14653 - small pointer problem in db/mork/src/morkDeque.cpp?

Status: VERIFIED FIXED

(Core Graveyard :: Tracking, defect, P3)

Description

[Colin Blake]

This belongs to Mork, but there's no component for Mork (or DB) yet!

In db/mork/src/morkDeque.cpp, there is the following code:

void morkNext::ZapOldNext(morkEnv* ev, nsIMdbHeap* ioHeap)
{
  if ( &ioHeap )
  {
    if ( this )
      ioHeap->Free(ev->AsMdbEnv(), this);
  }
  else
    ev->NilPointerError();
}

The OpenVMS C++ compiler complains that the last statement (the call
to NilPointerError) is unreachable.

Since ioHeap is a pointer, I guess that &ioHeap is the address of the
pointer. And since this is always non-zero, even for a null pointer,
the if test doesn't make sense???

Anyway, thought you might want to look at this and maybe give it
some consideration.

Colin.

Comment 1

[Colin Blake]

Same goes for morkLink::ZapOldLink. That makes two occurences of this
"problem" in morkDequeue.cpp.

Comment 2

[chris hofmann]

-> davidmc

Comment 3

[davidmc]

Okay, I'll fix it; originally the code would have passed nsIMdbHeap as
a reference, so that &ioHeap would yield actual the pointer.  All I can
say is I wrote the code in a hurry one day when making change lists for
incrementally writing tables.  Thanks for reporting this.

Comment 4

[David :Bienvenu]

these are mine now, I guess.

Comment 5

[David :Bienvenu]

accepting

Comment 6

[David :Bienvenu]

m20

Comment 7

[David :Bienvenu]

fix checked in - if ioheap were null, I think we would have crashed, so this was 
useful to fix.

Comment 8

[[e-mail] (gone - use [e-mail] instead)]

1.6 bienvenu%netscape.com Jun 22 2000 fix build warning 14653