C++ for Amiga?
21 messages in this thread
Is anyone using the the Lattice C++ – it was a "close-out" item around the
time that SAS took over the C compiler. Is there an update from SAS or
Lattice to C++ 2.0 or 3.0, including 2.0 streams?
Ken,
Comeau C++ seems to be the only current C++ for the Amiga. Version 3.0 was
recently released, which includes ANSI C++ Templates.
I'm not certain about 2.0 streams. Perhaps Greg Comeau can comment on that
aspect.
David
Yes, Comeau C++ 3.0 With Templates includes the "new" (which is a few years
now) streams.
When SAS took over the marketing/support/development of Lattice C, they did
not want nor take Lattice C++. SAS has said that they will come out with a
C++ for Amiga but to date they have not — and with the very strong C++
compiler from Comeau Computing, they probably won't.
-sja
Steve-
I'd be surprised if SAS didn't come out with a C++, since they've had one
in beta-test for a while now. Do you have any info from SAS which indicates
that they've dropped it in mid-test?
I unfortunately received but a single beta version, which crashed on a
plain 68000- never did receive an update which would work on my
unaccelerated machine. Of course, this would be consistent with a
discontinuation of the program…
Bob
– (!) –
Bob,
I think you answered your own question.
-sja
Naw- I *posed* a *possibility.* I just wanted to know if you have info on
this, or are speculating, that's all.
Bob
– (!) –
No speculation. SAS has been discussion an Amiga C++ for a number of years
now and their latest statement is that it wouldn't even be *started* until
the beginning of next year.
What's wrong with Comeau C++? It's pretty solid and is amongst the most
up-to-date compiler on any platform.
Greg-
Now see, that's the kind of statement which puzzles me. How could I ahve
received a beta version of their compiler a year ago if it "hadn't been
started?" Was "their latest statement" from an official spokesperson? It's
sort of hard to figure out what's up when very definite statements conflict
with what I personally know as fact. Delayed, postponed, deferred, dropped-
all are plausible. "Not started" ain't.
And I don't know why you think I implied anything is "wrong" with Comeau
C++, as I've never heard of it until one message back, and I know nothing
about it. I just answered a message concerning SAS C++. Best wishes for
both you and your compiler…
Bob
– (!) –
hello Bob I have a copy of Lattice C++ for the amiga it is VERSION 1.00 I
live in the UK and it was ordered from USA by my computer store. I have had
it about 6 months
DAVE
– via Whap!
Dave-
That is most likely the interpreter they've had for some time now. The SAS
product was to be a true (or "native") compiler- i.e., not a "front end"
like "cfront" which translates C++ into C prior to compiling. I was never
really interested in that- I've thankfully had few opportunities to debug
generated code.
Bob
– (!) –
Actually, I don't think that SAS has ever stated that they would have a
true compiler. The product being worked on was a cfront like product which
was just a little more closely integrated into the compiler product line.
Now it is not impossible for it to become a true compiler, but that was a
bit more work. Note that I have no association with SAS Institute.
SAS clearly indicated to me that the intended product was a native
compiler. Whether or not that was an accurate statement is, of course, a
completely different issue! Sounds like it's a moot point anyhoo- sigh.
Bob
– (!) –
>That is most likely the interpreter they've had for some time now.
The Lattice C++ 1.0 compiler (not interpreter) was a cfront 1.x.y compiler
ported by a 3rd party and published by Lattice that is something like 4 years
old and has never been upgraded.
>The SAS product was to be a true (or "native") compiler-
All are true compilers. There are different "object code" output's from
compilers though. Some produce asm code, some produce native object code,
some produce native object code for a different target CPU (including
"unreal" ones), some produce C.
>"cfront" which translates C++ into C prior to compiling. I was never really
>interested in that
Yes, cfront is a translating compiler (Comeau C++ is a licensed port of
cfront 3.0, the only one currently out there, and what Borland is claiming
compliance to). It has pros and cons just like the other flavors.
Note that the translation into C is not just text substitution but full
compilation (the C compiler is there only as a portable assembler in order to
produce the native object code and is not expected to produce any diagnostic).
Greg-
You're right about my incorrect use of the term "interpreter;" I'd claim
temporary insanity, only many would immediately argue the use of the
adjective "temporary…"
What I was referring to with respect to Lattice C++ is the disadvantage of
its production of C source code as its output, obscuring the constructs and
syntax of the original C++ source and thus adding greatly to the difficulty
of debugging. I know that Cfront is the de facto "standard," but I
personally don't feel a C++-to-C-to-binary is worth the trouble.
– (!) –
>hello Bob I have a copy of Lattice C++ for the amiga it is VERSION 1.00 I live
>in the UK and it was ordered from USA by my computer store. I have had it
>about 6 months
That is not a new product but an old product that is know to be very buggy
and extremely outdated.
>Now see, that's the kind of statement which puzzles me. How could I ahve
>received a beta version of their compiler a year ago if it "hadn't been
>started?" Was "their latest statement" from an official spokesperson? It's
>sort of hard to figure out what's up when very definite statements conflict
>with what I personally know as fact. Delayed, postponed, deferred, dropped-
>all are plausible. "Not started" ain't.
I'm not going to argue the definition of start. Use occur or proceed if
you like.
What I know is fact and official as well. That somebody sent something to
somebody else does not necessarily dispute that. No SAS/C++ work will occur
during 1992. SAS/C++ work may occur after that, or it may never occur.
These are not my words.
>>What's wrong with Comeau C++?…
>And I don't know why you think I implied anything is "wrong" with Comeau C++,
Sorry about that. I was intending to be rhetorical'ly and it obviously didn't
come out right.
Where does one get a copy of Comeau C++? I have never seen it in the
stores or advertised. Since it is 3.0, and SAS have said don't expect
anything until well after SAS/C 6.0, I would be interested in moving to
Comeau.
>Where does one get a copy of Comeau C++? I have never seen it in the stores
>or advertised. Since it is 3.0, and SAS have said don't expect anything until
>well after SAS/C 6.0, I would be interested in moving to Comeau.
Right here if you'd like. Or you can call 718-945-0009 and speak to one of
our knowledgeable sales people.
If you do it here, send e-mail containing your full name, full address,
telephone number, credit card number (all major cards accepted except
American Express), expiration date, full name on card, what product you want
(I suspect Comeau C++ 3.0 With Template for AmigaDOS in your case) and a
statement of ($US250 for ADOS) and your agreement to the price. Domestic
goes UPS Blue (2nd day air) at no extra charge. If you are outside the US,
add $US60 for international shipping. That goes UPS Red (next day air).
Well over a year ago, when I was at SAS, there was a trial Beta test of
C++, but the feedback was that there were fundamental problems that
required major work to correct. SAS has pretty publicly stated that there
will not be a C++ for the Amiga any time in 1992. Please note: I DO NOT
WORK FOR OR IN ANY WAY REPRESENT SAS INSTITUTE, INC.
Thanks, John. From the comments made by Steve and Greg, it sounded like the
project was at best on hiatus. Your input is consistent with both their
input and my own experience. Another best-laid plan of men gang astray…
Bob
– (!) –