#2004: Pattern matching against GADTs without -XGADTs has odd behavior.
-------------------------------------+--------------------------------------
Reporter: guest | Owner: simonpj
Type: bug | Status: new
Priority: normal | Milestone: 6.10 branch
Component: Compiler (Type checker) | Version: 6.8.1
Severity: normal | Resolution:
Keywords: | Difficulty: Unknown
Testcase: | Architecture: Unknown
Os: Unknown |
-------------------------------------+--------------------------------------
Comment (by simonpj):
Turns out that if `-XGADTs` was not set when pattern matching on a GADT,
we simply ignored the equational constraints. This is silly.
I've changed it so that `-XGADTs` is consulted only when deciding whether
to permit the ''definition'' of a GADT, not to permit its use. This is
consistent with (eg) overlapping instances.
Here's the note from TcPat
{{{
Note [Flags and equational constraints]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
If there are equational constraints, we take account of them
regardless of flag settings; -XGADTs etc applies only to the
*definition* of a data type.
An alternative would be also to reject a program that *used*
constructors with equational constraints. But want we should avoid at
all costs is simply to *ignore* the constraints, since that gives
incomprehensible errors (Trac #2004).
}}}
Probably worth merging into the 6.8 branch, but not a big deal:
{{{
Fri Jan 4 12:58:14 GMT 2008 [EMAIL PROTECTED]
* Do not consult -XGADTs flag when pattern matching on GADTs
}}}
Simon
--
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/2004#comment:2>
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
