[Haskell-cafe] Indentation Creep

[Thomas Conway]

On 7/16/07, Claus Reinke <claus.reinke at talk21.com> wrote:
> enjoy (i hope:-),

Very much. Work hasn't been so much fun since.... Well, for a long time. :-)

One small question, which might reveal if I've understood things aright.

>         (do (v,e) <- dmin' l
>             (do guard e
>                 me <- empty m
>                 re <- nullT r
>                 write m p (v,me && re))
>              `mplus` return ((v,e)))
>          `mplus` (do v <- readTVar' m
>                      re <- nullT r
>                      write m p (v,re))
>          `mplus` (do (v,e) <- dmin' r
>                      when e $ writeTVar' p Empty
>                      return ((v,e)))
>          `mplus` error "emit nasal daemons"

If I refactor this a little into dminLeft dminMiddle and dminRight, I
believe I should be able to replace this with:

    ...
    msum [dminLeft l m r, dminMiddle m r, dminRight r, error "emit
nasal daemons"]
    where
    dminLeft l m r = do
       (v,e) <- dmin' l
       (do guard e
           me <- empty m
           re <- nullT r
           write m p (v,me && re)) `mplus` return (v,e)
    dminMiddle m r = do
        v <- readTVar' m
        re <- nullT r
        write m p (v,re)
    dminRight r = do
        (v,e) <- dmin' r
        when e $ writeTVar' p Empty
        return (v,e)
    ....

Is this correct? And if so, is GHC likely to do constant folding on
msum over the list skeleton?

T.
-- 
Dr Thomas Conway
drtomc at gmail.com

Silence is the perfectest herald of joy:
I were but little happy, if I could say how much.