[AIPS] Patches to AIPS 31DEC06 [NRAO]

This page most recently updated on $Date: 2007/11/07 01:34:59 $ (UT)

You will find patches to known bugs in 31DEC06 here.

With previous releases, we did not change the frozen release tar ball; user sites were required to download the text files and do any compilations themselves. But, beginning with 31DEC04, we provide binary releases for which user sites are not required to have compilers. Therefore, we have reversed our policy. These patches have been applied to the official code of 31DEC06 and may be downloaded using a 31DEC06 "Midnight Job". Any tarball taken after the change date listed below will already have the patches applied. Of course, you may still download and compile the files as before.

The following is a list of pointers to the different patches in this document.

  1. CALIB handled scan times badly when averaging within a scan 2007-01-02
  2. UPDCONTROL in the MNJ used obsolete syntax for sort2007-01-10
  3. UVFIX did not contain the latest leap second 2007-02-27
  4. CLCOR operation SUND did not work 2007-02-27
  5. VBGLU did not handle the PC table STATE column correctly 2007-02-27
  6. CVEL had a bad call sequence which could cause an abort 2007-02-27
  7. SNSMO had a bad call sequence which could cause bad smoothing of rates including flagging them 2007-04-26
  8. UVFIX had a frequency error for uncompressed data only and CLCOR had a minor conceptual error both affecting phases after a position shift 2007-04-26
  9. BLCAL and UVFND set the integration time to one day, causing bad amplitude calibration when there were rates and delays 2007-04-26
  10. aips.l man page was lost 2007-05-04
  11. FILLM skipped a record at ends of file which could lose a data sample and possibly cause confusion if the mode changed 2007-05-24
  12. SPLIT lost the calibration flags when a source was not found so that later data did not have calibration applied 2007-06-10
  13. SNPLT lost data from phase plots of PC tables due to failure to check for wraps and got hour angles wrong by 6 hours 2007-06-14
  14. FILLM used a blank in the middle of some station names, confusing other software packages. 2007-06-16
  15. BLAVG scaled by the sum of the weights once too often, giving bad output amplitudes very dependent on whether the weights were calibrated 2007-09-05
  16. FILLM used 2 different defaults for a nominal sensitivity recorded as 0.0. This led to bad weights. 2007-09-28
  17. SPLAT set the wrong IF frequencies whenever channels were averaged to make a multi-channel output. SPLIT had a similar error, but only when the channel increments were different in different IFs. 2007-10-15
  18. SPLIT set the wrong frequency reference pixel when averaging channels and BCHAN was not 1 and messed up the FQ offsets when the IF increments differed. 2007-10-23
  19. SPLIT and SPLAT copied and computed on too many correlators, leading to buffer overruns and less than optimal scaling for compressed data. 2007-11-05 and 2007-11-06
  20. 2007-??-??

The remainder of this document is comprised of the patch descriptions pointed to by the above references.

Task:
CALIB
Summary: Scan times were rounded outward by nearly 5 seconds causing strange time intervals when averaging within a scan.
Files:
/31DEC06/Q/PGM/NOTST/CALIB.FOR
Fix:
  1. Fetch the file listed above and place it in the corresponding area ($QPGNOT) on your system.
  2. COMLNK $QPGNOT/CALIB.FOR
Documentor:
Eric W Greisen
Date:
2007-01-02
Task:
UPDCONTROL
Summary: used the obsolete syntax forms with sort. This would cause the PUTBCK.UPD list to be compromised, either being erased or left possibly containing removed files. PUTBCK.UPD is used only by secondary MNJs and so this error would only affect a few systems and many of those probably still allow the obsolete syntax.
Files:
/31DEC06/SYSTEM/UNIX/UPDATE/UPDCONTROL
Fix:
  1. Fetch the file listed above and place it in the corresponding area ($UPDUNIX) on your system.
  2. cp $UPDUNIX/UPDCONTROL $TST/$ARCH/UPDATE
Documentor:
Eric W Greisen
Date:
2007-01-10
Task:
UVFIX
Summary: The leap second routine LPSEC did not contain the latest (2006) leap second.
Files:
/31DEC06/APL/SUB/NOTST/LPSEC.FOR
Fix:
  1. Fetch the file listed above and place it in the corresponding area ($APLNOT) on your system.
  2. COMRPL $APLNOT/LPSEC
  3. COMLNK $APGNOT/UVFIX
Documentor:
Eric W Greisen
Date:
2007-02-23
Task:
CLCOR
Summary: OPCODE 'SUND' in CLCOR did not work correctly due to a missing branch.
Files:
/31DEC06/APL/PGM/NOTST/CLCOR.FOR
Fix:
  1. Fetch the file listed above and place it in the corresponding area ($APGNOT) on your system.
  2. COMLNK $APGNOT/CLCOR
Documentor:
Eric W Greisen
Date:
2007-02-23
Task:
VBGLU
Summary: The STATE column in PC tables was not handled correctly. Although this column is never used, its wrong structure confused PCCOR.
Files:
/31DEC06/APL/PGM/NOTST/VBGLU.FOR
Fix:
  1. Fetch the file listed above and place it in the corresponding area ($APGNOT) on your system.
  2. COMLNK $APGNOT/VBGLU
Documentor:
Eric W Greisen
Date:
2007-04-26
Task:
CVEL
Summary: had a bad call sequence at the end of normal operations (before copying extension tables) which caused an abort with some compilers.
Files:
/31DEC06/Q/PGM/NOTST/CVEL.FOR
Fix:
  1. Fetch the file listed above and place it in the corresponding area ($QPGNOT) on your system.
  2. COMLNK $QPGNOT/CVEL
Documentor:
Eric W Greisen
Date:
2007-02-23
Task:
SNSMO
Summary: The call sequence for smoothing rates had an error which caused odd smoothing of rates for the first IF and which could cause some to be flagged.
Files:
/31DEC06/APL/PGM/NOTST/SNSMO.FOR
Fix:
  1. Fetch the file listed above and place it in the corresponding area ($APGNOT) on your system.
  2. COMLNK $APGNOT/SNSMO
Documentor:
Eric W Greisen
Date:
2007-04-26
Task:
CLCOR and UVFIX
Summary: CLCOR computed the phase correction for a position shift from the new position as if it were the old position. UVFIX did not set the reference channel for uncompressed data and so made a frequency error when correcting positions.
Files:
/31DEC06/APL/PGM/NOTST/CLCOR.FOR
/31DEC06/APL/PGM/NOTST/UVFIX.FOR
Fix:
  1. Fetch the files listed above and place them in the corresponding area ($APGNOT) on your system.
  2. COMLNK $APGNOT/{CLCOR,UVFIX}
Documentor:
Eric W Greisen
Date:
2007-04-26
Task:
BLCAL and UVFND
Summary: The calibration software system has an integration time parameter which is expressed in days. These tasks assumed that its units were seconds apparently. When there is a rate or delay, such a long integration time will produce an enormous apparent amplitude correction which would always be wrong.
Files:
/31DEC06/Q/PGM/NOTST/BLCAL.FOR
/31DEC06/APL/PGM/NOTST/UVFND.FOR
Fix:
  1. Fetch the files listed above and place them in the corresponding areas ($QPGNOT and $APGNOT) on your system.
  2. COMLNK $QPGNOT/BLCAL
  3. COMLNK $APGNOT/UVFND
Documentor:
Eric W Greisen
Date:
2007-04-26
Task:
aips.l
Summary: The aips man page aips.l used to be a link to AIPS.l so that both "man aips" and "man AIPS" would work and say the same thing. Unfortunately, the Mac operating systems do not allow two files differing only in the case of their names. AIPS.l was removed, breaking the man entirely. An old copy was found and placed in the system as aips.l. "man AIPS" will still not work.
Files:
/31DEC06/SYSTEM/UNIX/manl/aips.l
Fix:
  1. Fetch the file listed above and place it in the corresponding area ($SYSMAN) on your system.
Documentor:
Eric W Greisen
Date:
2007-05-04
Task:
FILLM
Summary: issued two reads at ends of file which would drop a data sample and, if the observing mode changed, could possibly cause confusion in which output data set should be written next.
Files:
/31DEC06/APL/PGM/NOTST/FILLM.FOR
Fix:
  1. Fetch the file listed above and place it in the corresponding area ($APGNOT) on your system.
  2. COMLNK $APGNOT/FILLM
Documentor:
Eric W Greisen
Date:
2007-05-24
Task:
SPLIT
Summary: lost the calibration flags (DOCALIB, DOPOL, DOBAND, DOBL) when a source was not found under some circumstances. Sources in the list following this were not calibrated.
Files:
/31DEC06/APL/PGM/NOTST/SPLIT.FOR
Fix:
  1. Fetch the file listed above and place it in the corresponding area ($APGNOT) on your system.
  2. COMLNK $APGNOT/SPLIT
Documentor:
Eric W Greisen
Date:
2007-06-10
Task:
SNPLT
Summary: Plots of PC table phases did not check for phase wraps and so could fail to plot many of the points. The X axis hour angle parameter was in error by 6 hours.
Files:
/31DEC06/Y/PGM/NOTST/SNPLT.FOR
Fix:
  1. Fetch the file listed above and place it in the corresponding area ($YPGNOT) on your system.
  2. COMLNK $YPGNOT/SNPLT
Documentor:
Eric W Greisen
Date:
2007-06-14
Task:
FILLM
Summary: used a blank in the middle of some (old VLA) station names, confusing some other non-NRAO software packages. Changed it to use an underscore. Changed PRTAN, LOCIT, FRPLT, and POSSM to support this form of station name.
Files:
/31DEC06/APL/PGM/NOTST/FILLM.FOR
/31DEC06/APL/PGM/NOTST/PRTAN.FOR
/31DEC06/Y/PGM/NOTST/FRPLT.FOR
/31DEC06/Y/PGM/NOTST/LOCIT.FOR
/31DEC06/Y/PGM/NOTST/POSSM.FOR
Fix:
  1. Fetch the files listed above and place them in the corresponding area ($APGNOT for the first 2, $YPGNOT for next 3) on your system.
  2. COMLNK $APGNOT{FILLM,PRTAN}.FOR/
  3. COMLNK $YPGNOT{FRPLT,LOCIT,POSSM}.FOR/
Documentor:
Eric W Greisen
Date:
2007-06-16
Task:
BLAVG
Summary: The code summed individual sums scaling by the individual sum of weights a second time. This could make the output amplitudes very small and always made them depend on whether the weights were calibrated or not.
Files:
/31DEC06/APL/PGM/NOTST/BLAVG.FOR
Fix:
  1. Fetch the file listed above and place it in the corresponding area ($APGNOT) on your system.
  2. COMLNK $APGNOT/BLAVG
Documentor:
Eric W Greisen
Date:
2007-09-05
Task:
FILLM
Summary: used different defaults for weighting and scaling data when the nominal sensitivity was recorded as 0.0. This affects only post-ModComp data and causes the calibrated weights to be too high by sqrt(3) for each antenna affected.
Files:
/31DEC06/APL/PGM/NOTST/FILLM.FOR
Fix:
  1. Fetch the file listed above and place it in the corresponding area ($APGNOT) on your system.
  2. COMLNK $APGNOT/FILLM
Documentor:
Eric W Greisen
Date:
2007-09-28
Task:
SPLAT and SPLIT
Summary: Both tasks mis-computed the IF frequency for IFs greater than 1 when the tasks were producing a multi-channel output data set while averaging spectral channels. SPLIT got it wrong only when the IFs had different increments (e.g. of opposite sign). SPLAT got all IFs wrong.
Files:
/31DEC06/APL/PGM/NOTST/SPLIT.FOR
/31DEC06/APL/PGM/NOTST/SPLAT.FOR
Fix:
  1. Fetch the files listed above and place them in the corresponding area ($APGNOT) on your system.
  2. COMLNK $APGNOT/{SPLIT,SPLAT}
Documentor:
Eric W Greisen
Date:
2007-10-15
Task:
SPLIT
Summary: set the wrong frequency reference pixel when averaging spectral channels to a multi-channel output file when BCHAN was not one. In addition, if the IF increments differ, the offset frequencies in the FQ table would be wrong. The first error has persisted for a long time, the second first appeared in 2006.
Files:
/31DEC06/APL/PGM/NOTST/SPLIT.FOR
Fix:
  1. Fetch the file listed above and place it in the corresponding area ($APGNOT) on your system.
  2. COMLNK $APGNOT/SPLIT
Documentor:
Eric W Greisen
Date:
2007-10-23
Task:
SPLIT and SPLAT
Summary: Both tasks set the number of correlators 3 times too high. They then copied this number, time averaged this number, and compressed this number. The resulting buffer overruns must have been usually benign although not always. The scaling of compressed data would be based on more than real data and so be less than optimal.
Files:
/31DEC06/APL/PGM/NOTST/SPLIT.FOR
/31DEC06/APL/PGM/NOTST/SPLAT.FOR
Fix:
  1. Fetch the files listed above and place them in the corresponding area ($APGNOT) on your system.
  2. COMLNK $APGNOT/{SPLIT,SPLAT}
Documentor:
Eric W Greisen
Date:
2007-11-05, corrected 2007-11-06
Task:
Summary:
Files:
/31DEC06/??
Fix:
  1. Fetch the file listed above and place it in the corresponding area (??) on your system.
  2. COMLNK ??
Documentor:
Eric W Greisen
Date:
2007-??-??
Eric W. Greisen