Hi,<br><br>I'm not sure I fully understand what you're trying to do, but when I read<br><br><blockquote style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;" class="gmail_quote">
Ideally, I'd like to restrict my<br>
search to Located instances that wrap an instance of Outputable,<br></blockquote><br>I think this means SYB is not the right tool. Due to the way Typeable works (it's monomorphic), SYB doesn't really play well with adhoc cases for types which satisfy given class constraints. Your adhoc cases have to be listed per type, a bit like in the empty function [1], for instance.<br>
<br>Also, SYB and parametrized types don't really mix all that well either. This affects things like extracting all the `b`s from a `Located b`. Perhaps you could have a look at the new generics in GHC 7.2 [2]? You should be able to derive `Generic` for all the GHC API types (using standalone deriving), and this mechanism deals fine with (single) parametrized types. I'd be happy to hear what happens if you try this.<br>
<br><br>Cheers,<br>Pedro<br><br>[1] <a href="http://hackage.haskell.org/packages/archive/syb/latest/doc/html/src/Data-Generics-Builders.html#empty">http://hackage.haskell.org/packages/archive/syb/latest/doc/html/src/Data-Generics-Builders.html#empty</a><br>
[2] <a href="http://www.haskell.org/haskellwiki/Generics">http://www.haskell.org/haskellwiki/Generics</a><br><br><div class="gmail_quote">On Thu, Aug 11, 2011 at 17:56, JP Moresmau <span dir="ltr"><<a href="mailto:jpmoresmau@gmail.com">jpmoresmau@gmail.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">Hello, I'm banging my head against a wall there trying to use the syb<br>
generics schemes against the GHC API. I'm looking at implementing a<br>
search mechanism in the AST in the more direct way that what Scion<br>
does (creating an instance of a TypeClass for each AST node type).<br>
Since the GHC AST types derive from Data and Typeable, I thought it<br>
would be possible.<br>
So the problem is a follows: I have a start type, say<br>
TypecheckedSource, that derives from Data. What I'd like to do is to<br>
find all instances of the Located type inside that source that span a<br>
particular range. The problem is that Located is a parametrized type<br>
that wraps anything and just add location information. So the result<br>
of the search could be anything. Ideally, I'd like to restrict my<br>
search to Located instances that wrap an instance of Outputable,<br>
pretty print that and output only the result.<br>
So I'm looking to implement something that would have that signature:<br>
TypecheckedSource -> (Line,Column) -> [String]<br>
<br>
I have written code along these lines:<br>
everything (++) ([] `mkQ` overlap) ts<br>
where<br>
overlap :: forall b1 . (Outputable b1, Typeable b1) =>Located<br>
b1 -> [String]<br>
overlap (a::Located b1)= ... trivial code here finding if the<br>
location overlaps, and if it does, pretty print the object, otherwise<br>
returns []<br>
<br>
And GHC complains:<br>
Ambiguous type variable `b10' in the constraints:<br>
(Outputable b10) arising from a use of `overlap'<br>
at ...<br>
(Typeable b10) arising from a use of `overlap'<br>
at ...<br>
<br>
So it doesn't like the fact that I don't know which types my Located<br>
instances wrap. But that's the point, I don't care, I just want to<br>
restrict my search to the ones I can pretty print. If I just try the<br>
code against simple non parametrized types it of course works. If I<br>
add some forall b1 . (Outputable ... in my main function signature it<br>
complains that I never use b1 anywhere else in the signature, of<br>
course.<br>
<br>
Is there a way to achieve what I want<br>
(-XOhPleaseDoWhatIwantEvenIfIamNotSureItMakesSense or something)?<br>
<br>
Thanks a million!<br>
<font color="#888888"><br>
--<br>
JP Moresmau<br>
<a href="http://jpmoresmau.blogspot.com/" target="_blank">http://jpmoresmau.blogspot.com/</a><br>
<br>
_______________________________________________<br>
Haskell-Cafe mailing list<br>
<a href="mailto:Haskell-Cafe@haskell.org">Haskell-Cafe@haskell.org</a><br>
<a href="http://www.haskell.org/mailman/listinfo/haskell-cafe" target="_blank">http://www.haskell.org/mailman/listinfo/haskell-cafe</a><br>
</font></blockquote></div><br>