[daip] small problem

[Eric Greisen]

Mark Reid wrote:
> Dear Aipser's
> 
>      Here is a problem that I've encountered more than once (but far 
> enough apart in time to forget what happened).  It has cost me hours of 
> work each time.
> 
> When I run the following with default settings for VLBA data:
> 
> AVSPC: to make a 16 chan "continuum" UV file from a 256 chan line file,
>         the new parameter DOACOR=-1 removes the AC data.  Since older
>         versions of AIPS didn't have this parameter, my old scripts don't
>         know about this.
> 
> ACCOR: runs "successfully" and creates an SN1 table.  But, the SN1
>         table it creates is empty (as there is no AC data to work on).
> 
> CLCOR: with SNVER 1; INVER 1 gives the message that it can't find
>         SN2.  This is very puzzling, since I told it only to look at SN1.
> 
> 
> Anyway, this all could be fixed by changing the the following:
> 
>      AVSPC defaults so that DOACOR=+1.
>      ACCOR should give a warning if it has no data to work on.
>      CLCOR should not look for SN2 if told only to look at SN1

1. The change to AVSPC was done October 1, 2007.  All tasks that allow 
calibration were changed to support all standard 
calibration/data-selection adverbs.  DOACOR false is what most users 
want most of the time.  It is a global adverb and you can set it to true 
(> 0) in your script.

2. I have today modified ACCOR to trap the no data found case and set 
the error return.  Your script will detect this error if you run with 
DOWAIT=TRUE.  It will not see the error if you run instead with 
DOWAIT=FALSE with task='accor'; go; wait.

3. CLCOR does not modify CL table 1 - it is an error to tell it to do 
so.  Instead, by default, it copies the input CL table (GAINVER) to an 
output CL version (GAINUSE) and then modifies that version.  This would 
be CL table 2 in normal usage (if you start with only 1 CL table).  The 
error message suggests that it was unable to read that new table, but 
you have not given me enough information to tell what really went wrong.

Eric Greisen