Subject: Re: err(3) and error handling
To: None <mycroft@gnu.ai.mit.edu>
From: Ian Fitchet <I.D.Fitchet@fulcrum.co.uk>
List: current-users
Date: 02/06/1995 11:02:47
 On Sat, 4 Feb 95 18:17:22 est, mycroft@gnu.ai.mit.edu said:

>What would you rather it print?
>[...]
>machine, or ...  (Or maybe you ought to RTFM?)

 RTFM or better RTFS may well solve the problem, provided you can
understand it!

 Whilst the traditional *NIX voodoo obscurity may suit many people I
don't see why NetBSD can't/shouldn't break the mould and be a bit more
punter friendly with relatively little overhead.  I presume that in
the new all singing/all dancing OO OSs error recovery and descriptive
information is altogether easier.


 However, what about a string (or number of strings) analogous to
errno?  If a system call fails the system call can set errstr to point
to a more descriptive reason why.  What I was thinking was that each
layer on the way back out could then (ap|pre)pend more information.
That could get messy (and involve lots of alloc'ing) hence an array of
descriptive strings, say, up to ten (of pointers to static strings?).


 There again, perhaps it's not such a good idea.


 Doubtless this has all be debated before, in which case can somebody
ressurect the conclusions?

Cheers,

	Ian

-------------------------------------------------------------------------------
      Ian Fitchet			I.D.Fitchet@fulcrum.co.uk
  Fujitsu Telecommunications Europe ltd., Birmingham Business Park, B37 7YU