Date   : Thu, 13 Oct 1983 17:56:12 EDT (Thu)
From   : Mike Ciaraldi <ciaraldi@Rochester.ARPA>
Subject: Digital Research C

For those who have been following my travails with DRI C for
CP/M-86 (R): 
We are now using Version 1.03 of the compiler, with Version
1.02 of LINK86, and everything looks good. There
are problems with using the compiler with version 1.01 of the linkage
editor.
How I found this out:
When Version 1.0 of the compiler showed up several months ago, it
sat on a desk for a few days. Then the linker arrived in a separate
package, with a note saying it was needed to run the compiler
but had inadvertently not been shipped.
Both the compiler and linker had "READ.ME" files, and the one for the
compiler did say you had to use at least version 1.02 of the linker.
It turns out the new disk really had version 1.01 of the linker!
If I had paid more attention to the header which the linker prints
out when it runs, I would have noticed this. As it is, none of the
other users did either.  
In any event, there were compiler problems which all seem to be
fixed with the new version, 1.03. When that version arrived, someone
noticed the linker mismatch, and a call to DRI resulted in their sending
a copy of linker 1.02.  

Results:  It works!  It even compiles and runs under Concurrent CP/M,
on both the IBM PC and the Zenith Z-100. We have successfully run multiple
tasks, sent queue messages back and forth, and shared blocks of memory.
The sample programs compile and run in all 4 memory models, too.
One bonus: under CP/M-86, if you are using the compact, medium,
or big memory models, you have to invoke you .CMD file with the R command,
so it will do segment fixup after loading. Under CCPM this is
automatic.