User Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_10_5) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/47.0.2526.106 Safari/537.36 Steps to reproduce: I loaded a page with a <video> element. I added a <track> element as a child of the <video> element. Test case: http://jsfiddle.net/jxLf4cnv/1/ Actual results: video.textTracks.length === 0 after appending a new <track> element to the <video> Expected results: The <track> should show up in video.textTracks

This may be similar to another bug I recently reported: Bug 1242594 - textTracks lost when video removed from DOM

P2

line 1: video.appendChild(track); line 2: added.innerHTML = video.textTracks.length; Looks like the HTMLTrackElement::LoadResource is an asynchronous call in HTMLTrackElement::BindToTree. So line 2 will get 0 if HTMLTrackElement::LoadResource is not finished.

Review commit: https://reviewboard.mozilla.org/r/52710/diff/#index_header See other reviews: https://reviewboard.mozilla.org/r/52710/

Review commit: https://reviewboard.mozilla.org/r/52712/diff/#index_header See other reviews: https://reviewboard.mozilla.org/r/52712/

Comment on attachment 8752695 [details] MozReview Request: Bug 1242599 - Create TextTrack before LoadResource(). r=rillian https://reviewboard.mozilla.org/r/52710/#review49810

Comment on attachment 8752696 [details] MozReview Request: Bug 1242599 - testcase. r=rillian https://reviewboard.mozilla.org/r/52712/#review49812 ::: dom/media/test/test_bug1242594.html:34 (Diff revision 1) > > document.getElementById("content").appendChild(video); > video.appendChild(trackElement); > > -video.addEventListener("loadedmetadata", function run_tests() { > +// Bug 1242599, access video.textTracks.length immediately after > +// the track element bind into the media element. Removing the event trigger makes me a little nervous that this will be racy. I recall vaguely that we changed things so we don't need to wait for loadedmetadata.Is this correct behaviour by the spec? ::: dom/media/test/test_bug1242594.html:35 (Diff revision 1) > document.getElementById("content").appendChild(video); > video.appendChild(trackElement); > > -video.addEventListener("loadedmetadata", function run_tests() { > +// Bug 1242599, access video.textTracks.length immediately after > +// the track element bind into the media element. > - is(video.textTracks.length, 1, "Video should have one TextTrack."); > +is(video.textTracks.length, 1, "Video should have one TextTrack."); Should be 'element binds'. Third-person singular.

Comment on attachment 8752696 [details] MozReview Request: Bug 1242599 - testcase. r=rillian https://reviewboard.mozilla.org/r/52712/#review49814

(In reply to Paul Sayre from comment #0) > Test case: http://jsfiddle.net/jxLf4cnv/1/ BTW, I am astounded by your testcase accessing elements by id without going through querySelector or getElementById. How does that work?

bz pointed me at bug 622491 and https://www.w3.org/Bugs/Public/show_bug.cgi?id=11960 https://twitter.com/bz_moz/status/732332291842772993

Ralph, in the code where I discovered the issue, I don't actually know where the node comes from, so I was mostly lazy in the test case.

Ok, thanks. Seems to be IE 'feature'.

https://reviewboard.mozilla.org/r/52712/#review49812 > Removing the event trigger makes me a little nervous that this will be racy. > > I recall vaguely that we changed things so we don't need to wait for loadedmetadata.Is this correct behaviour by the spec? The spec only says the loadedmetadata fired after "the text tracks are ready". I don't see any relation between "loadedmetadata" and "video.textTracks".

Comment on attachment 8752695 [details] MozReview Request: Bug 1242599 - Create TextTrack before LoadResource(). r=rillian Review request updated; see interdiff: https://reviewboard.mozilla.org/r/52710/diff/1-2/

Comment on attachment 8752696 [details] MozReview Request: Bug 1242599 - testcase. r=rillian Review request updated; see interdiff: https://reviewboard.mozilla.org/r/52712/diff/1-2/

Review commit: https://reviewboard.mozilla.org/r/57150/diff/#index_header See other reviews: https://reviewboard.mozilla.org/r/57150/

Comment on attachment 8752695 [details] MozReview Request: Bug 1242599 - Create TextTrack before LoadResource(). r=rillian Review request updated; see interdiff: https://reviewboard.mozilla.org/r/52710/diff/2-3/

Comment on attachment 8752696 [details] MozReview Request: Bug 1242599 - testcase. r=rillian Review request updated; see interdiff: https://reviewboard.mozilla.org/r/52712/diff/2-3/

Comment on attachment 8759060 [details] MozReview Request: Bug 1242599 - [webplatformtest] Enable /html/semantics/embedded-content/media-elements/track/track-element/track-api-texttracks.html. r=rillian https://reviewboard.mozilla.org/r/57150/#review54052 Hooray for passing more tests!

Pushed by cbook@mozilla.com: https://hg.mozilla.org/integration/mozilla-inbound/rev/c9dd7a6b157b Create TextTrack before LoadResource(). r=rillian https://hg.mozilla.org/integration/mozilla-inbound/rev/9d3cf4d5824e testcase. r=rillian https://hg.mozilla.org/integration/mozilla-inbound/rev/318fae85147f [webplatformtest] Enable /html/semantics/embedded-content/media-elements/track/track-element/track-api-texttracks.html. r=rillian

https://hg.mozilla.org/mozilla-central/rev/c9dd7a6b157b https://hg.mozilla.org/mozilla-central/rev/9d3cf4d5824e https://hg.mozilla.org/mozilla-central/rev/318fae85147f