HP-UX 11.31 64-bit iTanium or Risc machine and MF Server Express V5.1.00

[Marrion Newsam Banks]

We are getting the following error when attempting a mass compile on an HP-UX 11.31 64-bit iTanium or Risc machine running MF Server Express V5.1.00. The compile blows up in the middle of the process on one of the programs in the stream:

cob -xv -PC " IBMCOMP noecho notrunc nobound confirm errlist assign=external"
-c -o pgmname.o pgmname.cob
cob64 -C nolist -xv -PC IBMCOMP noecho notrunc nobound confirm errlist assign=
external -c -o pgmname.o pgmname.cob
* Micro Focus Server Express V5.1 revision 000 Compiler
* Copyright (C) 1984-2009 Micro Focus (IP) Limited. URN RXCAQ/AA0/00000C
* Accepted - verbose
* Accepted - nolist
* Accepted - IBMCOMP
* Accepted - noecho
* Accepted - notrunc
* Accepted - nobound
* Accepted - confirm
* Accepted - errlist
* Accepted - assign(external)
* Accepted - list("pgmname.lst")
* Compiling pgmname.cob
* Total Messages: 0
* Data: 2760 Code: 26913

Execution error : file ''
error code: 114, pc=0, call=1, seg=0
114 Attempt to access item beyond bounds of memory (Signal 11)

* Micro Focus Server Express V5.1.00 Code Generator
* Copyright (C) 1984-2009 Micro Focus (IP) Ltd. All rights reserved
* Accepted - verbose
* Generating pgmname
cob64: error(s) in code generation: pgmname.int
*** Error exit code 1

Stop.

System default option values taken from: /oracle/product/10.2.0/precomp/admin/pc
bcfg.


The same code has been installed and successfully compiled using the same version of MF Server Express on Solaris and Linux machines.

Any insight on this issue would be much appreciated. Thanks in advance.

[James R. Ferguson]

Hi:

By using the 'nobound' option you are disabling the validation of 'occurs' subscripts at runtime. While this makes for code that executes faster, since no verification of the subscript is made to insure that it addresses only within the boundaries of an array, using this option means that you are subject to signal-11 termination when your subscript points you outside of the array being referenced.

Regards!

...JRF...

[Marrion Newsam Banks]

Thanks so much for your quick reply! I am testing this option now and will let you know how things turn out.

[Dennis Handly]

>successfully compiled using the same version of MF Server Express on Solaris and Linux machines.

All this means is that your source is probably legal.

>JRF: By using the 'nobound' option you are disabling the validation of 'occurs' subscripts at runtime.

Yes but Marrion is compiling, not running.

If the MF compiler aborts, you need to talk to them.

[Marrion Newsam Banks]

Thanks to both of you for your responses. I have posted this same inquiry to the MF user forum but have not yet received a response. I tried JRF's "bound" versus "nobound" solution with no joy, but I appreciate the suggestion. If I ever get this figured out, I'll post the solution here.

[James R. Ferguson]

Hi (again):

As Dennis suggested, I would open a case with MicroFocus. I mis-read the fact that your error was during compilation, not at runtime.

Regards!

...JRF...