[erlang-questions] I think I wish I could write case Any of whatever -> _ end.

[Eric Newhuis (personal)]

I am NOT proposing that the wild-card character behave selectively in pattern matching.  That just isn't part of what I am proposing.

As far as I can tell, the wild-card is not presently allowed at all in the context that I have proposed.  So there is no formal ambiguity as far as I can tell.  And it wouldn't break any existing code.   ...unless I am missing something basic.

I am not proposing that wild-card ('_') ever be reinterpreted on the left hand side of an arrow ('->').

To make this clearer one could introduce a special operator placeholder ('@').

>> case some_module:some_function(...) of
>>        {some, _, pattern} -> % _1
>>                case @ of ->  % _2
>>                        {some, great, pattern} ->
>>                                not_so_bad;
>>                        _ -> % _3
>>                                {_, Kind, _} = @, % _4, _5, _6
>>                                Kind
>>                end
>> end.


On May 17, 2010, at 9:36 AM, Robert Virding wrote:

> On 17 May 2010 16:00, Eric Newhuis (personal) <enewhuis@REDACTED> wrote:
>> ...
>> 
>> I guess where readability might break down is in nesting:
>> 
>> case some_module:some_function(...) of
>>        {some, _, pattern} -> % _1
>>                case _ of ->  % _2
>>                        {some, great, pattern} ->
>>                                not_so_bad;
>>                        _ -> % _3
>>                                {_, Kind, _} = _, % _4, _5, _6
>>                                Kind
>>                end
>> end.
>> 
>> Although I can still read the above once I learn that underscore ('_') is context sensitive.
>> 
>> _1 :: any()
>> _2 :: {some, any(), pattern}
>> _3 :: {some, any(), pattern}, not {some, great, pattern}
>> _4 :: some
>> _5 :: pattern
>> _6 :: _3
> 
> That is completely "unworkable". To put it very mildly. To have '_' in
> a pattern sometimes be a don't care which matches anything and is
> dropped and sometimes be a value which is matched against is just not
> acceptable. To make it even worse it changes its value in the body
> which would be the only case where this happens. Also the usage in
> your 4,5,6 case is inconsistent with the other examples as Pat = Expr
> is really just a shorter form of:
> 
> case Expr of
>    Pat -> <rest of clause>
> end
> 
> and completely equivalent to it except for the error message. So you really have
> 
> case _ of
>    {_,Kind,_} -> Kind
> end
> 
> which doesn't behave as you imply.
> 
> Sorry no!
> 
> Robert