Haskell 2010 libraries

[Brandon S. Allbery KF8NH]

On May 4, 2010, at 05:09 , Duncan Coutts wrote:
> On Fri, 2010-04-30 at 10:42 +0100, Simon Marlow wrote:
>> Bear in mind these goals: we want to
>>
>>   a. support writing code that is Haskell 2010 only: it only uses
>>      Haskell 2010 language features and modules.
>>
>>   b. not break existing code as far as possible

I'm going to dissent here:  current code assumes extensions, not a  
standard.  I think it's not outside the pale to have code that wishes  
to conform to Haskell2010 be modified to do so, and otherwise the code  
continues to be "extended nonstandard Haskell" if it is not already  
Haskell98-conformant.  After all, Haskell2010 doesn't quite include  
*all* of the extensions that are in common use.

So, I'm going to go out on a limb here and suggest that Haskell 2010  
code should specify

 > {-# LANGUAGE Haskell2010 #-}

to distinguish from Haskell '98 code (no LANGUAGE pragmas) and  
extension code (other LANGUAGE pragmas).  Users who wish to combine  
the above with nonstandard extensions can expect to do extra work.   
Existing code continues to work because it doesn't explicitly limit  
itself to Haskell2010.

A side benefit of this is that it requires the code (not a cabal file  
or etc.) to specify that it is Haskell2010 as opposed to Haskell98 or  
etc.  Unless cabal-install is a mandatory part of Haskell2010, relying  
on it to specify the language support level strikes me as not the best  
of ideas.

> If people think the order in the .cabal file is not sufficiently
> explicit then I'm sure we can concoct some more explicit syntax. We
> already need to add some syntax to allow a package to depend on  
> multiple
> versions of a single dependency.

We already have most of that, don't we?  There's the extension to  
allow you to specify the exact package to use for a given module; with  
"as" syntax one might presumably say something like

 > import "haskell-2010" Data.List
 > import "containers-ext" Data.List as L;

I understand this may require some work, but it seems a reasonable  
extension of existing syntax.  It also puts the onus of making things  
work together on the user, and (as mentioned above) I think that's  
eminently sensible for existing code that assumes that it uses  
extensions, not a standard.

> The advantage of the client doing it is it's quite general. The  
> downside
> is it's quite general: people can do it anywhere and can easily get
> incompatible collections of types. For example base:Prelude.Int would
> only be the same as haskell2010:Prelude.Int because it is explicitly  
> set
> up to be that way. Arbitrary shadowing would not be so co-operative.

ghc already lets you do this by rebinding syntax.  What happens if you  
rebind (>>=) in a way that isn't quite compatible with the monad  
laws?  Granted, right now you have to do fairly esoteric stuff to get  
yourself into that kind of trouble, whereas we're talking now about  
something likely to be more common.

> I'm not quite sure how it would be implemented but from the user's  
> point
> of view they just list the package dependencies as usual and get the
> sensible overlapping order. Presumably packages not designed to be  
> used
> in an overlapping way should still give an error message.

The problem here is, how do you know?  I recall suggesting some time  
back that dependencies without an upper bound were going to be a  
problem, and lo and behold, when base-4 came out they broke.  If a  
package declares itself to be capable of overlapping, does it apply  
only to Haskell2010 or is it assumed to also apply to Haskell2011  
unless specified otherwise?  Or do we try to figure it out  
automatically?  (Which I suspect would cause all sorts of problems.)

-- 
brandon s. allbery [solaris,freebsd,perl,pugs,haskell] allbery at kf8nh.com
system administrator [openafs,heimdal,too many hats] allbery at ece.cmu.edu
electrical and computer engineering, carnegie mellon university    KF8NH


-------------- next part --------------
A non-text attachment was scrubbed...
Name: PGP.sig
Type: application/pgp-signature
Size: 195 bytes
Desc: This is a digitally signed message part
Url : http://www.haskell.org/pipermail/libraries/attachments/20100504/86079807/PGP.bin