HTML: Extend Test Coverage for _parent named browsing context (7.1.5) #5150

lyzadanger · 2017-03-15T19:38:05Z

This PR extends some existing tests for the _parent named browsing context and adds some new ones.

The Webkit/Chrome case-sensitivity bug noted in #5145 is indeed manifest here, again, as expected. I will be opening bugs shortly and will link back.

There's also an interesting difference in browser behavior WRT choose-browsing-context-parent-003.html.

According to the spec:

The opener IDL attribute on the Window object, on getting, must return the WindowProxy object of the browsing context from which the current browsing context was created (its opener browsing context)

(emphasis mine)

Safari and Chrome are both steadfastly retaining a browsing context's original opener, even if another browsing context later opens a different document into that context/window. FireFox, however, will replace the window.opener of the _parent context with the window that invokes .open (assuming the opener isn't disowned, etc.).

I have two hunches here: Firefox's behavior is the errant one and this is likely going to affect more than just the named _parent context. But I'm not 100% certain on the former.

Supporting files have been lightly organized but there is still more work to do in this section as I proceed.

wpt-pr-bot · 2017-03-15T19:38:12Z

Notifying @ayg, @jdm, @jgraham, @zcorpan, and @zqzhang. (Learn how reviewing works.)

ghost · 2017-03-15T19:45:43Z

View the complete job log.

Firefox (nightly channel)

Testing web-platform-tests at revision 2f87ce3501be9c9e4dd48ba18f655065d40ea41b
Using browser at version BuildID 20170326110231; SourceStamp f5e214144799889e2408c4841351f4053f00544e
Starting 10 test iterations
All results were stable

All results

7 tests ran

/html/browsers/windows/browsing-context-names/browsing-context-choose-existing.html

Subtest	Results	Messages
	OK
`The browsing context must be chosen if the given name is same as its name`	PASS

/html/browsers/windows/browsing-context-names/browsing-context-choose-parent-001.html

Subtest	Results	Messages
	OK
`The parent browsing context must be chosen if the given name is _parent`	PASS

/html/browsers/windows/browsing-context-names/browsing-context-choose-parent-002.html

Subtest	Results	Messages
	OK
`choosing _parent context: multiple nested contexts`	PASS

/html/browsers/windows/browsing-context-names/browsing-context-choose-parent-003.html

Subtest	Results	Messages
	TIMEOUT
`_parent should reuse window.parent context`	TIMEOUT	`Test timed out`

/html/browsers/windows/browsing-context-names/browsing-context-choose-parent-004.html

Subtest	Results	Messages
	OK
`choosing _parent context should be case-insensitive`	PASS

/html/browsers/windows/browsing-context-names/browsing-context-choose-self-1.html

Subtest	Results	Messages
	OK
`The current browsing context must be chosen if the given name is '_self'`	PASS

/html/browsers/windows/browsing-context-names/browsing-context-choose-self-2.html

Subtest	Results	Messages
	TIMEOUT
`The current browsing context must be chosen if the given name is empty string`	NOTRUN

ghost · 2017-03-15T19:45:44Z

View the complete job log.

Chrome (unstable channel)

Testing web-platform-tests at revision 2f87ce3501be9c9e4dd48ba18f655065d40ea41b
Using browser at version 59.0.3047.0 dev
Starting 10 test iterations
All results were stable

All results

7 tests ran

/html/browsers/windows/browsing-context-names/browsing-context-choose-existing.html

Subtest	Results	Messages
	OK
`The browsing context must be chosen if the given name is same as its name`	PASS

/html/browsers/windows/browsing-context-names/browsing-context-choose-parent-001.html

Subtest	Results	Messages
	OK
`The parent browsing context must be chosen if the given name is _parent`	PASS

/html/browsers/windows/browsing-context-names/browsing-context-choose-parent-002.html

Subtest	Results	Messages
	OK
`choosing _parent context: multiple nested contexts`	PASS

/html/browsers/windows/browsing-context-names/browsing-context-choose-parent-003.html

Subtest	Results	Messages
	OK
`_parent should reuse window.parent context`	PASS

/html/browsers/windows/browsing-context-names/browsing-context-choose-parent-004.html

Subtest	Results	Messages
	TIMEOUT
`choosing _parent context should be case-insensitive`	TIMEOUT	`Test timed out`

/html/browsers/windows/browsing-context-names/browsing-context-choose-self-1.html

Subtest	Results	Messages
	OK
`The current browsing context must be chosen if the given name is '_self'`	PASS

/html/browsers/windows/browsing-context-names/browsing-context-choose-self-2.html

Subtest	Results	Messages
	TIMEOUT
`The current browsing context must be chosen if the given name is empty string`	NOTRUN

zcorpan · 2017-03-16T11:55:52Z

html/browsers/windows/browsing-context-names/browsing-context-choose-parent-001.html

+  window.addEventListener('message', t.step_func_done(e => {
+    assert_equals(e.data.name, 'parentWin');
+  }));
+  window.frames['embedded'].src = 'resources/parent-iframe-1.html';


I get confused about the intent here, if it's trying to set src on the iframe or on its WindowProxy. In Safari, if the iframe happens to not have a name attribute, this would return the WindowProxy. I think that is a bug per https://html.spec.whatwg.org/#named-access-on-the-window-object - but testing that should be separate.

Here it would be clearer to just use document.getElementById('embedded').

zcorpan · 2017-03-16T12:03:38Z

Are the TIMEOUT/NOTRUN results in /html/browsers/windows/browsing-context-names/browsing-context-choose-parent-004.html and /html/browsers/windows/browsing-context-names/browsing-context-choose-self-2.html expected?

lyzadanger · 2017-03-16T12:52:16Z

@zcorpan

Are the TIMEOUT/NOTRUN results in /html/browsers/windows/browsing-context-names/browsing-context-choose-parent-004.html and /html/browsers/windows/browsing-context-names/browsing-context-choose-self-2.html expected?

For 004, yes! I'm about to open bugs against both Chrome and Webkit regarding that—both are erroneously (I believe) treating named browsing contexts (e.g. _top, _parent) as case-sensitive. The spec says they should be case-insensitive.

For self-2.thml...that one's a pre-existing test. I have tackled _blank and _parent named contexts so far: _top and _self are next. I'll take a look at that test when I work on the _self coverage!

lyzadanger · 2017-03-16T13:30:56Z

As promised, bugs opened on case-sensitive _parent issue:

ghost · 2017-03-23T15:56:00Z

These tests are now available on w3c-test.org

zcorpan · 2017-03-24T10:01:58Z

html/browsers/windows/browsing-context-names/browsing-context-choose-parent-001.html

+<script src="/resources/testharness.js"></script>
+<script src="/resources/testharnessreport.js"></script>
+<div id="log"></div>
+<iframe id="embedded" src="resources/parent-iframe-1.html" name="parentWin" style="display:none"></iframe>


This should be moved to after the script. In principle there is a race condition otherwise.

Hah! That's the reason I had the hard-to-read assignment of src to that iframe in the script and not in the markup initially. OK, I'll rectify.

zcorpan · 2017-03-24T14:56:01Z

So in Gecko for /html/browsers/windows/browsing-context-names/browsing-context-choose-parent-003.html I get a TIMEOUT, that is because of the behavior you describe in the OP, correct?

lyzadanger · 2017-03-24T15:00:30Z

@zcorpan Yep! I had to go remind myself by re-reading my original opening comment. Yes, this looks potentially buggy in Gecko.

zcorpan · 2017-03-27T11:24:23Z

Filed https://bugzilla.mozilla.org/show_bug.cgi?id=1350858

zcorpan · 2017-03-27T14:54:20Z

Not sure why 39948b2 is now in this PR. I suppose "squash and merge" does the right thing, but leaving this open if you want to clean up this branch before landing @lyzadanger

lyzadanger · 2017-03-27T14:58:16Z

@zcorpan @jgraham asked me to rebase onto #5196 before it was merged because the CI was wedging. I can re-rebase my branch to get rid of it. Will do shortly.

- Add tests for reused `parent` context - Add tests for case-sensitivity - Organize supporting files

bzbarsky · 2017-05-05T19:09:17Z

Safari and Chrome are both steadfastly retaining a browsing context's original opener

Do they? Simple testcase:

<!DOCTYPE html>
<input type="button" onclick="alert(opener)" value="check opener">
<input type="button" onclick="window.open(location.href, '_self')" value="do open">

Click the "check opener" button, observe null, click the "do open" button, click the "check opener" button, observe non-null. In both Safari and Chrome.

bzbarsky · 2017-05-05T20:54:00Z

I filed whatwg/html#2636 to get this sorted out in the spec. Given the behavior on the testcases there, I have no idea how Blink or WebKit pass html/browsers/windows/browsing-context-names/browsing-context-choose-parent-003.html from this PR, unless they somehow special-case the "opener would be one of our descendants" case or something.

wpt-pr-bot added the html label Mar 15, 2017

zcorpan reviewed Mar 16, 2017

View reviewed changes

annevk mentioned this pull request Mar 20, 2017

Should target=_blank and window.open('...', '_blank') be case sensitive? whatwg/html#2443

Closed

zcorpan reviewed Mar 24, 2017

View reviewed changes

zcorpan mentioned this pull request Mar 24, 2017

Stability bots run tests that are not changed in PR #5218

Closed

zcorpan added the status:needs-spec-decision label Mar 24, 2017

lyzadanger force-pushed the browsing-context-names-parent branch from 769fecc to f125f12 Compare March 24, 2017 16:38

wpt-pr-bot added the infra label Mar 24, 2017

zcorpan removed infra status:needs-spec-decision labels Mar 27, 2017

zcorpan approved these changes Mar 27, 2017

View reviewed changes

HTML: Extend test coverage for named contexts: _parent

070a1f4

- Add tests for reused `parent` context - Add tests for case-sensitivity - Organize supporting files

lyzadanger force-pushed the browsing-context-names-parent branch from f125f12 to 070a1f4 Compare March 27, 2017 16:10

lyzadanger merged commit e27e750 into web-platform-tests:master Mar 27, 2017

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

HTML: Extend Test Coverage for _parent named browsing context (7.1.5) #5150

HTML: Extend Test Coverage for _parent named browsing context (7.1.5) #5150

lyzadanger commented Mar 15, 2017

wpt-pr-bot commented Mar 15, 2017

ghost commented Mar 15, 2017 •

edited by ghost

Loading

ghost commented Mar 15, 2017 •

edited by ghost

Loading

zcorpan Mar 16, 2017

zcorpan commented Mar 16, 2017

lyzadanger commented Mar 16, 2017

lyzadanger commented Mar 16, 2017

ghost commented Mar 23, 2017 •

edited by ghost

Loading

zcorpan Mar 24, 2017

lyzadanger Mar 24, 2017

zcorpan commented Mar 24, 2017

lyzadanger commented Mar 24, 2017

zcorpan commented Mar 27, 2017

zcorpan commented Mar 27, 2017

lyzadanger commented Mar 27, 2017

bzbarsky commented May 5, 2017

bzbarsky commented May 5, 2017 •

edited

Loading

HTML: Extend Test Coverage for _parent named browsing context (7.1.5) #5150

HTML: Extend Test Coverage for _parent named browsing context (7.1.5) #5150

Conversation

lyzadanger commented Mar 15, 2017

wpt-pr-bot commented Mar 15, 2017

ghost commented Mar 15, 2017 • edited by ghost Loading

Firefox (nightly channel)

All results

ghost commented Mar 15, 2017 • edited by ghost Loading

Chrome (unstable channel)

All results

zcorpan Mar 16, 2017

Choose a reason for hiding this comment

zcorpan commented Mar 16, 2017

lyzadanger commented Mar 16, 2017

lyzadanger commented Mar 16, 2017

ghost commented Mar 23, 2017 • edited by ghost Loading

zcorpan Mar 24, 2017

Choose a reason for hiding this comment

lyzadanger Mar 24, 2017

Choose a reason for hiding this comment

zcorpan commented Mar 24, 2017

lyzadanger commented Mar 24, 2017

zcorpan commented Mar 27, 2017

zcorpan commented Mar 27, 2017

lyzadanger commented Mar 27, 2017

bzbarsky commented May 5, 2017

bzbarsky commented May 5, 2017 • edited Loading

ghost commented Mar 15, 2017 •

edited by ghost

Loading

ghost commented Mar 15, 2017 •

edited by ghost

Loading

ghost commented Mar 23, 2017 •

edited by ghost

Loading

bzbarsky commented May 5, 2017 •

edited

Loading