1448181 - media readyState behavior changed

Status: RESOLVED DUPLICATE of bug 1457661

(Core :: Audio/Video: Playback, defect, P2)

Description

[ddrogahn]

User Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:61.0) Gecko/20100101 Firefox/61.0
Build ID: 20180322100349

Steps to reproduce:

In current versions: a force-preloaded media file does not go past readyState 3 HAVE_FUTURE_DATA.
Did work in previous versions, and does work in other browsers.

https://jsfiddle.net/ddrogahn/sdjadd2s/

Force-Preload: fetch(url), URL.createObjectURL(response.blob), audio.src = objUrl, wait for readyState or event


Actual results:

oncanplay event
readyState = 3

59.0 and 59.0.1 and Nightly 61.0a1 (2018-03-22) have this behavior


Expected results:

oncanplay event
oncanplaythrough event
readyState = 4

Previous FF version: v58.0.2 is OK
Other browsers: Edge 16, Chrome 65.0.3325.181 are OK.

Comment 1

[Alice0775 White]

Regression window:
https://hg.mozilla.org/integration/autoland/pushloghtml?fromchange=772fef647fd7750f8c87dfd4f55bbd85046a555a&tochange=821560cd716d50fbb0c49d0170bd351255e55bc3

Regressed by: Bug 1423465

@:jw_wang,
Your bunch of patch seems to cause the regression, can you look into this?

Comment 2

[dustin.kerstein]

This also affects video playback. See the console output here - https://jsfiddle.net/h21qrzno/

Comment 3

[dustin.kerstein]

Any updates on this? This is definitely breaking a few video implementations I've seen.

Comment 4

[Ryan VanderMeulen [:RyanVM]]

Gerald, do you have cycles to look at this by chance?

Comment 5

[Gerald Squelart (he/him) (not at Mozilla since 2022-09-15)]

JW is gone, and officially I'm not in media anymore, all my cycles are spent in layout.

Jean-Yves, you're probably the best person left! Please contact me if you'd like to discuss.

Comment 6

[Jean-Yves Avenard [:jya]]

:bryce, is this something you could take?

Comment 7

[Andreas Pehrson [:pehrsons]]

Seems like a fairly high-impact regression that already made it to release. Making it a high P2 but feel free to adjust.

Comment 8

[Bryce Seager van Dyk [:bryce] (he/him) - Not reading bugmail]

Looks like this is a dupe of bug 1457661. Have verified the fix there has fixed this in all versions where the fix was applied. 1457661 landed test coverage, but I'm going to take a look at if there's anything extra we should be covering that's different for this case. If so, I will raise a follow up bug for further tests.

Comment 9

[Bryce Seager van Dyk [:bryce] (he/him) - Not reading bugmail]

Following up from comment 8: I think the already landed test provides sufficient coverage of the underlying issue for both bugs.