[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: IDL and Microsoft Object Linking and embedding OLE
- Subject: Re: IDL and Microsoft Object Linking and embedding OLE
- From: Richard Younger <younger(at)ll.mit.edu>
- Date: Fri, 26 Jan 2001 15:10:12 -0500
- Newsgroups: comp.lang.idl-pvwave
- Organization: MIT Lincoln Laboratory
- References: <3A64B0EA.C58A0C24@email.sps.mot.com>
- Xref: news.doit.wisc.edu comp.lang.idl-pvwave:23235
Peter Brooker wrote:
>
> I have a software package called Prolith that allows "Visual Basic,
> C++, or other programming languages to control PROLITH from other
> OLE-compliant applications."
>
> You can write a program in C++ and make calls to Prolith routines.
> ProLith then return the data back to the C++ program.
>
> This sounds allot like the "call_external" command in IDL.
>
> They go on to define OLE as object linking and embedding that is a
> "strategic technology from Microsoft that is the standard for
> application integration and interchange in Windows."
>
> Is there anyway that an IDL program can access the OLE ProLith
> routines the same way a C++ program can?
>
> thanks-Peter Brooker
Sorry for the delay in my reply. I've been doing data acquisition and
device control through IDL for a little while, but I hardly would
consider myself an expert on the subject. Consequently, I wanted to
mull it over a bit before I stuck my head out too far. :)
Just about every time, the answer to "Can I call X from IDL" has been
"Write a C wrapper function." Occasionally for routines that return no
data, just do something like manipulate files, "Make it into an
executable and use SPAWN" has been the answer of choice, if you don't
mind a little extra overhead.
As to the question if you can use Your Favorite Language (YFL), it
depends. AFAIK, Call_External works acceptably if you can get YFL to
accept C-style calling conventions. There is an example floating around
of calling a Fortran routine. But generally, accepting C calling
conventions means some facility for dereferencing pointers, a compiled
(not interpreted) language, and some sort of typing.
Linkimage and DLMs prefer C natively, as you need to manipulate C-style
IDL internals and there are C libraries that you need to call in your
routines. In my experience (admittedly only with MS VC++) C++ will work
too, as long as you make sure to set C calling conventions and prefix
all the IDL-callables with the (export.h) IDL_CDECL macro. Other
languages would be much more work.
There is one major caveat, however. Especially when you start thinking
towards callable packages and away from programming languages. Scope
and data persistence. Every time you go back to IDL, you loose all your
local YFL memory, all your variables. You can play games with
dynamically allocating things or global memory blocks and passing
pointers to them, or passing data back and forth through IDL variables,
but it gets klugey if you've got too many things up in the air. If you
are trying to use IDL as an interface interactively with ProLith, you're
in for some work. If each call is independent of every other call, it
will go more smoothly.
To sum up, I would think that you could write a C++ Call_external
function that would call the Prolith routines. Although the package
itself might do funky things that IDL doesn't like in its memory space.
I know that IDL, for one, doesn't support (nor will it soon support)
threaded code. Maybe someone else knows more about whether Prolith will
actually interfere with IDL. No guarantees here. I think you'd have to
try it to know. Look to the External Development Guide and the online
help get you started.
If you haven't used them before, I would guess it would take a few days
of playing to get a rudimentary call_external interface up for a few
functions, and maybe a few weeks to get a more sophisticated DLM
interface going. If you take the DLM route, get Ronn's new book, as it
has lots of helpful examples that will save you no end of frustration.
best of luck,
Rich
--
Richard Younger