On 05/04/16 18:53, Davide Andreoli wrote: > 2016-04-05 9:34 GMT+02:00 Tom Hacohen <[email protected]>: > >> On 01/04/16 18:13, Davide Andreoli wrote: >>> 2016-03-29 17:04 GMT+02:00 Tom Hacohen <[email protected]>: >>> >>>> There you go. In, and fully tested (eo_suite_add_fallback suite) as >>>> promised. Please let me know if the tests fail on your box, though it's >>>> very unlikely because it's almost a direct copy from the previous eo_do >>>> implementation. >>>> >>> >>> hmmm, I have a problem with this change >>> EAPI Eo * >>> -_eo_add_internal_start(const char *file, int line, const Eo_Class >>> *klass_id, Eo *parent_id, Eina_Bool ref) >>> +_eo_add_internal_start(const char *file, int line, const Eo_Class >>> *klass_id, Eo *parent_id, Eina_Bool ref, Eina_Bool is_fallback) >>> >>> in bindings (both python and lua) we are directly using the >>> _eo_add_internal_start() >>> function instead of eo_add() because it's impossible (please prove I'm >>> wrong) to bind >>> so complex macros in cffi. >>> So the question is: what should I pass to the "is_fallback" param? >> >> You should pass false. We need this because of a limitation in C that >> prevents us from having a context variable in macros that return a >> value. Since you are reimplementing this macro you are able to have a >> such a context variable, so you don't need the fallback support. See the >> non-fallback implementation for how it should be in python/lua. >> > > Perfect thanks, > then I have already done it correctly.... it was 50% probability :) >
Both would work, it's just that no fallback is faster and more correct. -- Tom. ------------------------------------------------------------------------------ _______________________________________________ enlightenment-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
