This is the mail archive of the
libffi-discuss@sources.redhat.com
mailing list for the libffi project.
Re: minimal closure usage
On Tue, 2002-04-30 at 02:37, Sam Liddicott wrote:
> Is there a simpler way to preserve all the old arguments but prepend or
> append the user_data as an additional one? Or does the wrapped function
> have to have a calling signature of (ffi_cif*, void*, void**, void*)?
Not that I'm aware of. One hack might be to simply shove the user_data
in some form of thread local data instead of passing it in as an arg.
> (BTW: Is there an array/structure/function/full-set-of-rules to convert
> between "int" and "ffi_type_sint" etc? SWIG would need to refer to this as
> it generates the sample code below)
No, not really. I think, however, that "int" always ==
"ffi_type_sint". It's just the ffi_type_sint32, etc, which are platform
specific.
AG