dp tests

[Denys Duchier]

I have modified gentraverser.hh so that no datastructure needs to
resize, yet the problem persists.  What is really odd is that the 1st
remote server behaves just fine.  The bug only shows up in the 2nd and
the 3rd ones (3 remote servers are created by dp_cell).  I logged the
opcodes being executed and the tail of the log for both is (I am showing
1 page worth of it):

...
OPCODE: moveYX
OPCODE: inlinePlus1
OPCODE: moveMoveYXYX
OPCODE: callBI
OPCODE: return
OPCODE: taskLock
OPCODE: deAllocateL5
OPCODE: return
OPCODE: moveYX
OPCODE: moveGX
OPCODE: testBI
OPCODE: inlinePlus1
OPCODE: moveYX
OPCODE: moveXX
OPCODE: moveMoveYXYX
OPCODE: moveMoveYXYX
OPCODE: deAllocateL7
OPCODE: fastTailCall
OPCODE: allocateL7
OPCODE: putRecordX
OPCODE: setValueX
OPCODE: setValueX
OPCODE: moveMoveXYXY
OPCODE: moveMoveXYXY
OPCODE: moveMoveXYXY
OPCODE: moveXX
OPCODE: createVariableMoveY
OPCODE: fastCall
OPCODE: moveGX
OPCODE: inlineDot
OPCODE: moveXX
OPCODE: moveXX
OPCODE: moveXX
OPCODE: tailCallX
OPCODE: taskCallCont
OPCODE: taskCallCont
OPCODE: moveYX
OPCODE: callBI
OPCODE: moveMoveYXYX
OPCODE: fastCall
OPCODE: allocateL5
OPCODE: matchX
OPCODE: getVarVarYY
OPCODE: lockThread
OPCODE: moveMoveXYYX
OPCODE: moveYX
OPCODE: createVariableMoveY
OPCODE: fastCall
OPCODE: allocateL2
OPCODE: putRecordX
OPCODE: setValueX
OPCODE: setConstant
OPCODE: setValueX
OPCODE: moveXY
OPCODE: moveXX
OPCODE: createVariableMoveY
OPCODE: fastCall
BAD OPCODE: 148792576

except that the last line for the 3rd remote server is:

BAD OPCODE: 152135008

So they seem to fail in the same place.


I could use some help debugging this.  If you don't have access to an
x86/Linux machine and need an account on my (home) laptop, I can provide
that too (just send me your ssh public key).

Cheers,

--Denys