Re: *arg* debugging..

From: Patrick Dughi (
Date: 02/07/00

> using the GDB debugger is a very nice way of finding bugs, no doubt about
> that, in 9 out of 10 cases the error lies just in front of you, and a
> simple "bt" will tell you more than much other things. *BUT* every now and
> then, gdb cannot tell you why something crashed and comes up with:
> #0  0x0 in ?? ()
> Perhaps the most interessting parts must be in "info", or as someone
> suggested, this is simply an error in either of the libraries I include,
> in this case I simply add -lcrypt and nothing else, I had -lefence on
> untill recently, but it has been removed prior to this error.
> Anyone with some idea as to where to look for an error?

        One of the interesting things that happened a while ago is that
RedHat linux was shipping all it's libraries compiled specifically for a
Cyrix 686 type processor.  While this was only of passing interest, since
I tend to upgrade gcc as soon as possible, it did create a slew of errors
of this sort.  Of course, when I rebuilt the libraries, I enabled
debugging in some of them, and all those ??'s went away.  As a matter of
fact, the majority of them went away when I switched to using a garbage
collection library instead of the standard *alloc + free functions.

        So, I recommend you recompile your libraries, and if worse comes
to worse, enable debugging on some of the main ones (you'll probably want
to remove it later though, just to be on the safe side)

                                        Patrick Dughi

     | Ensure that you have read the CircleMUD Mailing List FAQ:  |
     |  |

This archive was generated by hypermail 2b30 : 04/10/01 PDT