FTBFS with Pd>=0.55

[umlaeute]

With the recently released Pd-0.55, pd-lyonpotpourri no longer compiles (at least not for Pd64)

the reason is, that Pd is now stricter when it comes to types, and uses a static assert to ensure that the signal holder member in CLASS_MAINSIGNALIN is of type t_float.
Unfortunately, pd-lyonpotpourri uses type float in the following files:

• adsr~.c
• buffet~.c
• bvplay~.c
• channel~.c
• chopper~.c
• clean_selector~.c
• click2bang~.c
• click2float~.c
• clickhold~.c
• click~.c
• convolver~.c
• counter~.c
• distortion~.c
• dmach~.c
• epluribus~.c
• expflam~.c
• flanjah~.c
• granola~.c
• granulesf~.c
• granule~.c
• kbuffer~.c
• killdc~.c
• magfreq_analysis~.c
• markov~.c
• mask~.c
• npan~.c
• oscil~.c
• phasemod~.c
• player~.c
• pulser~.c
• rtrig~.c
• samm~.c
• sigseq~.c
• vdb~.c
• vdp~.c
• waveshape~.c

the solution is simply to switch to using t_float

[umlaeute]

in general, you should use t_float instead of float whenever possible (which i think is everywhere except for specialized standalone functions like the penrose generators)

[umlaeute]

related to ericlyon/pd-fftease#51