HP3000-L Archives

April 1998, Week 2

HP3000-L@RAVEN.UTC.EDU

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
John Zoltak <[log in to unmask]>
Reply To:
John Zoltak <[log in to unmask]>
Date:
Mon, 13 Apr 1998 15:13:50 -0400
Content-Type:
text/plain
Parts/Attachments:
text/plain (77 lines)
Greg,
I tried your example and the compiler gives the same error. As a
workaround you could do the following. The compile doesn't seem to like
to compare CONDITION-CODE in a perform loop. This version will compile
ok and does the same thing, but a little slower.

John Zoltak
North American Mfg Co

             77 C-C-RESULT PIC S9(4) COMP.

                   PERFORM WITH TEST AFTER
                 VARYING K-INDEX FROM 1 BY 1
                 UNTIL K-INDEX = 10
                          OR C-C-RESULT = ZERO
                      CALL INTRINSIC "FCONTROL"
                          USING INPUT-FILE-NUMBER \4\ DUMMY-WORD
                      END-CALL
                      IF C-C = ZERO
                          MOVE ZERO TO C-C-RESULT
                      ELSE
                          MOVE 1 TO C-C-RESULT
                      END-IF
                  END-PERFORM


> -----Original Message-----
> From: Stigers, Greg ~ AND [SMTP:[log in to unmask]]
> Sent: Monday, April 13, 1998 2:49 PM
> To:   [log in to unmask]
> Subject:      [HP3000-L] Codegen detected error (5705)
>
> Here's some fun. Under MPE 55 using COBOL II/iX HP31500A.04.16,
> compiling a program with the following inline perform, I get this
> Codegen error. Searching the LaserROM for 5705, it seems not to like
> the
> comparison, but I find no good solution for this. I have tried
> changing
> the comparison; for confirmation, I commented out 1048, and could
> compile. I edited to just test C-C =  ZERO, and got the same error. I
> tried ZEROES and 0, and enclosing in parentheses, without success.
> Getting clever and defining another identifier with a value of zero is
> not legal (EXPECTING ONE OF THE FOLLOWING: <NUMERIC-LIT> ZERO). OK...
> GOTO would be messy in this context. Ideas, anyone? Is this just a
> genuine bug?
>
>  01044                D          PERFORM WITH TEST AFTER
>  01045                *              K-INDEX isn't used anywhere else,
> so I'll use it here.
>  01046                D              VARYING K-INDEX FROM 1 BY 1
>  01047                D              UNTIL K-INDEX = 10
>  01048                D                  OR C-C = ZERO
>  01049                D              CALL INTRINSIC "FCONTROL"
>  01050                D                  USING INPUT-FILE-NUMBER \4\
> DUMMY-WORD
>  01051                D              END-CALL
>  01052                D          END-PERFORM
> ...
>
> COBOL ERRORS:
>
> LINE #   SEQ #  COL  ERROR  SEV   TEXT OF MESSAGE
> ------  ------  ---  -----  ---   ---------------
>
>  01048          80    384    Q    Codegen detected error (5705) at LOC
> 1048 in testprog Ustack underflow (5705).
>  01048          80    999    D    UNABLE TO CONTINUE. COMPILATION
> TERMINATED.
>
> 1 ERROR(s), 1 QUESTIONABLE, 0 WARNING(s)
>
> *** DISASTROUS ERROR NEAR LINE  01048 ***
> *** COMPILATION TERMINATED ***
>
>     CHECKED SYNTAX ONLY.
>     CPU TIME = 0:00:00.  WALL TIME = 0:00:05.

ATOM RSS1 RSS2