902611 - test_bug475156.html failing on b2g mochitest with (NS_NOINTERFACE) [nsISupports.QueryInterface]

Status: RESOLVED FIXED

(Core :: Networking, defect)

Description

[Martijn Wargers (dead)]

Follow-up from bug 901343, comment 7.
This is the code that's causing this error:
20   var ch = SpecialPowers.wrap(xhr).channel.QueryInterface(SpecialPowers.Ci.nsICachingChannel);

from bug 901343, comment 8:
"
Well, necko in a child process acts a bit differently... e.g. HttpChannelChild does not in fact QI to nsICachingChannel.  So at least that test will need fixing to work in an e10s environment...
"

Comment 1

[Martijn Wargers (dead)]

Attachment: 475156.diff (poc)

This is a poc patch that seems to work, not sure if this is the best way to fix this.

Comment 2

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 787112 [details] [diff] [review]
475156.diff (poc)

Honza?

Comment 3

[Martijn Wargers (dead)]

Fromm bug 901343, comment 10:
Yes, we deliberately split out a subset of the the nsICachingChannel features into nsICacheInfoChannel, which is all that you get on the child side at the moment.

Comment 4

[Martijn Wargers (dead)]

Attachment: 475156.diff

Ok, QI to nsICacheInfoChannel works very well, also works on desktop, so I can just replace the existing nsICachingChannel, I think.

Comment 5

[Jason Duell]

Comment on attachment 787371 [details] [diff] [review]
475156.diff

Review of attachment 787371 [details] [diff] [review]:
-----------------------------------------------------------------

glad that worked.

Comment 6

[Martijn Wargers (dead)]

Just to make sure it's not causing failures somehow on some platform, I pushed to try: https://tbpl.mozilla.org/?tree=Try&rev=16f49767dff7

Comment 7

[Martijn Wargers (dead)]

Attachment: 902611.diff (for check-in)

Everything green, this is for check-in.

Comment 8

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/integration/b2g-inbound/rev/2f8ad5343149

Comment 9

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/2f8ad5343149