Right you are, Glenn.

I'd forgot that SPEC calls something to scan a delimited string that
quietly gobbles up STRING to mean that it is a delimited string and
not test for hex or binary.

I'll revert SPEC, now that I am temporarily a bit wiser.  Right now
SPEC STRING STRING S 1 is broken.

Oh well.

   j.

2009/1/8 Glenn Knickerbocker <[email protected]>:
> On Thu, 8 Jan 2009 11:20:10 +0100, John wrote:
>>pipe literal | spec string 1-1 1 | cons
>>-
>
> This works as far back as z/VM 4.4.0 (and did any of us ever notice?).
> That's why I'm confused about what you're "introducing" now.
>
>>Yes, |specs #1:=1 print #1| is what I have in mind.  If you're telling
>>me it is a bad idea, I shall desist.
>
> I haven't come up with a case yet that's actually a valid spec today, but
> I think someone who accidentally left out an output column like that
> would be pretty astonished to have his pipeline run without errors and
> then find the string "1:=1 print " pasted over the front of all his
> records.  (If you ever added the convenience of filling in a default
> output column for any spec whose last item was a field, rather than just
> a single item, then this case would become ambiguous.)
>
> ¬R     "People late for work only deal with the essentials, cutting away
> uneccesary tasks like Occam's Razor with no time to shave." --oTTo--Bahn
>

Reply via email to