2012-01-04



I just converted a similar C routine to  the  standard  strchr().
Shortly  afterwards,  I discovered a bug: endless looping or seg-
fault.  It was likely related to the  strchr()  switch,  but  the
code  appeared to be fine.  The man page strchr(3) wasn't talking
about irregularities neither.  It was surprising to  me  to  find
the following note to strchr() in POSIX:

     The terminating null byte is considered to be  part  of
     the string.  [0]

Of course, this explained the bug.

But huh ... is this consistent? Well, likely it is,  although  it
surprises.  Seems I haven't understood the principles well enough
yet.


This is a starting point for a discussion on  the  idea  of  few,
consistent rules vs. the principle of least surprise.

[0]  http://pubs.opengroup.org/onlinepubs/007904975/functions/strchr.html


http://marmaro.de/lue/        markus schnalke <meillo@marmaro.de>