[Gtk-sharp-list] Memory leak in tree cell func

Matjaž Mihelič mlgs at arsis.net
Sat Jul 18 08:35:11 EDT 2009

On Fri, 2009-07-17 at 22:18 -0500, Mike Kestner wrote:
> On Sat, 2009-07-18 at 04:25 +0200, Matjaž Mihelič wrote:
> > Same example as provided for gtk, except added empty method to handle
> > cell information and connected it in constructor.
> > 
> > Memory leak is constant on every call and more cells you use, more it
> > leaks. Easiest way to see it in action is simply resizing this window
> > for few seconds.
> Please file a bug report.  Does it only happen with a managed custom
> tree model implementation?


and lol, I don't know, but I suspect they do, because cell func is
pinned on column. 

trouble is that leak is less visible on dead forms (less redraws), but
when data is alive... 

I've been tracking this leak for 1 week now just to find out it is not
part of any lib that I wrote. but then again, I profiled my work to
insanity because of it (I had few not so serious leaks too, but this one
was hard to find because your example worked without leak, so I kept
going back to profile my own work) I couldn't even find it until I
started to search for profiling differences between your sample and my


More information about the Gtk-sharp-list mailing list