That's what I had originally. However, some people have made critical
comments about CPP macros on this list and I thought that TH was
considered the better option.
I was one of those people advising against the use of CPP macros.
However, Template Haskell is ghc-only, and is unlikely ever
Hi Malcolm,
In this case, I am counting on GHC's
{-# LANGUAGE GeneralizedNewtypeDeriving #-}
feature to derive the instances for the classes I am including in the
deriving clause.
So perhaps portability is not a big issue here in any case.
I do think that
defObj(MyType)
looks a bit cleaner
On Wed, Sep 15, 2010 at 2:11 PM, Kevin Jardine kevinjard...@gmail.com wrote:
I do think that
defObj(MyType)
looks a bit cleaner than
$(defObj MyType)
I believe as of GHC 6.12 you no longer need the $() around top-level
splices. So that would just be:
defObj MyType
Hi Ben,
Good point! I can confirm that it does compile under GHC 6.12.
So really the same number of characters either way.
Kevin
On Sep 15, 4:49 pm, Ben Millwood hask...@benmachine.co.uk wrote:
On Wed, Sep 15, 2010 at 2:11 PM, Kevin Jardine kevinjard...@gmail.com wrote:
I do think that
On 9/15/10 1:31 AM, Malcolm Wallace wrote:
[...] However, Template Haskell is ghc-only, and is unlikely ever to
be implemented by any other Haskell compiler. [...]
Could it be implemented as a separate preprocessor?
Cheers,
Greg
___
Haskell-Cafe
It has been - there is a package called 'zeroth' on hackage.
It only works for top-level splices, the last I looked.
On Sep 15, 2010 11:20 AM, Gregory Crosswhite gcr...@phys.washington.edu
wrote:
On 9/15/10 1:31 AM, Malcolm Wallace wrote:
[...] However, Template Haskell is ghc-only, and is
On 9/15/10 9:11 AM, Kevin Jardine wrote:
Hi Malcolm,
In this case, I am counting on GHC's
{-# LANGUAGE GeneralizedNewtypeDeriving #-}
feature to derive the instances for the classes I am including in the
deriving clause.
So perhaps portability is not a big issue here in any case.
Yes, but
I supposed the simple solution might be CPP:
#define defObj(NAME) newtype NAME = NAME Obj deriving (A,B,C,D)
and then use
defObj (MyType)
I have heard some people, however, say that CPP macros are horrible in
Haskell, so is there a better solution?
Kevin
On Sep 14, 10:34 am, Miguel
Thanks Serguey!
The library code compiles, but when I try to use it in client code:
a. I get:
Not in scope: type constructor or class 'A'
and even stranger,
b. GHC cannot find any of my code after the
$(mkNewType A)
and claims that all the functions I defined there are also not in
scope.
Hmm - It seems to work if the code is defined before my main function
and not after it.
Does this have to do with TH being part of the compile process and so
the order matters?
Kevin
On Sep 14, 6:03 pm, Kevin Jardine kevinjard...@gmail.com wrote:
Thanks Serguey!
The library code compiles,
Yes, if you use template haskell, all top level functions and values
have to be defined before you use them.
Erik
On Tue, Sep 14, 2010 at 18:11, Kevin Jardine kevinjard...@gmail.com wrote:
Hmm - It seems to work if the code is defined before my main function
and not after it.
Does this have
OK, thanks for everyone's help!
Serguey's code works very well now.
Kevin
On Sep 14, 6:14 pm, Erik Hesselink hessel...@gmail.com wrote:
Yes, if you use template haskell, all top level functions and values
have to be defined before you use them.
Erik
On Tue, Sep 14, 2010 at 18:11, Kevin
Hi John,
That's what I had originally. However, some people have made critical
comments about CPP macros on this list and I thought that TH was
considered the better option.
What do other people think?
Serguey's code is great in any case as it gives me a clearer
understanding on how TH works.
13 matches
Mail list logo
