release notes 1.1: "A lot more"

General Discussion
#1

In the release notes for version 1.1.0 it mentions “a lot more” for the new features. I was wondering if someone could please list this out as much as possible.
Maybe part of “a lot more” is not fully functional, but I would still like some sort of unofficial mention of what to expect.

Thank you.

#2

OK, I’ll make a list here. This is far from complete. As I think of more, I’ll edit this post.

  • I did a lot of work on render-to-texture. You can now copy a non-power of two window to a texture, and it will automatically pad it. You can copy the depth buffer to a depth texture. You can create a parasite buffer that automatically tracks the size of the window of which it is a parasite. You can set up a buffer to render to texture, but only when you trigger the copy with a task (ie, not necessarily every frame). There is a method to ask a window to make a texture card, and it’s more convenient than CardMaker. Each of these things is a tiny change, but collectively, they make render-to-texture a lot more usable.

  • I fixed the “-ps” command line option for all the utility programs. You can now use -ps keep, -ps abs, -ps rel, etc, and they all work.

  • I modified x2egg so that it can handle the X-files that blender produces.

  • It’s now possible to store multiple UV-sets in an egg file, and access those in a shader.

  • The utility programs can compute tangents and binormals. Shaders can access those values. This is a necessary prerequisite for doing bump-mapping.

  • In principle, panda now supports hardware vertex animation. However, I’m not sure it’s entirely reliable yet, so I didn’t advertise it.

  • Panda now has a thing called “drawMaskAttrib,” which lets you control which cameras render which objects.

  • The whole “genpycode” subsystem, which used to be crucial to python-C++ interface, is much less important. It’s our hope that we can phase it out, in time.

  • The function “nodepath.setPythonTag(“key”, pythonobject)” lets you store a pointer to a python object (arbitrary user-data) in a panda node.

I’ll edit this post as I think of more.

#3

Thanks a lot Josh. And don’t worry about not remembering, thats why the post was started. So when it does come back in bits and pieces, you can update us all.

Thanks again.

#4

I think there is also the possibility to store any object in nodepath by
using setPythonTag (this way you can find back your python stuf…)
like:

x = Actor(yada yada)
x.setPythonTag("myactor", x)

Would you confirm Josh?

#5

confirmed, manakel. I just added it to the list above.

(BTW, “pandaForumAdmin” and “Josh Yelon” are the same person.)