[comp.lang.c++] C++, A/UX, and Mac OS

richer@well.UUCP (Mark H. Richer) (09/16/88)

Is anyone using C++ on A/UX?
So far I know of two sources of C++ for A/UX:
(1)  I spoke with someone from Oasys that said that they are shipping
their C++ 1.2 (binary) product for A/UX now. I believe it is $795. Has anyone seen
it on A/UX?

They claim that they have fixed some of the bugs in 1.2 and that they
have written a more efficient front-end.  Has your experience shown
that Oasys' version of C++ is substantially better than what AT&T
ships with source?

(2) For $2,000 we (as a commercial entity) could license the source
code from AT&T and port it ourselves. The main advantage is that we
could try to port it to the Mac OS as well as to A/UX. That would be
very important to us. Has anyone tried porting the AT&T sources to
A/UX or Mac OS? I would imagine that porting to A/UX should be easy
enough. Would it be a lot more difficult porting it to the Mac OS?  In
actuality, C++ on the Mac OS is more important to us (primarily
because there are 2 million machines out there running the Mac OS and
only a handful running A/UX right now), but we'll take what we can get
first.

Are there any other alternatives?

I know that Apple is working on MPW C++ for the Mac OS, but will we
see it this decade? Apparently the product release is delayed and last
I heard Apple might not even start alpha testing until the end of the
year. I suspect the delays are on the AT&T end. WE need another
alternative ---- and developing with C++ under A/UX is a
consideration, perhaps the only consideration at this point in time.

What are we doing?

We are developing an application framework and user interface toolkit
for the Macintosh (a commercial product).  Although somewhat similar
in concept and intent, there are some significant differences between
our framework and Apple's MacApp. Of course one major difference is
that our toolkit is written in object-oriented C rather than Object
Pascal (it is true that MPW C++ is supposed to interface with MacApp,
but we feel we can provide a better alternative).  And we plan to
offer a version that compiles under A/UX-- something you cannot do
with MacApp at present.  So far we have been using a "home-grown"
pre-processor to do object-oriented programming in C, thinking that
Apple's C++ would have been here long ago. Now we need to start
porting to C++ because we want to demonstrate the toolkit at MacWorld
in January. And we'd hate to have to demonstrate a version which is
not in C++ if that's what the product will ultimately use (and so far
I can't think of a better alternative). 

If you are interested in our toolkit send me a note --- I'll answer
whatever questions I can and keep you up-to-date on the status of the
project.  Don't expect to see it released before the second quarter of
1989, but I hope we can start outside testing by the end of the year.
We plan to alpha/beta test the C++ version so the timing depends
partly on how and when we can port our stuff to C++.

Mark Richer
Mountain Lake Software, Inc.
1041 Lake Street
San Francisco, CA 94118

beard@ux1.lbl.gov (Patrick C Beard) (09/16/88)

I am very interested in your toolkit based on C++.  Please keep
comp.sys.mac.programmer, and comp.lang.c++ informed about Apple's
C++ availability and how your progress is going.  I believe you
could do a LOT better than MacApp and that I could myself do a lot
better.  I would love to be able to write an AppleTalk class that
all I had to do was pass the constructor the Entity Name, and protocol
desired and off I would go on an appletalk application.  More power to 
you!

Patrick Beard
beard@ux1.lbl.gov

tjt@tis.llnl.gov (Tim Tessin) (09/17/88)

In article <7114@well.UUCP> richer@well.UUCP (Mark H. Richer) writes:
> 
> Is anyone using C++ on A/UX?

I have tried to port the ATT cfront 1.2.1 to A/UX.  I was not successful.
I had problems running the C compiler on my machine which is still at
2mb memory.  This may soon be fixed when my second-source memory 
comes in (in a few weeks) and I will try again.  I have had absolutely
no problems porting 1.2.1 to any machine I tried, so if the A/UX compiler
is reasonably robust, I don't forsee any problems.

mgchow@Apple.COM (Mike Chow) (09/17/88)

In article <22404@tis.llnl.gov> tjt@tis.llnl.gov (Tim Tessin) writes:
>In article <7114@well.UUCP> richer@well.UUCP (Mark H. Richer) writes:
...
>I had problems running the C compiler on my machine which is still at
>2mb memory.  This may soon be fixed when my second-source memory 
...

There is also a "large" version of the C compiler for A/UX; it has
much larger symbol tables then standard one.   You can invoke the
large C compiler by using the "-B /usr/lib/big/" option to CC.
It's fairly robust; programs such as Electric, Magic, X11, and
INGRES are all fairly large and have successfully been compiled
with the large C compiler (including a 160 line C macro).

You'll know when to use the large C compiler if the standard A/UX
C compiler dies on a Symbol Table Overflow error.

			Mike Chow
			Apple Computer, Inc.
			mgchow@apple.com

hugh@maths.tcd.ie (Hugh Grant) (09/24/88)

In article <7114@well.UUCP> richer@well.UUCP (Mark H. Richer) writes:
>
>Is anyone using C++ on A/UX?
>So far I know of two sources of C++ for A/UX:
>(1)  I spoke with someone from Oasys that said that they are shipping
>their C++ 1.2 (binary) product for A/UX now. I believe it is $795. Has
>anyone seen it on A/UX?

The A/UX version of designer C++ works quite well (allowing for the
overall performance of A/UX anyway). It works as well as on any other
system (being a straightforward port onto the A/UX Uniplus+ System V
O.S.)

>They claim that they have fixed some of the bugs in 1.2 and that they
>have written a more efficient front-end.  Has your experience shown
>that Oasys' version of C++ is substantially better than what AT&T
>ships with source?

Oasys actually distribute Glockenspiel's designer C++. Glockenspiel have
put a lot of effort into improving the AT&T code and have made over 300
bugfixes and have rewritten much of the code to achieve performance
gains. Also the Glockenspiel product has had a lot of work put into it
to make it portable across a wide variety of systems (from the PC to
VAX/VMS to Sun4). Glockenspiel C++ is *very much* better than the AT&T
version.

>(2) For $2,000 we (as a commercial entity) could license the source
>code from AT&T and port it ourselves.

Why bother when it's been done already?

>...The main advantage is that we
>could try to port it to the Mac OS as well as to A/UX....

That would cause problems, especially if you use one of the incremental
compilers such as Lightspeed. The Mac OS is quite different and would
require quite a bit of change to the user interface to C++.
I'm sure Glockenspiel would do a port if there was sufficient interest.
Why not contact them? There number is +353-1-364515

--
Hugh