[daip] 31DEC04 LINUX midnite job (dave, 20050816.230239)

[aipsmgr at aoc.nrao.edu]

dave update report for 31DEC04 on Tue Aug 16 23:02:53 UTC 2005 (UT)

Copies and compiles:

AIPSUPD - Using CVS for copying files.
UPDCVS - cvs update of 31DEC04 completed
UPDCVS - cvs update of TEXT completed
AIPSUPD - UPDCVS apparently worked

UPDCONTROL - /home/AIPS/31DEC04/LINUX/UPDATE/LASTREMOVE.DAT = 20050422.190608
UPDCONTROL - /home/AIPS/31DEC04/LINUX/UPDATE/LASTCOPY.DAT = 20050422.190608
UPDCONTROL - /home/AIPS/31DEC04/LINUX/UPDATE/LASTCOMRPL.DAT = 20050422.190608
UPDCONTROL - /home/AIPS/31DEC04/LINUX/UPDATE/LASTCOMLNK.DAT = 20050422.190608

UPDCONTROL - Everything seems in order

UPDCOMRPL: COMRPLs


UPDCOMRPL: currently in /home/AIPS/31DEC04/LINUX/UPDATE

UPDCOMLNK: COMLNKs


UPDCOMLNK: currently in /home/AIPS/31DEC04/LINUX/UPDATE


UPDCONTROL - Reasons for file updates on Tue Aug 16 23:02:52 UTC 2005 (UT)

16-AUG-2005 19:20:36 20050816 NEW APGNOT   BOXES.FOR     move to patch
16-AUG-2005 19:20:37 20050816 NEW QPGNOT   CALIB.FOR     move to patch
16-AUG-2005 19:20:39 20050816 NEW APGNOT   FITLD.FOR     move to patch
16-AUG-2005 19:28:09 20050816 NEW HIST     CHANGE.DOC    patches

CHANGE.DOC entries:


11842.  May 3, 2005                 BOXES                Eric
        This task, like FIXBX, could not handle really simple cases
        although it did the fancy BOXFILEs really well.  Added minor
        fixes to let it do simple BOXFILEs and single fields.
        Moved to 31DEC04 16 August, nowhere else..

11868.  June 9, 2005               CALIB                  Eric
        When CALIB created a new flag table (DOFLAG > 0) and a new SN
        table, the latter seemed to be forgotten.  This is probably
        because the task used two in-memory copies of the image
        header.  One was referenced for the SN and the other for the
        FG.  Changed to use the same one for both.  This has been
        working fine for single-source files but not for
        multi-source.
        Moved this fix only to 31DEC04 16 August, nowhere else.

11900.  August 4, 2005             FITLD                   Eric
        Phil code: Two JD dates obtained by dufferent means were
        differenced and the difference was required to be >= 0.  A
        difference of -1.0D-6 caused the task to skip the data file!
        To add more injury, more than one subroutine did not
        initialize the version number before calling the ext. table
        INI routine.  That was fine on the first call and if there is
        only 1 version but not if there are multiple subarrays or
        whatever.  But this is the first 2005 fix of FITLD!!!!!
        Moved to 31DEC04 16 August, nowhere else.

11907.  August 11-12, 2005          FITLD                  Eric
        Changed:
        (1) removed a large set of unused FORMATs.
        (2) Added code to recognize station Y with gain 1.0 as needing
        the Tant set to -1.0 rather than magic blank.  This tells
        later code that the Tsys is really Tsys/Tant (values less than
        1 rather than in 2 or 3 figures actually).
        (3) 8/12 changed to test sensitivity rather than gain curve
        since that is 1.0 frequently.
        Moved to 31DEC04 16 August, nowhere else.