Re: [GHC] #7983: Bug in hsc2hs --cross-safe

#7983: Bug in hsc2hs --cross-safe -------------------------------------+------------------------------------- Reporter: singpolyma | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: hsc2hs | Version: 7.6.3 Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: | Unknown/Multiple Type of failure: None/Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Comment (by rwbarton): There seem to be two issues here. 1. The program is accepted by `hsc2hs --cross-safe` even though it's rejected by `hsc2hs --cross-compile`. 2. The program is rejected by `hsc2hs --cross-compile` for reasons that seem to have to do with the use of doubles. The ticket seems to have been intended to be about 1, since it's titled "Bug in `hsc2hs --cross-safe`". In general this has nothing to do with the weirdness of point 2. For example, you could write something clearly non- constant like `(#const getpid())`. hsc2hs will accept this in direct mode, which is okay I think; but it certainly won't work in cross-compiling mode, so `hsc2hs --cross-safe` should reject it too. I think this should be easy to implement: just include the output of `CrossCodegen.outValidityCheck` in the direct code generation path as well, if `--cross-safe` is enabled. I will create a separate ticket for point 2. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/7983#comment:7 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler
participants (1)
-
GHC