Re: gperl_callback_invoke userdata copy

[muppet <scott asofyet org>]

On May 3, 2008, at 6:30 PM, Kevin Ryde wrote:

> While nosing around for the tree filter modify callback I saw the
> comment in gperl_callback_invoke about needing or not needing to copy
> the userdata arg.  I came up with the degenerate program below; it
> unhooks the GtkLinkButton callback from within a callback, and the
> userdata arg goes away,
>
>    userdata: ARRAY(0x83c131c)
>    Use of uninitialized value in print at t-callback-data.pl line 15.
>    userdata:
>
> I wonder if it does, alas, have to do a copy to preserve the args  
> for a
> case like that.
>
> <t-callback-data.pl><GClosure.xs.invoke-copy.diff>

Luckily, the copy is fairly cheap.

I tried setting up a similar case with GLog handlers to create a unit  
test, but GLog appear to keep the handlers alive across invocation to  
avoid this issue.  So, without writing C code for a test, i don't know  
a way to unit test this.

With GtkLinkButton's hook, i don't think anybody ever expected a  
callback would unregister itself.

It's safer and more correct to use the copy, so we should probably  
commit it.


--
The one difference between Dali and a crazy man is very simple: Dali  
is not crazy at all.
  -- Salvador Dali