[daip] SNEDT bug
Adam Deller
adeller at nrao.edu
Fri Dec 3 18:38:23 EST 2010
Hi Eric,
I wrote out a file from TASAV which exhibits the behaviour. You can get it
at /users/adeller/forsnedt_uv.fits. I double checked by loading the FITS
file back into AIPS, and the the first time I repeated the process it
happily saved the table. I then tried again with SN version 5 instead of 6,
and it crashed as before. The input/output from below is from when I ran
SNEDT before writing the file out and reading it back in - the only
difference is that I look an SN table 6 instead of 5.
In both cases, I paged through the first few antennas and set the display to
delay and rate before exiting. I believe SN table 6 was generated by CALIB
and SN table 5 was from FRING. Here are the inputs and output:
AIPS 1: SNEDT: Interactive SN/CL/TY/SY-table data editor using the TV
AIPS 1: Adverbs Values Comments
AIPS 1: ----------------------------------------------------------------
AIPS 1: INNAME 'BD141I_C' Input UV data (name)
AIPS 1: INCLASS 'TASAV' Input UV data (class)
AIPS 1: INSEQ 1 Input UV data (seq. #)
AIPS 1: INDISK 1 Input UV data disk drive #
AIPS 1: INEXT 'SN' Input 'SN', 'CL', 'TY' table
AIPS 1: INVERS 6 Input table file version no.
AIPS 1: DODELAY 1 > 0 => also edit delay/rate
AIPS 1: TIMERANG *all 0 Time range to include
AIPS 1: BIF 0 First IF to include
AIPS 1: EIF 0 Last IF to include
AIPS 1: ANTENNAS *all 0 List of antennas to include
AIPS 1: FREQID -1 Frequency ID; 0 => 1
AIPS 1: SUBARRAY 0 Subarray number 0 => 1.
AIPS 1: FLAGVER 0 > 0 FG version applied
AIPS 1: SOLINT 0 Times within SOLINT (min) are
AIPS 1: treated as equal.
AIPS 1: DETIME 0 Break interval (min)
AIPS 1: DOTWO 1 True => do second observable
AIPS 1: plot of main baseline
AIPS 1: EXPERT 0 > 0 start in expert mode
AIPS 1: CROWDED 0 > 0 => allow plots with all
AIPS 1: polarizations and IFs
AIPS 1: ANTUSE *all 0 Initial displayed antennas
AIPS 1: BADDISK *all 0 Disks to avoid for scratch.
and I get as before:
localh> SNEDT1: No data changes to write back to disk
localh> SNEDT1: REQUESTED ROW **** EXCEEDS MAX OF 1280
localh> SNEDT1: ERROR FOR TABLE OBJECT: Output SN/CL/TY/SY table to be e
localh> SNEDT1: EDGTSN: ERROR READING REC 2 ROW****** COL WEIGHT 1
localh> SNEDT1: EDITSN: ERROR WITH OBJECT Output SN/CL/TY/SY table to be e
localh> SNEDT1: ERROR EDITING Output SN/CL/TY/SY table to be e
Hope this helps,
Adam
On Fri, Dec 3, 2010 at 2:51 PM, Eric Greisen <egreisen at nrao.edu> wrote:
> Adam Deller wrote:
>
>> Hi daip,
>>
>> It seems there is a bug that has been introduced into SNEDT sometime in
>> the last couple of months. If I open an SN table to edit it, page through
>> some antennas and change what is being displayed (i.e. put up delay and
>> rate, for example) and then exit without making an edit, SNEDT crashes
>> complaining about a row ****. See output below.
>> localh> SNEDT1: Task SNEDT (release of 31DEC10) begins
>> localh> SNEDT1: Doing no flagging this time
>> localh> SNEDT1: Copied SN file from vol/cno/vers 125148 3 to 125148
>> 5
>> localh> SNEDT1: Allocate 22799 words at 89068349
>> localh> SNEDT1: All data will reside in memory
>> localh> SNEDT1: Found 6 times, stored on a time axis of length 16
>> localh> SNEDT1: Applying FG table to the data
>> localh> SNEDT1: Plotting:
>> localh> SNEDT1: Press buttons A, B, or C to choose an operation
>> localh> SNEDT1: Press button D for on-line help
>> localh> SNEDT1: Plotting:
>> localh> SNEDT1: Press buttons A, B, or C to choose an operation
>> localh> SNEDT1: Press button D for on-line help
>> localh> SNEDT1: Plotting:
>> localh> SNEDT1: Press buttons A, B, or C to choose an operation
>> localh> SNEDT1: Press button D for on-line help
>> localh> SNEDT1: Plotting:
>> localh> SNEDT1: Press buttons A, B, or C to choose an operation
>> localh> SNEDT1: Press button D for on-line help
>> localh> SNEDT1: Plotting:
>> localh> SNEDT1: Press buttons A, B, or C to choose an operation
>> localh> SNEDT1: Press button D for on-line help
>> localh> SNEDT1: Computing difference data, then plotting
>> localh> SNEDT1: Press buttons A, B, or C to choose an operation
>> localh> SNEDT1: Press button D for on-line help
>> localh> SNEDT1: No data changes to write back to disk
>> localh> SNEDT1: REQUESTED ROW **** EXCEEDS MAX OF 57
>> localh> SNEDT1: ERROR FOR TABLE OBJECT: Output SN/CL/TY/SY table to be e
>> localh> SNEDT1: EDGTSN: ERROR READING REC 2 ROW****** COL WEIGHT 1
>> localh> SNEDT1: EDITSN: ERROR WITH OBJECT Output SN/CL/TY/SY table to be e
>> localh> SNEDT1: ERROR EDITING Output SN/CL/TY/SY table to be e
>> localh> SNEDT1: Purports to die of UNNATURAL causes
>> localh> SNEDT1: marathon 31DEC10 TST: Cpu= 2.9 Real= 17 IO=
>> 2
>>
>> This is running 31DEC10 which is MNJ'd every night. I don't have any
>> other versions to test, but I had not seen this bug before a month or two
>> ago. Initially I thought it was related to some external software I was
>> using, but now I've seen I can reproduce the bug in AIPS directly (change
>> displayed quantity, change antennas, exit without any edits).
>>
>>
> I cannot make it do this. It is clear that it rereads the table at the end
> but your messages suggest that the info in core is messed up before
> that happens and I cannot get it to do that with any of the trials
> I did. Is there some way you could do a TASAV and make a FITS file which I
> could read from somewhere? There aren't a lot of adverbs but it might help
> to know what they were also.
>
> Eric
>
>
--
!=============================================================!
Dr. Adam Deller Ph +1 575 418 0121
NRAO Jansky Fellow Fax +1 510 642 3411
Astronomy Department, 601 Campbell Hall, UC Berkeley
Berkeley, CA 94720-3411 USA
!=============================================================!
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listmgr.nrao.edu/pipermail/daip/attachments/20101203/06c27dbc/attachment.html>
More information about the Daip
mailing list