WG15 Defect Report Ref: 9945-2-21
Topic: sh


This is an approved interpretation of 9945-2:1993.

.

Last update: 1997-05-20


								9945-2-21

	Class: Editorial defect


 _____________________________________________________________________________


	Topic:			sh
	Relevant Sections:	3.13.1


Defect Report:
-----------------------
 
          In Section 3.13.1 - Patterns Matching  a  Single  Character, 
          the standard states that ``special characters can be escaped 
          to remove their special meaning by  preceding  them  with  a 
          <backslash>.''  [Draft 12  of  ISO/IEC  9945-2:1993  (July 
          1992), p. 157, lines 1397-1399] The standard  defines  these 
          ``special  characters''  to  be  ``?,''  ``*,''  and  ``[.'' 
          [Ibid., p. 157, lines 1386-1389] 
 
          Thus, it appears that special characters within []  are  not 
          within the scope of the statement on lines 1397-1399.  Thus, 
          the statement in Section 2.8.3.2, ``[t]he special characters 
          .  *  [  \ (period, asterisk, left-bracket,  and  backslash, 
          respectively) shall lose  their  special  meaning  within  a 
          bracket expression'' [Ibid.,  pp  82-83.  lines  2892-2895], 
          presumably holds. 
 
          This implies to  me  that  the  pattern  bracket  expression 
          ``[a\-z]''  should  match  the  letter  a,  and  <backslash> 
          through  z,  as  it  does  in  regular  expressions.    This 
          disagrees  with  historical  practice,  where  this  pattern 
          matches a, -, and z. Is the  intent  to  disable  historical 
          practice, or can we interpret  lines  1397-1399  as  reading 
          ``...  special  characters,  including  special   characters 
          within a pattern bracket expression?'' 
 

WG15 response for 9945-2:1993 
-----------------------------------


The standard is specific in its requirement that RE behavior apply to
bracket expressions with the one noted exception. The example in the
request does not fulfill this exception, hence it must be interpreted
according to the RE rules (2.8.3.2). Concern over the wording of this
section has been forwarded to the Sponsors of the standard.

Rationale for Interpretation:
-----------------------------
None.
 _____________________________________________________________________________