If you DupBorder the surface (already done in the file, cyan) then explode the border, then run SelDup, you will see the two linear edges circled above select. The line end points are identical out to 12 digits past the decimal.
This means that the part of the surface defined by those edges has zero (and I mean ZERO) area. I fail to see why it should be considered valid.
I have asked more or less the same question a couple of times in the last couple of months.
If you extend this surface ExtendSrf_bug.3dm (49.3 KB) you get a polysurface that is obviously not valid. In this case when you run DupBorder Rhino does recognize the border as invalid, but how can a polysurface with an invalid border be considered valid?
