Comment # 20
on bug 7833
from Roger
(In reply to comment #19)
> (In reply to comment #18)
>
> That's with Wireshark showing the welcome screen, with the interface list,
> list of recently opened files, etc., rather than showing the packet
> list/packet details/hex dump pane,
It occurs identically in both cases. However, when I do have a capture file
open, I also notice the memory usage increase a little when I expand a segment
of a frame (fair enough), but then it increases further when I collapse that
segment. If I do nothing else but re-expand the segment, memory usage increases
further, and again when I collapse it. If I sat here all day simply expanding
and collapsing one single segment of one single frame, I believe that I would
eventually induce the crash.
>
> Please file that as a separate bug; it *might* be the same as this bug, but
> it might not be. Your bug is clearly *not* a bug in the dissecting code, as
> there aren't any captured packets; it's probably either a bug in the Windows
> version of GTK+ or in whatever Wireshark code gets invoked on resizes and
> moves (it saves size and possibly position information to a file, but I have
> difficulty seeing that chew up memory that fast; we have some of our own
> widgets that might be handling those events, but I'm not sure any of them
> would be active if you're on the welcome screen.
>
OK, will do. For the record, my system is a Gigabyte M/B with a Core i7 and 6GB
of RAM. Graphics is an ATI Radeon HD4870, with driver version 8.861.0.0. All
disks are 1TB Western Digitals. I also run 64-bit Windows 7 on this same
hardware, so I'll fire that up and install Wireshark. I'll let you know how it
goes.
> I wasn't able to reproduce it with the same Wireshark version on my 32-bit
> XP SP3 virtual machine; perhaps it's some issue with the graphics adapter
> driver or an issue with GTK+ that only shows up with some graphics adapters?
Hmmmm....
You are receiving this mail because:
- You are watching all bug changes.