204659 - Security manager check needs adjusting

Status: VERIFIED FIXED

(Core :: XPConnect, defect)

Description

[David Bradley]

I didn't get the security manager checking logic quite right. What should be
happening is if the security manager pointer returned is null, we shouldn't fail
the check, but allow it to pass.

Patch coming up

Comment 1

[David Bradley]

Attachment: Corrects the security manager checks

Comment 2

[David Bradley]

Comment on attachment 122610 [details] [diff] [review]
Corrects the security manager checks

Small tweak to the security manager checks. The end result is that this allows
COM objects to be instantiated inside XPCShell and brings the check in line
with the rest of XPConnect security checks.

Comment 3

[Alec Flett]

Comment on attachment 122610 [details] [diff] [review]
Corrects the security manager checks

sr=alecf

Comment 4

[David Bradley]

I'll see if I can get this in for 1.4. I don't think it's that critical, as it
only affects chrome and xpcshell. But it's a simple change and only affects
IDispatch code.

Comment 5

[Adam Lock]

Comment on attachment 122610 [details] [diff] [review]
Corrects the security manager checks

r=adamlock

Comment 6

[David Bradley]

Comment on attachment 122610 [details] [diff] [review]
Corrects the security manager checks

This is small two liner that I'd like to try and get in for 1.4. It will make
QA's life easier, since it allows GeckoActiveXObject to be called from XPCShell
for unit testing. Currently it blocks all components from instantiation due to
the errant security check logic.

Comment 7

[(not reading, please use [e-mail] instead)]

Comment on attachment 122610 [details] [diff] [review]
Corrects the security manager checks

a=sspitzer for 1.4 final

Comment 8

[David Bradley]

Patch checked in

Comment 9

[Phil Schwartau]

Checkin verified -