jit.catch and jit.graph dim jumps to a very high number (even if I send it a correct dim message)
Hello, I'm baffled by the behaviour of jit.graph. I am hoping to keep it at a smaller dimension - a fraction of the jit.world screen size. However, the dim seems to jump to a high value and I do not want this. to happen, because I have 16 of these objects in my patch, and I want to save on processing.
The problem doesn't come from [jit.graph] but from [jit.catch~] which by default is set to @mode 0 which outputs everything from the last bang, which isn't exactly regular especially when coming from [jit.world], hence the varying matrix dim. By the way [attrui] isn't very good at catching quick changes of matrix dims in that case, [jit.fpsui] gives a better hint of what is happening.
Check out @mode 2 which allows you to define a fixed @framesize.
More generally, when you are baffled by the behavior of an object, select it, open the reference in the right side bar, and go through the attributes and see what can be changed.
Also, I saw that you were using [jit.pworld] for previewing, this is a bad idea since you are already using a [jit.world]. You can just use a [jit.pwindow] instead. You can think of [jit.pworld] as the combination of a [jit.world] embedded inside of a [jit.pwindow].
Hi, Martin.
The key factor missing was to adjust the framesize in jit.catch~.
That's where the "width" is defined, not in jit.graph.
By the way, I would use jit.gl.graph instead of jit.graph for most tasks.
@tfl and @pedro - thanks!!
I completely overlooked the modes for jit.catch, I was not looking at the right place when troubleshooting
@tfl thanks for the hint on jit.pworld vs jit.pwindow - I thought jit.pwindow is used for jitter and jit.pworld is used for textures, but it usually works in any case for troubleshooting I was unaware that using a jit.world and a jit.pworld together could cause issues, but looking back to previous issues, I think this makes sense!
also for @pedro's comment on jit.gl. graph, I would love to convert to the GL version if i can find an alternative to jit.glue (a pseudo jit.gl.glue) maybe ill digg into this later!
Otherwise here is a corrected version of this patch.