ciaraldi@Rochester.ARPA (01/24/84)
From: Mike Ciaraldi <ciaraldi@Rochester.ARPA> The new version of the Digital Research C compiler for CP/M-86 is in. This is the first one to use their new front-end/back-end technique. It shares the back-end (code generator) with the Fortran 77 compiler, and has a different front-end (parser). We received a letter in early December saying the new release would be sent in mid-December. After Christmas we called to ask where it was. They said we had to send in the form enclosed with the letter. There had, naturally enough, been no form with the letter, and no mention of a form in the letter. DRI took the information over the phone (address, serial number, etc., all of which they knew already, since they had sent the letter only to registered owners). The first week in January we called again, and found they were not actually shipping the update yet. It arrived last week and has most of the old bugs. Specifically, the floating-point I/O and floating-point to integer conversions are all messed up. So, we are using our own routines for these, which we had written for use with the old release. If anyone knows of a C compiler that runs under Concurrent CP/M-86 and handles more than 64K of data, we would appreciate knowing about it. Mike Ciaraldi ciaraldi@rochester
jeff@heurikon.UUCP (01/27/84)
Your report was on CP/M-86 'C' compiler bugs and DRI's failure
to fix them in the second release. I don't think DRI concentrates
too much on their compilers. We were a beta test site for CP/M-68K.
The O/S worked fine, but wow(!) were there ever a lot of 'C' compiler
bugs. They've acknowledged all of them and pointed out that they
are relying on their supplier to fix them, since they didn't write
the thing. (I'll tell you who did if you ask, but not on the net.)
There's only been one release of CP/M-68K so far, and I'm waiting
to see if the bugs are fixed in the next release. But, frankly,
I won't be too surprised it they aren't.
Fortunately, all of the bugs have work-arounds. But I'm slightly
bald now from pulling hair, and I've gotten very good at reading
the cryptic *.s files produced by the compiler to find out what
the 68K is *really* doing.
--
/"""\ Jeffrey Mattox, Heurikon Corp, Madison, WI
|O.O| {harpo, hao, philabs}!seismo!uwvax!heurikon!jeff (news & mail)
\_=_/ ihnp4!heurikon!jeff (mail)
ciaraldi@Rochester.ARPA (01/31/84)
From: Mike Ciaraldi <ciaraldi@Rochester.ARPA> Whoops! Preceding message was supposed to have been private.