Thus it was written in the epistle of Larry Boyd,
>
> Okay, maybe I'm missing something in your email, Ted.

Actually, I think you're missing something not in my email ;-)

>                                                        The HPFOPEN
> requires a starting and ending character for the name.  In your original
> HPFOPEN, you apparently met this requirement, since the correct file was
> opened.  In the second HPFOPEN, it doesn't appear that the ";" was
> placed in the first character - only the last.  Therefore, when HPFOPEN
> saw /SYS/ASHTED/TI; it properly started with the first "/" and ended
> with the second "/", deciding the name was SYS.  With ;/SYS/ASHTED/TI;
> it should work as you would like.

It seems I made the mistake of reading the manual.  If I may quote it here,

2  CA  Formal designator:

       Passes a formal file designator that is interpreted according to
       MPE-escaped semantics (unless another syntax has been chosen via item
       41).  The file name must be terminated by a nonalphanumeric character
       other than a period (.), a slash (/), a hyphen (-), and an underscore
       (_).  Use of matched starting and ending name delimiters (a quoted name)
       alleviates the need for a terminating character other than the quote
       characters.

Now there are some things which seem obvious from the above and which I now
believe to be false:
  1) Quoting characters are not necessary, so long as there is a terminating
     nonalpha-nonperiod-nonslash-nonhyphen-nonunderscore character.
  2) Alphanumeric characters and the slash are not valid quoting characters.

Those two beliefs can get you in trouble.  As far as I'm concerned, the
documentation is buggy.  All of that blither about a terminating character is,
so far as I can tell at the moment, rot, probably left over from some previous
definition.  If the first character of the character array IS GOING TO BE USED
AS A QUOTING CHARACTER NO MATTER WHAT and if using quoting characters
"alleviates the need for a terminating character" then why tell me about a
terminating character?!?!?

As I said at some length to the long-suffering Jeff Vance (see below), the
slash (and, I think, all alpha characters and probably the period) should be
eliminated as valid quoting characters.  That would allow FFILEINFO output to
be fed into HPFOPEN and work as expected.  That is, if the array starts with
any of ['A'..'Z','a'..'z','.','/'], the intrinsic should decide that we've
gone with the terminating character option instead of the matching delimiters
option.

> > It's almost 2am and I'm going home.  Tomorrow I'll put in the
> > ugly code to hand
> > it a pac delimited with "&"s or some such.  But as far as I'm
> > concerned, that
> > first behavior is ridiculous and the second bizarre.
>
> Ah!! 2am - BTDT and have made several minor sins of omission at this
> time of the morning with a deadline pending.  Typically, after a few
> hours of sleep I can look back at the code and see my omission
> immediately.  Hopefully, you'll have the same good luck today.

With some excellent help from the wonderful Jeff Vance and a little sleep, I
seem to be back on track, thanks.

Ted
--
Ted Ashton ([log in to unmask]), Info Serv, Southern Adventist University
          ==========================================================
The imaginary number is a fine and wonderful recourse of the divine spirit,
almost an amphibian between being and not being.
                        -- Leibniz, Gottfried Whilhem (1646-1716)