23 Dec
2013
23 Dec
'13
10:47 p.m.
nice. btw, what's the cpu usage of those features, compared to normal data structures? still high, or lower than usual?
Check it out: http://www.jonathanwilkes.net/more-ds.webm
- transformations with a matrix
- svg path support for all drawing commands
- everything plays nice with Pd's subpatch x/y-range and size (not sure about margins), making it almost >like an svg viewport (without clipping, unfortunately)
Finally-- I'm also working on a patch that will instantiate certain scalars just by typing the struct name in an >object box. This means people will be able to use data structures for simple (or even complex) prototyping >without having to send a pointer anywhere. If you're just using a single scalar like my demo, you could just >as easily send methods to the parent [draw] objects. (Though if you understand how pointers work it can >be handy.)
Problems
- getting the bbox for a path is _hard_. I can't figure out a good way to cache it without requiring the user to >send a message.
- matrix transformations currently do an end run around all of Pd's glist_xtopixels/basex junk and send >directly to tkpath. This makes it way too complex to do stuff like figure out the center of a rotation. I don't >know a good way to deal with this, other than giving a high level interface for translation/rotation/etc.
-Jonathan