Advertisement
kilon

Untitled

Mar 16th, 2012
99
0
Never
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
text 6.23 KB | None | 0 0
  1. [12:01:49] kilon [~kvirc@178.59.17.196] has joined #blenderpython
  2. [12:01:49] Channel topic is: Welcome to blenderpython, http://wiki.blender.org/index.php/Dev:Py/Sharing , http://wiki.blender.org/index.php/Extensions:Py/Scripts , https://projects.blender.org/projects/bf-extensions/
  3. [12:01:49] Topic was set by stiv!~steve@blender/coder/stivs on Δευ 15. Μαρ 17:22:50 2010
  4. [12:01:49] Channel synchronized in 0.0 seconds
  5. [12:01:56] adams.freenode.net [*@*] has set channel mode +cn
  6. [12:01:56] Channel was created at Τετ 25. Νοε 15:24:23 2009
  7. [12:04:26] frigge [~quassel@p3EE2CB7F.dip.t-dialin.net] has joined #blenderpython
  8. [12:22:55] <rking> kilon: Were you in on that whole BlenderArtists.org "Python should have Nodes" thread?
  9. [12:24:14] <kilon> rking: i post alot, but yeah I have been talking with lukas_t and his custom nodes features that is planned for future releases
  10. [12:24:30] <kilon> he seems to be close to completion
  11. [12:24:41] <rking> Wait, what is his project?
  12. [12:24:59] <rking> kilon: This was the thread: http://blenderartists.org/forum/showthread.php?247968-Random-thought-nodal-python&highlight=python+nodes
  13. [12:25:02] <kilon> that will open the door to addons that will give python developers the abilities to use nodes for anything
  14. [12:26:47] CoDEmanX [~CoDEmanX@p54BE497C.dip.t-dialin.net] has quit IRC: Quit: Leaving.
  15. [12:27:19] kromar59 [~Miranda@213-193-81-34.static.cablecom.ch] has joined #blenderpython
  16. [12:27:32] <rking> I was thinking a cool set of keybindings for a graphical programming interface would be actual code. Like you type if bpy.data.objects['Cube'].location.x < 20: ... then it creates the nodes that represent that block.
  17. [12:28:08] kromar5959 [~Miranda@213-193-81-34.static.cablecom.ch] has quit IRC: Ping timeout: 252 seconds
  18. [12:28:13] <rking> And vice-versa, being able to view the nodes as code -- complete round-tripping for the best of both worlds.
  19. [12:28:44] <kilon> rking: https://www.gitorious.org/~lukastoenne/blenderprojects/blender-lukastoenne/commits/custom_nodes
  20. [12:28:57] vantage|work [~chatzilla@d5152EDDB.static.telenet.be] has joined #blenderpython
  21. [12:29:53] <rking> Ahh, cool, thanks.
  22. [12:30:26] <kilon> the visual programming language i am working is abit like custom nodes but pack the full power of lisp programming which means that the user can define his own syntax and workflows
  23. [12:31:25] <kilon> it could eventually be integrated with blender , probably not inside blender but it could be made to compile to python code and refactor all blender python libraries, but still we are very long way from there
  24. [12:32:41] <rking> If it could compile to python then why not take the additonal step of putting it inside Blender?
  25. [12:32:47] <kilon> but when we do it, it going to explode the abilites of the blender user
  26. [12:33:20] <kilon> well a very good reason
  27. [12:33:25] <kilon> because its python and not lisp
  28. [12:33:26] <kilon> :D
  29. [12:34:11] <kilon> lisp come with an awesome debugger , that everytime you screw up it does not crash it ask you what to do next, give you full bactrace and even allows you to change your code before resuming
  30. [12:34:24] <rking> Ahh, I see.
  31. [12:35:08] <kilon> this is the sole reason I am interested in lisp, i think user are going to love the fact that they cant crash blender with their scripts or get their scripts exited and trying to figure out what went wrong
  32. [12:35:26] <lukas_t> kilon, rking: there are a few things still to be done, but it's well underway
  33. [12:35:31] <kilon> but as i said, there is tons of works before we get there
  34. [12:35:46] <kilon> lukas_t: hey thanks for all the effort i can wait as much as it takes
  35. [12:35:57] <lukas_t> custom data storage is the biggest headache atm
  36. [12:36:09] <rking> It's kind of weird how the old languages like Lisp and Smalltalk had such amazing blending between "programming" and "using", while so-called modern languages are slowly creeping in this direction.
  37. [12:36:14] <lukas_t> how to allocated data in python and store it in blend files
  38. [12:36:31] <rking> lukas_t: You can't slap things into custom properties?
  39. [12:37:08] <lukas_t> rking: yes that will be the initial solution, but it still doesn't allow pointers and structure such as color ramps and curves
  40. [12:37:18] <kilon> rking: its beyond the realm of idiocy, but that is what the world we live it , where hip is above progress
  41. [12:37:23] <lukas_t> general problem with id properties though, not specific to custom nodes
  42. [12:37:28] <kilon> its not just coding its everywhere
  43. [12:37:42] <rking> lukas_t: Have you ever looked at YAML?
  44. [12:37:52] <kilon> custom data being ?
  45. [12:38:05] <kilon> like python dictionaries ?
  46. [12:38:37] <lukas_t> rking: no, what would be the benefit for blender?
  47. [12:39:09] <rking> YAML is a human-readable (text-only) serialization language. It is most human-readable when you're dealing with dicts/lists/scalars -- but it has features to support objects and references.
  48. [12:40:08] <rking> lukas_t: So basically I'm wondering if you can flatten your objects to strings using YAML then put them on Custom Properties
  49. [12:40:24] <lukas_t> kilon: any data associated with a node, e.g. radius in a blur node, pointer to an object, a name string, etc.
  50. [12:40:39] <kilon> have dont the same with json and eval rking if i am understanding what you are saying
  51. [12:40:52] <kilon> *have done the same...
  52. [12:40:57] <rking> kilon: Yes, you are very much so. Only I think JSON is simpler.
  53. [12:41:28] <kilon> i used json to serialise blender python types / properties to a string
  54. [12:41:31] <rking> Does it have references?
  55. [12:41:52] <lukas_t> there are a number of options i think, but in the end it will have to be integrated into existing blender features (id properties)
  56. [12:42:00] <kilon> and then stored that string into a string property in a blender scene to be saved with my blendfile
  57. [12:42:49] <lukas_t> you also want to be able to use blender ui after all, it's not purely internal data
  58. [12:43:20] <lukas_t> so blender has to understand it too
  59. [12:43:46] <lukas_t> and not sure this would work well for pointers to objects, etc.
  60. [12:44:28] <kilon> yeah thats c stuff, you can mess with the same easy as with python objects
  61. [12:44:33] <kilon> *you cant
  62. [12:46:20] <kilon> but i dont know how to handle c well so
  63. [12:46:30] <kilon> it may be just my ignorance
Advertisement
Add Comment
Please, Sign In to add comment
Advertisement