Re: findall with limited solution count
- From: Joachim Schimpf <j.schimpf@xxxxxxxxxxxxxxx>
- Date: Fri, 03 Mar 2006 12:38:59 GMT
Jan Wielemaker wrote:
> Nice! B.t.w. SWI-Prolog offers change_arg/3 as nb_setarg/3. It shouldn't
> be exposed to much to the normal Prolog programmer, but I find it an
> increasingly handy building tool for things that simply _need_ to
> survive backtracking and where assert is the only traditional way to
> realise your goals. Mostly because it avoids all the pain to make code
> reentrant, cleanup after exceptions and thread-safe that is associated
> using permanent shared resources such as dynamic predicates.
>
> GNU-Prolog also offers somethings similar.
In ECLiPSe, rather than clobbering standard Prolog stuctures
in that way, we use anonymous objects called shelves, see
http://eclipse.crosscoreop.com/eclipse/doc/bips/kernel/arrays/shelf_create-2.html
--
Joachim Schimpf /
ECLiPSe Team / http://eclipse.crosscoreop.com
CrossCore Optimization / mailto:j.schimpf@xxxxxxxxxxxxxxx
.
- References:
- Re: findall with limited solution count
- From: binnetcorp
- Re: findall with limited solution count
- From: Jan Wielemaker
- Re: findall with limited solution count
- Prev by Date: comp.lang.prolog Frequently Asked Questions
- Next by Date: [ICLP'06] Call for Posters - Deadline: March 14, 2006
- Previous by thread: Re: findall with limited solution count
- Next by thread: comp.lang.prolog Frequently Asked Questions
- Index(es):