[daip] BPASS

[Eric Greisen]

Lynn D. Matthews writes:

 > I tested the fixed BPASS on my VLBA data.
 > Seems to be all okay.
 > 
I had forgotten a trick in which the INIT flag is set to false only
after a good solution which may not come for a few channels.  So I
reset it to false for channel 2 no matter what and then the solution
never did a proper initialization.  For some reason this worked in my
data even when I flagged channel 1 until I tried the debug version.
Then it failed with NaNs.

Thanks for following everything up so carefully.  I depend on users
like you.

Eric Greisen