|Compilers digest Vol. 1 Number 2 compilers@ima.UUCP (1986-01-03)|
|Re: Compilers digest Vol. 1 Number 2 compilers@ima.UUCP (1986-01-03)|
|Relay-Version:||version B 2.10.2 9/12/84; site mit-hermes.ARPA|
|Posting-Version:||Notesfiles $Revision: 184.108.40.206 $; site ima.UUCP|
|Date:||3 Jan 86 19:51:00 GMT|
|Posted:||Fri Jan 3 14:51:00 1986|
|Date-Received:||4 Jan 86 15:21:47 GMT|
|Nf-From:||ima!compilers Jan 3 14:51:00 1986|
[from darryl at ISM780C (Darryl Richman)]
Has anyone looked at some of the newer, more automated methods of code
generation. I built a small code generator based on some of the ideas by
Graham and Glanville, supporting Unary *, =, +, and Binary -. I built it
for the 68k in about 5 hours. This technique presupposes a linear
intermediate language, which you then parse with, say, a LALR(1) parser,
emitting coe (or not as appropriate) with each reduction. The productions
represent addressing modes and operations.
Oh, BTW, although YACC itself is owned by AT&T, isn't it the case that
the output from yacc owes nothing to them? I believe that you owe no
royalties for the use and distributions of /lib/yaccpar.
--Darryl Richman, INTERACTIVE Systems Corp.
The views expressed above are my opinions only.
Return to the
Search the comp.compilers archives again.