> >If it creates DOS, Windows, OS/2, Unix, Hurd, CP/M or VMS
> >executables, that isn't the issue. If it can compile Clipper syntax
> >then that's as close as you get to being 100% Clipper.
> Hmm, so if it's 100% clipper compatible then where would the following
> output go ?
> #include <simpleio.ch>
Nowhere, this line should break the compiler.
> function main()
> ? "Hello World"
> return NIL
Thinking of a theoretical Windows oriented Clipper compiler (because
the one we are talking about doesn't "exist" at the moment), it could
do what a lot of C/WinSDK systems have done in the past and work
inside a simple window. In fact, thinking about it a little more, it
should do exactly what the above code says (when the bugs are fixed)
and write "Hello World" to stdout for the application and place "HELLO
WORLD!" at position "10,10", whatever the meaning of that position is
in the current environment.
> different meaning inside a windows GUI app than it does inside a DOS
> text mode application.
Correct, but that has nothing to do with the compiler as such and
everything to do with what you do with the compiler. The important
thing is that the compiler is 100% language compatible. A C compiler
that compiles for DOS, Windows or Unix won't guarantee that the code
you've written will run in the new environment, or that it will even
make sense in the new environment. It's important to make the
distinction between 100% language compatible and providing a
multi-platform framework. They are two different things.
> Now unless it creates a Win32 CONSOLE app, then we have code that
> would not work or not do the same as it did in its CA-Clipper
> counterpart, so is that 100% clipper compatibility??
See above. Besides, Clip4Win (and possibly FiveWin, can't remember)
> I know I may be "attempting" to stretch a point out of proportion
> here, but in my view the Syntax of a language is just as important
> as the functions bound to that syntax,
I'd sort of agree. However, there is no "standard" for the language
and no standard what functions are part of "Clipper", so you have to
think a little about what makes sense. Don't forget, it would be more
than acceptable to have a function "do nothing" given the environment
in which it's being used. Anyone who has ever used DJGPP will
understand that concept.
> another words, if all it claims to be is a 32-bit compiler that can
> compile 100% CA-Clipper apps, but decides to alter the meaning of
> the functions being called, is this 100% Clipper???
Depends on what you mean by "the meaning of the functions". If a
version of this mythical compiler "transformed" a function call into
something that made sense for the current platform I'd call that
correct, change of meaning or not. FlagShip, for example, appears to
do a very good job of this.
> Whilst Multi-threading is important and very handy, surely that
> should be viewed as an EXTENSION. What Five claims to be is a 100%
> 32-bit clipper compiler.
However, you have to consider that the target platform is Win95/NT. If
that is the intended target it makes sense that you should have access
to the abilities of that target environment.
> >I'm not so sure that creating 32bit DOS apps is that important at
> >this late stage, creating Win32 console apps could turn out to be
> Perhaps not, but I wonder just how much work would have been
> involved in allowing the creation of 32bit DOS apps ....
I would hazard a guess that there would be more work than payback. One
would assume that the market for 32bit DOS compilers it pretty
thin. The market for Win32/x-platform xBase/Clipper compilers is
probably a lot bigger.
> Would that sort of functionality not broaden the martket for such a
> compiler, would it not assist the migration of systems from a 16bit
> DOS world, upwards??
Up to what? Take a DOS application and make it a DOS application when
most people are now screaming out for Windows applications (for
reasons both good and bad)? Naa, sorry, as much as I like DOS I'd be
suprised if anyone thought it a good idea to introduce a DOS based
non-CA Clipper compiler. If someone does do it, more power to them,
but I can't imagine that they'd shift than many units.
Personally, I'm looking forward to the first Clipper compatible
compiler that compiles down to Java PCode.... :-)
Take a look in Hagbard's World: | w3ng - The WWW Norton Guide reader.
http://www.acemake.com/hagbard/ | ng2html - The NG to HTML converter.
Also available in the UK: | eg - Norton Guide reader for Linux.
http://www.hagbard.demon.co.uk/ | dgscan - DGROUP scanner for Clipper.