Re: flexfax: HylaFax bug in SNPPserver

[thoth@purplefrog.com]

Matthias.Apitz@SOFTCON.de ,in message <9710091807.AA07194@kant.SOFTCON.de>, wro
	te: 

> 
> BTW: Do you have some information why some systems map a NULL pointer
> to a 0x00 and others don't?

  Perhaps I should explain.  The null pointer is always comparable to zero in
an ANSI or C++ system.  I'm fairly certain they defined it that way.  The bit
battern in memory or register may be non-zero, but when you convert it or
compare it to an int, it must look like (int)0.

  But, I was talking about something else.  I have heard legends that if you
dereference a NULL pointer on some systems (I believe HPUX had this "feature")
it gives you a NUL character: '\0' .

  Consider the code

  char *a = 0;
  char c = *a;                /* SEGV on most systems */
  printf("%d", (int)c);       /* prints out 0 on the wacky systems */

  Yes, this can lead to sloppy coding by people weaned on such systems.  I
object to the very concept that you could dereference a NULL pointer and keep
running.

-- 
Bob Forsman                                   thoth@gainesville.fl.us
           http://www.gainesville.fl.us/~thoth/