[daip] RESEQ
Eric Greisen
egreisen at nrao.edu
Sun Aug 28 18:23:42 EDT 2005
Craig Walker writes:
> RESEQ dies with
>
> ZOPEN: ERROR INVALID LUN = 0
>
> This happened to Mark Reid at SAO and to me on my Mac and on my Linux
> box. For me it happened when trying to use an external file (Mac) and
> when trying just to combine 2 antennas into 1 with the normal inputs
> (Linux in NEW and TST). The latter test was done on a randomly chosen
> uv data set and with randomly chosen antennas, which suggests to me that
> this will be the result of any attempt to run the program.
>
> Let me know if you need more info.
>
> Also note that the first line that shows up with INP RESEQ indicates
> that it is a task that sorts data. I presume that this is a holdover
> from whatever RESEQ was adapted from.
>
I do not know how the authors of RESEQ thought that it worked. Here
is the resulting change.doc entry. Available by MNJ tomorrow.
11914. August 28, 2005 RESEQ, MATCH, PRTAN Eric
Changed MATCH help file to make it clear that it effectively
concatenates the AN, FQ, and SU information. Thus, careful
applications of MATCH will allow separate VLBI data sets (each
missing some antennas, sources, and/or FQIDs) to be made so
similar that they can be safely merged.
Task RESEQ is meant to renumber the various ground station
"antenna numbers" for the space antenna into a single antenna
number. It had the attributes:
1. It used a non-standard way to read and write UV files which
failed when I tried it and died with no message in its set-up
phase when Craig tried it. A FUDGE clone would have worked.
2. The tables subroutine did not set a LUN.
3. The polarization checking code used a wrong reference
pixel.
4. The subarray number was dropped from the output baseline.
5. The antenna number column in AN tables was not recognized
as such.
Took a FUDGE clone to copy the UV and made RESEQD (DIDDLE) do
the UV copy wit antenna renumber and possible baseline
direction swap. Copied the TABLE handling routine from the
old RESEQ and cleaned it up and made it work. The new task
will be able to renumber antennas in complicated schemes,
completely swapping things around if desired.
PRTAN confused AN row number with antenna number, causing the
VLA display to be messed up when the 2 are not equal.
Moved nowhere.
Cheers,
Eric
More information about the Daip
mailing list