a Google account.

CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel;master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel

TBR=hablich@chromium.org,machenbach@chromium.org,littledan@chromium.org,vogelheim@chromium.org

Review-Url: https://codereview.chromium.org/2339063003
Cr-Commit-Position: refs/heads/master@{#418492}

This is regression from
http://crrev.com/dac0c7af14c29c9d8509b775d5044e5448b6263b
The show_ime_if_needed is one shot message so we can't use it as a
trigger of monitoring composition bounding box, since we need
to keep monitoring during the text input session is active.

BUG=640012

Review-Url: https://codereview.chromium.org/2330423002
Cr-Commit-Position: refs/heads/master@{#418491}

Revert of cc: Plumb the monitor color profile to renderer for rasterization (patchset #6 id:100001 of https://codereview.chromium.org/2325773003/ )

Reason for revert:
Suspected cause of the following MSAN failure.

https://build.chromium.org/p/chromium.webkit/builders/WebKit%20Linux%20MSAN/builds/12218

19:03:27.700 29717   ==4==WARNING: MemorySanitizer: use-of-uninitialized-value
19:03:27.700 29717       #0 0x7853dbe in computeTypeMask third_party/skia/src/core/SkMatrix44.cpp:59:23
19:03:27.700 29717       #1 0x7858825 in getType third_party/skia/include/core/SkMatrix44.h:208:31
19:03:27.700 29717       #2 0x7858825 in setConcat third_party/skia/src/core/SkMatrix44.cpp:378:0
19:03:27.700 29717       #3 0x76f8840 in operator*= ui/gfx/transform.h:253:5
19:03:27.700 29717       #4 0x76f8840 in ColorSpaceToColorSpaceTransform ui/gfx/color_transform.cc:563:0
19:03:27.700 29717       #5 0x76f78de in NewColorTransform ui/gfx/color_transform.cc:703:15
19:03:27.700 29717       #6 0x767f315 in GetColorSpace ui/gfx/icc_profile.cc:142:45
19:03:27.700 29717       #7 0xb2263ae in initializeLayerTreeView content/renderer/render_widget.cc:1141:61
19:03:27.700 29717       #8 0xb1e68fd in initializeLayerTreeView content/renderer/render_view_impl.cc:1956:17
19:03:27.700 29717       #9 0xc1c6b06 in initializeLayerTreeView third_party/WebKit/Source/web/WebViewImpl.cpp:4341:19
19:03:27.700 29717       #10 0xc1c57bf in WebViewImpl third_party/WebKit/Source/web/WebViewImpl.cpp:471:5
19:03:27.700 29717       #11 0xc1c29ff in create third_party/WebKit/Source/web/WebViewImpl.cpp:342:25
19:03:27.700 29717       #12 0xc1c29ff in create third_party/WebKit/Source/web/WebViewImpl.cpp:336:0
19:03:27.700 29717       #13 0xb19b457 in Initialize content/renderer/render_view_impl.cc:717:7
<snip>

Original issue's description:
> cc: Plumb the monitor color profile to renderer for rasterization
>
> This adds the output device color profile to display::Display, and
> populates it correctly on Mac. We will want to do this for all
> platforms.
>
> The color profile is then plumbed through the same IPCs that take
> the device scale factor, to get to the renderer process'
> RenderWidgetCompositor.
>
> Note that we are sending the full ICCProfile this far. This is
> important, because the renderer process will be setting the ICCProfile
> of its rasterized IOSurfaces, and there is a power impact of even slight
> differences between the monitor profile and the IOSurface profile.
>
> The ICCProfile is then sent as a gfx::ColorProfile (which internally
> references the ICCProfile, for the above purpose) to the
> RenderWidgetCompositor, from where it will be pushed to cc.
>
> In the next step, we will (under a flag) specify the color space
> for rasterization.
>
> BUG=44872
> CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel
>
> Committed: https://crrev.com/f1fef7489c29deeee2a29c41ce4851e9ad1bd67b
> Cr-Commit-Position: refs/heads/master@{#418422}

TBR=enne@chromium.org,clamy@chromium.org,dcheng@chromium.org,ellyjones@chromium.org,ccameron@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=44872

Review-Url: https://codereview.chromium.org/2336113003
Cr-Commit-Position: refs/heads/master@{#418490}

This cl splits the big CL with changes to net/ folder and
has changes only to net/sdch according to the changes
to URLRequest API.

What the big cl does:
It modifies net/ clients that use URLRequest API as long as
URLRequest::Read returns int net errors and
URLRequest::Delegate and NetworkDelegate methods
(for example, OnResponseStarted or OnCompleted) have int
net_error in the arguments now.

The reason behind splitting the CL into small one is that
an android bot started to be unstable and unittests became
flaky. It was not possible to locate the problem and the
decision was to split the CL and upload small parts with a
6+ hours interval in order to make it possible to locate
the problematic code.

The big CL is located here -
https://codereview.chromium.org/2265873002/

BUG=423484

Review-Url: https://codereview.chromium.org/2329903002
Cr-Commit-Position: refs/heads/master@{#418489}

The text autosizer is used for both making desktop pages more legible
on mobile, and for applying the accessibility font scale factor. When
text-size-adjust support was added, pages were able to override the
accessibility setting in addition to the autosizing multiplier.

This patch ensures accessibility font scale settings are respected even
with text-size-adjust. A followup bug has been filed (crbug.com/645717)
for moving the accessibility font scale factor out of the autosizer.

This patch also re-enables support for text-size-adjust which was
disabled temporarily due to breaking accessibility settings.

This was reverted due to a forgotten "!" in checking whether the
viewport was specified by the author in TextAutosizer.cpp. A new test
has been added to catch this in the future. The unneeded meta viewport
declarations in all other tests have been removed because they had
no effect with the meta viewport setting disabled.

Original review:
https://codereview.chromium.org/2329733002

BUG=645269,646342

Review-Url: https://codereview.chromium.org/2340553003
Cr-Commit-Position: refs/heads/master@{#418488}

The following APIs are whitelisted:
- imageWriterPrivate
- chromeosInfoPrivate
- the feedback API
- fileSystem

BUG=642141
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:closure_compilation

Review-Url: https://codereview.chromium.org/2297113002
Cr-Commit-Position: refs/heads/master@{#418487}

We add WebGLImageConversionTest in blink_platform_unittests.
This covers following conversion tests:
- ConvertRGBA4444toRGBA8
- ConvertRGBA5551toRGBA8
- ConvertRGBA8toRA8
- convertBGRA8toRGBA8
- ConvertRGBA8toR8
- ConvertRGBA8toRGBA8
- ConvertRGBA8ToUnsignedShort4444
- ConvertRGBA8ToRGBA5551
- ConvertRGBA8ToRGB565

R=kbr@chromium.org
BUG=642653

Review-Url: https://codereview.chromium.org/2333233002
Cr-Commit-Position: refs/heads/master@{#418486}

The problem was that the path to the filter is different depending on whether the test is swarmed or not. Also make sure to send the filter file for the tests as part of the isolate data.

BUG=645983
TBR=scottmg@chromium.org

Review-Url: https://codereview.chromium.org/2337383002
Cr-Commit-Position: refs/heads/master@{#418485}

https://chromium.googlesource.com/external/github.com/catapult-project/catapult.git/+log/b86cfef4b9da..c42044026b81

$ git log b86cfef4b..c42044026 --date=short --no-merges --format='%ad %ae %s'
2016-09-13 benjhayden Add BarChart.

TBR=catapult-sheriff@chromium.org

Review-Url: https://codereview.chromium.org/2333373005
Cr-Commit-Position: refs/heads/master@{#418484}

Also, perform some clean-up.

BUG=592873

Review-Url: https://codereview.chromium.org/2338803003
Cr-Commit-Position: refs/heads/master@{#418483}

Summary of changes available at:
https://chromium.googlesource.com/v8/v8/+log/72ed9eb5..0ea5d982

Please follow these instructions for assigning/CC'ing issues:
https://github.com/v8/v8/wiki/Triaging%20issues

Please close rolling in case of a roll revert:
https://v8-roll.appspot.com/
This only works with a Google account.

CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel;master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel

TBR=hablich@chromium.org,machenbach@chromium.org,littledan@chromium.org,vogelheim@chromium.org

Review-Url: https://codereview.chromium.org/2339063002
Cr-Commit-Position: refs/heads/master@{#418482}

BUG=621841
TEST=dbus_unittests
TBR=steel@chromium.org for device/bluetooth/dbus

Review-Url: https://codereview.chromium.org/2337893002
Cr-Commit-Position: refs/heads/master@{#418481}

Update recent QUIC flag changes.

Deprecate FLAGS_quic_limit_frame_gaps_in_buffer.

Merge internal change: 132467804
https://codereview.chromium.org/2341483002/

Deprecate flag FLAGS_quic_sequencer_buffer_retire_block_in_time.

Merge internal change: 132461150
https://codereview.chromium.org/2331343003/

Make DiversificationNonce a std::array

Changing a char[32] to a std::array<char, 32>.  No functional change
intended, not flag-protected

Merge internal change: 132456567
https://codereview.chromium.org/2340583002/

remove unused methods: num_early_frames_received() from the shared code which was used by chromium only.

Triggered by ref chromium cl: https://codereview.chromium.org/2250713004.

Merge internal change: 132449425
https://codereview.chromium.org/2337063004/

Use unique_ptrs to manage callback lifetimes in some QUIC code

Follow-up to cl/132102408.

This CL eliminates the self-deleting callback pattern from a chunk of
QUIC code and replaces it with explicit unique_ptr ownership of all
callbacks.

Replace self-deleting callbacks with unique_ptr owned callbacks in some
QUIC code.  No functional change intended, not flag-protected.

Merge internal change: 132439857
https://codereview.chromium.org/2339433004/

Simplify lifetime management of ValidateClientHelloResultCallback::Result objects

The lifetimes of ValidateClientHelloResultCallback::Result objects are
managed by bare new/delete.  These objects are also passed around by
pointer which gives no indication as to who is managing the lifetime,
and many methods are passed both a pointer to a Result object as w
ell as a reference to a sub-object.  This CL eliminates these patterns.

Migrating from bare pointers to unique_ptr in some quic codepaths.  No
functional change intended, not flag-protected.

Merge internal change: 132359769
https://codereview.chromium.org/2338013004/

Use the largest received packet instead of the most recently received packet to calculate the sequence number from.

Merge internal change: 132350039
https://codereview.chromium.org/2333153003/

Deprecate FLAGS_quic_close_stream_after_writing_queued_data

Merge internal change: 132335255
https://codereview.chromium.org/2338713003/

Deprecate FLAGS_quic_remove_obsolete_forward_secure.

Merge internal change: 132331281
https://codereview.chromium.org/2336803004/

Deprecate FLAGS_quic_change_alarms_efficiently.

Merge internal change: 132330925
https://codereview.chromium.org/2338013003/

deprecate FLAGS_quic_enable_app_limited_check

Merge internal change: 132318066
https://codereview.chromium.org/2340563002/

Review-Url: https://codereview.chromium.org/2334363002
Cr-Commit-Position: refs/heads/master@{#418480}

BUG=646683

Review-Url: https://codereview.chromium.org/2334243003
Cr-Commit-Position: refs/heads/master@{#418479}

Refactor disk_mount_manager_unittest by replacing gmock with a dedicated mock class of DiskMountManagerObserver.

This change is based on
https://codereview.chromium.org/2292473002/
which was once merged but reverted by
https://codereview.chromium.org/2340433002/

BUG=641943

Review-Url: https://codereview.chromium.org/2333983004
Cr-Commit-Position: refs/heads/master@{#418478}

DOMError is only used in quota, so may as well move it there
while waiting for the deprecated parts of quota to be removed.
This patch also removes unneeded DOMError includes.

BUG=460725

Review-Url: https://codereview.chromium.org/2335073003
Cr-Commit-Position: refs/heads/master@{#418477}

TBR=lazyboy@chromium.org
BUG=632548,636222,638621

Review-Url: https://codereview.chromium.org/2332313003
Cr-Commit-Position: refs/heads/master@{#418476}

Android’s InputConnection API specifies a way to commit (confirm) the
current composition and move the cursor at the same time:
https://developer.android.com/reference/android/view/inputmethod/
InputConnection.html#commitText(java.lang.CharSequence, int)

Currently, commitText doesn't conform to the documentation nor to
Android TextView's behavior when newCursorPosition is not 1.
newCursorPosition
is a relative value that depends on the location of the text being committed.

When |newCursorPosition| < 1, the cursor should move to the left of
the committing text by |newCursorPosition| characters. If the cursor
exceeds the left boundary, it will clamp to  the left boundary.

When |newCursorPosition| >= 1, the cursor will move to the right of
the committing text by |newCursorPosition - 1| characters. If the cursor
exceeds the right boundary, it will clamp to the right boundary.

(Rebased version of yabinh@'s http://crrev.com/1995333002 which received
lgt'ms from TBRed reviewers below, I'm landing this on his behalf.)

TBR=tkent,yosin,changwan,sievers,yabinh
NOTRY=true
BUG=570920

Review-Url: https://codereview.chromium.org/2339793002
Cr-Commit-Position: refs/heads/master@{#418475}