 |
lua-l archive |
|
|
||
|
On 04/08/2008 20:09, A.S. Bradbury wrote:
I hate to ask a question I could answer myself, but I'm going to do it anyway - have you benchmarked luaffi vs compiled lua bindings at all? Clearly for some sorts of library most of the work will be being done on the C code, so the cost of the function call is negligible but for other applications (maybe a datastructure implementation?) the cost of the call would be more important.
I don't think this kind of library is designed to replace a good native wrapper. Although it can be used to prototype one. The way I see its main use is to allow easy access to a large set of API functions which are rarely used. An obvious (for me!) use is to access the very big Win32 API: you cannot imagine wrapping each and every function there. For example you might write a GUI wrapper to do common, usual base window tasks (creating base components, getting and setting base information from/to them) but you might want to leave some obscure, rarely used functions, letting users access them with a FFI call if needed: less work, less bloat, more flexibility.
-- Philippe Lhoste -- (near) Paris -- France -- http://Phi.Lho.free.fr -- -- -- -- -- -- -- -- -- -- -- -- -- --