On 2016-06-09 17:20, Lizette Koehler wrote:
> Cross posting to IBM Main and TSO-REXX
>
> I have just created this RFE, so you may or may not be able to vote on it.
> But
> when available, I hope you will vote for it
>
> http://www.ibm.com/developerworks/rfe/execute?use_case=viewRfe&CR_ID=89817
> ...
> Also asked to add regular expressions ...
>
Good idea. Why not?
> DEL HLQ.%[LX9]%A.LIST MASK This would request a delete only if the file had
> somewhere in the second node LX9 and ends in A.
>
Where does that use of '%[' and ']%' come from? Is it customary syntax
in a facility I'm unfamiliar with? Otherwise, it's quite confusing to
those already familiar with regular expressions.
> HLQ.LX9A.LIST would delete
> HLQ.LX9.LIST would not delete
> HLQ.ABLX92A.LIST would delete.
>
If I understand, the regex for this in ISPF Edit notation would be:
DELETE r'^HLQ\..*LX9.*A\.LIST$'
> If you like it, please vote for it.
>
I'll go further. The convention should be made universal. Once, naïvely
I asked in ISPF-L (IIRC) why I can't use delimited pattern-strings in
DSLIST or member lists? "Of course not! Those work only in Edit (and
perhaps Browse)!" But why not? Think reusable code. Write the code
once (it's already been done; Edit uses it) and use it everywhere that
strings are matched.
I know; in UNIX filename matching patterns are different from regular
expressions. But here's an opportunity for z/OS to be more UNIform
than UNIX!
Alas, while matching in ISPF panels/lists is done by ISPF which can
make its own rules, TSO commands such as DELETE are a different matter
and would need to be addressed separately.
One of the great strengths of UNIX is that filename matching is done
by the shell, not by the individual commands, so it behaves UNIformly
over all commands.
-- gil
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
