version 1.2

using character expressions such as [ac-z] or [ace-z] seem to lead to
broken regexps.  the following two tests for RETest should show the
problem:
#149
[ac-z]+
ace
YES
ace

#150
[ace-z]+
ace
YES
ace

the results that i get are:
149. [ac-z]+

   Match against: 'ace'
   Match: YES
   Paren count: 1
   Paren 0 : ce


*******************************************************
*********************  FAILURE!  **********************
*******************************************************


Register 0 should be = "ace", but is "ce" instead.

150. [ace-z]+

   Match against: 'ace'
   Match: YES
   Paren count: 1
   Paren 0 : a


*******************************************************
*********************  FAILURE!  **********************
*******************************************************


Register 0 should be = "ace", but is "a" instead.

it would seem that the range is somehow invalidating the previous simple
character.

on a related note, the only way to match '-' in a character class is to 
escape it and add it at the end.  that seems wrong.