I tried to run fieldsign calculation using surfaces created with the new recon-all but the results look wrong, with some vertices that have a salt-and-peppery look to them. The code I'm using is ripped out of an older version of tksurfer and works fine on old surfaces.
It seems that there is a lack of correspondence between vertex numbers in the flattened surface and orig or something. Another odd thing is that the ripflag is set for all vertices in the flat patch, but that is probably unrelated. We noticed that the file size of the different surfaces (e.g. orig, smoothwm, inflated) varies. In the older version, the surfaces all had an identical size. This suggests that the different surfaces have different number of vertices now, or you've changed the file format somewhat. Can you tell me what's going on there?
My current work-around is to use recon-all up to the point of fill, taking advantage of the new aseg, but then I'm using the old programs to generate surfaces compatible with my fieldsign code.