home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.lang.rexx
- Path: sparky!uunet!wrkgrp!ets
- From: ets@wrkgrp.COM (Edward T Spire)
- Subject: Re: NOVALUE (was: Blanks, REXX, and portability...)
- Message-ID: <1992Sep15.212407.12086@wrkgrp.COM>
- Organization: The Workstation Group
- References: <2427.imc@uk.ac.ox.prg>
- Date: Tue, 15 Sep 92 21:24:07 GMT
- Lines: 31
-
- imc@comlab.ox.ac.uk (Ian Collier) writes:
- : And in article <1686052F.GILMART@LSTC2VM.stortek.com>, he quoth:
- :
- : >But it would be a real pain if this were reported as an error unless there
- : >were some way to use explicit constants -- better by far, general expressions
- : >-- as compound variable indices. Namely:
- :
- : > say a.'X' /* so I don't have to assign 'X' to dummy */
- : > /* and even: */
- : > say a.(substr('UVWXYZ',4,1)) /* Why not, haven't you wanted it? */
- :
- : >Is anyone proposing this extremely desirable extension?
- :
- : You can already do this on REXX/imc. Of course, I'm not asking anyone
- : actually to use that feature! :-)
-
- So in REXX/imc, what does this say:
-
- a.=1
- a.x=2
- say a.'X'
-
- I got the impression from the above that it would say 2.
- Shouldn't it say 1X (i.e., abbutal concatenation)?
-
- I don't think you can extend the language in this specific manner
- without breakage. And yet this requirement has been discussed,
- I'm just not up on the details of any proposed syntax. Maybe they
- exhibit the same problem...
-
- -Ed
-