[gvd-devel] GVD problem

[MALERBA_Vivien@stna.dgac.fr]

Hi!

First thanks for GVD, it's really nice. I've had some problems with it
though, and here is a bug report/comment for new feature.

Version of GVD: 1.1.0-i686-pc-linux-gnu
Running on a Linux Mandrake 7.2

* When I want to view the memory at an adress (from a variable in a
struct on the canvas), the memory window is opened without mention of
the adress in the 'Location' field, and the adress of the variable I
asked for is not displayed (I have to go up for that by one row).

* When I debug a program using GTK+ (compiled with -g), I simply run it
and then interrup it. At this time if I ask for the current file and
line number, I get the first line of instruction of main(). Is it a
normal behaviour?

* If I display on the canvas a struct having a GList pointer, then when
I want to explore the GList, I can only get pointers to the data fields
of each node of the list, because they are declared as void*. Would it
be possible to tell GVD what kind of cast to perform on this void*
pointer to actually display the structure pointer by the data field (it
would be of much help to debug lists)?

* I got a GVD bug dialog (actually two) when I tried to modify the
memory (from the memory view window): I tried to modify the memory
directely from the HEX view (entering a 31 instead of 30 within what is
a string). Maybe it is the wrong way to do it? The two errors I got are
(in that order):

Version: 1.1.0
Date: 20010307
Target: i686-pc-linux-gnu
Exception name: STORAGE_ERROR
Message: stack overflow (or erroneous memory access)
(skipped the Call stack traceback locations because too many hex to copy
by hand I don't think usefull since the bug is easy to reproduce;
however if you really need them I can send you the sources of my
program, and the hex list).

Version: 1.1.0
Date: 20010307
Target: i686-pc-linux-gnu
Exception name: PROGRAM_ERROR
Message: unhandled signal
(same remark for the call stack traceback locations)


I hope this will help.

thanks again,

Vivien Malerba