I also do that other workaround, but it often needs undoing and is time consuming to figure out what’s the neat distribution of the newly added spans.
The biggest flaw of the current “Match surface” is the inability to match with G2 by moving the 3rd row of control points in the normal direction only, thus preserving the intended flow of the surface.
Another missing feature is the ability to individually match the start edge and/or end edge to inherit tangency from the target surface’s side edges, especially if they are trimmed.
Individuall matching to U direction or V direction is another missing feature. Currently, the “Match surface” tool has a common option for UV direction that produces ugly and unusable output surfaces.