Fionn Malone
Fionn Malone
From #651 we need updates for: 1. QFxp 2. BoundedQInt 3. QIntOnesComplement
No, I need to profile to see why Fxp is so slow.

@mpharrigan / @tanujkhattar what are your thoughts on adding a ControlledHadamard Bloq, which possibly delays the above implementation and just tags the cost? It would be nice to have this...

Modifying the bloq directly with the generalizer is definitely not the way to go.
A maybe better way to do it is ``` python from qualtran.bloqs.and_bloq import And and_cv0 = ssa.new_symbol('cv0') and_cv1 = ssa.new_symbol('cv1') def generalize(bloq): if ('cirq.And' in bloq.pretty_name()) and (len(bloq.gate.cv) == 2):...
@mpharrigan is this last way something sensible, or is there a better way?
Could you just add a CX gate controlled on the target of the original inequality target?
Are you running ./check/pylint?