1210294 - Remove the release-mode IsCallerChrome assertions from nsDOMWindowUtils

Status: RESOLVED FIXED

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

Description

[Bobby Holley (:bholley)]

I think it's time for these things to go away. Many of these methods have infrequently-used C++ callers, and it doesn't seem worth converting all of them to LegacyIsCallerChromeOrNativeCode.

See https://crash-stats.mozilla.com/report/index/b07e8ec3-2344-4f8b-8761-8cacc2150927 and similar.

Comment 1

[Bobby Holley (:bholley)]

Attachment: Remove the release-mode IsCallerChrome assertions from nsDOMWindowUtils. v1

Comment 2

[Boris Zbarsky [:bzbarsky]]

I thought we were waiting on XPConnect not being exposed to the web, so we're very sure that this stuff is not ever being touched by content, before doing this?

Comment 3

[Bobby Holley (:bholley)]

(In reply to Boris Zbarsky [:bz] from comment #2)
> I thought we were waiting on XPConnect not being exposed to the web, so
> we're very sure that this stuff is not ever being touched by content, before
> doing this?

I kinda-sorta shifted my opinion to deciding that it probably doesn't matter. There's no classinfo for DOMWindowUtils anywhere, which means that we'll refuse to create a reflector for it in any content scope, even if chrome explicitly passes it to content (I just tested this to be extra sure).

So I more or less decided that there's no particular reason why we should be worried about that. But I'm also totally fine to be safe and just replace all of these to LegacyIsCallerChromeOrNativeCode. Your call.

Comment 4

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 8668274 [details] [diff] [review]
Remove the release-mode IsCallerChrome assertions from nsDOMWindowUtils. v1

r=me

Comment 5

[Pulsebot]

https://hg.mozilla.org/integration/mozilla-inbound/rev/63496344479e

Comment 6

[Phil Ringnalda (:philor)]

https://hg.mozilla.org/mozilla-central/rev/63496344479e