[Haskell-cafe] ANN: vector-buffer package 0.1

Alexander McPhail haskell.vivian.mcphail at gmail.com
Tue Feb 15 05:35:56 CET 2011


Hi Luke,

Apologies, I think I got bitten by unsafePerformIO.

The reason it wasn't in the IO monad is that the Vector created from the
Buffer is supposed to be immutable.

I've changed the API so that:

main = do
   buf <- newBuffer 10 :: IO (Buffer Int)
   pushNextElement buf 1
   v1 <- toVector buf
   v2 <- toVector buf
   print $ V.toList v1
   pushNextElement buf 2
   print $ V.toList v2

does as expected.


Thanks,

Vivian


On 15 February 2011 17:08, Luke Palmer <lrpalmer at gmail.com> wrote:

> This interface is an outlaw.
>
> main = do
>    buf <- newBuffer 10 :: IO (Buffer Int)
>    pushNextElement buf 1
>    let v1 = V.toList (toVector buf)
>        v2 = V.toList (toVector buf)
>    print v1
>    pushNextElement buf 2
>    print v2
>
> Despite v1 and v2 being defined to equal the exact same thing, this
> program prints two distinct lines.
>
> toVector depends on the current state of the buffer.  If this is to be
> a law-abiding interface, toVector must return a value in IO:
>
>    toVector :: (Storable a) => Buffer a -> IO (Vector a)
>
> Luke
>
> On Mon, Feb 14, 2011 at 7:28 PM, Alexander McPhail
> <haskell.vivian.mcphail at gmail.com> wrote:
> > Hi,
> >
> > I have uploaded a small package, vector-buffer, to hackage.  It provides
> a
> > buffer that can be turned into a Data.Vector.Storable.  The mapM*
> functions
> > map from the oldest element, not the first.  Similarly for the derived
> > Vector.
> >
> > Feature requests etc. welcome.
> >
> > Vivian
> >
> >
> > _______________________________________________
> > Haskell-Cafe mailing list
> > Haskell-Cafe at haskell.org
> > http://www.haskell.org/mailman/listinfo/haskell-cafe
> >
> >
>



-- 
---
yolar et elver.
---

DISCLAIMER

This transmission contains information that may be confidential. It is
intended for the named addressee only. Unless you are the named addressee
you may not copy or use it or disclose it to anyone else.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.haskell.org/pipermail/haskell-cafe/attachments/20110215/880217a9/attachment.htm>


More information about the Haskell-Cafe mailing list