Recursion in user patches causes earth-shattering kaboom

[bgribble]

Recursion should be detected at load or edit time and cause an error. It can't be meaningfully supported in mfp. Currently it's not all that hard to make a patch that will blow up the Python stack on creation by including itself.

C-s boom.mfp RET
p boom RET
C-s boom.mfp RET
C-f 
p boom RET

Any solution to this needs to also detect mutual recursion at depth, i.e. a -> b -> c -> a

What does this say about the metaphor of processors as function/method handlers? I think it's more like the problems of module inclusion in Python source...

[bgribble]

Saving this for later. Still needs to be fixed though.