Re: Integrating string types into EUPHORIA handling
- Posted by ghaberek (admin) May 05, 2014
- 1401 views
jimcbrown said...
This is doable without adding a new builtin type. We'd just add a new C type, say C_STRING in addition to C_POINTER and C_CHAR. The user passes that type into define_c_func/proc and then later when the user calls c_func/c_proc, the callc code would take a sequence, validate it (so no subsequences in that sequence, for sample), and convert it to a NUL-terminated char array.
Or, for C_WSTRING, do conversion into UTF-16. For rare cases, C_WWSTRING would convert to UTF-32, while C_USTRING would do UTF-8.
euadvlib does this. 
override function c_func( integer rid, sequence args = {} ) -- get the argument types and return type sequence arg_types = map:get( m_args, rid, {} ) atom return_type = map:get( m_retv, rid, 0 ) -- make sure we have the correct number of arguments if length( args ) = length( arg_types ) then for i = 1 to length( args ) do -- allocate strings as necessary if sequence( args[i] ) and arg_types[i] = C_STRING then args[i] = allocate_string( args[i], 1 ) elsif sequence( args[i] ) and arg_types[i] = C_WSTRING then args[i] = allocate_wstring( args[i], 1 ) end if end for end if -- call the function object result = eu:c_func( rid, args ) -- fetch strings as necessary if return_type = C_STRING then result = peek_string( result ) elsif return_type = C_WSTRING then result = peek_wstring( result ) end if return result end function
-Greg
