[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Avoiding name clashes between different scripts
From: |
Chusslove Illich |
Subject: |
Avoiding name clashes between different scripts |
Date: |
Fri, 1 Apr 2005 12:55:10 +0200 |
User-agent: |
KMail/1.7.1 |
Hi folks,
I am trying to introduce Guile for scripting some things within, eh, a
certain C++ application. Some of these scripts, written by different
people, would come bundled with the app, and users could build their
scripts on top.
One (of many :) questions that I haven't been able to resolve yet, is what
mechanism to use to avoid name clashes in top level defines? I would like
to avoid forcing script writers to use unique prefix for defines as in C,
something like namespaces in C++ would be preferable.
Some specialties of the intended use might be helpful:
(1) Distributed scripts can be of arbitrary complexity, but user scripts
would normally be just calls to single functions within distributed
scripts, and in any case no more than few lines of code. In that light, I
wouldn't like heavy syntax for users to use namespace separations.
(2) The app itself can provide a sensible default namespace for each user
script, so ideally users wouldn't have to bother with namespaces at all,
unless they need something from other namespace. But when they do need it,
I'd like that they must name it (no automatic resolution in case of
unambiguous situation).
(3) I wouldn't even mind completely disabling users to use other than
default namespace if that would cut down the syntax in user scripts, since
users and writers of distributed scripts would actually be well connected.
(4) While namespaces would be fine, consider this about just prefixing top
level defines: prefixes could be very well defined (no need for anyone to
think what to use) and consisting of only two to three letters. Together
with (2) and (3), I might settle to a universal single character prefix,
to be replaced by suitable real prefix by the application before user
script is executed.
This was my attempt to shrink the problem to general considerations, but if
you would like the full story, it is here:
http://caslav.gmxhome.de/writings/ktranscript.html
--
Chusslove Illich (Часлав Илић)
pgpDln8kxaJQb.pgp
Description: PGP signature
- Avoiding name clashes between different scripts,
Chusslove Illich <=