Reply-To: netnews@gclare.org.uk
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 8bit
Lines: 56
Organization: Zen Internet
NNTP-Posting-Host: 217.155.142.121
X-Trace: 1170092462 prichard.zen.co.uk 27098 217.155.142.121:42612
X-Complaints-To: abuse@zen.co.uk
Xref: number1.nntp.dca.giganews.com comp.unix.shell:193671

I wrote, on Thu, 25 Jan 2007:

> Stephane CHAZELAS <this.address@is.invalid> wrote, on Tue, 23 Jan 2007:
> 

You might want to hold off on those bug reports.  I have been
doing some digging, and it may be that the situation on this
issue is not as clear cut as I previously thought.

When this came up in the Austin Group last March, I expressed
the opinion that I thought it was a defect in the standard, but
a decision was made in a teleconference in July which (I thought)
upheld the POSIX requirement.  Looking at it again, there is
something odd about the formal response, and I have queried it
on the Austin Group mailing list.  So there is still a possibility
that this might get treated as a defect in the standard.
[vbcol=seagreen] 
>
> Not exactly "notified" as such, but he was one of the technical
> reviewers of POSIX.2b (which added symlinks to POSIX.2, among
> other things; that is where the requirement for the shell to
> initialise PWD to a value that does not include any symlink
> components comes from).  So he should have been aware of it.
>
> The fact that ksh93 doesn't implement the requirement is
> certainly a bit of conundrum.

Once I dug back and found the old discussion of the issue, I found
some mail from David supporting my opinion that this is a defect in
the standard.  So that explains why current ksh93 still doesn't
implement the requirement (although the historical situation
regarding ksh93 and POSIX.2b is still a mystery).

--
Geoff Clare <netnews@gclare.org.uk>

[ Post a follow-up to this message ]