From mboxrd@z Thu Jan  1 00:00:00 1970
From: Jon Schutz <Jon.Schutz@youramigo.com>
To: Jochen Heinzmann <jochen@seeingmachines.com>
Cc: insight@sources.redhat.com
Subject: Re: Insight seg fault at startup
Date: Sun, 26 Nov 2000 22:14:00 -0000
Message-id: <14881.64447.623461.499310@gargle.gargle.HOWL>
X-SW-Source: 2000-q4/msg00275.html

Jochen,

    Jochen> I downloaded, compiled and installed insight successfully
    Jochen> from insight+dejagnu20001115 but I have problems starting
    Jochen> gdb. It starts okay with the -nw option, but otherwise
    Jochen> creates a segfault after a second of thinking.  I dumped
    Jochen> the core and had a look, it discovers some error, tries to
    Jochen> print the error with vprintf and stuffs up the argument
    Jochen> list and crashes WITHOUT printing the error.  Attached is
    Jochen> the text of the stack. If anybody could give me a hand
    Jochen> figuring out what my problem is it would be much
    Jochen> appreciated.

I didn't see any other responses on the list, so here's my offering:

I had the same problem with insight+dejagnu20001118 compiled with
gcc version 2.96 20000731 (Red Hat Linux 7.0).  My quick and dirty fix
was to change gdb/symtab.c:120 from

char no_symtab_msg[] = "No symbol table is loaded.  Use the \"file\" command.";

 to

const char *no_symtab_msg = "No symbol table is loaded.  Use the \"file\" command.";


For reasons that escape me, the pointer to no_symtab_msg was being
mangled in the call to error(const char *string,...) with the first
declaration.  An issue with the compiler, I expect.  What worries me
about the quick fix above is that there are many other message
declarations in the source with a similar declaration that may or may
not cause the same fault.  Unfortunately I don't have the time to look
at the problem any deeper, but so far the debugger has been working OK.

Regards,

Jon