Re: Compilier Construction Tools?

dwight@pentasoft.com (Dwight VandenBerghe)
27 Jul 1998 11:41:04 -0400

          From comp.compilers

Related articles
Compilier Construction Tools? shindle@toocool.com (1998-07-26)
Re: Compilier Construction Tools? dwight@pentasoft.com (1998-07-27)
Re: Compilier Construction Tools? wclodius@aol.com (1998-07-27)
Re: Compilier Construction Tools? vmakarov@cygnus.com (Vladimir Makarov) (1998-07-27)
Re: Compiler Construction Tools? friwi@prosun.first.gmd.de (1998-07-28)
| List of all articles for this month |

From: dwight@pentasoft.com (Dwight VandenBerghe)
Newsgroups: comp.compilers
Date: 27 Jul 1998 11:41:04 -0400
Organization: http://www.supernews.com, The World's Usenet: Discussions Start Here
References: 98-07-183
Keywords: tools

On 26 Jul 1998 00:33:49 -0400, shindle@toocool.com wrote:
> PCCTS:
> tried, true, supported, and apparently fairly popular. Unfortunately,
> our parser guy thinks the C support of the new java based ANTLR needs
> some more testing before we should try it.


Nah, it's a great tool. Top drawer, should be in your A list.


> Eli:
> This looks REAL nice with some builtin symbol management, and large
> library of builtin functions. Appearently, spec files can be used
> to combine multiple lexer/parsers combinations into 1 executable
> (so we could combine Cobol, SQL, etc). On the downside, it uses
> LALR(1?) parsers. I have just got it built on my Linux box and am
> trying to learn it, but its HUGE :-). Is it likely to meet our
> needs ??


It's HUGE. Very complex, and last I heard it only works under unix.
Major learning curve, but once you're through it, three years from
now, you'll be able to write great compilers ... that only other
Eli initiates will be able to understand.


> GRDP/RDP:
> GRDP isn't finished yet..but it promises to be compatible with RDP
> grammars. The parser generator sounds like it would be good for us
> novices, but what other tools does it provide ??


It's simple and clean and understandable. For your task, I think it
might fit the bill really well. The parsers that it generates are
VERY understandable. It has new graph support that is pretty cool.
C only, not set up well for C++.


> TXL:
> I played with ver. 7.4 a little, maybe a year or two ago. At the
> time, there was no way to store the AST's - has this changed in v8 ?
> The source to source nature of it seemed like it might be useful in
> our work, but we also need to track var. usage for metrics, etc.
> Can it be used for this ??


Didn't he go commercial with the real version, and hobble the old one?


> Cocktail:
> Another one I ran into a couple of years ago, but doesn't seem to
> be used much anymore ?? I remember it was taken commercial - is
> freeware version still used ?? Would the it or the commercial version
> be good for our uses ??


Same thing - old version not supported any more, new version costs big
bucks.


> Gentile
> This is a new one to me - the page I saw was very sparse in terms of
> what tools it provides. Would it be a good fit for our uses ??


It's fabulous. This is a wonderful system, but it takes a little
getting used to. Solid as a rock, in my experience, and the model
for compilation (structural induction) is great. I think you might
have to pay for the commercial version, but it's the same as the
free version; so you could try it out and see if you like it.


> SUIF
> Another one that I can't seem to find much info on. I realize its
> designed for parellizing compiliers, but would its tools be useful
> to us ??


Nope, stay away. This is big-time optimization, very slow, made
to prototype state-of-the-art compiler backends. This is the
last thing in the world you need right now.


> Scorpion
> An appearently dead project at Arizona State University and an
> University in Carolina. It looked real ambitious, but seems to
> have died ??


No info on it.


> The list goes on and on...Lisa, CoCo, BTYACC..Any help in sorting
>all this out would be appreciated !! If anyone wants to reply
>directly, I'm user shindle on toocool.com


Actually, Steve, as strange as it seems, I think the best way to go
with compiler generation is not to find a tool that front-ends C/C++
for you, but rather, to move into a language that is made from the
ground up to handle compiler issues elegantly. I'm talking about ML
and Objective Caml, which both have astoundingly complete facilities
for writing and maintaining compilers. I use Ocaml every chance I
get.


On the other hand, this may not be an option for you. If you're in
C++, then look at the best available tool: Yacc++ by Computer
Resources. I have a copy, it's the best lexer/parser generator system
I've ever seen. If you're in C, then take a look at Gentle, and see
if you can make it through the learning curve. It generates out
lex/yacc front-ends, but has great support for the backend.


And, you can always contract with one of us in the group here
to give you a hand...


Take care, Dwight
--


Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.