Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10 5/3/83; site umcp-cs.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!harpo!seismo!rlgvax!cvl!umcp-cs!zben
From: zben@umcp-cs.UUCP
Newsgroups: net.lang.c
Subject: Re: Univac 1100 "c" compiler...
Message-ID: <6119@umcp-cs.UUCP>
Date: Sun, 25-Mar-84 00:42:11 EST
Article-I.D.: umcp-cs.6119
Posted: Sun Mar 25 00:42:11 1984
Date-Received: Mon, 26-Mar-84 07:38:34 EST
Organization: Univ. of Maryland, Computer Science Dept.
Lines: 24

Ron, it may be that this person wants a "c" compiler to execute in the
Native 1100 operating system environment.  While this would be a very
useful compiler, we have not seen such a beast.  We had a sideline project
for awhile to try to transport the ATT compiler out to the native
environment, but the project died due to the following factors:

1).  The compiler does not produce anything like the standard assembly
     language.  Instead, it produces a "Bell Univac Assembly Language"
     with local labels and other thingies that would be difficult or
     impossible to fake, even with the power of the Univac meta-assembler.

2).  The time to compile and link a copy of the Bell Univac C compiler on
     our (as usual) hopelessly overloaded 1100/82 approached a half an
     hour of clock (real) time.

3).  Even if we had been successful in doing so, we would still have to
     code a set of compatable runtimes (STDIO et al).  While (ahem) my
     Software Tools runtimes would be a very good start on such a beast,
     the project died before we even got a chance to think about it.

Even so, I think the project worthy of work.  Wish I had the time to 
take it on (after the BITNET <=> ARPANET gateway?  Sigh).

Ben Cranston         ...seismo!umcp-cs!zben        zben@umd2.ARPA