920586 - Allows chrome usages of TCPSocket

Status: RESOLVED FIXED

(Core :: DOM: Core & HTML, defect)

Description

[Alexandre Poirot [:ochameau]]

For the Fx OS Simulator addon, we are using TCPSocket as an xpcom:
  let TCPSocket = Cc["@mozilla.org/tcp-socket;1"]
                    .createInstance(Ci.nsIDOMTCPSocket);
  let socket = TCPSocket.open("127.0.0.1", 5037, {});

But the thing is that, in order to do that, we are toggling the dom.mozTCPSocket.enabled pref, that also enables it for content.

If I remember correctly nsIDOMGlobalPropertyInitializer behavior, given that we returns null in `init` method when the pref is false, we shouldn't need to disable open and listen methods as mozTCPSocket will be null or undefined in content.

Comment 1

[Alexandre Poirot [:ochameau]]

Attachment: Allows chrome usages

Here is a patch to remove the two checks in open and listen, but I'm also fine keeping checks as soon as we can somehow toggle chrome usage without enabling content one.

Comment 2

[Boris Zbarsky [:bzbarsky]]

>as soon as we can somehow toggle chrome usage without enabling content one.

Move to WebIDL?  ;)

Comment 3

[(no longer active)]

(In reply to Boris Zbarsky [:bz] from comment #2)
> >as soon as we can somehow toggle chrome usage without enabling content one.
> 
> Move to WebIDL?  ;)

Yes, seriously!

Comment 4

[Alexandre Poirot [:ochameau]]

Hum sure, WebIDL, but it sounds way more complex than removing 4 lines :p
It's not very clear what is the state of bug 885982. The patch is most likely bitrot, but is it really about only keeping test_tcpsocket_ipc.js in some way or another? (It looks like we should just be doing a mochitest-plain, don't we? Mochitest-plain would run in a child on b2g.)

I can surely give a hand about testing, as I have some background on mochitests and b2g,
but what concerns me is that I'd need to uplift to aurora... I'm not sure the move to WebIDL is a good candidate for uplift.

Comment 5

[Boris Zbarsky [:bzbarsky]]

Yes, if you need to uplift then "use WebIDL" is not viable.

Comment 6

[Alexandre Poirot [:ochameau]]

Comment on attachment 809950 [details] [diff] [review]
Allows chrome usages

bent looks busy, trying another potential review for this file.

Comment 7

[Ed Morley [:emorley]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/7dbd88b7f43b

Comment 8

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/7dbd88b7f43b