Re: cu coredumps

[Timour Ezeev <timour@xxxxxxxxxxxxxxxxxxx>]

	I don't have the code before me right now and made my comments based on
	the diff you proveded. You sure can free() cp before assigning it to
	tmp_cp. That should work too. Unless it points to a static memory block
	or memory on the stock. As I said, I don't have a source code here. But
	should be able to determine if it is safe to free() cp considering the
	above.

	Regards,

--
Timour Ezeev
Pivotal Dynamics
timour@xxxxxxxxxxxxxxxxxxx
V: 305.406.9904
F: 305.406.9689

On Wed, 27 Oct 2004, Mark Cullen wrote:

> Mark Cullen wrote:
> > Timour Ezeev wrote:
> >
> > > 	I think you got small memory leak there IMHO (either tmp_cp is not 
> > > used
> > > 	but allocated or cp is lost).
> >
> >
> > I think I see what you mean with the tmp_cp allocated but not used. I 
> > don't think I get you on the cp is lost part though. Is the attached 
> > better?
> >
> > Also, where would I put the free() exactly? I forgot about that..
>
> Actually, would I free(cp) just before reassigning cp to tmp_cp? Then 
> assuming cp got free'd later on in the code would that then be free'ing 
> tmp_cp? I'm guessing it wouldn't work quite like that, would it?