Michael:
Per Dave Jone's earlier comment, I believe I'm actually running the
newer cut of the compiler - what I'm using came off the RSU tape along
with the other components of the z/VM 5.1 development system that has
the problem.
Maybe a better question is: Does RPCGEN support the newer C/C++ compiler?
Thanks to all for feedback received thus far,
-dan.
Michael Donovan wrote:
Dan,
Ah. One problem I can see is the level of your C compiler. The V1R2
C compiler is quite old. Beginning with z/VM 4.4.0, TCP/IP uses the
"IBM C/C++ for z/VM V1R1" compiler. Product id is 5654-A22.
RPCGEN, being part of TCP/IP, also needs this level of the compiler.
I'm willing to wager that the older C compiler is doing something
which contributes to the ABEND.
As for CP's EMSG being turned off, this is an artifact of any IBM C
compiler. At some point in the C compiler's history they ran into
something which always produced an error message. The "fix" for
that became "CP SET EMSG OFF" inside the C run time library code.
You can get around this by setting a global variable to tell the C run time
code to leave the EMSG setting alone. Issue the following command:
GLOBALV SELECT CENV SET _EDC_KEEP_EMSG Y
Issuing the GLOBALV command this way will place that in your
LASTING GLOBALV which will stick around "forever". So, you
only need to do this once.
Cheers!
Mike Donovan
"Daniel P.
Martin"
<[EMAIL PROTECTED] To
ks.com> [email protected]
Sent by: The IBM cc
z/VM Operating
System Subject
<[EMAIL PROTECTED] Re: RPCGEN misery on z/VM 5.2 +
ARK.EDU> RSU05104
03/28/2006 11:11
PM
Please respond to
The IBM z/VM
Operating System
<[EMAIL PROTECTED]
ARK.EDU>
Kurt:
Thank you very much for your reply. I've got access to an apparently
working C compiler, but appear to be missing some crucial component in
the environment. Here's another screen shot:
--- tear here ---
q loadlib
LOADLIB = SCEERUN
Ready; T=0.01/0.01 21:50:49
q txtlib
TXTLIB = SCEELKED
Ready; T=0.01/0.01 21:50:51
cc ccncivp
Ready; T=0.20/0.24 21:50:57
load ccncivp (clear nomap start
Product Name: 5694A01
Version 1 Release 2 Modification 00
Text Creation Date: 06:087
VALIDATION SUCCESSFUL
Ready; T=0.01/0.01 21:51:03
l * expand
DMSVSMA EXPAND A1
Ready; T=0.01/0.01 21:51:12
erase dmsvsma expand
Ready; T=0.01/0.01 21:51:15
rpcgen dmsvsma
DMSITP142T Operation exception occurred at 8214DAB0 in routine RPCGEN
during SPIE exit routine
21:51:20 * MSG FROM MAINT : DTCINI004S Abend detected by DTCINI
DMSABE148T System abend 0C1 called from 0214DAB0 reason code 00000000
CMS
l rpcgen * * (l
FILENAME FILETYPE FM FORMAT LRECL RECS BLOCKS DATE
TIME LABEL
RPCGEN MODULE Z1 V 65535 4 18 7/22/04
12:40:40 TCM592
Ready; T=0.11/0.13 21:54:44
l cc exec * (l
FILENAME FILETYPE FM FORMAT LRECL RECS BLOCKS DATE
TIME LABEL
CC EXEC Y2 V 72 150 2 6/17/03
15:38:55 MNT19E
Ready; T=0.01/0.01 21:54:49
--- tear here ---
As a free bonus grumble, somewhere in the course of running rpcgen, it
appears that a SET EMSG OFF is issued, thus stifling potentially helpful
error messages.
*sigh* I was so sure that Chuckie would immediately dash to the
forefront and point out the Obvious Necessary Thing that I'd missed... ;-)
-dan.
Kurt Acker wrote:
A note passed to me on this with a request to post:
Hi Dan,
In the TCP/IP Programmer's Reference, Chapter 5 "Remote Procedure
Calls",
there is a section called "RPCGEN Command" which indicates that the
CC EXEC is used.
Please try accessing the disk containing the C compiler. Then
erase the DMSVSMA EXPAND which was created by any RPCGENs that failed
and try the RPCGEN again.
Best Regards,
Bill
yes, its still ok to shot the messenger if needed!
Kurt Acker
