On Wed, Apr 2, 2008 at 9:51 PM, ere wer <[EMAIL PROTECTED]> wrote: > ...But my main point is - do we really need clutter_shader_bind to be exposed at all. Its very very low level function (much like glBlend or > -- snip -- > In the end, from end user perspective , actor_set_shader should be enough: > Its sets some "property" on my actor, that changes the way it looks. Just > like set_size or set_position.
The clutter_shader_bind is the only call that allows the user of the API to check for compilation/linking etc errros in the shader source code, thus it is needed assuming that the shaders used during development might contain programming errors. > PS In the test-shader example, is it really necessary to call > clutter_actor_set_shader (<actor>, NULL), which destroys the previously set > shader, before setting a new one? Looking at the set_shader code, its seams > not only unnecessary, but even faster if you call it the second time because > it skips the creation of the params hashtable. The tests in clutter are not neccesarily examples of how to use the APIs they might as well be samples that test corner cases or different ways of using the API. /Øyvind K. -- «The future is already here. It's just not very evenly distributed» -- William Gibson http://pippin.gimp.org/ http://ffii.org/ -- To unsubscribe send a mail to [EMAIL PROTECTED]
