Only execute default actions on trusted events.

Only execute default actions on trusted events.

Add a RuntimeFeature to try out only allowing default actions to
execute on trusted events.

This was previously tried in https://codereview.chromium.org/894913002 in
a slightly different approach; this attempt is based on the isTrusted
support. Also this time it is hidden behind a feature flag that
isn't enabled as it has some web compatibility issues that needs to
be quantified.

BUG=334015
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=199346

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=199380

[dtapuska, 2015-07-15 23:20:57]

dtapuska@chromium.org changed reviewers:
+ tkent@chromium.org

[tkent, 2015-07-16 23:39:45]

https://codereview.chromium.org/1234613004/diff/1/Source/platform/RuntimeEnab...
File Source/platform/RuntimeEnabledFeatures.in (right):

https://codereview.chromium.org/1234613004/diff/1/Source/platform/RuntimeEnab...
Source/platform/RuntimeEnabledFeatures.in:160: TrustedEventsDefaultAction
Do we need a separated runtime flag?

[dtapuska, 2015-07-17 02:46:10]

https://codereview.chromium.org/1234613004/diff/1/Source/platform/RuntimeEnab...
File Source/platform/RuntimeEnabledFeatures.in (right):

https://codereview.chromium.org/1234613004/diff/1/Source/platform/RuntimeEnab...
Source/platform/RuntimeEnabledFeatures.in:160: TrustedEventsDefaultAction
On 2015/07/16 23:39:45, tkent wrote:
> Do we need a separated runtime flag?

Rick wanted two separate flags because this change is a web compatibility
breaking change we might land the TrustedEvents but hold back on the default
action change.

The trusted events change is needed for the password extensions so they properly
deal with injected content from a page.

The idea was the the TrustedEvents experimental flag would be removed once the
extensions developers did a sanity check on their end and security reviewed the
implementation but that we might not do the default action just yet.

The problem I see with this and wanted to discuss with Rick (and value your
opinion) is that if a site wanted to detect that the browser does the default
action on the untrusted events they couldn't just test the Event.isTrusted
property exists. But not sure if someone would take advantage of that or not.
Not sure what they would do in the case of they do detect that it supports
trusted events.

[tkent, 2015-07-17 05:50:42]

https://codereview.chromium.org/1234613004/diff/1/LayoutTests/fast/events/dis...
File LayoutTests/fast/events/dispatch-synthetic-keyboardevent-no-action.html
(right):

https://codereview.chromium.org/1234613004/diff/1/LayoutTests/fast/events/dis...
LayoutTests/fast/events/dispatch-synthetic-keyboardevent-no-action.html:17: var
event = new KeyboardEvent('keydown', { keyIdentifier: 'Down' } );
On Mac, Down key opens the popup.  We should dispatch a key event for 'B' key to
change the selected option.

https://codereview.chromium.org/1234613004/diff/1/Source/platform/RuntimeEnab...
File Source/platform/RuntimeEnabledFeatures.in (right):

https://codereview.chromium.org/1234613004/diff/1/Source/platform/RuntimeEnab...
Source/platform/RuntimeEnabledFeatures.in:160: TrustedEventsDefaultAction
On 2015/07/17 02:46:10, Dave Tapuska wrote:
> On 2015/07/16 23:39:45, tkent wrote:
> > Do we need a separated runtime flag?
> 
> Rick wanted two separate flags because this change is a web compatibility
> breaking change we might land the TrustedEvents but hold back on the default
> action change.
> 
> The trusted events change is needed for the password extensions so they
properly
> deal with injected content from a page.
> 
> The idea was the the TrustedEvents experimental flag would be removed once the
> extensions developers did a sanity check on their end and security reviewed
the
> implementation but that we might not do the default action just yet.
> 
> The problem I see with this and wanted to discuss with Rick (and value your
> opinion) is that if a site wanted to detect that the browser does the default
> action on the untrusted events they couldn't just test the Event.isTrusted
> property exists. But not sure if someone would take advantage of that or not.
> Not sure what they would do in the case of they do detect that it supports
> trusted events.

ok, I got it.

[dtapuska, 2015-07-21 21:30:05]

https://codereview.chromium.org/1234613004/diff/1/LayoutTests/fast/events/dis...
File LayoutTests/fast/events/dispatch-synthetic-keyboardevent-no-action.html
(right):

https://codereview.chromium.org/1234613004/diff/1/LayoutTests/fast/events/dis...
LayoutTests/fast/events/dispatch-synthetic-keyboardevent-no-action.html:17: var
event = new KeyboardEvent('keydown', { keyIdentifier: 'Down' } );
On 2015/07/17 05:50:41, tkent wrote:
> On Mac, Down key opens the popup.  We should dispatch a key event for 'B' key
to
> change the selected option.

Done.

[dtapuska, 2015-07-22 21:01:50]

On 2015/07/21 21:30:05, Dave Tapuska wrote:
>
https://codereview.chromium.org/1234613004/diff/1/LayoutTests/fast/events/dis...
> File LayoutTests/fast/events/dispatch-synthetic-keyboardevent-no-action.html
> (right):
> 
>
https://codereview.chromium.org/1234613004/diff/1/LayoutTests/fast/events/dis...
> LayoutTests/fast/events/dispatch-synthetic-keyboardevent-no-action.html:17:
var
> event = new KeyboardEvent('keydown', { keyIdentifier: 'Down' } );
> On 2015/07/17 05:50:41, tkent wrote:
> > On Mac, Down key opens the popup.  We should dispatch a key event for 'B'
key
> to
> > change the selected option.
> 
> Done.

PTAL

[tkent, 2015-07-22 23:15:03]

https://codereview.chromium.org/1234613004/diff/80001/LayoutTests/fast/events...
File LayoutTests/fast/events/dispatch-synthetic-keyboardevent-no-action.html
(right):

https://codereview.chromium.org/1234613004/diff/80001/LayoutTests/fast/events...
LayoutTests/fast/events/dispatch-synthetic-keyboardevent-no-action.html:12:
function test()
You don't need to define the function |test()|.

https://codereview.chromium.org/1234613004/diff/80001/LayoutTests/fast/events...
LayoutTests/fast/events/dispatch-synthetic-keyboardevent-no-action.html:15:
testRunner.dumpAsText();
This is unnecessary.  js-test.js does it.

https://codereview.chromium.org/1234613004/diff/80001/LayoutTests/fast/events...
LayoutTests/fast/events/dispatch-synthetic-keyboardevent-no-action.html:18: var
client = node.getBoundingClientRect();
|client| is not used.

https://codereview.chromium.org/1234613004/diff/80001/LayoutTests/fast/events...
File LayoutTests/fast/events/frame-tab-focus.html (right):

https://codereview.chromium.org/1234613004/diff/80001/LayoutTests/fast/events...
LayoutTests/fast/events/frame-tab-focus.html:92: modifiers[0] = "shiftKey";
modifieers.push("shiftKey")  would be simpler.

https://codereview.chromium.org/1234613004/diff/80001/LayoutTests/fast/events...
LayoutTests/fast/events/frame-tab-focus.html:94: modifiers[modifiers.length] =
"altKey";
modifiers.push("altKey") is simpler.

https://codereview.chromium.org/1234613004/diff/80001/LayoutTests/fast/events...
File LayoutTests/fast/events/option-tab.html (right):

https://codereview.chromium.org/1234613004/diff/80001/LayoutTests/fast/events...
LayoutTests/fast/events/option-tab.html:17: if (window.testRunner)
should test window.eventSender.

https://codereview.chromium.org/1234613004/diff/80001/LayoutTests/fast/events...
LayoutTests/fast/events/option-tab.html:28: if (window.testRunner)
Ditto.

https://codereview.chromium.org/1234613004/diff/80001/LayoutTests/fast/forms/...
File LayoutTests/fast/forms/listbox-onchange.html (right):

https://codereview.chromium.org/1234613004/diff/80001/LayoutTests/fast/forms/...
LayoutTests/fast/forms/listbox-onchange.html:127: modifiers[0] = "shiftKey";
modifiers.push("shiftKey")

https://codereview.chromium.org/1234613004/diff/80001/LayoutTests/fast/forms/...
LayoutTests/fast/forms/listbox-onchange.html:129: if
(navigator.userAgent.search(/\bMac OS X\b/) != -1)
We can simplify this block by
    modifiers.push("addSelectionKey");

"addSelectionKey" is metaKey on Mac, and "ctrlKey" on other platforms.

https://codereview.chromium.org/1234613004/diff/80001/LayoutTests/fast/forms/...
LayoutTests/fast/forms/listbox-onchange.html:130: modifiers[modifiers.length] =
"metaKey";
modifiers.push("metaKey")

https://codereview.chromium.org/1234613004/diff/80001/LayoutTests/fast/forms/...
LayoutTests/fast/forms/listbox-onchange.html:132: modifiers[modifiers.length] =
"ctrlKey";
modifiers.push("ctrlKey")

https://codereview.chromium.org/1234613004/diff/80001/LayoutTests/fast/forms/...
LayoutTests/fast/forms/listbox-onchange.html:140: debug("FAIL: This test
requires window.eventSender.");
should be
    testFaiiled("This test requires window.eventSender.");

https://codereview.chromium.org/1234613004/diff/80001/LayoutTests/fast/forms/...
LayoutTests/fast/forms/listbox-onchange.html:148: modifiers[0] = "shiftKey";
modifiers.push("shiftKey")

https://codereview.chromium.org/1234613004/diff/80001/LayoutTests/fast/forms/...
LayoutTests/fast/forms/listbox-onchange.html:150: if
(navigator.userAgent.search(/\bMac OS X\b/) != -1)
modifiers.push("addSelectionKey");

https://codereview.chromium.org/1234613004/diff/80001/LayoutTests/fast/forms/...
LayoutTests/fast/forms/listbox-onchange.html:161: debug("FAIL: This test
requires window.eventSender.");
testFailed("This test ...");

https://codereview.chromium.org/1234613004/diff/80001/LayoutTests/fast/forms/...
LayoutTests/fast/forms/listbox-onchange.html:171: if
(navigator.userAgent.search(/\bMac OS X\b/) != -1)
modifiers.push("addSelectionKey");

[dtapuska, 2015-07-23 01:28:10]

https://codereview.chromium.org/1234613004/diff/80001/LayoutTests/fast/events...
File LayoutTests/fast/events/dispatch-synthetic-keyboardevent-no-action.html
(right):

https://codereview.chromium.org/1234613004/diff/80001/LayoutTests/fast/events...
LayoutTests/fast/events/dispatch-synthetic-keyboardevent-no-action.html:12:
function test()
On 2015/07/22 23:15:02, tkent wrote:
> You don't need to define the function |test()|.

Done.

https://codereview.chromium.org/1234613004/diff/80001/LayoutTests/fast/events...
LayoutTests/fast/events/dispatch-synthetic-keyboardevent-no-action.html:15:
testRunner.dumpAsText();
On 2015/07/22 23:15:03, tkent wrote:
> This is unnecessary.  js-test.js does it.

Done.

https://codereview.chromium.org/1234613004/diff/80001/LayoutTests/fast/events...
LayoutTests/fast/events/dispatch-synthetic-keyboardevent-no-action.html:18: var
client = node.getBoundingClientRect();
On 2015/07/22 23:15:02, tkent wrote:
> |client| is not used.

Done.

https://codereview.chromium.org/1234613004/diff/80001/LayoutTests/fast/events...
File LayoutTests/fast/events/frame-tab-focus.html (right):

https://codereview.chromium.org/1234613004/diff/80001/LayoutTests/fast/events...
LayoutTests/fast/events/frame-tab-focus.html:92: modifiers[0] = "shiftKey";
On 2015/07/22 23:15:03, tkent wrote:
> modifieers.push("shiftKey")  would be simpler.

Done.

https://codereview.chromium.org/1234613004/diff/80001/LayoutTests/fast/events...
LayoutTests/fast/events/frame-tab-focus.html:94: modifiers[modifiers.length] =
"altKey";
On 2015/07/22 23:15:03, tkent wrote:
> modifiers.push("altKey") is simpler.

Done.

https://codereview.chromium.org/1234613004/diff/80001/LayoutTests/fast/events...
File LayoutTests/fast/events/option-tab.html (right):

https://codereview.chromium.org/1234613004/diff/80001/LayoutTests/fast/events...
LayoutTests/fast/events/option-tab.html:17: if (window.testRunner)
On 2015/07/22 23:15:03, tkent wrote:
> should test window.eventSender.

Done.

https://codereview.chromium.org/1234613004/diff/80001/LayoutTests/fast/events...
LayoutTests/fast/events/option-tab.html:28: if (window.testRunner)
On 2015/07/22 23:15:03, tkent wrote:
> Ditto.

Done.

https://codereview.chromium.org/1234613004/diff/80001/LayoutTests/fast/forms/...
File LayoutTests/fast/forms/listbox-onchange.html (right):

https://codereview.chromium.org/1234613004/diff/80001/LayoutTests/fast/forms/...
LayoutTests/fast/forms/listbox-onchange.html:127: modifiers[0] = "shiftKey";
On 2015/07/22 23:15:03, tkent wrote:
> modifiers.push("shiftKey")

Done.

https://codereview.chromium.org/1234613004/diff/80001/LayoutTests/fast/forms/...
LayoutTests/fast/forms/listbox-onchange.html:129: if
(navigator.userAgent.search(/\bMac OS X\b/) != -1)
On 2015/07/22 23:15:03, tkent wrote:
> We can simplify this block by
>     modifiers.push("addSelectionKey");
> 
> "addSelectionKey" is metaKey on Mac, and "ctrlKey" on other platforms.

Done.

https://codereview.chromium.org/1234613004/diff/80001/LayoutTests/fast/forms/...
LayoutTests/fast/forms/listbox-onchange.html:130: modifiers[modifiers.length] =
"metaKey";
On 2015/07/22 23:15:03, tkent wrote:
> modifiers.push("metaKey")

Done.

https://codereview.chromium.org/1234613004/diff/80001/LayoutTests/fast/forms/...
LayoutTests/fast/forms/listbox-onchange.html:132: modifiers[modifiers.length] =
"ctrlKey";
On 2015/07/22 23:15:03, tkent wrote:
> modifiers.push("ctrlKey")

Done.

https://codereview.chromium.org/1234613004/diff/80001/LayoutTests/fast/forms/...
LayoutTests/fast/forms/listbox-onchange.html:140: debug("FAIL: This test
requires window.eventSender.");
On 2015/07/22 23:15:03, tkent wrote:
> should be
>     testFaiiled("This test requires window.eventSender.");

Done.

https://codereview.chromium.org/1234613004/diff/80001/LayoutTests/fast/forms/...
LayoutTests/fast/forms/listbox-onchange.html:148: modifiers[0] = "shiftKey";
On 2015/07/22 23:15:03, tkent wrote:
> modifiers.push("shiftKey")

Done.

https://codereview.chromium.org/1234613004/diff/80001/LayoutTests/fast/forms/...
LayoutTests/fast/forms/listbox-onchange.html:150: if
(navigator.userAgent.search(/\bMac OS X\b/) != -1)
On 2015/07/22 23:15:03, tkent wrote:
> modifiers.push("addSelectionKey");
> 

Done.

https://codereview.chromium.org/1234613004/diff/80001/LayoutTests/fast/forms/...
LayoutTests/fast/forms/listbox-onchange.html:161: debug("FAIL: This test
requires window.eventSender.");
On 2015/07/22 23:15:03, tkent wrote:
> testFailed("This test ...");

Done.

https://codereview.chromium.org/1234613004/diff/80001/LayoutTests/fast/forms/...
LayoutTests/fast/forms/listbox-onchange.html:171: if
(navigator.userAgent.search(/\bMac OS X\b/) != -1)
On 2015/07/22 23:15:03, tkent wrote:
> modifiers.push("addSelectionKey");

Done.

[tkent, 2015-07-23 01:42:22]

The CQ bit was checked by tkent@chromium.org

[tkent, 2015-07-23 01:42:22]

lgtm

[commit-bot: I haz the power, 2015-07-23 01:42:44]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1234613004/100001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1234613004/100001

[commit-bot: I haz the power, 2015-07-23 02:55:09]

Committed patchset #6 (id:100001) as
https://src.chromium.org/viewvc/blink?view=rev&revision=199346

[alancutter (OOO until 2018), 2015-07-23 07:23:46]

A revert of this CL (patchset #6 id:100001) has been created in
https://codereview.chromium.org/1251043004/ by alancutter@chromium.org.

The reason for reverting is: This change is suspected to be causing the
following failures on the Mac builders:

https://build.chromium.org/p/chromium.webkit/builders/WebKit%20Mac10.10/build...
virtual/trustedeventsdefaultaction/fast/events/autoscroll.html
virtual/trustedeventsdefaultaction/fast/events/context-no-deselect.html
virtual/trustedeventsdefaultaction/fast/events/reveal-link-when-focused.html

http://build.chromium.org/p/chromium.webkit/builders/WebKit%20Mac10.6/builds/...
virtual/trustedeventsdefaultaction/fast/events/hit-test-cache-scrollbar-no-crash.html
virtual/pointerevent/fast/events/hit-test-cache-scrollbar-no-crash.html

https://build.chromium.org/p/chromium.webkit/builders/WebKit%20Mac10.6%20%28d...
fast/events/hit-test-cache-scrollbar-no-crash.html
virtual/pointerevent/fast/events/hit-test-cache-scrollbar-no-crash.html.

[dtapuska, 2015-07-23 15:07:29]

The CQ bit was checked by dtapuska@chromium.org

[dtapuska, 2015-07-23 15:07:29]

The patchset sent to the CQ was uploaded after l-g-t-m from tkent@chromium.org
Link to the patchset: https://codereview.chromium.org/1234613004/#ps120001
(title: "Add Yosemite exceptions")

[commit-bot: I haz the power, 2015-07-23 15:07:39]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1234613004/120001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1234613004/120001

[commit-bot: I haz the power, 2015-07-23 16:21:53]

Committed patchset #7 (id:120001) as
https://src.chromium.org/viewvc/blink?view=rev&revision=199380