Conference turris::debug

1851.0. "%DEBUG-F-ACCVIO w C++ V5.5, Alpha V6.2, DEBUG V7.1C" by CSC32::D_SANFORD () Wed Mar 26 1997 12:06

    OpenVMS Alpha V6.2
    DEC C++ V5.5
    OpenVMS Alpha DEBUG Version V7.1C-000

    The following example will demonstrate a %DEBUG-F-ACCVIO error.

    Regards, Drew Sanford
    Customer Support Center
    C970324-3493 

$ ! test.com
$ create test.cxx
#include stdio
class A {
public:
  A();
  ~A();
  void foo1();
  void foo2();
};

A::A() { }

A::~A() { }

void A::foo1() {  printf("foo1\n"); }

void A::foo2() {  printf("foo2\n"); }

int main() {

  A anA;
  anA.foo1();
  anA.foo2();

  return 3;
}
$ cxx/noopt/debug test
$ cxxlink/debug test
$ create test.ini
set mode symbol,screen
display/push inst
go
set break A::foo1
set break A::foo2
go
$ define/nolog/user dbg$init test.ini
$ run test
$ exit


@sys$startup:debugsetup
Please specify setup type; /JOB or /SYSTEM: /JOB
Please specify debugger name; VMS or CXX or UNSET: CXX

$ @test
         OpenVMS Alpha DEBUG Version V7.1C-000

%DEBUG-I-INITIAL, Language: C++, Module: TEST
%DEBUG-I-NOTATMAIN, Type GO to reach MAIN program
)0
=
)0
)0
q SRC -scroll-sourceqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqq OUT -outputqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqq PROMPT -error-program-promptqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqq


break at routine TEST\main691:   A anA;


break at routine A::foo1685: void A::foo1() {  printf("foo1\n"); }


%DEBUG-E-NOACCESSR, no read access to address 0000000000012000
: module TEST -scroll-source
%DEBUG-F-ACCVIO, access violation, reason mask=00, virtual address=00000004, PC=00153E00, PS=0000001B


Logging enabled - this is being written to log file DBGERR.LOG

The following information may be helpful to the DEBUG maintainers.

If you wish to report this problem please include:

     o The stack dump that follows.
     o The debugger version number.
     o The program being debugged, both source and image.
     o The last, if not all, commands entered to the debugger.
     o Any other information that may be helpful.

The version of the debugger is: OpenVMS Alpha DEBUG Version V7.1C-000

  image    module    routine             line      rel PC           abs PC      
 CXXDEBUGSHR                                0 00000000002EBE3C 00000000002EBE3C
 CXXDEBUGSHR                                0 00000000002EAF94 00000000002EAF94
 CXXDEBUGSHR                                0 000000000029FA58 000000000029FA58
 CXXDEBUGSHR                                0 000000000029F688 000000000029F688
                                            0 FFFFFFFF80495DB8 FFFFFFFF80495DB8
----- above condition handler called with exception 0000000C:
%SYSTEM-F-ACCVIO, access violation, reason mask=00, virtual address=00000004, PC=00153E00, PS=0000001B
----- end of exception message
                                            0 FFFFFFFF832C82BC FFFFFFFF832C82BC
 CXXDEBUGSHR                                0 0000000000153E00 0000000000153E00
 CXXDEBUGSHR                                0 0000000000317760 0000000000317760
 CXXDEBUGSHR                                0 00000000003238F8 00000000003238F8
 CXXDEBUGSHR                                0 0000000000118B34 0000000000118B34
 CXXDEBUGSHR                                0 00000000001B7DD4 00000000001B7DD4
 CXXDEBUGSHR                                0 0000000000109B70 0000000000109B70
 CXXDEBUGSHR                                0 0000000000108F28 0000000000108F28
 CXXDEBUGSHR                                0 0000000000108E00 0000000000108E00
 CXXDEBUGSHR                                0 0000000000430074 0000000000430074

Due to this internal error this debugging session may be unreliable.


DBG> exit

Yup, it's broken. 

I'm fixing it now.  The fix will appear in the next release.

    Any workaround?  Will an ECO be released?  Just in case this
    is a can't use product issue...
    
    -drew

The problem occurs when you try to display, in screen-mode, the instructions
of a C++ member function.

So the problem can be avoided by leaving out any one of those elements.

Well, truth be known, the problem will likely show up in other scenarios. 
The important element is the member functions.

In this particular scenario, the debugging session is fine in spite of the
problem.  The most immediate work-around is to ignore the accvio and carry on.

ECO?  Ahh, well, let's see what Mary says.

From:	CSC32::HIPOWR::SANFORD      "SET SECURITY/PROTECTION=(WORLD:DREW)"  2-APR-1997
18:52:11.94
To:	CSC32::SSPADE::DIETER
CC:	SANFORD
Subj:	RE: Note 1851.3 %DEBUG-F-ACCVIO w C++ V5.5, Alpha V6.2, DEBUG V7.1C

Hi Mary,

>Is the customer asking for an ECO or were you asking, as a proactive measure?
>i.e. what is the real demand/reason for asking?

  Customer did not ask, I was being proactive so you wouldn't get asked 
  over and over.

>We could ECO it but I hesitate for two reasons...
>1) You can continue, without problems, despite the ACCVIO (as Mark notes below)
>2) The processes/people that we deal with to get ECOs out the door seem are 
>   very busy working on the 10k ECOs.  I don't think we/they could provide a
>   timely ECO in this cas (but, perhaps, I may be wrong...)

  Considering we have a workaround I don't see any reason to produce
  an ECO...  I might ask for engineering to review our article should 
  we get called again about this, just want to have a formal closure to 
  this.

  -drew