758481 - When seeking in an unbuffered range after having ended a media, networkState is NETWORK_IDLE

Status: RESOLVED FIXED

(Core :: Audio/Video, defect)

Description

[Paul Adenot (:padenot)]

STR :
1. Load a page containing a video ;
2. Start playing the video, the networkState is NETWORK_LOADING ;
3. Seek near the end of the video, the video will buffer from that point, and then the download finishes. The networkState is now NETWORK_IDLE ;
4. Seek in an unbuffered range, in the middle of the video.

Expected result :

Per spec [1], the networkState should be NETWORK_LOADING :

> NETWORK_LOADING (numeric value 2)
>     The user agent is actively trying to download data.

Actual result :

networkState is NETWORK_IDLE.

I could reproduce this on current release, today's nightly. Not sure if it is a regression or it has always been like that.

Comment 1

[Paul Adenot (:padenot)]

Attachment: Patch v0 : When opening a channel, set the networkState to NETWORK_LOADING.

I'm currently also (re)writing test_networkState.html to be sure not to regress here (considering the incoming stop download / resume download feature), but I need other change to land before I can write a reliable test.

Comment 2

[cajbir (:cajbir)]

Comment on attachment 628233 [details] [diff] [review]
Patch v0 : When opening a channel, set the networkState to NETWORK_LOADING.

Review of attachment 628233 [details] [diff] [review]:
-----------------------------------------------------------------

::: content/html/content/public/nsHTMLMediaElement.h
@@ +155,5 @@
>    // asked the decoder to resumed the download.
> +  // If aForce is True, ignore the fact that the download has previously
> +  // finished. We are downloading the middle of the media after having
> +  // downloaded the end, we need to notify the element a download in ongoing.
> +  void DownloadResumed(bool aForce = false);

Can we call this 'aForceNetworkLoading' or something like that? 'aForce' doesn't really explain what it does.

Comment 3

[Paul Adenot (:padenot)]

Attachment: Addressed Chris' comments.

Comment 4

[cajbir (:cajbir)]

https://hg.mozilla.org/integration/mozilla-inbound/rev/28fe5eec1ca3

Comment 5

[cajbir (:cajbir)]

My apologies Paul, I had to fix a merge conflict and when committing and I forgot to set the author information for the patch to refer to you.

Comment 6

[Ed Morley [:emorley]]

https://hg.mozilla.org/mozilla-central/rev/28fe5eec1ca3