Xref: utzoo comp.lang.eiffel:475 comp.lang.c++:5291
Path: utzoo!utgpu!watmath!att!dptg!pegasus!psrc
From: psrc@pegasus.ATT.COM (Paul S. R. Chisholm)
Newsgroups: comp.lang.eiffel,comp.lang.c++
Subject: Re: Interface between Eiffel and C++
Summary: bad news
Message-ID: <4210@pegasus.ATT.COM>
Date: 29 Oct 89 21:41:10 GMT
References: <1989Oct27.205448.8516@mentor.com>
Organization: AT&T Bell Laboratories
Lines: 35

In article <1989Oct27.205448.8516@mentor.com>, lisch@mentor.com (Ray Lischner) writes:
> How easy is it to design an interface between C++ and Eiffel?

Not very.  Bear in mind that both usually produce C as an "assembler
language".  For better or worse, C++ usually produces C code that is
relatively similar in structure to the C++ code; I don't think this is
the case for the current Eiffel implementation.

> Specifically, we would need to call C++ functions from Eiffel,
> so how can one get Eiffel to understand C++ name mangling?
>...
> We would also need to call Eiffel routines from C++, and I know
> that extern "Eiffel" does not exist.

So far as I know, the only way to do either would be through an
intermediary C function.

> The big problem, I suppose, would be making class definitions from
> one language available in the other, but there is the problem of C++
> virtual functions and base classes, etc., etc., etc.

(No kidding.-)  C++ programs can directly use C header files with
struct, union, and typedef definitions.  The early implementations of
Eiffel only allowed passing primitive types (integers, maybe char
pointers) to C functions.  I think the current implementation promises
to do a little more, but I don't know the details.

No C compiler can understand C++ or Eiffel class definitions.  That's
what "upward but not downward compatible" means, right?

> Ray Lischner        UUCP: {uunet,tektronix,decwrl}!sequent!mntgfx!lisch

Paul S. R. Chisholm, AT&T Bell Laboratories
att!pegasus!psrc, psrc@pegasus.att.com, AT&T Mail !psrchisholm
I'm not speaking for the company, I'm just speaking my mind.