winds from sink particles should be implemented in the main code

[danieljprice]

currently winds from sink particles are implemented in plug-and-play injection routines, but it would make sense for this to be a generic property of sink particles and wrapped into the main code, where Mdot and v_inject are properties inside the xyzmh_ptmass array.

this would enable applications like:

• double wind interaction
• star with a wind interacting with the ISM
• star with a wind interacting with a disc
• binary wind accretion where the wind is from one or both stars