Two bugs fixed. First one was in printing, pmsg was spelled as psmg. Second one is more of a temporary fix. neg(x) was optimized to -1*x, where was dtype int which ended up in a dtype error in dot22scalar. I changed the optimization to create a -1 of same dtype as x but fancier solution possible.

Dimensions where the broadcastable flag is added are marked denoted by '1',
dimensions where it is removed are displayed as '0',
the ones left unchanged are marked by '?'.
This list is implicitly extended to the right by as many '?' as needed.

Make the op Rebroadcaste work for the gpu. Miss the opt that wil move the HostFromGpu after this op.

Modified realted calls as well.

 Fixed a bug in pfunc ( bug =  when you replaced a shared variable with givens, the shared variable did not got replaced, and the update rule was executed ) that resulted in fixing the failing tests of scan. I also did a bit of cleaining in scan tests and code, and fix an unobserved bug in inplace computation of scan plus made sure scan knows (once the optimization is written) to only store the last k steps of an output