On Mon, Apr 21, 2008 at 4:27 AM, iain <[EMAIL PROTECTED]> wrote: > On Fri, 2008-04-18 at 17:04 -0700, Blake Ramsdell wrote: > > Another omitted WEBKIT_API problem. I think there's a bunch more of > > these, but I just fixed this one. > > In what situation do you need this? I was hoping that WebkitActor was > going to remain a completely private object.
Well, I may not understand the magic tools for generating Python bindings to GObject objects, so this may be a problem on my end. But when I got to the end of the code generation and build my Python binding, I got: ~$ python -c "import clutter; from clutter import clutterwebkit;" python: symbol lookup error: /usr/local/lib/python2.5/site-packages/clutter/clutterwebkit.so: undefined symbol: webkit_actor_get_type So fine, I made it public and it got better. I also tried commenting this out: ;; (define-function webkit_actor_get_type ;; (c-name "webkit_actor_get_type") ;; (return-type "GType") ;; ) But he still felt a need to track it down, presumably due to the fact that he knows it's a GObject-type thing and he made a Python Actor class (which I presume I still need in order for the object inheritance to work, since the inheritance order seems to be ClutterCairo -> WebkitActor -> WebKitWebView). If there's a way to get h2def or related tools to relax the need to know about the parent classes (which indeed, I really don't care about), I haven't found it yet. This is more due to my inexperience and ignorance than anything else. But I do feel like I'm forcing something to fit here, and I'm not sure if it's an inconsistent application of the GObject inheritance model, an incorrect application of the bridge code generation tools, or what. But that's why I felt making it public was a good idea. Believe me, I'm certainly not smart enough to call it myself, but someone in the clutter-webkit Python bridge code that I generated thinks it's a good idea. Blake -- Blake Ramsdell | http://www.blakeramsdell.com -- To unsubscribe send a mail to [EMAIL PROTECTED]
