[commit: ghc] master: Build fix for dyn way on Windows; patch from nus (c39d315)
Bill Tutt
bill at tutts.org
Sun Feb 17 00:26:35 CET 2013
On Sat, Feb 16, 2013 at 5:04 PM, Ian Lynagh <igloo at earth.li> wrote:
> diff --git a/rts/Prelude.h b/rts/Prelude.h
> index dcd7b94..89e80a0 100644
> --- a/rts/Prelude.h
> +++ b/rts/Prelude.h
> @@ -93,6 +93,7 @@ PRELUDE_INFO(base_GHCziStable_StablePtr_con_info);
>
> #define runSparks_closure
> DLL_IMPORT_DATA_REF(base_GHCziConcziSync_runSparks_closure)
> #define ensureIOManagerIsRunning_closure
> DLL_IMPORT_DATA_REF(base_GHCziConcziIO_ensureIOManagerIsRunning_closure)
> +#define ioManagerCapabilitiesChanged_closure
> DLL_IMPORT_DATA_REF(base_GHCziConcziIO_ioManagerCapabilitiesChanged_closure)
> #define runHandlers_closure
> DLL_IMPORT_DATA_REF(base_GHCziConcziSignal_runHandlers_closure)
>
> #define flushStdHandles_closure
> DLL_IMPORT_DATA_REF(base_GHCziTopHandler_flushStdHandles_closure)
>
> diff --git a/rts/Schedule.c b/rts/Schedule.c
> index f39ef96..ea779dd 100644
> --- a/rts/Schedule.c
> +++ b/rts/Schedule.c
> @@ -2070,7 +2070,7 @@ setNumCapabilities (nat new_n_capabilities
> USED_IF_THREADS)
> // Notify IO manager that the number of capabilities has changed.
> rts_evalIO(
> &cap,
> - &base_GHCziConcziIO_ioManagerCapabilitiesChanged_closure,
> +
> DLL_IMPORT_DATA_REF(base_GHCziConcziIO_ioManagerCapabilitiesChanged_closure),
> NULL);
>
Shouldn't this use the new #define you added to Prelude.h just above?
This bit in rts\Win32\ThrIOManager.c should also use the Prelude.h macro:
#if defined(COMPILING_WINDOWS_DLL)
# if defined(x86_64_HOST_ARCH)
rts_evalIO(&cap,__imp_base_GHCziConcziIO_ensureIOManagerIsRunning_closure,NULL);
# else
rts_evalIO(&cap,_imp__base_GHCziConcziIO_ensureIOManagerIsRunning_closure,NULL);
# endif
#else
rts_evalIO(&cap,&base_GHCziConcziIO_ensureIOManagerIsRunning_closure,NULL);
#endif
Bill
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.haskell.org/pipermail/ghc-devs/attachments/20130216/d91b0cde/attachment.htm>
More information about the ghc-devs
mailing list