WG15 Defect Report Ref: 9945-2-134
Topic: wordexp() and IFS

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


Last update: 1997-05-20



	Topic:                  wordexp() and IFS
	Relevant Sections:      B.9 and 3.6

Defect Report:

	From: andrew@novell.co.uk (Andrew Josey)
	Date: Mon, 24 Jul 1995 17:02:52 +0100

The following is a request for interpretation regarding ISO/IEC 9945-2-92 ,

Just as is expected by the accompanying rationale, our implementation
of wordexp() executes a shell subprocess to do the word expansions.
We use the Korn shell (the one used as the basis for the POSIX.2 shell)
which resets its IFS on startup, as permitted by subclause

The test assertions for wordexp() presume that a modified IFS setting
will be reflected in its field splitting, and that a reset IFS is non-
conformant.  Is this the expected POSIX.2 behavior for wordexp()?

thanks in advance

Interpretation response

The standard clearly states the requirements for word expansion
and conforming implementations must conform to this. The fact that
IFS can be reset does not preclude implementations using the shell.

Forwarded to Interpretations group: Jul 25 1995
Recirculated for 30 day review: Oct 19 1995
Finalised: Nov 20 1995