understood. that's probably very similar to the "voltage controlled current source" mentioned by NiHaoMike. the "charme" of such a solution is the inherent short-circuit protection (iout = 0 @ vmains= 0). but then you are losing the info about fmains and the "shape" of vmains when mains gets shorted, i.e., in case of f.r.t.. that's why i said that a PLL is required. leaving f.r.t. aside, the VCCS is definitely a viable solution, given that it manages to synthesize a good sine-shape from vmains.
you could argue though that using the shape of vmains directly to synthesize the iout waveshape would be sufficient and "ideal" in a sense that your inverter would basically represent a negative resistor to the grid. however, that wouldn't support the grid in terms of reducing harmonic contents in the mains current, as opposed to sine-shaped iout.