On 3 Apr, 2008, at 15:46, has wrote:


One of the nice things of the current Carbon bindings
is that a large portion of the API is just there and you probably
don't have
to write C code when you want to use an API.

Sadly enough that's only true of API's that were present in OS9, but
the
idea stands: it would be nice to have complete bindings to the bits of
Carbon that still make sense.

Yes, although I'd repeat my earlier suggestion that the most
economically viable way to provide Carbon bindings would be to create
ObjC wrappers for the Carbon APIs of interest. That way, any language
with ObjC bindings, not to mention ObjC itself, can take advantage of
these bindings for no extra effort.

I'm far from convinced that this is true. The bridgesupport tools don't target Objective-C exclusively but can also be used to wrap C API's. The only problem w.r.t. Carbon might be the arcane memory- management rules of Carbon. I'm not sure if that's relevant for the non-GUI bits of Carbon. Even then it should be possible to extend the bridgesupport tools a little with annotations that tell bridges that a specific type is a Carbon-style handle.

Even fixing bgen isn't that much work, once you understand the code. The problem is that bgen is a nearly vertical learning curve and Jack seems to be the only person that understands enough of bgen to be able to hack on it. I'm definitely in favor of ditching bgen and moving toward PyObjC-based wrappers.

Ronald


Attachment: smime.p7s
Description: S/MIME cryptographic signature

_______________________________________________
Pythonmac-SIG maillist  -  Pythonmac-SIG@python.org
http://mail.python.org/mailman/listinfo/pythonmac-sig

Reply via email to