[Haskell-cafe] Why were unfailable patterns removed and "fail" added to Monad?

[Gregory Crosswhite]

On 01/20/12 14:52, Michael Snoyman wrote:
> Essentially, I would want:
>
> SomeConstr args <- someAction
>
> to be interpreted as:
>
> temp <- someAction
> case temp of
>     SomeConstr args ->

I completely agree;  perhaps what we really want though is something
more akin to a language extension --- say,
DisableMonadFailForRefutablePatterns?

Cheers,
Greg