230368 - better refcount logging for nsXULPrototypeNode subclasses

Status: RESOLVED FIXED

(Core :: XUL, defect, P1)

Description

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

nsXULPrototypeNode and its subclasses are reference counted.  I have a patch to
hook them up to refcount logging.

Comment 1

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Attachment: patch

Comment 2

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Attachment: patch

sizeof(*this), not sizeof(this)

Comment 3

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Attachment: patch

Plus a change to fix some wacky code in the XBL content sink.

Comment 4

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Comment on attachment 138596 [details] [diff] [review]
patch

Well, I can no longer reproduce the leak I was trying to find using this patch,
but I think it's a good patch anyway.

(The leak was that I was seeing 3 nsXULPrototypeScript and 4
nsXULPrototypeElement leaked, with one leaked reference each.  The former held
JS GC roots that held a bunch of other stuff.  The whole thing was just doing
"./mozilla -mail".)

Comment 5

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Attachment: patch

cleaner approach

Comment 6

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Attachment: patch

ok, virtual functions aren't good things to call from base class constructors

Comment 7

[Brendan Eich [:brendan]]

Comment on attachment 138599 [details] [diff] [review]
patch

Drive-by: looks good to me, just lose the extraneous ; on the nearby line shown
in context:

     virtual void ReleaseSubtree() { Release(); };

/be

Comment 8

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Fix checked in to trunk, 2003-01-08 16:01 -0800.

Comment 9

[Stephen Donner [:stephend] Not actively reading bugmail]

I don't have time/access to run Purify anymore, but a lot of leaks that I've
filed in the past should still be valid...