[daip] UVSUB

[Eric Greisen]

Bryan Butler writes:
 > >    Nothing except a change of the version ID string.  At least
 > >nothing intentional.  There was a change some time ago that should
 > >have applied to UVSUB but may not have been applied.  It was a
 > >correction for an error handling multiple frequencies.
 > >
 > >I will need details to be able to say  more....
 > 
 > well, i don't think that it's a "problem", but just an oddity.
 > i've been working on a bunch of 4-band data, and had been going
 > along with various runs of CALIB, IMAGR, UVSUB, etc... and had
 > been allowing them all to choose the type of FT (gridded vs.
 > DFT).  CALIB and UVSUB had always been choosing gridded, and
 > then, for some reason unknown to me, on one dataset, they started
 > choosing DFT.  it might be related to the number of visibilities
 > changing, but i didn't think that that had happened.  so, i just
 > wondered if something had been changed yesterday (because this
 > change happened between morning and late afternoon) to change
 > the way that the gridded vs. DFT was chosen...

Did you switch from new to tst?  I changed how that choice is computed
many months ago and I know that CALIB at the very least has been
linked many times since.  It is sensitive to the number of vis and the
image size so any change in those could switch it.  WE only linked
things in the last few days - no rebuilding of subroutine libraries -
so any subroutine change that failed to be compiled would still be in
that failure mode.  I am guessing that you were near the break point
and the number of vis dropped a little.

Eric