Date   : Thu, 09 Dec 1982 02:49:33 EST (Thu)
From   : Ben Goldfarb <goldfarb.ucf-cs@Udel-Relay>
Subject: Re: 8080 to 8086 translation

Further to my complaint about XLT86, I finally got around to reading the
manual from DRI.  (This is always the final step around here).  The 
memory overflow problem is indeed documented, to wit:

(page 27)
4.3 Memory Overflow

       "The XLT86 program occupies approximately 30K bytes of main 
memory.  The remainder of memory, up to the base of CP/M, stores the 
program graph that represents the 8086 program being translated [sic].
The error message:

         ERROR  (7)  "Free Space Exhausted"

is issued if the program graph exceeds available memory.  A 64K CP/M 
system allows translation of 8080 programs of up to approximately 6K. [!!!!]

       The above error causes XLT86 to terminate.  To continue, you
must divide your source program into smaller modules and retry the
translation."


Well, if DRI is trying to impose modular programming practices upon us,
that is just about as bad as Godbout "requiring" CP/M purchasers to 
purchase a Sorcim assembler to maintain their BIOS.  While DRI used
some nice dataflow analysis in XLT86, that is of little consequence 
when most of my programs are larger than "6k" and, therefore, can't
be translated.  I feel like I got "soaked" for $150 at this point.
                                       Ben