ups for Linux 7.1?

From: Guo, Liang (guol_at_cs.bu.edu)
Date: Thu Oct 04 2001 - 15:42:56 BST

  • Next message: Andrew Lister: "Re: can ups be operated from the keyboard?"
    Hi, folks:
    
    I recently upgraded my Linux to Redhat 7.1 with a gcc-2.96-85
    package installed. Then I downloaded ups-3.37-beta2 as well
    as ups-3.36 and tryed to install them on my machine.
    
    The frustrating thing is, now whenever I want to trace the
    value of some variable (by clicking on the variable), I come
    across the following error message and the debugger simply
    crashes:
    
    ***
    Fatal internal error: bad number in parse_num (aborting) ...
    Dumping core ... Abort
    ***
    
    So I went back and took a look at the source code, and tried to
    determine what's happened. Here's what I did: I printed the
    symbol ups debugger is trying to parse by replacing:
        if (!parse_number(stf, sr, p_s, &res)) {
    		panic("bad number in parse_num");
        }
    with:
        if (!parse_number(stf, sr, p_s, &res)) {
                    char buffer[80];
                    sprintf(buffer, "bad number in parse_num %s", *p_s);
                    panic(buffer);
        }
    
    Then the error message becomes:
    
    ***
    Fatal internal error: bad number in parse_num
    ;0;1;(0,1),0,64;; (aborting) ...
    Dumping core ... Abort
    ***
    
    That means the message passed to "parse_num" call is neither
    a string started with '\\' nor a number, that's why it returns
    a panic call.
    
    I think this bug should be related to Redhat kernels since I've
    never seen it when I use old version of Redhat.
    
    Did anyone come across the same problem? How did you fixed it?
    Thanks a lot! I like ups debugger a lot, the only problem is,
    once you get used to it, you start to hate all other debuggers
    like xxgdb.:)
    
    cheers.
    


    This archive was generated by hypermail 2.1.4 : Wed Feb 13 2002 - 21:58:06 GMT