#5616: TH type quotes cannot contain free type variables
-------------------------------+--------------------------------------------
Reporter: Lennart | Owner:
Type: bug | Status: closed
Priority: normal | Milestone:
Component: Template Haskell | Version: 7.2.1
Resolution: invalid | Keywords:
Testcase: | Blockedby:
Difficulty: | Os: Unknown/Multiple
Blocking: | Architecture: Unknown/Multiple
Failure: None/Unknown |
-------------------------------+--------------------------------------------
Comment(by simonpj):
I asked in email: Can’t you pass the whole type?
{{{
inst t = [d| instance $t |]
}}}
You replied. No, you can’t. Both of those give you the error Malformed
instance header.
Indeed! How could we possibly type check the declaration in
{{{
inst t = [d| instance $t where
op x = x+2
|]
}}}
All the usual stuff for type checking instance declarations assumes that
we know the class, and indeed the instantiating types, of an instance
declaration. Else how would we know whether ‘op` came from the right
class, or `op x = x+2` has the right type.
I think the real answer here is in my
[http://hackage.haskell.org/trac/ghc/blog/Template%20Haskell%20Proposal
blog post about Template Haskell].
We can’t reasonably typecheck such declarations, but it is very convenient
to use the quotation syntax. So support the quotation syntax, but don’t
typecheck it. Restrict typechecking of Template Haskell to
''expressions'', and do the Full Job a la !MetaMl. See my blog post.
Here’s another example, even more extreme:
{{{
foo :: Name -> Q [Dec]
foo t = [d| $t = MkT Int |]
}}}
Here I want to splice in the ''name'' of the type to be declared.
Reasonable enough, and you can do it if you use explicit constructors, but
totally impossible using quotation syntax.
Do you like the blog post ideas? How much does it matter to you?
Simon
--
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/5616#comment:3>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler
_______________________________________________
Glasgow-haskell-bugs mailing list
[email protected]
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs
