Path: utzoo!mnetor!tmsoft!torsqnt!news-server.csri.toronto.edu!cs.utexas.edu!uunet!cbmvax!daveh
From: daveh@cbmvax.commodore.com (Dave Haynie)
Newsgroups: comp.sys.amiga.programmer
Subject: Re: Lattice C++
Keywords: C++
Message-ID: <18020@cbmvax.commodore.com>
Date: 24 Jan 91 20:08:00 GMT
References: <2875@dsacg3.dsac.dla.mil> <1991Jan18.120415.22068@zorch.SF-Bay.ORG>
Reply-To: daveh@cbmvax.commodore.com (Dave Haynie)
Organization: Commodore, West Chester, PA
Lines: 61

In article <1991Jan18.120415.22068@zorch.SF-Bay.ORG> xanthian@zorch.SF-Bay.ORG (Kent Paul Dolan) writes:
>nfs1675@dsacg3.dsac.dla.mil ( Michael S Figg) writes:

>I'd wait.  It is a port of the C++ 1.0 release, which is a long, long
>way from current C++ 2.1, and you really wouldn't be learning enough
>useful to do you good later.

Well, actually, it's C++ 1.1a, as I recall.  It is sufficient to get you
started on C++, but it does have a number of missing pieces when compared
to modern C++.

>And, it really isn't a good stand alone package; though C++ 1.0 is
>supposed to be a completely upward compatible revision of C, and this
>package nominally comes with the 4.0 C compiler, they left out so much
>of the C stuff that you can't compile even crufty old rev. 3.6 C code
>with it, much less start with that and build in C++ functionality.

According to Lattice, they didn't intend the C++ distribution to be C++ with
the revision 4.0 compiler, they intended it simply to be C++.  The fact that
they used the 4.0 compiler as a back end, and based their C++ 1.0 release on
AT&T's "cfront", was just an implementation detail.  They could very easily,
at any time, offer an upgrade which didn't use the C compiler as a back end.
So, no documentation on the 4.0 compiler, and none of the 4.0 extras that
would not be useful under C++.  

They did, however, include the full C library with the C++ program.  When it
first shipped, they hadn't considered the library a supported part of C++.
Based mainly on user feedback, I suppose, they did change their mind on this,
rather early in the life of the product, too, and shipped out a manual to all
registered C++ owners detailing the C function library stuff.

>On the other hand, Dave Haynie long ago reported doing some good work
>with it, so gather other opinions.

I did, though I haven't used it recently.  They have only themselves to blame.
I found the C++ distribution, initially, to have two annoying problems.  The
first was speed -- going through cpp, cfront, LC1, LC2, and BLINK to produce
code takes longer than you might like.  I suspect the real problem is the first
three phases -- each of these requires a lexical analysis of ASCII input, which
is often one of the most time consuming phases of any compiler.  Faster Amigas
and disks solved this problem for me, basically.  The next problem was the
compiler's tendency to crash on syntax errors.  Large stack (a meg or so)
helped, but didn't eliminate, this problem.  I suppose the typically sloppy
coding (from AT&T, not Tim's fault) of the originally UNIX based C code in 
cpp and cfront is more than likely the problem here, but it put a damper on
things.  

What finally got me to stop using it much, though, was neither of these.  When
Lattice released CodeProbe, I got spoiled.  It didn't read C++, and I found 
I could get much more done with C and source level debugging than with C++
and my older, more primitive debugging techniques.  I still write better code
in C++, and would pay to switch back if I can have C++ source level debugging
along with it.

>Kent, the man from xanth.


-- 
Dave Haynie Commodore-Amiga (Amiga 3000) "The Crew That Never Rests"
   {uunet|pyramid|rutgers}!cbmvax!daveh      PLINK: hazy     BIX: hazy
	"What works for me might work for you"	-Jimmy Buffett