Re: SSP
- From: "Mark Yudkin" <DoNotContactMe@xxxxxxxxxxxxxx>
- Date: Tue, 15 Aug 2006 18:53:03 +0200
It is important: consider Interlanguage Communication with
options(nodescriptor), especially with languages like (older) Fortran which
supported this construct.
BTW, I got the opposite case fixed: declaring an entry with * bounds and
options(nodescriptor), as well as the ability to pass a scalar to an array
with * bounds and options(nodescriptor). The second case is for C
compatibility.
But since I can't open a PMR / Requirement any more, you'll have to do it.
"Peter Flass" <Peter_Flass@xxxxxxxxx> wrote in message
news:G1EDg.21420$8j3.19004@xxxxxxxxxxxxxxxxxxxxxxx
Thanks Mark; is it even legal PL/I? It's a rather obscure "feature", and
I'd rather just change it going forward. Sometimes being bug-for-bug
compatible is not a good thing.
Mark Yudkin wrote:
IIRC, what you coded used to work, but has now been eliminated. If you
have a support contract, I'd suggest your opening a PMR for this
"incompatibility with old host compiler". If the PMR is rejected, you can
then turn it into a requirement. I was able to get a few issues addressed
in this manner, but can no longer do so (I have changed jobs).
"glen herrmannsfeldt" <gah@xxxxxxxxxxxxxxxx> wrote in message
news:xsSdnW1Q_5SZFkTZnZ2dnUVZ_qSdnZ2d@xxxxxxxxxxxxxx
Peter Flass wrote:
The thing I found, compiling with Enterprise PL/I, after getting rid of
the 48-char set nonsense were a couple of routines that had something
like this:
ABCD: PROC(AA,BB);
DCL AA FIXED BINARY(15),
BB (AA)FLOAT BINARY;
Changing them to "...BB (*)FLOAT BINARY" worked for me.
It makes sense for Fortran 66 or Fortran 77, but not for PL/I.
It might be that someone wanted to keep them similar.
-- glen
.
- References:
- Prev by Date: Re: How IBM recovered from the PL/I fiasco
- Next by Date: Re: SSP
- Previous by thread: Re: SSP
- Next by thread: Re: SSP
- Index(es):
Relevant Pages
|
