[erlang-questions] Erlang elseif

damien morton <>
Thu Nov 20 14:08:24 CET 2008


traditionally, at least in C, this stuff is done with a bitvector, or
rather a vector of bits.

you would declare

#define CAPS 1
#define SMALL 2
#define NUMS 4

const unsigned bits[256] = { ..., CAPS, CAPS, CAPS, ..., SMALL, SMALL,
SMALL, ... }

#define IS_CAPS(x) (bits[x] & CAPS)
#define IS_ALPHANUM(x) (bits[x] | | CAPS | SMALL | NUMS)

Cant see any reason why you wouldnt do the same in erlang, but perhaps
you are only using characters as an example?

2008/11/20  <>:
> Thanks for your solutions, they give me good insight in how the Erlang world
> would solve this. I have one more way of solving this that I think is kind
> of clean and it would be efficient because the lists:member() would need to
> be evaluated once a match is found.
>
> elseif5(A) ->
>     try
>         lists:member(A,?CAPS) andalso throw(caps),
>         lists:member(A,?SMALL) andalso throw(small),
>         lists:member(A,?NUMS) andalso throw(nums),
>         throw(none)
>     catch
>         throw:caps -> io:format("Member of capital\n");
>         throw:small -> io:format("Member of small\n");
>         throw:nums -> io:format("Member of nums\n");
>         throw:none -> io:format("Not a member\n");
>         X:Y -> io:format("Bad exception ~p:~p\n",[X,Y])
>     end.
>
> However, some might think this is the wrong way of handling exceptions. What
> do you think?
>
> BR,
> Daniel Rönnqvist
>
>
> 2008/11/20 Richard Carlsson <>
>>
>>  wrote:
>> > I wrote a couple of examples how to solve this but none of them really
>> > feels right, the logic in what I'm trying to do is really REALLY not
>> > that complicated. The functions takes an argument and prints something
>> > based on what list the argument is a member of (if any). Like this:
>> >
>> > 42> e:elseif4($Q).
>> > Member of capital
>> >
>> > I would appreciate any input on this, even if it means you're bashing
>> > all my examples below :-)
>>
>> If you want a reusable pattern, rather than a one-off hack,
>> the best thing is to start by separating the logic and the
>> side effects (as in your 4th example).
>>
>> member_of(X, Lists) ->
>>  case lists:dropwhile(fun ({A,L}) -> not lists:member(X,L) end, Lists) of
>>    [{A,_}|_] -> A;
>>    [] -> none
>>  end.
>>
>> print_class(C) ->
>>  case member_of(C, [{caps, ?CAPS}, {small, ?SMALL}, {nums, ?NUMS}]) of
>>    caps -> io:format("Member of capital\n");
>>    small -> ...;
>>    nums -> ...;
>>    none -> ...
>>  end.
>>
>>
>> For one-off things, prefer the threaded functions solution if there
>> are more than 2 levels.
>>
>>    /Richard
>
>
> _______________________________________________
> erlang-questions mailing list
> 
> http://www.erlang.org/mailman/listinfo/erlang-questions
>



More information about the erlang-questions mailing list