864775 - Use proper compartment when deserializing ArrayBuffer from parent process

Status: RESOLVED FIXED

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

Description

[Josh Matthews [:jdm]]

Without this, the email app remains broken.

Comment 1

[Josh Matthews [:jdm]]

Attachment: Use unwrapped content window object to determine compartment when creating new ArrayBuffers for use by content from TCPSocket. r=

Comment 2

[:gerard-majax]

Thanks, I can confirm the patch fixes email on my Nexus S too.

Comment 3

[Josh Matthews [:jdm]]

Comment on attachment 740818 [details] [diff] [review]
Use unwrapped content window object to determine compartment when creating new ArrayBuffers for use by content from TCPSocket. r=

For reference, useWin comes from unwrapping (http://hg.mozilla.org/mozilla-central/annotate/acf388eaf9e9/dom/network/src/TCPSocket.js#l306) the window provided as part of nsIDOMGlobalPropertyInitializer.init.

Comment 4

[Andrew Sutherland [:asuth] (he/him)]

Yeah, confirmed on unagi for v1-train too.  Thanks, Josh!

Comment 5

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 740818 [details] [diff] [review]
Use unwrapped content window object to determine compartment when creating new ArrayBuffers for use by content from TCPSocket. r=

r=me

Comment 6

[Josh Matthews [:jdm]]

https://hg.mozilla.org/projects/birch/rev/5a62b421e490

Comment 7

[Josh Matthews [:jdm]]

Build fix for non-b2g18: https://hg.mozilla.org/projects/birch/rev/86018e20f5c8

Comment 8

[Josh Matthews [:jdm]]

Final build fix: https://hg.mozilla.org/projects/birch/rev/4676b8c35f46

Comment 9

[Josh Matthews [:jdm]]

Attachment: Use unwrapped content window object to determine compartment when creating new ArrayBuffers for use by content from TCPSocket.

b2g18 patch.

Comment 10

[Josh Matthews [:jdm]]

Attachment: Use unwrapped content window object to determine compartment when creating new ArrayBuffers for use by content from TCPSocket.

Updated b2g18 patch.

Comment 11

[Josh Matthews [:jdm]]

And https://hg.mozilla.org/projects/birch/rev/761a78c8ab03 because, yeah, running tests is good.

Comment 12

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/5a62b421e490
https://hg.mozilla.org/mozilla-central/rev/86018e20f5c8
https://hg.mozilla.org/mozilla-central/rev/4676b8c35f46
https://hg.mozilla.org/mozilla-central/rev/761a78c8ab03

Comment 13

[Andrew Overholt [:overholt]]

Needed for bug 863770 which is tef+ so tef+ here, too.

Comment 14

[Andrew Sutherland [:asuth] (he/him)]

I uplifted patched 1, 3, and 4 to mozilla-b2g18 which are equivalent to the updated b2g18 patch here on the bug after running tests.  (Patch 2 was a fix for m-c for a patch made for 18.)  This was done for maximum correspondence, but done as a push so push-aware bisection should be happy.

https://hg.mozilla.org/releases/mozilla-b2g18/rev/6809f965cdd7
https://hg.mozilla.org/releases/mozilla-b2g18/rev/d51b1bb27054
https://hg.mozilla.org/releases/mozilla-b2g18/rev/2376197bf4de

I'll uplift to v1.0.1 once it's clear that a git mirror derived build is 100% happy with this and TBPL remains equivalently happy.

Comment 15

[Andrew Sutherland [:asuth] (he/him)]

pushed to mozilla-b2g18_v1_0_1:
https://hg.mozilla.org/releases/mozilla-b2g18_v1_0_1/rev/18351f56f9d5
https://hg.mozilla.org/releases/mozilla-b2g18_v1_0_1/rev/ae7f06523b43
https://hg.mozilla.org/releases/mozilla-b2g18_v1_0_1/rev/250b88651582