Re: Supress Option List

I agree with Colin, that actually having "non-changing" compiler options
PROBABLY means that your shop has other (serious?) problems. In addition to
what he mentions, I would ask what 3rd party products you use for COBOL
development? Some require NOOPT (which I certainly HOPE you don't use in
production). Same for TEST/NOTEST.

I *can* imagine (but not recommend) a shop that doesn't want application
programmers to SEE what options are in effect. OTOH, if you work with IBM to
ever report a "problem" they will want to KNOW what options are in effect and
will (in some cases) ask you to run with different options.

Can you tell us WHY the "lengthy" list of options in effect is a problem? Do
programmers have problems "scrolling" to the identification division when
viewing output? Do you use the FLAG(I,I) (now the default - but previously not)
option to see compiler messages by source code causing the problem? Is the
problem, that every listing (in hard-copy) is ONE page longer?

Again, simply WHY do you want this "feature"?

--
Bill Klein
wmklein <at> ix.netcom.com
"Colin Campbell" <cmcampb@xxxxxxxxxxxx> wrote in message
news:fcCdnRDnC8AiJpfeRVn-hQ@xxxxxxxxxxxxxxx
> bfwd wrote:
>
>>IGYCRCTL lists all the options in effect for compile. For us, this list
>>will never change. Is there a way to suppress this option listing so
>>the resulting source listing begins with the IDENTIFICATION DIVISION?
>>
>>
> I assume you are referring to IBM mainframe COBOL, such as COBOL for z/OS and
> OS/390.
>
> It seems pretty unlikely that the option list would NEVER change.
> Doesn't anyone in your shop compile for testing, as opposed to production?
> Will you never need to see the generated code for a program (LIST vs. OFFSET)?
>
> Will you always (or never) compile programs for 31-bit mode execution and
> shared use under IMS or CICS?
>
> Will there never be a time when you want to use either the CICS or SQL
> pre-processor/compiler?
> Will you never write any international applications (which might use different
> currency symbols)?
> Do you always (or never) use extended arithmetic?
> Will you never need to check that a program meets the applicable COBOL
> standard?
> Is there no chance that you would ever want to switch between ANSI and Lillian
> date formats?
>
> I could go on....
>
> I hope you have made sure that the options such as BUFSIZE, COMPILE, NUMPROC,
> OPTIMIZE, TEST, and the like have been optimized to provide the best compiler
> and run unit performance, and the most useful output in case of an abnormal
> termination.
>
> Bill Klein's response that you could write a PRTEXIT routine to control what
> parts of the listing actually get printed is true; of course, that would
> change the option list when it is in use.
>
> Another simple option is to write the listing to a temporary data set, and
> write a simple program to read that, and omit the portions of the listing you
> don't want to see.
>
> If you use a tool such as CA-Endevor to install your application programs, it
> is likely that you'll have different compiles for batch programs, IMS
> transactions, CICS transactions, 31-bit programs, etc, with different options.
> Having the options listed would be one way to check that a program was
> compiled the right way.


.

Relevant Pages

  • Re: The Lisp CPU
    ... compile a function to be a list at runtime. ... proceeding to the next instruction. ... might compile into (where are opcode primitives, ... by itself is a symbol reference, and I'm using dotted lists for each ...
    (comp.lang.lisp)
  • Re: pctel HSP56 modem compilation problem
    ... > I have a HSP56 MR modem and I had problems to compile the driver. ... lspci lists PCI devices, it is in /sbin/, so you may need to run the ... > checking for running kernel version...2.4.22 ... Your modem is a sis type modem. ...
    (Fedora)
  • Re: Haskell optimization (mandelbrot generator)
    ... result list, and that had no affect on the speed, I decided the lists ... I saw the section on unboxed types in the GHC manual and that sounds ... but I haven't gotten the app to compile using them yet. ... Anyone have any good references on using their unboxed type extensions? ...
    (comp.lang.functional)
  • Re: Python and generic programming
    ... implements a simple insertion sort algorithm. ... sort can actually apply to lists of any type, ... The appropriate type checking is done at compile time. ...
    (comp.lang.python)
  • Re: [Nbd] [ANNOUNCE] Portable NBD server
    ... but this is relevant to all three lists. ... > Please don't cross-post followups (unless they're relevant to several ... intention, in any case, is for it to compile on anything resembling GCC ... send the line "unsubscribe linux-kernel" in ...
    (Linux-Kernel)