users-prolog
[Top][All Lists]
Advanced

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

Bit flags...


From: emacstheviking
Subject: Bit flags...
Date: Tue, 21 Oct 2014 23:54:19 +0100

I really didn't think this would be so difficult! I have tried for several hours now to find an elegant solution to my problem: given a 32 bit value, create a list of atoms that represent the names of those bit positions of interest.

The problem domain: decoding a Java .class file into a meaningful compound term because I can. 
Well, I thought I could!

I initially started with the idea of using findall/3 to find all solutions etc. but really didn't get anywhere with that so I went down the more conventional recursive route

%% these define the bit positions of interest
accflag(0x0001, public).
accflag(0x0010, final).
accflag(0x0020, super).
accflag(0x0200, interface).
accflag(0x0400, abstract).

%% predicate to convert Value into a list of accflag/2 names.
access_flags(Value, Output) :-
mapflags(Value,
[0x001:public, 0x0010:final,
 0x0020:super, 0x0200:interface,
 0x0400:abstract],
Output).

mapflags has taken many incarnations!!! None of which made me happy at all so here is the question: what is the most elegant way to process an input value and create an output list of atoms for each bit that was set?

Example: access_flags(0x0601, [public,interface,abstract]).

I will continue to try to find my own way but if anybody could show me how this can be dine with findall/3 I'd be very pleased. It just feels right that there should be a solution that reads something like "for all solutions to accflag(Bit, Name) compose a list of Name-s when Value /\ Bit is true". It mushroomed because as well as checking for the end of loop condition, which I couldn't decide was best to use either an empty list of bits (or no more solutions to accflag/2) or if the input Value reaches zero because then there would be no more bits to test!

AAARGGGGH! Next month will be my 30th year in software and it gets better every day!
Once again list, thanks!
Sean
:)


reply via email to

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