[GHC] #13649: RebindableSyntax causes type errors when 'fail' is not defined, even if not used.
GHC
ghc-devs at haskell.org
Sat May 6 16:44:05 UTC 2017
#13649: RebindableSyntax causes type errors when 'fail' is not defined, even if not
used.
-------------------------------------+-------------------------------------
Reporter: AaronFriel | Owner: (none)
Type: bug | Status: new
Priority: normal | Milestone:
Component: Compiler | Version: 8.2.1-rc1
Resolution: | Keywords:
| RebindableSyntax
Operating System: Unknown/Multiple | Architecture:
Type of failure: GHC rejects | Unknown/Multiple
valid program | Test Case:
Blocked By: | Blocking:
Related Tickets: | Differential Rev(s):
Wiki Page: |
-------------------------------------+-------------------------------------
Comment (by nomeata):
That is indeed a good question. It seems that the type signature of `fail`
is ignored completely. This might indicate that the renamer requires
`fail` to be present, but that the type-checker does not see it. And
`compiler/deSugar/DsExpr.hs` definitely does not emit a call to `fail` if
the match can fail.
In function `tcMonadFailOp` the `fail` part of a bind is only type-checked
if the pattern can fail, using `isIrrefutableHsPat`.
Potentially, the local function `getFailFunction` in
{{{
rnStmt ctxt rnBody (L loc (BindStmt pat body _ _ _)) thing_inside
}}}
could do the same.
Is that enough pointing-the-way for you to try to fix this?
If you do, make a good note somewhere explaining the interplay of how the
renamer, the type checker and the desugarer decide whether to assume a
call to the `fail` function or not.
--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13649#comment:3>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler
More information about the ghc-tickets
mailing list