https://bugs.wireshark.org/bugzilla/show_bug.cgi?id=6457
--- Comment #16 from Michael Mann <mmann78@xxxxxxxxxxxx> 2011-10-14 05:45:31 PDT ---
(In reply to comment #14)
> Created an attachment (id=7233)
--> (https://bugs.wireshark.org/bugzilla/attachment.cgi?id=7233) [details]
> Add cf_callback_invoke to cf_open
> My novice analysis is that frame_data is being destroyed underneath the GUI
> when switching files and that is causing the crash (as cells try to redraw data
> that doesn't exist).
> Looking through the comments of various cf_ functions, it appears the GUI needs
> to "cleanup" before the data is cleared (so it doesn't try to reference that
> data) and cf_callback_invoke(cf_cb_file_closing, cf) is just the function to do
> it.
> The attached patch stopped the crashes on my setup, but someone with more
> experience with the GUI should evaluate its validity.
Just noticed that the patch forces a close which can briefly bring up the
"start up" screen before opening the second file. Comments within file.c
suggest the goal is 'to go from "old file open" to "new file open and being
read"', which this patch breaks, but I don't see a way around the original
issue of the GUI still trying to read data from the "old file" that has since
been freed. The GUI wanting to still read data from the "old file" appears to
be a timing issue as I can't always reproduce the bug (and others have had a
difficult time reproducing it).
--
Configure bugmail: https://bugs.wireshark.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are watching all bug changes.