Operating System - OpenVMS

ACCVIO error within C++ Compiler

Occasional Contributor

ACCVIO error within C++ Compiler

I'm using CXX V7.2-018 on OpenVMS IA64 V8.3 and I'm getting the below Compiler ACCVIO when compiling a xercesc source file. Could you please tell me what maybe causing this problem and how it could be resolved ?
Thanks in Advance.

Building XERCES-C$ROOT:[000000.src.xercesc.dom.impl]
Compiling XERCES-C$ROOT:[000000.src.xercesc.dom.impl]DOMNamedNodeMapImpl.cpp
because XERCES-C$ROOT:[000000.src.xercesc.obj]DOMNamedNodeMapImpl.obj not found.
%SYSTEM-F-ACCVIO, access violation, reason mask=00, virtual address=0000000000000001, PC=00000000015EA460, PS=0000001B
%TRACE-F-TRACEBACK, symbolic stack dump follows
image module routine line rel PC abs PC
CXX$COMPILER TYPES EXPR_Dtype 130759 0000000000026E80 00000000015EA460
129586 00000000000034D0 0000000001F6C0D0
CXX$COMPILER CSI_NOT phase2_Csi_Part2b
173421 000000000000EE10 0000000002AC01C0
172387 0000000000006D90 0000000002AB8140
180831 0000000000005200 0000000002ACCCF0
CXX$COMPILER IL0_FE il0_routine_end 129794 0000000000001460 000000000204ED20
CXX$COMPILER EDG_CODE dump_routine 114925 000000000000D900 00000000014B9510
CXX$COMPILER EDG_CODE dump_all_routines
115050 000000000000E520 00000000014BA130
CXX$COMPILER EDG_MAIN back_end 94797 00000000000033E0 00000000014F23C0
CXX$COMPILER CFE PHASE1_Begin 131922 0000000000000400 0000000000F465F0
CXX$COMPILER MAIN IVMS_phase1main 125587 0000000000013030 0000000002AA9F20
185940 0000000000003DD0 000000000254F790
CXX$COMPILER GEM_CP_VMS GEM_CP_MAIN 2630 00000000000027A0 0000000002EFA990
CXX$COMPILER IVMS_IS_DCL main 20174 0000000000000290 00000000025520B0
CXX$COMPILER IVMS_IS_DCL __main 20160 00000000000000E0 0000000002551F00
394 0000000000000250 0000000002F546A0
CXX$COMPILER 0 0000000002F55530 0000000002F55530
DCL 0 000000000006B9F0 000000007AE279F0
%TRACE-I-END, end of TRACE stack dump

Volker Halle
Honored Contributor

Re: ACCVIO error within C++ Compiler


welcome to the OpenVMS ITRC forum.

Please consider to make sure, that you've installed the most recent compiler, which would currently be C++ V7.3

If the compilation still fails, consider to raise a call with HP. An ACCVIO at CXX$COMPILER TYPES EXPR_Dtype does not seem to be a publicly documented problem.

Consider to check the compile qualifiers used and try to compile the failing module with slightly modified qualifiers, like more or less optimizations.

Honored Contributor

Re: ACCVIO error within C++ Compiler

Rule of Thumb: an access violation (ACCVIO) in any HP OpenVMS component or in any HP layered product or in any HP compiler is always an error within the tool. Yes, this error may be secondary to an error in the input, but it's still indicative of an error in the tool.

If you're not current with your compiler and with your associated and your mandatory OpenVMS ECO kits, then get there first.

If you are current on your compiler and your ECO kits, then report the error (with enough source code to reproduce the problem) directly to HP.

We do thank you for letting us know about this compiler error, but -- as a general rule -- HP engineers do not follow the ITRC forums, nor comp.os.vms newsgroup, nor OpenVMSHobbyist.org nor any of the other discussion forums around the Internet. So long as you always expect all problem reports in any of these informal forums and areas to be ignored (by HP), you'll only be pleasantly surprised if somebody does notice and does take ownership of the informal problem report. But if you want this addressed, then get current and then use the formal HP channels.

John Reagan
Respected Contributor

Re: ACCVIO error within C++ Compiler

Try upgrading to the latest compiler. You'll find V7.3-032 on the ITRC ECO pages. There is yet another ECO on the way (-039) but none of the problems fixed for -039 seem to match your ACCVIO.
Dennis Handly
Acclaimed Contributor

Re: ACCVIO error within C++ Compiler

>Hoff: this error may be secondary to an error in the input, but it's still indicative of an error in the tool.

For HP-UX, we solve that by not printing any signals/tracebacks if there were any previous user errors.

>John: Try upgrading to the latest compiler.

I guess I don't need to pass this on to you, especially since we now work for the same boss. :-)