[daip] [Fwd: Re: small problem]

[Eric Greisen]

Mark Reid wrote:

>    It is easy to simulate this problem.  Just take a multi-source uv 
> file that has no AC data and run ACCOR.  It will create an SN table that 
> has no rows (with no warnings).  Then run CLCAL, requesting that SN 
> table.  Note that it says it is looking for SN2 when only SN1 was 
> requested.
> 
> Here are the inputs and messages I just generated.
> 
> Mark

>  >go
> CLCAL1: Task CLCAL  (release of 31DEC07) begins
> CLCAL1: Using interpolation mode 2PT
> CLCAL1: Processing SN table    1
> CLCAL1: TABINI: REQUESTED SN FILE   2 DOES NOT EXIST
> CLCAL1: SNINI: ERROR   2 FROM TABINI OPCODE = READ
> CLCAL1: SNINI: ERROR INITIALIZING SOLUTION TABLE FOR READ
> CLCAL1: ZCLOSE: LUN = 29 ALREADY CLOSED IN FTAB ON SEARCH
> CLCAL1: Purports to die of UNNATURAL causes
> CLCAL1: mjr          31DEC07 TST: Cpu=       0.0  Real=       1
> AIPS 1: Resumes
>  >
> 
> 
primat> ACCOR1: Task ACCOR  (release of 31DEC08) begins
primat> ACCOR1: Writing to SN table   5
primat> ACCOR1: ACRCOR: NO VALID AC DATA FOUND
primat> ACCOR1: Purports to die of UNNATURAL causes
primat> ACCOR1: primate 31DEC08 TST: Cpu=      0.1  Real=      1  IO= 
       51

So an up to date ACCOR returns an error code although it does not delete 
the empty SN table as it should.

What CLCAL does is to gather SN tables SNVER through INVER, merging and 
smoothing etc to a new temporary SN table which will be SN ver max+1 or 
2 in your case.  What heppens is that this new table is actually never 
created when there is no valid input.  CLCAL clearly needs to have its 
error detection updated.  The TABINI error, while effective, is less 
than obvious to the user.

Eric Greisen