[Webkit-unassigned] [Bug 16401] [GTK] GObject/C DOM binding

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=16401





------- Comment #82 from lkcl at lkcl.net  2008-10-17 09:41 PDT -------
> (O) As Luke states in comment #48, these changes were needed for the
> patch to run at all (he didn't receive any answer to that, by the way).

 yes, i didn't.  (that is the key reason why i was annoyed)

> The way it looks, there seems to be a refcounting problem somewhere,
> either introduced, or exposed by the patch.

 well, if i remove the code from GDOMBinding.cpp which is designed to
 help debug refcounting problems, we can't FIND OUT, CAN WE!!!

 :)

 and yes, at the request of the reviewers, i have been forced to remove it,
 because it had "#ifdef 0" around it.

 explanations in the code from which it was cut-and-paste - from JSBindings -
 state that the code is [even #ifdef'd out there] necessary to check for
 refcounting errors.

 leaving that aside:

 i am reasonably confident that, having used RefPtr in the right places,
 and having done ref increasing in the constructors, and ref decreasing
 in the destructors, i am reasonably confident that there are not any
 refcount errors.

 certainly, the code can be utilised for exceptionally large and
 complex applications for prolonged periods of time, and
 "nothing bad happens", leading me to conclude, empirically, that at least
 there are no _zero_ referencing errors or _negative_ referencing errors.

 whether there are any "over" referenced errors, resulting in memory leaks,
 will need some investigation.


 i forget, unfortunately, how and where the refcounting is increased and
 decreased.  certainly, the code to do both is auto-generated by the gobject
 bindings perl module, so you will need to run it and then look in
 ....

 damn i've even forgotten the name of the directory where the generated
 sources go.


-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.