projects / ghc-hetmet.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 8ae8c86) | patch
Undo the fix for Trac #3772 and do it a new way
authorsimonpj@microsoft.com <unknown>
Tue, 5 Jan 2010 10:16:00 +0000 (10:16 +0000)
committersimonpj@microsoft.com <unknown>
Tue, 5 Jan 2010 10:16:00 +0000 (10:16 +0000)
commit9a4c93a59e008ddc376fde5f9eb468b762f0d0a7
tree757600c7543f7fd7322ac16a278ea92654105aa0tree | snapshot
parent8ae8c865716e0ee03383ddfae58ab778ce527412commit | diff
Undo the fix for Trac #3772 and do it a new way

The main idea is that I'm now treating a single-method dictionary very
much like a multi-method dictionary.  In particular, it respond to
exprIsConApp_maybe, even though newtypes aren't *really* proper
constructors.

See long comments with Note [Single-method classes] for why
this slight hack is justified.
compiler/basicTypes/MkId.lhs diff | blob | history
compiler/typecheck/TcInstDcls.lhs diff | blob | history
Glasgow Haskell Compiler with heterogeneous metaprogramming (generalized arrows and modal types)