1388974 - canvas.captureStream throws NS_ERROR_NOT_INITIALIZED if a 2d context hasn't been made

Status: RESOLVED DUPLICATE of bug 1257440

(Core :: WebRTC: Audio/Video, defect, P3)

Description

[Andrew Morris]

User Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_12_6) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/60.0.3112.90 Safari/537.36

Steps to reproduce:

Ran https://output.jsbin.com/yayati


Actual results:

Got what appears to be an internal error:

Error: [Exception... "Component not initialized"  nsresult: "0xc1f30001 (NS_ERROR_NOT_INITIALIZED)"  location: "JS frame :: http://output.jsbin.com/yayati :: <TOP_LEVEL> :: line 32"  data: no]
.message: 
.stack: @http://output.jsbin.com/yayati:32:18


Expected results:

Output should be a list of rtpmap lines from an sdp.

Comment 1

[Daosheng Mu[:daoshengmu]]

Andrew, your user agent seems to be Chrome's. Could you go to http://www.whoishostingthis.com/tools/user-agent/ to get your real agent?

Besides, I didn't see your source code. Could you give more information about your code?

IIUC, RTP is not related with Canvas. Right?

Comment 2

[Andrew Morris]

@Daosheng: Ah yeah, I used Chrome to file the bug but the output is from my Firefox testing. I just ran it again in Firefox and got the same result with this user agent:

Mozilla/5.0 (Macintosh; Intel Mac OS X 10.12; rv:54.0) Gecko/20100101 Firefox/54.0

Also, yeah RTP isn't the issue, that's just the expected output of the program. It's using the canvas stream as a video source for WebRTC, but it fails to get the canvas stream.

Comment 3

[Daosheng Mu[:daoshengmu]]

I think you need to call canvas.getContext('2d'); before you start to captureStream(). Otherwise, it will throw error message here, http://searchfox.org/mozilla-central/rev/999385a5e8c2d360cc37286882508075fc2078bd/dom/html/HTMLCanvasElement.cpp#740.

By the way, it should be belong to WebRTC module. Let me ni? pehrson, probably, he would have some feedback.

Comment 4

[Andreas Pehrson [:pehrsons]]

Thanks. Yeah this is a known spec-compliance issue we have in the canvas captureStream implementation. The spec says, as your bug description, that captureStream() should just work. Until we fix this the workaround is to create a context by requesting one (2d or webgl) first. Not sure if there is an existing bug for this, so I'll keep this one.

Comment 5

[Bulk Bug Changes for mreavy's org]

Mass change P2->P3 to align with new Mozilla triage process.

Comment 7

[Andreas Pehrson [:pehrsons]]

There's even the older bug 1257440 for this.