| Newsgroups |
gmane.comp.emulators.hercules390.mvs |
| Message-ID |
<[email protected]> |
Mmm ... interesting Dave.
The error is about "fragmented core".
GETMAIN badly fails on a memory allocation, I bet.
WHY?
Peppe.
On Wed, 11 Sep 2019, 'Dave Wade' [email protected] [H390-MVS] wrote:
> Folks
>
>
>
> Why does the same compiler work in VM?
>
> I havenÿÿt tried a large program but Jayÿÿs simple ÿÿhello Worldÿÿ compiles with XREF set in the options.
>
> In fact the EXEC in VM always seems to set XREF so I assume that it will XREF larger programsÿÿ
>
>
>
> So I have :-
>
>
>
> 001 IDENTIFICATION DIVISION.
>
> 002 PROGRAM-ID. 'HELLO'.
>
> 003 ENVIRONMENT DIVISION.
>
> 004 SOURCE-COMPUTER. IBM-360.
>
> 005 OBJECT-COMPUTER. IBM-360.
>
> 006 SPECIAL-NAMES.
>
> 007 CONSOLE IS CNSL.
>
> 008 DATA DIVISION.
>
> 009 WORKING-STORAGE SECTION.
>
> 010 77 HELLO-CONST PIC x(12) VALUE 'HELLO, WORLD'.
>
> 100 PROCEDURE DIVISION.
>
> 110 000-DISPLAY.
>
> 120 DISPLAY HELLO-CONST UPON CNSL.
>
> 130 STOP RUN.
>
>
>
>
>
> And in the listing, I get :-
>
>
>
> CB545 V2 LVL78 01MAY72 IBM OS AMERICAN NATIONAL STANDARD COBOL
>
>
>
> And later the cross reference is simple but complete
>
>
>
> - CROSS-REFERENCE DICTIONARY
>
>
>
> DATA NAMES DEFN REFERENCE
>
> HELLO-CONST 00010 00013
>
> PROCEDURE NAMES DEFN REFERENCE
>
> 000-DISPLAY 00012
>
>
>
> So is there something wrong in way the compiler is built in MVS, is my simple program too simple, or something missing in MVS?
>
>
>
> The COBOL compiler in VM appears to have been generated from source and appears to be an un-adulterated copy of the MVS COBOL
>
> compiler, in that there is no custom VM front end. It is run using the ÿÿRUNPARMÿÿ module which simple generates a OS parmlist and loads the program..
>
>
>
> Thoughts anyone?
>
>
>
> Dave
>
>
>
>
>
> From: [email protected] <[email protected]>
> Sent: 10 September 2019 12:19
> To: [email protected]
> Subject: Re: [H390-MVS] MVT COBOL under MVS3.8j.
>
>
>
>
>
>
>
>
> Peppe,
>
>
>
> Many people have had this issue. The only resolution is to remove XREF from the PARM.
>
>
>
> Joe
>
>
>
> On Tue, Sep 10, 2019 at 7:13 AM Gerhard Postpischil [email protected] <mailto:[email protected]> [H390-MVS] <[email protected] <mailto:[email protected]> > wrote:
>
>
>
> On 9/10/2019 6:59 AM, Giuseppe Vitillaro giuseppe-yuD/ahkh7LvrZ44/[email protected] <mailto:giuseppe-yuD/ahkh7LvrZ44/[email protected]>
> [H390-MVS] wrote:
>
>> By the way, just tested under TK4-, with
>> the TK4- COBOL V2 version, SYS2.PROCLIB(COBUCLG),
>> using PARM='LOAD,SUPMAP,SIZE=8192K,BUF=2048K,LIB,XREF'.
> I wonder whether you can reduce the buffer size - if it's for I/O
> buffers 256K or so should suffice. I doubt that it's a sysgen related
> problem, but it could be something to do with the compiler generation
> options?
>
> Gerhard Postpischil
> Bradford, VT
>
> ---
> This email has been checked for viruses by AVG.
> https://www.avg.com
>
>
>
>
>
>
>
>
>
--
Giuseppe Vitillaro | E-Mail : giuseppe-yuD/ahkh7LvrZ44/[email protected]
CNR - ISTM | 06123 Perugia Phone:+39.075.585-5518
-----------------------------------------------------------------------------