i never use the trimmed edge from the extend option in _filletSrf.
most times i do
_untrimAll
(optional _extendSrf) _split Isocurve Shrink - to get the desired section of the new surface
many times the _split is at one end - exactly at the position where the un-extended surface would have been.
this is also the workflow i teach.
having this inside the command would allow nicer Knot-Spacing.
i do not see any useful usage of the provided trim of the extend Option and totally agree with jim.
even worth - the provided trim might even be a pitfall - see above example - the lower edge is not planar - see the green boundingbox of it:
i think the extend option should be replaced with some more advanced “set surface / fillet length”
default behaviour - ask the user to set the startpoint (like split at isocurve to get less of the fillet, or like extendSrf to get more of it) press enter to get a non-extended start.
Same dialog for the end, press enter to get a non-extended end.
Option / Optional for both: select a spliting / target Object.
it would be really powerful if this function is combined with the continuous fillet option, allowing to snap to points where the next surface pair is choosen - resulting in a partial continous fillet-stripe.
What you are describing is what the user is forced to do in Rhino versions 1-8. In Rhino9 all of that should be completely unnecessary. With the ContinueAcrossFaces=Yes option, FilletSrf already knows which end of the Fillet to extend and which end to not extend. The program should be able to determine where the ends of the Fillets go (so that the ends are made compatible with the base surfaces) and do it more reliably and accurately than most users can.
add an extension length option, with zero being the length that is currently returned from the extended option (but without trimming across the fillet).
For me adding something like 0.1mm just to make sure that the fillet fully intersects whatever surface is beyond where the fillet ran into a non-tangent edge.
maybe this would also be a great place where to add a “force pointy filelt / end - pick point” step.
sure this is the users view - the option will target 2 different algorithms:
… rail finding forcing to a point
… vs “just” splitting on top of the current algorithm.
