[GHC] #13609: regression: `ANN` pragmas for TH-less GHCs no longer ignored in GHC 8.2.1
GHC
ghc-devs at haskell.org
Mon Apr 24 10:37:31 UTC 2017
#13609: regression: `ANN` pragmas for TH-less GHCs no longer ignored in GHC 8.2.1
-------------------------------------+-------------------------------------
Reporter: hvr | Owner: (none)
Type: bug | Status: new
Priority: high | Milestone: 8.2.1
Component: Compiler | Version: 8.2.1-rc1
Keywords: | 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:
-------------------------------------+-------------------------------------
Consider
{{{#!hs
module M where
{-# ANN myId "HLint: ignore" #-}
myId :: a -> a
myId x = x
}}}
GHC 8.0.2 would simply ignore them:
{{{
$ ghc-8.0.2 -c M.hs
M.hs:3:1: warning:
Ignoring ANN annotation, because this is a stage-1 compiler or doesn't
support GHCi
}}}
however, with a recent GHC 8.2 snapshot this just fails with
{{{
$ ghc-8.2 -c M.hs
ghc: this operation requires -fexternal-interpreter
}}}
This is a serious problem because `hlint` annotations are quite popular,
and this would require us to retrofit `other-extensions: TemplateHaskell`
into the meta-data of quite a few packages on Hackage to declare that a
package no longer works with a TH-less GHC (which would basically mean
that ports of GHC such as the one for AIX that don't support TH will
become very inconvenient to use and require to pester package authors to
explicitly guard `ANN` pragmas with CPP... which I'm not looking forward
to ;-) ).
--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13609>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler
More information about the ghc-tickets
mailing list