|Re: Thompson's Plan 9 C compiler email@example.com (1991-08-12)|
|Re: Thompson's Plan 9 C compiler firstname.lastname@example.org (1991-08-13)|
|Re: Thompson's Plan 9 C compiler email@example.com (1991-08-13)|
|Re: Thompson's Plan 9 C compiler firstname.lastname@example.org (1991-08-13)|
|Re: Thompson's Plan 9 C compiler email@example.com (1991-08-14)|
|Re: Thompson's Plan 9 C compiler firstname.lastname@example.org.COM (1991-08-15)|
|giving away the store by Factors of Two email@example.com (1991-08-16)|
|Re: Thompson's Plan 9 C compiler firstname.lastname@example.org (1991-08-16)|
|[1 later articles]|
|From:||email@example.com (Lars Svensson)|
|In-Reply-To:||firstname.lastname@example.org's message of 12 Aug 91 16:39:05 GMT|
|Organization:||IMEC vzw, Leuven, Belgium|
|Date:||13 Aug 91 07:50:04 GMT|
[From comp.arch -John]
In article <GW5DMC@xds13.ferranti.com> email@example.com (peter da silva) writes:
The Thompson compiler is extremely hardware-independent and generates
good 68020, MIPS, and Crisp code.
Note that the code generator proper is hand-written for each of these
architectures. To quote the paper: "There is a considerable amount of
talk in literature about automating this part of the compiler with a
hardware description. Since this code generator is so small (less than
500 lines of C) and easy, it hardly seems worth the effort." I guess
this is less than a gcc hardware description anyway.
It also implements full ANSI standard C, not a subset.
Not correct. To quote the paper again: "The compiler implements ANSI C
with some restrictions and extensions." "Several of the poorer
features were left out."
This is just nit-picking. I also really liked the paper.
Return to the
Search the comp.compilers archives again.