[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Issue: PATHNAME-STREAM (Version 5)

    Date: 23 Oct 87 09:52 PDT
    From: Masinter.pa@Xerox.COM

    In this version, I explicitly disallow synonym streams (and two-way,
    echo, string, broadcast streams), and explicitly allow streams created
    with open and with-open-file. 

I believe synonym streams should pass the PATHNAME operation on to their
target, and therefore a synonym stream should be acceptable for coercion
to a pathname if its current target is acceptable.  This seems to me to
be more consistent with the intent of p.329, although since the term
"operations" [on a stream] does not seem to be defined anywhere, it's
hard to be sure.  Also it seems useless to have a restriction against
synonym streams here.

				  I enhanced the discussion section, and
    added a reference to p. 417  "the name returned represents the name used
    to open the file, which may not be the actual name of the file".

    I'm actually puzzled now as to whether it is legal to do with Xerox
    Common Lisp does, which is only remember the TRUENAME of a stream once
    the stream is open. The language on p. 417 seems to imply that it is
    necessary also to remember the pathname supplied to open. Clarification?

The second paragraph under namestring on p.417 seems very explicit that
the pathname that was supplied to OPEN must be remembered.  Of course
file systems are by nature so implementation-dependent that it would be
difficult for someone to construct an unequivocal test case that would
catch XCL out.  The relation between a pathname and its truename cannot
really be defined in an implementation-independent way, I think.

Version 5 of PATHNAME-STREAM:FILES-ONLY is fine with me otherwise.


NAMESTRING etc. (p. 417) should be added to the references.
The edit history has two entries for version 4.
Right at the end, "currently no way portable" has two words