Re: vTable offsets

• Previous message: Mike H: "Re: vTable offsets"
• In reply to: Mike H: "Re: vTable offsets"
• Next in thread: Mike H: "Re: vTable offsets"
• Reply: Mike H: "Re: vTable offsets"
• Messages sorted by: [ date ] [ thread ]

Date: Sat, 31 Jul 2004 19:52:13 -0700

What about starting by posting the interface in question you are
referencing and the client code you used to obtain it? I suspect
you obtained a pure IDispatch* and tried to address it as a dual
interface, but this is a mere guess. At any rate, the problem is
somewhere with your perception not corresponding with the
reality. A dual interface doesn't have a gap in the vtable as you
are trying to explain it. Ah, details as the server type (in-proc vs
out-of-proc), threading model, etc might be of further help too.

-- 
=====================================
Alexander Nickolov
Microsoft MVP [VC], MCSD
email: agnickolov@mvps.org
MVP VC FAQ: http://www.mvps.org/vcfaq
=====================================
"Mike H" <mikehg1@removethis foraccess-4-free.com> wrote in message
news:zLWOc.148454$%t6.5656@roc.nntpserver.com...
>
> "Alexander Nickolov" <agnickolov@mvps.org> wrote in message
> news:uq$RdoldEHA.3612@TK2MSFTNGP09.phx.gbl...
> > Your assertions fall out from thin air, you need to provide some
> > hard details before we can even consider them... Working in C
> > is one such detail - you can get things wrong easily in C, since you
> > have to build all vtables by hand. (And I'm probably not saying
> > anything new here...)
>
> I'm not sure how much more detail I can give without answering my own
> question?
> I would have thought this would be an easy question for someone with a
> strong knowledge of this stuff. I have read Inside OLE, it does not
provide
> much detail on the subject.
>
> This seems to be common to Controls made with VB. The vTable offset starts
> at 0 for IUknown, but the first exposed function starts at 1956. So we
have:
>
> offset zero:
> IUknown
> IDispatch <--This would end at offset 27
> 28 through 1955 missing ...If I insert dummy placeholders here everything
> works as expected
> Foo Interface <--The offset starts at 1956
>
> So I guess my next 2 questions would be why the gap? and is it possible to
> specify an offset using C instead of placing 482 dummy placeholders?
>
> Thanks,
> Mike H.
>
> >
> > As a general advice, read the first 6 chapters of "Inside OLE"
> > by Kraig Brockschmidt, MS Press (available on old MSDN
> > Library CDs, otherwise long since out of print). Chapter 5 or 6
> > contains an example of building a COM object in C.
> >
> > -- 
> > =====================================
> > Alexander Nickolov
> > Microsoft MVP [VC], MCSD
> > email: agnickolov@mvps.org
> > MVP VC FAQ: http://www.mvps.org/vcfaq
> > =====================================
> >
> > "Mike H" <mikehg1@removethis foraccess-4-free.com> wrote in message
> > news:J0vOc.118584$%t6.34381@roc.nntpserver.com...
> > > I'm still in the learning stage of the COM world and have a question
> about
> > > vtables. Using an Object browser I notice the vTable offset is 28 for
> the
> > > first function, which seems to make sense if you calculate the size of
> the
> > > Dispatch interface first. But some start with a vTable offset of
usually
> > > 1956 and these are usually the interface functions that I can't seem
to
> > > access. Most of the time it results in a GPF or a No Interface return
> > value.
> > > Can someone shed some light on how these vTables need to be
implemented
> > > and/or called? I am using plain C btw.
> > >
> > > Thanks,
> > > Mike H.
> > >
> > >
> >
> >
>
>

• Previous message: Mike H: "Re: vTable offsets"
• In reply to: Mike H: "Re: vTable offsets"
• Next in thread: Mike H: "Re: vTable offsets"
• Reply: Mike H: "Re: vTable offsets"
• Messages sorted by: [ date ] [ thread ]