Re: lynx-dev 2-8-1 bug: print vs partial

[Klaus Weide]

On Thu, 3 Dec 1998, Philip Webb wrote:

> 981203 KW & TD commented on the print/e-mail bug: 
> >> it's something to do with a `sigbus', which is beyond me.
> > Yes, if it is memory corruption (in this case, freeing memory
> > that another part of the program may still access later),
> > then it is unpredictable where a problem will occur (if at all).
> > SIGSEGV or SIGBUS may be generated in a place that looks totally unrelated.
> -- snip --
> 
> that matches the result i described in `Further (2)'.
> it is weird that the problem arises
> only when  p  is invoked PART-WAY INTO the document (not at the top):
> does that suggest anything?

Maybe, but I don't think it is worth pursuing.  But if your urge for
detective work is too strong to resist, run the program in a debugger...

> i'm due to fall asleep soon & will try patches & further research thereafter.
> 
> BTW for my continuing education:
> what is a segmentation error (SIGSEGV) & why is it so named?

It's usually enough to know that it is Something Bad, that Should Not
Happen, and in most cases caused by errors with malloc(), free(), and
other cases of pointers to invalid memory.  But your system should have
man pages, see man signal (I have signal(2) and signal(7)) and so on.

Somebody else may know (or speculate about) a good reason for the name...

   Klaus