Hi I am looking into the ghc source code for the first time in order to see if I can understand why type synonyms are not correctly parsed with the ghc extension for infix :-*-: etc. I note that the following is present in the file Parser.y in ghc-6.4.20050216/ghc/compiler/parser and I assume it here ghc is parsing the type declarations? tycl_decl :: { LTyClDecl RdrName } : 'type' syn_hdr '=' ctype -- Note ctype, not sigtype. -- We allow an explicit for-all but we don't insert one -- in type Foo a = (b,b) -- Instead we just say b is out of scope { sL (comb2 $1 $>) $ let (tc,tvs) = $2 in TySynonym tc tvs $4 } on the next line happy is probably given instructions how to parse data declarations? | 'data' tycl_hdr constrs deriving { L (comb4 $1 $2 $3 $4) (mkTyData DataType $2 Nothing (reverse (unLoc $3)) (unLoc $4)) } Now, can someone explain what happens here, e.g. what is the difference between parsing of data and type synonyms? I guess (!) that the handling of infixes is merely an issue in the parser, and would not affect much else in ghc? Johan Glimming.