[daip] [bananas] VBGLU and 31DEC09

[Eric Greisen]

Michael Bietenholz wrote:
> Happy New Year!
> 
> About the VBGLU bug
> 
> 1) is this bug in 31DEC07 as well?
> 
> 2) am I safe if the data-set says "TB" sort order after loading?
> 

(1) Yes the bug is in 31DEC07 and maybe even older than that - probably 
made 2006/08/07 or even 2006/03/06.

(2) NO.  The TB is probably a lie and it is actually T?.  Do a PRTUV and 
you will probably see the baselines are not 
1-2,1-3,1-4,1-5,....,2-3,2-4,...9-10.  A proper UVSRT and probably MSORT 
would make a correct baseline order for a true TB.

I tested things in part by doing a VBGLU and then UVCOP to pull them 
apart again.  The 2 data sets then have to be UVSRTed since VBMRG will 
mess with order before a UVDIF can be run to compare.

I am amazed that this bug was not reported long ago - a user hit a case 
where autocorrelations ended up in cross-correlation baselines so it was 
very obvious.

Eric Greisen