guile-user
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Determining programatically whether the interpreter is Guile or Clis


From: 白い熊
Subject: Re: Determining programatically whether the interpreter is Guile or Clisp or Emcs
Date: Thu, 01 Aug 2013 16:42:31 +0400
User-agent: Roundcube Webmail/0.9.2

On 2013-08-01 05:12, Nala Ginrut wrote:
On Wed, 2013-07-31 at 12:20 +0200, Ralf Mattes wrote:
Isn't the main problem here that the OP assumes that all three languages
have "the same syntax"? This isn't true at all. They share some basic
syntax but any "real" CL/Elisp/Scheme code will use more than this basic subset. And even within this limited syntactic subset, while one syntax will work the same syntactic consgruct will have different _semantics_.


Yes, I'm afraid you're right, so I said pre-process is more portable.
Anyway, the project based on such an assumption will be very fragile.

I disagree. I'm quite aware of the differences and that is exactly the reason, I'm doing this. Do have to say that originally I assumed the syntax between Lisp and Scheme is much closer, that's right.

Anyhow, if it's of any interest, what I'm doing is: I'm developing http://kumatux.org/ What's on the web is the old version, I've converted about half the code to Lisp already and have made it fold both ways Lisp and Elisp. What this means is that I can call any function in Emacs or Clisp, and it'll automatically decide what the interpreter is and then modify the execution of the functions and code, so that the syntactic set is respected.

I realize now that this is completely not very feasible - to merge it with Scheme, however would like to at least keep playing with this for petty parts of the project. Just like like the idea of having one .lisp file, that can be called without any preprocessing within Emacs, Clisp, or Guile and it'll run. Currently I like playing with Clisp more, what I like about Guile is that it's a new project, so would like to incorporate it into the game a little too.

Anyhow, I've been experimenting with the latest recommendation of: (with-output-to-string (lambda... and seems I'm close, however it coughs up errors in Clisp, so still no go.

So if anyone would have alternate ideas, I'd be very much obliged. Thanks for helping me.
--
白い熊



reply via email to

[Prev in Thread] Current Thread [Next in Thread]