[erlang-questions] Getting rid of some repetition.

[Kenneth Lundin]

On 3/18/08, Mats Cronqvist <mats.cronqvist@REDACTED> wrote:
> Colin Z wrote:
> > Say I have a process loop with a receive block that receives a bunch
> > of different messages...typical scenario:
> >
> > loop(A, B, C)->
> > receive
> >  {msg1, A1}->
> >   loop(A1, B, C)
> > ;
> > {msg2, B1}->
> >   loop(A, B1, C)
> > ;
> > {msg3, C1}->
> >   loop(A, B, C1)
> > ;
> > {msg4, A2, C2}->
> >   loop(A2, B, C2)
> > end
>  this idiom would reduce the typing;
>
> loop(A0, B0, C0)->
>  receive
>   {msg1, A1}-> B1=B0,C1=C0;
>   {msg2, B1}-> A1=A0,C1=C0;
>   {msg3, C1}-> A1=A0,B1=B0;
>   {msg4, A1, C1}-> B1=B0;
>  end,
>  loop(A1, B1, C1).
>
>  the new syntax you propose is a bit too perly for my taste.
>
>  mats
>
I think this is even better:

loop(A0, B0, C0)->
  {A,B,C} =
    receive
     {msg1, A1}-> {A1,B0,C0};
     {msg2, B1}-> {A0,B1,C0};
     {msg3, C1}-> {A0,B0,C1};
     {msg4, A1, C1}-> {A1,B0,C1};
  end,
  loop(A1, B1, C1).

Of course you could have called loop directly from the case alternatives
as in the original example, but if each alternative need to do
something more complicated I think this is clear and easy to
understand.

Mats example is actually more complicated in my view with all the ugly
B1=B0, ... stuff.

/Kenneth