Path: utzoo!attcan!uunet!pilchuck!dataio!bright
From: bright@Data-IO.COM (Walter Bright)
Newsgroups: comp.lang.c++
Subject: Zortech and 3rd party libraries
Message-ID: <1800@dataio.Data-IO.COM>
Date: 3 Jan 89 22:40:36 GMT
References: <3792@midas.TEK.COM> <4800048@m.cs.uiuc.edu>
Reply-To: bright@dataio.Data-IO.COM (Walter Bright)
Organization: Data I/O Corporation; Redmond, WA
Lines: 39

In article <4800048@m.cs.uiuc.edu> rchen@m.cs.uiuc.edu writes:
<I wish Zortech could generate (as an option) Turbo C compatible object code
<so I can link it with my Turbo C graphics libraries.  I consider it as a great
<drawback that Zortech is incompatible with any of the existing well known
<C compilers on MS-DOS.  Zortech is a good choice if you want to learn C++
<but it is too primitive to do anything real.

Lots of people don't understand why they can't link libraries intended for
compiler x with code generated from compiler y. The problems are:

1. The compiler 'helper' functions (like floating point, long divide, etc.)
   are different.
2. The 3rd party libraries frequently use undocumented internal compiler
   library functions (the printf formatting routine is a popular example).
3. The 3rd party library makes use of nonstandard obscure functions that
   are peculiar to a specific compiler.
4. The 3rd party library was compiled with a different set of .h files,
   thus making things like the layout of the stdio _iob[] struct different.
5. You get my drift.

To be 100% compatible requires exactly copying someone else's library.
Obviously, this is stealing. The closer you match somebody else, the
more you risk being sued. It's a fine line.

Most of these 3rd party libraries can be made compatible with Zortech C
simply by recompiling them. If you don't have the source, give the vendor
a call and say you want a Zortech C version. Let them know it's worth
their while, they're probably not reading usenet.

If you really gotta have absolute MSC or TC compatibility, you need to
buy those products, not something else. One of the reasons I wrote my
own compiler is because I wanted to do it my way, not Phillipe's or Bill's.

As far as it being 'too primitive to do anything real', you need to be
more specific. The compiler has been in use for years for lots
of 'real' projects (including itself!). Perhaps your idea of a 'real'
project is different from mine.

** Turn up the volume, and rip off the knob! **