[Haskell-cafe] Conditionals and case expressions in do blocks

[aditya siram]

Yep, I do this often and it's pretty nasty. It's especially
inconvenient that you don't have access to computations before the if,
for example:
do
 x <- something
 y <- something-else
 case y of
    Foo a -> do
                ....   <--- I want access to 'x' here
    Bar b -> ...

-deech

On Wed, Jan 26, 2011 at 2:51 PM, Edward Amsden <eca7215 at cs.rit.edu> wrote:
> Those functions are useful if I'm more interested in actions (for
> instance putStrLn) in IO, but they restrict the monad's return type to
> ().
>
> I'm often writing code like the following:
>
> do
> ...
> case ... of
> Foo x -> do ...
>                  ...
>                  return foobar
> Bar x -> do ...
>                 ...
>                 return foobar
>
> or worse:
>
> if cond
>  then throwError "BBQ!"
>  else do
>         ...
>         ...
>         return ...
>
> (which works btw because throwError is polymorphic in the return type)
>
> So I end up with a nasty set of nested do statements for case expressions.
>
>
> On Wed, Jan 26, 2011 at 3:32 PM, aditya siram <aditya.siram at gmail.com> wrote:
>> I think that for monads the cleanest way of doing conditional
>> execution is using  'when' and 'unless' [1] and 'guard' if your type
>> is a Monoid.
>>
>> -deech
>>
>> [1] http://www.haskell.org/ghc/docs/latest/html/libraries/base/Control-Monad.html#6
>>
>> On Wed, Jan 26, 2011 at 2:22 PM, Edward Amsden <eca7215 at cs.rit.edu> wrote:
>>> Inside a do block, I can very conveniently substitute
>>>
>>> let x = <pure exp>
>>> <continued monadic code>
>>>
>>> for either
>>> x <- return <pure exp>
>>> <continued monadic code>
>>>
>>> or
>>> let x = <pure exp>
>>> in do
>>> <continued monadic code>
>>>
>>> However, I can't do anything similar (that I know of) with if or case
>>> expressions. If I use if or case inside a do block, it's likely that
>>> I'm still using monadic expressions (at least in my experience). Is
>>> there some nasty semantic ambiguity that comes of this? Is it just not
>>> in the standard yet? Is there some extension that I'm not aware of?
>>>
>>> --
>>> Edward Amsden
>>> Undergraduate
>>> Computer Science
>>> Rochester Institute of Technology
>>> www.edwardamsden.com
>>>
>>> _______________________________________________
>>> Haskell-Cafe mailing list
>>> Haskell-Cafe at haskell.org
>>> http://www.haskell.org/mailman/listinfo/haskell-cafe
>>>
>>
>
>
>
> --
> Edward Amsden
> Undergraduate
> Computer Science
> Rochester Institute of Technology
> www.edwardamsden.com
>