141356 - Add JS_[Un]LockGCThingRT()

Status: VERIFIED FIXED

(Core :: JavaScript Engine, defect)

Description

[Brian Ryner (not reading)]

Given that LockGCThing and UnlockGCThing only need a JSRuntime, there should be
a version of this that doesn't require a JSContext.

Comment 1

[Mike Shaver (:shaver emeritus)]

I got yer back.  Patch coming up.

Comment 2

[Mike Shaver (:shaver emeritus)]

Attachment: add JS_{Lock,Unlock}GCThingRT

This builds and the shell starts, and it looks "obviously right" to me,
following the pattern established by the Root variants.  I haven't tested the
browser yet, but my build progresses...

Bryner, does this ease your pain?

Comment 3

[Brian Ryner (not reading)]

Well, not quite:

JS_PUBLIC_API(JSBool)
JS_UnlockGCThingRT(JSRuntime *rt, void *thing)
{
    return JS_UnlockGCThingRT(rt, thing);
}

is a good exercise in blowing out the stack :)

With that fixed, things do seem to work.

Comment 4

[Mike Shaver (:shaver emeritus)]

Did I say "obviously right"?  Who's using my bugzilla account?

*ahem*

Comment 5

[Brendan Eich [:brendan]]

Dudes, cc me next time (I'm trying to catch up on my bugmail).

/be

Comment 6

[Mike Shaver (:shaver emeritus)]

Attachment: More patch. Less duh.

Comment 7

[Brian Ryner (not reading)]

Comment on attachment 82045 [details] [diff] [review]
More patch. Less duh.

r=bryner

Comment 8

[Brendan Eich [:brendan]]

Comment on attachment 82045 [details] [diff] [review]
More patch. Less duh.

If you're game, make the parallel name change for js_RemoveRoot (that is,
suffix RT to that name).  More files to touch, but it's a safe substitution.

sr=[e-mail]

/be

Comment 9

[Brian Ryner (not reading)]

Checked in on the trunk and MOZILLA_1_0_0_BRANCH.

Comment 10

[Phil Schwartau]

Patch verified on trunk and branch -