;----------------------------------------------------------------------- ; Copyright (C) 2011-2013 ; Associated Universities, Inc. Washington DC, USA. ; ; This program is free software; you can redistribute it and/or ; modify it under the terms of the GNU General Public License as ; published by the Free Software Foundation; either version 2 of ; the License, or (at your option) any later version. ; ; This program is distributed in the hope that it will be useful, ; but WITHOUT ANY WARRANTY; without even the implied warranty of ; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ; GNU General Public License for more details. ; ; You should have received a copy of the GNU General Public ; License along with this program; if not, write to the Free ; Software Foundation, Inc., 675 Massachusetts Ave, Cambridge, ; MA 02139, USA. ; ; Correspondence concerning AIPS should be addressed as follows: ; Internet email: aipsmail@nrao.edu. ; Postal address: AIPS Project Office ; National Radio Astronomy Observatory ; 520 Edgemont Road ; Charlottesville, VA 22903-2475 USA ----------------------------------------------------------------------- This file must be used to record any changes made to the Master version of AIPS kept in SOCORRO. An entry is: line 1: Date Program name Your name line 2: What was done (may span multiple lines) line n-1: Where has this change been moved (e.g. 15JUL94, nowhere) line n: Blank The details of line 1 are important since programs are used to find specified strings and do conversions on them. It MUST be typed as follows: 1 blank, 4 digits, a period, two blanks, then the Month, day, comma, year, some space, a title, some more space, your name. *********************************************************************** *********************************************************************** Changes while 31DEC12 is TST *********************************************************************** *********************************************************************** 13719. December 12, 2011 code rollover Eric ZPBUFSZ small buffers 3 times larger, large ones 6 times larger PMAD buffers 4 times larger Change release date in GETRLS, INSTEP1, install.pl, update.pl, ZXHLP2.PL LaCook.PL, UPDXCONTROL, AIPSPATH.CSH. AIPSPATH.SH. Moved nowhere. 13720. December 12, 2011 help files Eric UVFIT: correct limits on visibilities and parameters, fix some bad grammar REWAY: correct minor typos and things easily mis-read SNIFS needed details for SNIFS rather than SNPLT in the summary sections. Moved to 31DEC11 this date. 13721. December 12, 2011 TV issue Eric Some plots try to finish on the TV when they never opened in the first place. Changed ZDCHIN to set TVBROK = -1, TVOPEN to set TVBROK = max (0, TVBROK), and TVCLOS to set to -1 if 0. The tasks would hang on IO to the TV trying to remove any HOLD condition - now the Z routine that does the IO will not attempt it if the TV is not open. Moved nowhere. 13722. December 14, 2011 APCAL Eric APCAL did not check if the TY antenna record had SUBARRAY <= 0 which is supposed to mean "any". The FQ tests were also in error on a similar point. Moved to 31DEC11 this date. 13723. December 14, 2011 FARS Eric Changed HI writing to explain all 9 of the APARM parameters. Changed handling of some as well - there were cases that would escape the code. Moved to 31DEC11 this date. 13724. December 14, 2011 REWAY Eric Added BADDISK Moved to 31DEC11 this date. 13725. December 16, 2011 CLIP Eric The history file using a generalized CALHIS recorded only the last source considered. It was prepared to do otherwise but did not. Took advantage of the preparation to prepare the source names and numbers for the history routine. Moved to 31DEC11 this date. 13726. December 19, 2011 SPLIT, SPLAT Eric Changed code to supress messages on UVGET CLOSe calls but to do the close calls even when no data were found. The FG table was opened and left open although other files were not open - a close is needed but it is best to avoid the file not open in FTAB messages. Also moved the RNXCLS to make sure that an NX table is closed even on error. Blocked continuing on the next source. Moved to 31DEC11 this date. 13727. December 20, 2011 FRING, KRING, UVFLG Olaf Wucknitz/Eric Using LOFAR data, some limits have been found that are incorrect. Changed the history writing in FRING and KRING to handle ANTWT is a general fashion. Changed the baseline limits in UVFLG to MXBASE (= 90*91/2.) rather than 500. Moved to 31DEC11 this date.. 13728. December 20, 2011 SLICE Eric Added X, Y, Z physical coordinates to the output text file. Allow multiple outputs to go to the same text file and added a header line at the start. Moved to 31DEC11 this date. 13729. December 20, 2011 SQASH Eric Changed to allow it to sum even axis 2, producing a 1-row image if summing all axes >= 2 or an image with 1 point on the 2nd axis otherwise. A trivial TRANS may be needed on the latter to make it more visible. Moved to 31DEC11 this date. 13730. December 21, 2011 DATCAL, DSMEAR, FXSEG Amy For VLBA data UVGET (through DATCAL) attempts to apply decorrelation corrections caused by averaging in the correlator. The info needed to compute this correction is in the CQ table. FXSEG, the subroutine that determined the segmetation loss factor only worked for FFT sizes of 2048 and smaller and the warning messages from DATCAL and DSMEAR about the FXSEG failure were incorrect and confusing. Changed: FXSEG Fixed so that will work for any FFT size with UNIFORM taper. Still a problem with HANNING taper. DSMEAR When FXSEG returns an error code prints a message that segmetation loss factor cannot be calulated. Also added an error message if there is some other problem CQ table. DATCAL Clarified warning message, it now says there is a problem with the CQ table (rather than there is no CQ table) and therefor no decorrelation corrections will be applied. Moved nowhere. 13731. December 23, 2011 LISTR Eric Changed to do SY tables (Psys and Pgain) in the GAIN listing routines (2 formats). Required correcting the handling of new scans. Added 0.1s of seconds to the times as well. Moved nowhere. 13732. December 27, 2011 SLICE Eric Made a bad array reference - the subscript got set to 0 by the coordinate call - so the text file outputs had zero for slice value. MOved to 31DEC11 this date. 13733. December 27, 2011 IM2CC Eric Changed to $APGNOT so that TABSRT can be called to sort the CC tables into descending order by flux. Moved nowhere. 13734. December 30, 2011 AFARS Leonia/Eric A small problem has existed near BLC, TRC. It has been fixed. The fitting of parabola to determine position of maximum (with beter accuracy) is not carried out now if maximum is too close to the edges of data. Changed coordinate computation to use header parameters rather than make assumptions about the coordinates. Moved to 31DEC11 on Jan 2. 13735. December 31, 2011 ANFIND Eric Added ARRAY 'NORH' for Nobeyama Radioheliograph. In time 31DEC12 tasks will recognize this. Moved nowhere. 13736. December 31, 2011 EVLA.ant.2011 Eric Moved 2011 EVLA antenna location corrections to $AIPSTARS. Set up analysts for 2012. Moved nowhere. 13737. December 31, 2011 SPCOR Eric NEW TASK: does PBCOR operation plus spectral index correction to data cubes. This may be needed especially for inputs to FARS. Moved nowhere. 13738. January 2, 2012 UNIXSERVERS Eric Removed debug message inadvertently left in Moved to 31DEC11 this date. 13739. January 2, 2012 IM2CC Eric Changed it to center each piece of the input image in an output image that is a power of 2 pixels on a side. Moved to 31DEC11 this date. 13740. January 3, 2012 FARS Eric FARS had the meaning of APARM(4) backwards for the HI file. It also did wierd things (nothing???) when APARM(4) > 1.5 - a holdover from when FARS had an AFARS kind of output. Moved to 31DEC11 this date. 13741. January 3, 2012 Imaging Eric Changed GRDAT, GRDSET, SETGRD to insist on powers of 2, rounding up as needed. The first allowed images up to and including 8192 to be doublesd in gridded subtraction, while the second limited them to 4096 and the third to 2048. Fixed the latter two to allow 8192 doubling. Changed SETDO3 to understand DO3D false imaging - it forbad multiple facets in 3D false. Also have it set FLDSZ which needs defaulting sooner than later. OOSUB was failing badly. Moved to patch January 16. 13742. January 4, 2012 OOCAL Eric New RUN file and procedure to do OOSUB to apply channel-dependent primary beam and spectral index options followed by CALIB self-cal with a TACOP to return the SN table to the original input file. Changed POPSDAT and NEWPARMS to create a new adverb DOKEEP. Moved nowhere. 13743. January 4, 2012 FARS Leonia FARS has had several mistakes near conversion RE,IM to AMP/PHASE especially when RESIDUAL sent to the output Moved to patch January 16. 13744. January 4, 2012 Imaging Eric Added progress messages to VISDFT, ALGMEM, ALGSTB. Moved to patch January 16. 13745. January 5, 2012 FARS Leonia Add shifting residual phase to the initial lambda^2 set. Moved to patch January 16. 13746. January 5, 2012 FARS Leonia Misprint at the previous installation Moved to patch January 16. 13747. January 6, 2012 Miscellaneous Eric DOFARS Procedure set default OUTCLASSes ignoring the amplitude/phase versus Real/Imaginary option. Fixed to be FARSam/FARSph or FARSre/FARSim and TRANam/TRANph or TRANre/TRANim depending on APARM(5). AFARS Help file: Added sentence about axis order - the input images are from FARS (which it already said) not the tranposed images from the DOFARS procedure. FITLD Help file: amplified the statements about FITS IDI only adverbs and re-ordered them so that all adverbs after some point are IDI only FITLD Fortran: reordered the adverbs expected to match the help file change. TGET will be a mess. Moved nowhere, the FARS ones maybe should go to patch. 13748. January 6, 2012 FITAB Eric Did not write the TNULL keyword for compressed uv data. It is -32767 rather than the obvious -32768 so it is important. Moved to patch January 16. 13749. January 6, 2012 SQASH Eric A WRITE statement for a history line had too many arguments for its format. Dropped the unwanted argument. Moved to patch January 16. 13750. January 7, 2012 UVFIT Eric Add option to have a comment string following the parameters in the INLIST file. This comment appears at the end of the line in the FITOUT file. Moved to patch January 16. 13751. January 9, 2012 Lustre issues Eric Lustre file systems lie about file sizes - not just saying 0 but syaing real sizes that are wrong. Added code to ZMI2 and ZFI2 so that, when a request exceeds the stated file size, the routine issues a message, delays 1 second, asks the system again about the file size. If that also fails, then it does the whole message/delay/ask over again. If it still fails,a fatal task error is generated and the task dies. Moved nowhere. 13752. January 9, 2012 AFARS Eric Found several errors in the code: (1) If the quadratic maximum fit failed or was too close to the edge, the BLC(1) correction was not added. (2) If the fit was once too close to the edge, it no longer did the quadratic fit at all (IRET never reset to 0). (3) If the input row was all zero or blank, the task used a pixel value that was not properly tested. Now it will use the blanking value. Moved to patch January 16. 13753. January 9, 2012 UVFIT Eric Found lots of places still dimensioned 4 rather than XMXCMP (20) so the commputation of lots of things should have gone into the weeds and did when there were > 4 model components. Adde a couple of spaces to the output file format and changed how time is handled so the actual minimum and maximum times used are what is reported. Moved to patch January 16. 13754. January 16, 2012 CLCOR Eric The GMRT uses a backwards phase convention to the rest of us. Changed a few places - esp the place that moves antenna and source positions - to reverse sign for GMRT. Moved to patch this date. 13755. January 16, 2012 CLCOR Eric Tediously removed all of the horribly non-standard illegal Fortran from the code. Corrected bug put in previously today. Moved to patch this date. 13756. January 19, 2012 UVFIT Eric Changed it to allow up to 60 components in a fit. A user with that many had better have really good data. Moved nowhere. 13757. January 20, 2012 TARS Leonia The new task is to test the Faraday rotation synthesis task FARS. The input file includes 3 columns: frequencies, Q, and U. The result is sent to the output file. The options are described by the same inputs used by FARS. Moved nowhere. 13758. February 21, 2012 Double precision AP Eric When INTEGERs count, they overflow at 2^31-1. When REALs count they cease to increment after 8 million counts or so. Our data can exceed this now. Computational inaccuracy also leads to errors, especially for example in the corners of FFTed images where the gridding function correction is large. It is important to get right answers not overflowed or undercounted ones and it is nice to have more accurate answers as well. So we are switching to double precision in the pseudo AP and making other corrections. Changed DAPC.INC Changed REAL to DOUBLE PRECISION and INTEGER to INTEGER*8. Our 32-bit systems (LINUX, SUL) handle this correctly although it is not legal. Q1GRD Local variables INTEGER = APCORI(..) to call another Q routine QGRD2 Changed INTEGER to LONGINT for numerous pointers! Also REAL -> DOUBLE QINIT Convert PSAPOF to a double precision index - ask for twice as many words QMAXV Corrected answer to remove offset to dynamic memory. Subscript was not used I guess. QMINV Corrected answer to remove offset to dynamic memory. Subscript was not used I guess. QNGRD Use local INTEGER = APCORI(..) variables to call another Q routine, REAL -> DOUBLE QUVIN use local INTEGER = APCORI(..) variables to call another Q routine, REAL -> DOUBLE QXFOUR Data to FFT changes from REAL to DOUBLE UVTBUW Fix call to QGET to ask for integer when integer is needed. UVWAIT REAL -> DOUBLE, drop integer count AUC Verb SETMAXAP - set KAPWRD to half as much per Mbyte. SETMAXAP Help file - add info about using 8 bytes per word ARRAY New read/write images to double array APUTIL For images, use new ARDRED and ARDWRI routines. For UV data read/write to local variables and then copy to/from AP. QCLEAN Direct uses of APCORE: changed CC one to read flux into local variable and then scale to APCORE. Changed the other to use ARDRED and ARDWRI. DEFLG Uses APCORE - changed to a loop to load the AP memory and some doubles to compute on it SDIMG Convert to DFILL with 0.0D0. UVCON Convert to double precision in local QXXPTS CXCLN Correct QGET call to ask for integer - equivalence tricks are not allowed! Changed it to disk I/O to a local variable and then use QPUT and to use QGET to a local variable and then disk I/O from there. Direct reading to the AP is not allowed. Simple changes of variables from REAL to DOUBLE: Q1FIN QBOXSU QCGAUS QCLNSU QCSQTR QCVEXP QCVMMA QCVMUL QCVSDI QCVSMA QCVSMS QDFT QFINGR QGADIV QGASUB QGRD1 QGRD3 QGRD4 QGRD5 QGRD6 QGRD7 QGRD8 QGRDMI QGRIDA QHIST QINTG QINTP QLVGT QMAKMS QMAXMI QPHSRO QPOLAR QPRJM QPTADC QPTDIV QRECT1 QRECT QRFFT QRFT QSPDIV QSPSUB QSVE QSVESQ QTAPER QUCONJ QVCLIP QVCLR QVFILL QVFIX QVFLT QVIDIV QVRVRS QVSADD QVSMA QVSMAF QVSMSA QVSMUL QVSWAP QVTRAN QVTSMU QWTCNT QWTGRD QWTSUM QXXPTS Change of comments - they had real errors in QPUT, QSORT. Also changes in typing QCRVMU QGET QGETIN QMULCL QPUT QPUTIN QSORT Moved nowhere. 13759. February 21, 2012 ZMEMRY Eric Changed the NWORDS argument to KWORDS - number of kilowords (kilobytes = 4 * KWORDS) in ZMEMRY. Changed calls to ZMEMRY: BPASET CALADJ POLSET TABSRT INTERPLATE TVUTIL QINIT QCLEAN QEDIUTIL ACFIT ACLIP BPSMO CLIP CORER CUBIT EVASN FGDIF FLGIT GETJY HGEOM LGEOM MSORT OTFUV PGEOM REFLG REWAY RLDIF SNFLG SPLAT STFUN TRUEP UVAVG UVHIM UVLIN UVMLN UVMTH UVRFI VBGLU XTRAN BLAPP CCEDT CCRES FILIT FIXWT FXPOL SABOX RFARS SDMOD SPCOR SQASH TRANS BLCAL BLCHN BPASS CALIB CVEL DEFLG FRING KRING PCAL RLDLY BLING BSAVG IMAGR OMFIT SCIMG WFCLN EVAUV BSCLN IMEAN RSPEC ANBPL DELZN ELINT IBLED OBPLT RFLAG SNIFS VPLOT WIPER Moved nowhere. 13760. February 21, 2012 Dispersive delay Eric There is no dispersive delay in the SN table, just the CL table. GAININ correctly deals with this but CSLGET failed to test for the pointer being < 0. Corrected to test pointers for ionosphere and dispersion before setting the variables from the data records. Moved nowhere. 13761. February 21, 2012 Source table Eric Added observed RA and Dec Changed includes: PUVD.INC Change number SU columns from 19 to 21. DSOU.INC Add RAOBS, DECOBS DSOV.INC Add RAOBS and DECOBS column pointers and variables, comment about the limits in row size Changed subroutines: SUINI New routine is SOUINI with added columns TABSOU Added RAOBS, DECOBS to call seq, added pointers for them and moved pointers for columns that follow. SOUINI Changed to call SUREFM and the real SUINI. SUREFM Reformats the SU table, taking the epoch positions to be the pointing positions SOINI Added new columns SOTAB Add new columns with EPO values if OBS columns missing. DGHEAD Change call sequence to TABSOU, use the observed coordinates to go into the header rather than epoch GETSOU Change call sequence to TABSOU, fill common from that or from header in single source files TABLEUTIL Changed call sequence to TABSOU and to OTABSU. Subroutines changed call sequence to TABSOU only ACTFRQ COPTAB FACSET FNDSOU SOUFIL SOUFND SOURNU SUSEL Tasks changed ATLOD Changed call to TABSOU, set RAOBS/DECOBS to epoch ones. FITLD Changed numerous calls to TABSOU. Pick up pointing position from SO table with IDI data. MATCH Fixed calls to TABSOU, made pointing as well as phase stopping positions match MULTI Fixed call sequences, get pointing from header UV2MS Fixed call sequences, added code to get pointing position from single-source header Tasks changed for call sequence only ACLIP ATMCA AVSPC CLCOR CLIP DBCON DFCOR DQUAL DTSIM EVASN FGDIF FILLM FILLR FIXAL FLOPM GETJY HAFIX LISTR M3TAR MK3IN NOIFS PRTUV QUOUT REFLG RLDIF SETJY SPLAT SPLIT SUFIX TI2HA TRUEP UVFIX UVFLG UVLOD UVMLN VLBIN BLAPP COHER DTSUM FXPOL SPECR BLCAL BLCHN BPASS CALIB CPASS CVEL DEFLG FRING LPCAL PCAL RLCAL OMFIT OMFIT DELZN ELINT FRMAP POSSM SOUSP Moved nowhere. 13762. February 21, 2012 SETGDS Eric SETGDS can loop to look over the facets twice if there is some issue of geometry. It was resetting the CCVER array to the input value which meant 0 in most cases. The image header that is used is not that of the facet when multiple facets are being gridded so a 0 can get the wrong thing easily. Moved nowhere. 13763. February 21, 2012 FILLM, XAS Eric Bryan Butler poited out errors in the list of frequencies - the exponent was D0 when it should have been D9 for two of the new EVLA bands in FILLM. Martin Hardcastle pointed out the need to set AF_INET for the server family in XAS.SHR subroutine comm.c. It caused certain Linux kernals to fail to run XAS in the Inet (default) mode. Moved nowhere. 13764. February 21, 2012 SU table some more Eric TABSOU used hard-coded column numbers in filling the arrays and some off them changed. Numerous routines used 19 for the number of SU columns rather than the parameter MAXSUC. Changed them to use the parameter: ACLIP, BPASS, CLIP, CPASS, DEFLG, DFCOR, FILLR, LISTR, SETJY, SPLAT, SPLIT, UVMLN, FACSET, FNDSOU. FILLR required MAXIF local variable to be renamed. Moved nowhere. 13765. February 22, 2012 TKSRV Eric The logic of include files in ZTKSRV.C differs from that in ZMSSRV and ZTVSRV - and a gfortran had troubles. Changed it to match. Moved nowhere. 13766. February 22, 2012 TVFLG, SPFLG Eric These tasks used LUN 29 for the output FG table and in some cases that LUN was still in use by something else. Changed to set the LUN with LUNTMP and so never have this conflict. Moved nowhere. 13767. February 22, 2012 ELINT Leonia Rick Perley asked to add to the plot headers the normalized RMS of residuals and the RMS of the pointing offset which is related with the RMS of residuals. Rick has provided the equation of the relation. It's done. Moved nowhere. 13768. February 22, 2012 ORIGIN Eric This keyword has a standard meaning in FITS and so should not be used in a table for another purpose. It is in SN tables and I changed it to SNORIGIN. Changed CLCAL, ACCOR, UVCRS, ACFIT, PCCOR, FRING, CALIB, KRING. Moved nowhere. 13769. February 22, 2012 Imaging Eric SETGDS used FLDSZ to decide whether to force DFT before it was set and so did not force it in some calibrator model cases. Reduced the progress messages from every 50000 to every 200000 in VISDFT, ALGMEM, and ALGSTB. Moved nowhere. 13770. February 23, 2012 ELINT Eric Changed: 1. Loop reading the OT table used the record number as the loop variable. In general, this meant only every other line was read since the read routine ups the record number variable. Changed to control things better. Added error handling because in one compiler's version it tried to read the right number of records but with record numbers twice too large. 2. The source finding routine fill a variable with only 30 values with all of the source names in the SU table when SOURCES or CALSOUR was not specified. This blasted pieces of other commons that control file access, messages, etc. Raised MAXSOU to 500 because of this routine. 3. Moved MAXFIT to a separate local include so that it could be put in ELFIT where local variables were of size 10 and needed to be MAXFIT (200) or MAXFIT*MAXFIT. 4. Added a comma to a FORMAT statement. 5. Changed LEASQR to allow for bigger problems - its limit was 50. Moved nowhere. 13771. February 23, 2012 FITLD Eric SOIO was not recompiled when it had to be making FITLD mess up some parameters in the SO table. The output data were fine. Moved nowhere. 13772. February 24, 2012 ZMEMRY Eric Several routines used the ZMEMRY call argument after the call to ZMEMRY and did not have the units properly restored. They included: BPASET INTERPLATE OTFUV UVRFI FIXWT SNIFS VPLOT UVMLN CVEL Moved nowhere. 13773. February 27, 2012 COMB Eric The actual OUTSEQ used was picked up from the wrong header after the MCREAT call and so ended up in the HIstory file with the INSEQ value. Moved nowhere. 13774. February 27, 2012 IMAGR Eric Created QDGET and QDPUT to transfer double precision values from/to pseudo-AP memory. Changed UVWAIT to compute all things in double precision and use QDGET and QDPUT for things as well. It was getting a few round-off failures that caused it to fail to get the sum of weights in some cells correctly. Also added code to handle such edge effects if they arise - but they went away in my data with the double precision change. Moved nowhere. 13775. February 27, 2012 POSSM Eric Added another counter through both plots with data and ones without. This lets us know when we are nearly done with NPLOT > 1 plots. Changed labeling indicators to get numeric labels with frequency and velocity in all rows of the plots and characters (e.g. FREQ MHZ) only in the bottom plot of each column. Moved nowhere. 13776. March 1, 2012 MULTI Eric Added auto creation of an NX table. Moved nowhere. 13777. March 1, 2012 VLANT Eric The code reversed the sign of the Y correction for EVLA for the delay computations. It also neglected to shift the hour angles from local to Greenwich which it must do for the EVLA. I guess these errors compensated in tests in 2009. Moved to patch this date. 13778. March 3, 2012 SNREF Eric NEW TASK: Takes an SN or CL table and tries to see which reference antenna minimizes the rms in the R-L phases. Moved nowhere. 13779. March 3, 2012 LISTR Eric Added DPARM(9) which if > 0 converts all times to LSTs. The labeling tries hard to make this clear. This works in every OPTYPE. Moved nowhere. 13780. March 6, 2012 SCMAP Eric Changes to Fortran and help to 1. Add BOXFILE to define initial Clean boxes 2. IM2PARM to allow auto-boxing including with TELL. 3. Changed APARM(8) to allow negative values to permit the use of negative summed Clean components in the model and in the restart of the next Clean. 4. Also put in code to insist that the input file be a single-source file. 5. Previously, uncompressed input files were fed directly into UVMDIV for model division. This meant that flag tables were not applied. Compressed data were copied first to make them uncompressed but also to apply flags. Changed two places to do this for both compressed and uncompressed data. Moved fix 5 to patch March 7. 13781. March 7, 2012 SCIMG Eric Made the same changes to SCIMG as to SCMAP except that BOXFILE was already in SCIMG. Moved fix 5 to patch this date. 13782. March 8, 2012 CLIP Eric Corrected formats in TABIO to encompass the full range of I*4 numbers (I12 with a space in front). Changed CLIP to issue progress messages every 100000 vis rather than 10 times more often. Moved nowhere. 13783. March 11, 2012 SETJY, CVEL Eric Changed: FRQUPD To return error code -1 if sample for the current time not found DAYNUM To use the arrays of numbers without erasing one in leap years. Two calls to the routine, the second after the first was in a leap year, would get the wrong answer in non leap years. It will work now. CVLDOP Had it set the year and day number if they were 0 on input. Changed to use ANT+1 as the antenna number in the arrays. Previously it allowed ANT=0 but did not handle times correctly in that case. That case is used in CVEL for the VLBA and EVLA. SETJY Help file: add OPTYPE='VCAL' to put velocities of the first observation into the SU table. Increased details of when adverbs are used and when they are not. SETJY Changed internal spelling of some variables to allow use of standard commons. Made velocity type and dwefinition be specified correctly including defaults. Add code to do VCAL - subroutine GETV gets antenna array info, gets the time of the first scan on each source from the NX table, calls FRQUPD and CVLDOP to get frequencies and Doppler offset. LISTR It used VELDEF not VELTYP for 'BARYCENTR'. Moved nowhere. 13784. March 12, 2012 TARS Leonia The model of rotation measure components are added into the input file. The model rows are labeled by the symbol "M" at the first position. Number of model components is determined by the relevant nuber of the model rows at the input file and is not limited. Moved nowhere. 13785. March 12, 2012 UVPRT Eric Changed to have DPARM(6) > 0 indicate that phases should only use 3 of the 4 spaces allocated in the format and thus range from -90 to 270. This will aid parsing the output. Moved nowhere. 13786. March 12, 2012 DOPOL Eric The ORI-ELP mode of PCAL requires access to the antenna file for phase difference information. POLSET was supposed to provode a 512-word buffer for PDRGET but gave it a 16-word local variable instead = causing stuff to get walked upon. Moved nowhere (so far - patch??). 13787. March 13, 2012 OMFIT Eric Changed with help from Michael Bietenholz: KEYIN The temporary storage for symbols wa 8 characters - not enough for an 8 character string plus quotes or for long numeric parameters either. Changed it to 16 internally and took steps to insure that the end mark test is also 16 (w a local variable). Put in a test for excessively long symbols with a message and fixed the formats from A8 to A. OMFIT Help file: (a) changed comment about Ketan. (b) Clarified the OUTFILE info. (c) Added option WEIGHTIT. (d) Corrected explanation of the 2nd parameter of the INFILE cards (it is source number for multi-source files as well as an identification number. (e) Add minimal documentation of spherical shell model type. OMFIT (a) Add WEIGHTIT adverb and apply it to data weights. (b) Changed OUTPRINT to append to exsiting files. (c) Dropped or changed version info (was 15OCT98) (d) In reading INFILE, initialized a variable so that checking of source IDs does not happen in single-source files. Otherwise if BCHAN > 0, single-source files could not have worked. (e) Add message when no model components are found. (f) Phase-only self-cal needed normalization added by MFB. (g) Add display of corrected "real" Chi-squared and reduced chi squared per MFB. (h) Increased error estimates by sqrt(2) per MFB. (i) Dropped warning about sources for single-source files. (j) In source reading routine added check for single-source files to avoid all the SU table does not exist messages. (k) MFB added sphere models. (l) Converted 3 subroutines written in lower case letters to the required upper case. The code is a mess in some ways but seems to work (i.e it writes directly to text files ignoring ZTXIO) so I am going to put it back for now so people can try it. Moved nowhere. 13788. March 14, 2012 SUREFM Eric SUREFM would clear only one READ status on the input file and then try to put a WRIT status on it. Changed it to loop clearing as many READ statuses as there are and then put the READ status on. Moved nowhere. 13789. March 14, 2012 NAMEGET Eric New verb - fills in INNAME, INCLASS, INSEQ, INTYPE, INDISK to match all wildcards and defaults, starting with whatever is in those adverbs to begin with. Changed AU8, POPSDAT, NEWPARMS. Moved nowhere. 13790. March 15, 2012 SETJY Eric The code put back a couple of days ago initially used RA and Dec values in degrees where radians were expected. The change to fix that caused all coordinates in the SU table to divided by 180/pi for each execution. The whole thing is corrected for now but we do need to think about whether the EVLA velocity is wrt center of the Earth or each antenna. Moved nowhere - early MNJ prep. 13791. March 15, 2012 IMEAN Eric Changed histogram fit fails message - happens when histogram has a bad shape. Moved nowhere. 13792. March 15, 2012 TVFLG, SPFLG Eric When a multi-source file was read into these tasks for a single SOURCE, the source number was lost. This made it flag all sources no matter what the user requested. Some work seemed to guarantee that the source number made it to the work file, but it still seemed to be zero reading it back. Added more code to cover that issue too. In TVFLG, corrected a STRING length that was at least 1 too short giving ? on the TV screen. Moved nowhere. 13793. March 16, 2012 SETJY Eric Changed it to avoid using COMMON except where necessary. The version of yesterday should work but this is safer. Moved nowhere. 13794. March 16, 2012 FILLM Eric Changed to display zenith opacity at IF 1 (plus K band) in the new opacity method. Forced change to KBOPAC to return the zenith opacities (IF) and to INDXR to deal with the changed call sequence. Moved nowhere. 13795. March 16, 2012 OMFIT Eric Changed: 1. "Real CHI2" to be scaled by actual variance when NOISE(1) is not specified. 2. Use G format to do CHI2 displays since weights are dubious in many cases. 3. Label Ketan's "scaled" CHI2 as such and display them only when PRTLEV >= 4. Display "Real CHI2" when PRTLEV >= 2. 4. Remove all direct writes to files and (*) = the terminal, replacing them with calls to ZTXIO and MSGWRT. The subroutine SGET is left as an exception since it is doing binary reads and writes to a file. Moved nowhere 13796. March 19, 2012 FRING, RLDLY Eric Changed both tasks to offer the ability to divide the spectrum into quarters for finding solutions. The code generalized nicely, actually simplifying the code structure. Corrected FRING to use the right frequency for a reference in each IF in these modes, but the answers did not change meaning that the reference was taken out correctly in the end. Moved nowhere. 13797. March 12, 2012 ELINT Eric Added QUAL and CALCODE adverbs to the task to separate the desired cal sources from the junk left behind by the EVLA. Moved nowhere. 13798. March 21, 2012 FARS Leonia Change the logic of terminating the clean under condition of NITER and FLUX. This change sents zero into the output, if the clean is terminated just before start of iteration. Before this change, the output was not zero. 13799. March 22, 2012 IMAGR Eric Add the option to also average channels together when doing the on-the-fly, baseline-length time averaging while preparing the work file which is imaged. This should speed up imaging significantly. Changed: IM2PARM Help file: add (13) to describe this and (8) and (9) which had been forgotten earlier. IMAGR Help file: add IM2PARM(13) in inputs and help sections and advice to be conservative with IM2PARM(12). UVWAIT Fixed format that printed # vis as ten times the right number. CHNNCP New subroutine - version of CHNCOP with channel averaging added to the call sequence. TABLEUTIL Changed OCHNCO to call CHNNCP rather than CHNCOP. UVUTIL In UVFRQS - added AVERAGEF keyword and change of freqencies out based on it when the input UVFREQ is not < -999. In UVFCOP - call OCHNCO with new argument of the value of the AVERAGEF keyword. UVDATA In OUVCLN - added AVERAGEF keyword and change of header number points, reference pixel, and increment due to frequency averaging. IMAGR Fortran: In IMLEAN: added pickup of IM2PARM of 11-13, using new routine FINDAV to choose the IM2P(13) when the user requests it (=1), and setting the AVERAGEF keyword in UVDATA. Set different values for AVERGAEF and CHINC in UVWORK (after clone) since the averaging is by then done. Added the BLIMIT values to IMACPY call (since the code picking them up had to move higher on the food chain). In IMACPY, protect the reference pixel and increment from a parameter copy of UVDATA -> UVWORK. IMBAVG gets changed summing code to sum over channels as well. Moved nowhere. 13800. March 23, 2012 IMAGR Eric Nothing filled in the defaults for the TV timeouts. Changed QCLEAN, where it gets the keyword values, to apply defaults as needed. Moved nowhere. 13801. March 23, 2012 CVEL Eric The source number tables were way too small and a format lacked a space to make it readable. Moved nowhere. 13802. March 28, 2012 LINIMAGE Eric Changed AU3A to allow DOCONFRM=-2 to mean no confirmation at all on ALLDEST and SAVDEST. Changed those 2 help files. Wrote new RUN file procedure (and help file) to do a loop over IF, copying the IF from the full data set, running IMAGR and then FLATN. When the loop finishes the proc decides whether MCUBE will work and then runs either MCUBE or FQUBE. It cleans up after itself fully. Moved nowhere. 13803. April 2, 2012 Slices Eric When making plots of data along an FQID axis, it would be nice to convert the thing to frequency either in labeling non-linearly or in making a liner axis and laying the points down depending on the actual frequencies. Changed: SLBINI Call sequence to include a frequency at the start of the slice and an increment when an FQID axis is converted to linear frequecies in the slice (0 -> not). Trap this and set the plot values and increments accordingly for both traditional (in x-y plane) and z axis slices. MV2C06SL Document added info in the SL table TKSLIN Pick up FQFREQ and FQFINC from slice header, put in plot image header and call to SLBINI. TVSLIN Pick up FQFREQ and FQFINC from slice header, put in plot image header and call to SLBINI. SLOCIN Pick up FQFREQ and FQFINC from plot catalog header and put in call to SLBINI AU6E Pick up FQFREQ and FQFINC from plot catalog header and put in call to SLBINI FRQGET New subroutine to read freqencies of a FQID axis. ISPEC Changed to plot FQID axis plots on a linear frequency axis when LTYPE=3. The data samples then fall irregularly. Changed SLice creation to interpolate the irregularly sampled FQIDs onto a regular frequency axis slice, recording the resulting initial frequency and increment in the slice file. Uses FRQGET. RSPEC As ISPEC BLSUM Changed to recognize FQID axes and write the SL file appropriately. SLFIT Pick up FQFREQ and FQFINC for SL file and pass to SLBINI. Correct code to honor magic blanks and to allow longer slices. SL2PL Pick up FQFREQ and FQFINC for SL file and pass to SLBINI. SLICE Changed to recognize a pure FQID slice in the X or Y axis and make an interpolated frequency slice instead. Uses FRQGET. Added capability to do a slice along the 3rd axis so long as BLC(1)=TRC(1) and BLC(2)=TRC(2). ISPEC Help: note regarding FQID axis plot and slice. RSPEC Help: note regarding FQID axis plot and slice. BLSUM Help: note regarding FQID axis plot and slice. SLICE Help: notes about use of BLC, TRC Also put no-op (FQFREQ=0, FQFINC=0) into calls to SLBINI in XPLOT, XBASL, XGAUS, PFPL2, and PLCUB. All of these could get fancy if there is a request for it (XGAUS would be the hardest). Moved nowhere. 13804. April 3, 2012 POSSM Eric Restored SOLINT for use with BP tables. Changed lots of places where it could complain about missing data to complain only when pieces (not all) are missing. Others of the complaints were already surpressed when SOLINT not zero. Moved nowhere. 13805. April 3, 2012 Velocity header parms Eric Changed: DGHEAD To fill in the alternate reference values from the source table including corrections for BCHAN when a multi-source file is going to a single-source file. Correct the alterante reference pixel in single-source input files for BCHAN. UVUTIL Changed handling of ALTRPIX in U2IDES because it is fixed down below in UVGET/DGHEAD. UVPHAS Dropped writing the ALTRPIX with the minimum coherence time. (???!!!) SPLIT Fixed to account for BCHAN being applied in lower level routines even in single-source cases. SPLAT Fixed to account for BCHAN being applied in lower level routines even in single-source cases. Moved nowhere. 13806. April 4, 2012 UVFIT Eric Changed it to use correct formulae for phases and derivatives. Unfortunately, this involves lots of double precision sines and cosines which are notably slower. Sources a long way from the phase center are fit much better however. Moved nowhere. 13807. April 5, 2012 ZCREA2 Eric ZCREA2 computed the size in bytes doing one multiply by 2 in 4-byte integer before switching to 8-byte values and this caused an overflow at 1 Terabyte! Corrected. Moved nowhere. 13808. April 5, 2012 INDXR Eric It neglected to write INFILE, CALIN, and most importantly BPARM to the history file. Moved nowhere. 13809. April 6, 2012 SETMAXAP Eric Changed AUC: SETMAXAP was lying about the current number of megabytes and handled other computations confusingly. The new number of megabytes requested did actually make it to the parameter which is used by AP code. Moved nowhere. 13810. April 6, 2012 ELINT Eric ELINT did not check the solution weights and so included flagged solutions in the fit. Moved nowhere. 13811. April 6, 2012 SNFLG Eric Added OPTYPE='A&P' to write flags whenever gain amplitudes, phases, and/or weights are out of user-specified ranges. Clarified usage of CUTOFF in OPTYPE 'AMP' and added a history writing routine as well. This should reduce the onerous usage of EDITA. Moved nowhere. 13812. April 6, 2012 UVUTIL Eric Added comments to U2IDES where the header frequency axis is set. The code looked strange, but it is not since lower level routines handled the complexities. Moved nowhere. 13813. April 6, 2012 AIPS Memo 117 Eric Put back corrections to a typo that listed some standard character-valued keywords as double floats. Moved nowhere. 13814. April 9, 2012 SNFLG Eric Corrected two places where FORMAT errors could arise. Moved nowhere. 13815. April 9, 2012 LISTR Eric Added XINC adverb to affect LIST and GAIN. Moved nowhere. 13816. April 10, 2012 LINIMAGE Eric Changed proc to make it use the OUTNAME from the beginning. In that way, users can set unique OUTNAMEs and so can run more than one LINIMAGE at the same time (from different POPS numbers on the same computer or using the same data areas from different computers). Moved nowhere. 13817. April 23, 2012 Slices Eric The call sequence to SLBINI was not changed for ISPEC and RSPEC and was changed in only one of two places in XPLOT, XGAUS, and XBASL. Moved nowhere. 13818. April 23, 2012 VBGLU Eric Changed: 1. Improved missing column message to say what type 2. Corrected declaration of character strings for ID columns which had length 1 rather than 24. 3. Corrected list of IF-type columns for BP tables - did not have WEIGHTs and had some bad column label, Moved nowhere. 13819. April 23, 2012 RFLAG Eric Shot itself in the foot when writing a flag table followed by writing plot files. Dropped the resetting status to READ so that there would be no status when the plot parts happen. Moved nowhere. 13820. April 24, 2012 CONVL Eric Changed code to have it change units from JY/PIXEL to JY/BEAM when the axes are celestial coordinates, Gaussian convolution, and FACTOR <= 0. Corrected help file which mis-spoke about FACTOR in this case (stating that the users FACTOR was changed even if > 0 to start with). Moved nowhere. 13821. April 25, 2012 CALSEL Eric Certain tasks call CALSEL to drop some of the records in an SN or CL table to make room for new ones. CLCAL in particular usually makes a new CL table and then drops pieces of it before interpolating the new values into it. Added remark to this effect ahead of the existing messages (either deleted file or deleted n records from file...). Moved nowhere. 13822. April 25, 2012 SAD Eric Changed the RMS determination to use first a robust RMS computation over the full image. That fills the histogram sensibly rather than depending on wild guesses from image max/min or elsewhere. If the histogram fits works, fine, otherwise use the rms found from the robust computation. Moved nowhere. 13823. April 26, 2012 VLANT Eric Added adverb DOINVERS which will let you undo the change to the AN file and do an oppposite correction in the CL file (if you ask for the correct CL files of course). Moved nowhere. 13824. April 26, 2012 COPTAB Eric Again: COPTAB should not copy CL tables to single source files ever. Changed to make an FO file in the case of a single-source output file when there is a cL file in the input data set. Uses CL2FO like in SPLIT. Changed local copies of COPTAB in NOIFS, UVCOP, and AVSPC. Changed SPLAT to avoid making a null CL table in single-source case. No changes needed in UVCOP and UJOIN versions of COPTAB. Moved nowhere. 13825. April 27, 2012 MORIF Eric NEW Task to break IFs into pieces each with fewer channels. The tables part required help from VBGLU. Moved nowhere. 13826. May 2, 2012 SETRGBL Eric Changed RUN file SETRGBL, procedure FLAMLEVS so that the first contour would not be zero, but instead a dark red. Moved nowhere. 13827. May 2, 2012 QUACK Eric The REASON adverb had no default, leaving blank which cannot be un-flagged. Added a default (task/date/time) and corrected the history writing section that messed up this snippet of code. Moved nowhere. 13828. May 3, 2012 RFLAG Eric Changed it to use a sliding median window to do the spectral function rather than a single window using all channels. The user may set the width of the window and also set how many channels are examined in each window (more makes it go faster since it slides the window by the number examined in the last window). The options to flag full spectra because their spectral rms was too low or too high are now gone. Added code in the plot routine to skip full flagged spectra. That code was present in the flagging side of the task and can make a real performance improvement. OPTYPE was changed to AVGCHAN, the width of the median filter, and selection of the method of spectral averaging was dropped. The time rms computation is now not done if the cutoff exceeds 1.E6 in abs value. Moved nowhere. 13829. May 4, 2012 RFLAG Eric Changed to add "grow" option FPARM(6) to expand any flags by + and - FPARM(6) channels. Moved nowhere. 13830. May 4, 2012 DTSIM Eric Changed DTPARM to enter antenna coordinates in right-hand form and KEYIN to parse things correctly (required calling strings to be put into local ones due to KEYIN stupidity). Changed some of the errors in the help file. I think it still does odd things with noise and the u,v,w values seem wrong to me. Moved nowhere. 13831. May 9, 2012 ANTENNAS, BASELINE Eric Added some examples to the help files for these two adverbs. Moved nowhere. 13832. May 9, 2012 header displays Eric Changed AIPS, AIPSB, AIPSC to allow for full verb number range in AU3B. Changed AU3B to implement IM2HEAD, IM3HEAD, IM4HEAD, IMOHEAD, Q2HEADER, Q3HEADER, Q4HEADER, and QOHEADER header displays for IN2NAME, IN3NAME, IN4NAME, and OUTNAME, resp. Changed POPSDAT and NEWPARMS to create these verbs and made new help files for each. Moved nowhere. 13833. May 9, 2012 DTSIM Eric DTTIME Changed this debug routine - not currently used anywhere - so that it contains its own initialization either automatic at the beginning or requested. DTINIT Removed this old subroutine from $APLSUB; it did the inits for DTTIME before. DTPM.INC Added INCLUDE 'INCS:PUVD.INC' to make it self-sustaining and added a set of default values in their own extra COMMON. DTPARM (1) Updated all standard station coordinates from the latest VLBI SCHED list. The VLA coordinates were in many cases VERY wrong. Added the HSA elements and Y1, adjusting pointers for the VLA. (2) Pick up and use default year, month, day in schedule, BP, and FG sections. (3) Change typing to standards and use more IF-THEN-ELSE to avoid excessive GO To's. DTCALS (1) Changed to use default values when the antenna had no CAL_ERR specification and when it did but with no specification for the GAIN(ifnum) annd/or the Tsys(ifnum). DTSIM: (1) Changed axis order to the standard one - Stokes before FREQ. (2) Added NX table creation and filling from the data (3) Put default year, month, day, gain, Tsys in the DTPM.INC common for later use. Added DPARM, FPARM, and VPARM for this purpose. (4) Dropped unused IN2FILE adverb. (5) Fixed Y coordinates so RIGHT assumed in antenna separation and in computing elevation of source. (6) Corrected history routine and added info about defaults and listing contents of INFILE. DTSIM.HLP (1) Dropped IN2FILE, added DPARM, FPARM, VPARM. (2) Clarify grammar on AN_TABLE KEY, add HSA antennas in list of VLBA_xx, add ALL_HSA. (3) SU_TABLE: EPOCH=2000 is default (4) FQ_TABLE: USB is the default (5) CAL_ERR: STATION required, FQID=1 is default, GAIN_ is defaulted with FPARM(ifnum), TSYS_ is defaulted with VPARM(ifnum) as constant (6) SU_MODEL: SOURCE if required (7) SCHEDULE: DAY, MONTH, YEAR required but default to previous starting with DPARM; SUBA and FQID default to 1, TINT defaults to 1.0 sec. (8) Document the FLAG option. (9) Document the BP_TABLE option. (10) Example - replaced with real one including adverb values. Two sources, one with 3-compnent model. Takes the VLA C array. Uses default time, gain, and Tsys values and changes gain for one IF on the VLA north arm. DTCHK Added default adverbs and setting them in the DTPM.INC common. DTCHK.HLP Added DPARM, FPARM, VPARM. Moved nowhere. 13834. May 11, 2012 POSSM Eric Added more support for linear polarization STOKES values. It just died before that when asking for 'XXYY' etc. Moved nowhere. 13835. May 11, 2012 CASA prompted Eric Changed ANFIND to add array name ALMA with +,-,- coords being right handed. Changed MULSDB to print the message about SU table with no source random parameter only twice. Moved nowhere. 13836. May 14, 2012 RFLAG Eric Previously, it flagged all channels not selected by ICHANSEL in the spectral flagging process. Changed it to avoid doing so - those channels will remain. Changed the default for ICHANSEL since we are now doing a narrower, sliding median window. The new default is to do the full range of channels. Moved nowhere. 13837. May 14, 2012 INDXR Eric Changed code to read EVLA.GAINS with up to 4 frequencies in each band. The gains are then interpolated to the actual IF frequencies in the data. EVLA.GAINS is also put to $AIPSIONS. Moved nowhere. 13838. May 14, 2012 PCAL Eric Changed the reference frequencies used so that they would be in the middle of each IF as in SETJY. This makes the fit spectral indices approximate those of SOUSP - or at least come closer. Moved nowhere. 13839. May 15, 2012 POSSM Eric Added note to help file portion to recommend specifying adverbs to limit searches when SOLINT is not zero. SOURCES, and/or ANTENNAS and/or other adverbs can be a great help to limit what is searched. Moved nowhere. 13840. May 15, 2012 IMEAN Eric Added DOPRINT adverb. DOPRINT=0 turns off OUTTEXT in all cases. DOPRINT=-3 turns on a special print of blc(3-7), trc(3-7), minimum, maximum, flux, and rms all in one line. Running IMEAN on each plane of a cube in turn gets a list of the min, max, flux, and rms spectra. Moved nowhere. 13841. May 16, 2012 RFLAG Eric Discovered that the channel flag array was not properly initialized. The flagging subroutine assumed that previously flagged channels were marked with a value of 1, but in fact they were all 0. So the task repeated all input flags more or less. Also changed: 1. The internal working array (one baseline, one IF) is now initialized as 0 or 2 by the external flag array and that is copied to all polarizations. Then the data weights are examined for the relevant times and any with weight <= 0 get 1's in this working array. 2. The test for some data valid now goes over all data since it has to set the working array. 3. The test to flag channels between flagged channels (FPARM(8)) now flags only if the working array has value 2 meaning that we are flagging in this pass of RFLAG. It will ignore previously flagged partially. This test is done before other global flag tests. 4. The section of code that actually does the flagging would do things even though no channels in the working array had value 2 - generating flags matching previous flags. 5. That section of code contained a second implementation of the FPARM(8) code has been dropped, replaced with a better algorithm looking to find consecutive channels to flag to reduce the number of output flags. Moved nowhere. 13842. May 17, 2012 AVSPC Eric In the COPTAB subroutine, CLNULL was called when CL2FO should have been (with the CL2FO call sequence) and then CLNULL was called with the wrong call sequence when it should have been (when DOCAL>0). The other COPTABs are okay. Moved nowhere. 13843. May 18, 2012 IMAGR & FG messages Eric Previously, when UVGET ('INIT'...) was called, a message about flag tables would come out on every 20'th such call. When IMAGR makes spectral cubes, UVGET ('INIT' is called for the input data set and then for the work file multiple times per channel. Doing a message only every 20'th time led to odd messages - possibly saying "flagging with vers n" sometimes and "doing no flagging" other times depending on whether it was the input or work file for which UVGET was called. This is confusing! So changed SELINI Made the initial FGVER=-100. The default is therefore no flagging which is a change from before. Tasks that want flagging should set FGVER. CALEDIT Changed SECINI to default FLAGVER in UVDATA objects to -100. It was -1, so no change on default flagging behavior. UVGET Changed to drop the every 20'th business and instead to remember the disk and catalog number of the last file for which UVGET was called. The first call to a new data set gets a flag message unless FGVER <= -99. This means no messages when FGVER was not deliberately set (e.g. scratch/work files). IMAGR Changed to set QCREAT to -1 when it has to blank an image plane and it fails to do so. This insures that files are filled with zeros. Moved nowhere (major link edit). 13844. May 18, 2012 COMB Eric Corrected to scale any image in JY/BEAM to the same beam as the JY/BEAM of the first image. Otherwise one is subtracting, ratioing, etc etc images in different units. Added mention of this in the help file. Moved nowhere. 13845. May 18, 2012 SNPLT, SNREF Eric Correct typos in SNREF help file and clarify wording. Mention SNREF with DOKEEP in SNPLT help file. Moved nowhere. 13846. May 21, 2012 RFLAG Eric FPARM(13) was changed by the history-writing that occurs after writing a flag table. As a consequence, if plotting followed writing of a flag table, the clip level FPARM(13) was used as the square root of the user's value of FPARM(13). Made the history routine restore the value of FPARM(13) for later use. Moved nowhere. 13847. May 22, 2012 UVHOL Eric The frequency averaging routine only zero'ed 3 of the 6 summing locations, making the last half of the polarizations bad on output. Error was put in August 2011 when the option to display real and imaginary and rms was added. Moved nowhere. 13848. May 23, 2012 PDINI, CPINI Eric The precursor comments of these were incorrect. Moved nowhere. 13849. May 23, 2012 POSSM Eric BP and BD plots need to allow the use of SOLINT although scan average is sufficient. Changed POSSM to force SOLINT not zero to SOLINT = -1. For scan average, POSSM uses the NX table to give the list of time ranges. This does not work for BP amd BD tables since they may be averaged in such a way that the average time does not fall within a current scan or they may have come from a source no longer included in the data set. Changed the NXSET routine to find the times in the BP or BD tables and make a list of "index table" times for the rest of the plot routines. Corrected the last time in the data set in case the BP record occurs after it now. Moved nowhere. 13850. May 25, 2012 FQ table Eric We need to keep track of which band was used so I am dding a column BANDCODE to the FQ table. It is 8 characters long with one value per IF (Spectral window). Changed FQINI to make the new table, not complain if missing on read FQTAB call sequence changed to include BNDCOD, get column values if present, else blank CHNDAT call sequence changes to add BNDCOD GETFQ call sequence changes to add BNDCOD PUVD.INC MAXFQC changes Subroutines changed for the changed call sequences: FRQGET ACTFRQ BPASET CALADJ CHNCOP CHNNCP CLREFM COPTAB CQMAKE DGHEAD FQFTAB FQMATC FRQTAB GAININ GETFRQ SN2CL SNREFM TABLEUTIL UVUTIL MAKMAP Tasks changed for call sequence changes MCUBE RFARS SPCOR SPIXR ANTAB ATLOD ATMCA AVSPC BPCOR CLCAL CLCOR CLSMO CSCOR DBCON DFCOR DFQID DTCHK DTSIM FARS FGDIF FILLM FILLR FITAB FITLD FITTP FLGIT FLOPM FQUBE FXVLA HAFIX LISTR M3TAR MATCH MK3IN MORIF NOIFS PCCOR PCLOD QUFIX QUOUT REFLG RLDIF SDCAL SETJY SHOUV SNCOR SNREF SNSMO SPLAT SPLIT TBAVG TI2HA TRUEP TYAPL TYSMO UJOIN UVCOP UVDEC UVFIT UVFIX UVFLG UVHIM UVIMG UVLIN UVLSD UVLSF UVMOD UVPRM VBGLU VLANT VLBIN VLOG APGPS CL2HF FXPOL MBDLY MULIF SPECR BPASS CALIB CONVL CPASS CVEL FRING KRING PCAL RLCAL RLDLY SDIMG ANBPL APCAL CAPLT CLPLT DELZN DFTPL ELINT FRMAP FRPLT LOCIT POSSM SNIFS SNPLT SOUSP UVHGM UVPLT VPLOT WIPER Other changes: ATMCA removed the assumption that FQ tables have only one record by using GETFQ DELZN As ATMCA MORIF Corrected the FG handling to do channels correctly and all versions of FG table. SPLAT Changed to copy CT tables and to copy GC and TY tables even if DOCAL is true. INDXR Added GETBND, an elaborate routine to use the FQ table BANDCODE and all observed frequencies to get the band identifier. Changed gain interpolation in frequency to extrapolate as well. Moved nowhere. 13851. May 31, 2012 SETGDS, FACSET Eric Changed SETGDS to allow a facet to not have CC files. The number of components for that facet is set to zero - previously it forced over to using images as models. Now it will do that only if not facets have CC files. Changed FACSET to open only those facets with components to be subracted, divided or whatever. Previously, it failed when one did not have the CC version wanted, whether or not components were wanted. Moved nowhere. 13852. May 31, 2012 CCNTR Eric Fixed the reading of the CC file to ignore flagged rows rather than dying. Added adverbs to limit the fluxes plotted and to allow the scale of the plot symbol to change with flux. Moved nowhere. 13853. June 4, 2012 INDXR Eric Corrected error when the number of IFs was 1 - it looked at NUMIF/2 which is 0 in this case when trying to find the band ID. Made the search center IF 1 in this case. Moved nowhere. 13854. June 6, 2012 ZEXIT Eric Wrote new Z routine ZEXIT to call the C function exit (n) where n > 0 => error which is communicated the higher level process (e.g. procedure). Changed $SYSUNIX/ZTRLOP.C to include a copy of ZEXIT. All of this to change the CALL EXIT (n) in $SYSUNIX/PP.FOR. The old Fortran semi-standard library routine has gone missing in several compilers. Changed $AIPNOT/HLPB to use ZEXIT (it complained that it wanted it but did not have a legitimate one). Moved nowhere. 13855. June 7, 2012 PRTSY Eric NEW TASK to print statistics from the SY table. Moved nowhere. 13856. June 8, 2012 FQ table Eric The following tasks dimensioned FQKOLS and FQNUMV as 5 rather than with the MAXFQC parameter and hence did not work right: FQUBE FARS FILLM RFARS MCUBE SPCOR SPIXR CONVL Moved nowhere. 13857. June 8, 2012 LINIMAGE Eric The algorithm checking for axis increments was not general enough so that something at very slightly below N times the axis increment triggered use of FQUBE when it should not. Moved nowhere. 13858. June 11, 2012 PRTSY Eric Added SOURCES, QUAL, CALCODE, SELBAND, SELFRQ, FREQID, and SUBARRAY to allow the user to restrict the statistics in reasonable ways. Moved nowhere. 13859. June 12, 2012 FIXAN Eric NEW TASK: should not normally be needed but can change an AN file between Earth centric and correct site-centric forms and also correct the naive and incorrect form in which Xsite(i) = Xcent(i) - X0 etc. Moved nowhere. 13860. June 12, 2012 PRTAN Eric Changed the DOBTWEEN option to function only with Earth centric systems and to properly correct for the longitude of the site. Previously that display was in the naive/wrong system mentioned above. Moved nowhere. 13861. June 13, 2012 IMFIT, JMFIT, SAD Eric The attempt to copy keywords was missing in SAD and in the wrong place (before the disk numbers were known) in IMFIT and JMFIT. Affected things only when the output residual file is created. Moved nowhere. 13862. June 13, 2012 TBIN Eric Raised limit from 1000000 to 100000000 rows. Moved nowhere. 13863. June 15, 2012 Y2K Eric The change to ALLDEST some time back broke the Y2K procedures. Changed Y2KLOAD.001 to set DOCONFRM=-2 everywhere. Moved nowhere. 13864. June 15, 2012 LISTR Eric The LST display option assumed that antenna 1 had a legitimate longitude. That may not be the case - added code to find and remember an antenna with a non zero longitude. Corrected GETANT - it required X not zero when either X or Y not zero in ATAN2(Y,X) is all that is required. Moved nowhere. 13865. June 15, 2012 Mountain Lion Eric The system command ps in Mountain Lion now issues an error message if there is an LD_LIBRARY_PATH set (which is required in binary installations). Changed $SYSUNIX files STARTPMON, TVALT, UNIXSERVERS, and XASERVERS to send the error output to /dev/null and to test for succesful completion rather than for error output. Moved to patch 19-November. 13866. June 15, 2012 FIXRL Eric NEW TASK: On occasion, receivers are hooked up incorrectly with the R channel labeled L and vice versa. Wrote a clone of the old (pre-calibration) FUDGE to do swaps for a user-supplied list of antennas. Moved nowhere. 13867. June 18, 2012 CC modeling Eric Changed GRDCRM which found no components beyonf the first flagged Clean component in the CC file (record pointer did not incremet). Affected UVSUB, CALIB, etc etc Moved nowhere. 13868. June 18, 2012 FIXRL Eric Moved from $APLPGM to $APGNOT, add automatic filling of the NX table. Moved nowhere. 13868. June 18, 2012 FRING Eric Minor re-coding to be more readable and efficient. Moved nowhere. 13869. June 20, 2012 EVLA.GAINS Bryan Butler 1 - normalize the gain curves to the 0 coefficient, so that they are 1.0 at zenith. We haven't done this before because the numbers have all always been _nearly_ normalized in this way without jiggering, but the ELINT results that rick gave me this time result in 0 coefficients that are significantly different than 1.0. rick and i agree in the end that just normalizing by that coefficient should be sufficient. 2 - for the lower bands (4, P, L, S, C), i've put in an explicit flat gain curve (1.0 at all elevations). 3 - for antennas which are not measured, i've put in the average antenna (with a comment to that effect). note that i left antenna 29 (PT) in, just in case we ever get it back into the array. i also tried to make some sensible extensions of dates for antennas not measured in one or the other of the runs (and for some frequencies, which were not repeated). 4 - combined the January 2010 and January 2012 results into a single file. Moved nowhere. 13870. June 20, 2012 UVGET Eric Changed UVGET ('INIT' to keep track of the FG version number last used and to report "using flag table version" when that changes even if the disk/cno does not change. This is appropriate for RFLAG. I also discovered that RFLAG was not working correctly (ignoring FLAGVER) because I had not relinked the last change. Therefor I relinked all tasks that use UVGET (have adverb DOCALIB). Moved nowhere. 13871. June 20, 2012 RFLAG Eric Changed plotting to avoid plots of time values when they are not computed. Fixed histogram of spectral plot which no longer has a negative range of values. Added labels to plots to say what is plotted a little more clearly. Moved nowhere 13872. June 21, 2012 BPASS Eric Changed the inputs and code to drop the BIF, EIF option. There were numerous errors in the application of this option - way too many to try and fix. Moved nowhere. 13873. June 21, 2012 VLBAEOPS Amy I bowed to pressure and removed the INFILE input for VLBAEOPS. It is more consistent this way and prevents people from making stupid mistakes. Moved nowhere. 13874. June 26, 2012 ELINT Leonia ELINT has skipped calculation of variance of the residuals, if only one calibrator is used. This error is corrected. Moved nowhere. 13875. June 28, 2012 ZLASIO Eric The error messages could overflow causing an abort instead of a normal quit in TVCPS and others. Moved nowhere. 13876. June 28, 2012 Memo 117 Eric Revised again: the source and freqsel columns in a UV-table format are required only when there is > 1 source and > 1 frequency setup, respectively. Added verbiage on pages 10 and 11 about this. Moved nowhere. 13877. June 29, 2012 AIPSLetter Eric Put back LET12A TEX, PS, PDF in $AIPSPUBL. Moved everywhere. 13878. June 29, 2012 REFLG Eric Added option to omit certain attempts to optimize the flags more globally. They work fine usually, but odd data sets such as a one baseline data set can confuse it badly. Moved nowhere. 13879. July 4, 2012 Flag application Eric Changed PFLG.INC to allow 60000 flags at any one time and TYAPL, UVCOP, RFLAG to handle 300000. Moved nowhere. 13880. July 4, 2012 REFLG Eric There is a gotcha in the task - if there are flags that apply to times other than those of the requested sources, those flags will be lost. Corrected this in part - if the flags specify a source number > 0 but not in the requested list they will be copied in the the initial pass throuigh the input FG table. Moved nowhere. 13881. July 4, 2012 XAS Eric Added option to set the character size multiplication. Fixed XAS.SHR to compute the default with roundoff and to read an .Xdefaults option. Added this to XAS help file and removed very obsolete information. Moved nowhere. 13882. July 5, 2012 CHARMULT Eric Created NEW TV VERB to change the TV's character size multiplier. Values from 1 through 5 are accepted. This changes AU5, POPSDAT, NEWPARMS, and XAS.SHR and creates new $YGEN and $YSS versions of YCMULT and a CHARMULT help file. Moved nowhere. 13883. July 5, 2012 SOUSP Eric Added PIXRANGE adverb to task to allow aligning 2 data sets. Moved nowhere. 13884. July 6, 2012 plots Eric Subroutine GETOFM read one too many points from the OFMFILE and so made all of the top levels 0 rather than what they should be. AffectEd GREYS, KNTR, PCNTR, LWPLA. Moved nowhere. 13885. July 7, 2012 GREYS, KNTR, PCNTR Eric When given an RGB image in axis three, GREYS, KNTR, and PCNTR are supposed to make a 3-color plot if so asdked. GREYS failed to set the corners correctly and so failed. KNTR and PCNTR set the corners without examining the image header values. Changed all to use the axis values to set the corners for such images. GREYS: also added option to have different FUNCTYPs for the gren and blue images and to have APARM(1-4) apply even when doing an RGB image (default in that case is PIXRANGE). Moved nowhere. 13886. July 9, 2012 TVHUI Eric Changed code to support transfer functions in the hue image. Changed HLPTVHUI.HLP to reflect this. Changed code and help to offer the option of writing out a step wedge image. It uses the pixrange of the input intensity and hue images as its range. The user controls the size and which axis the intensity and hue are on. Moved nowhere. 13887. July 9, 2012 Coordinate verbs Eric Several coordinate verbs allowed the COORDINA adverb to have the declination ahead of the RA when the image was rotated a lot. This confuses users and code like STARS, Changed AU5B, AU6D, and AU9 to not swap the two axes. All tasks and some verbs already assumed that this adverb was always longitude and then latitude. Moved nowhere. 13888. July 10, 2012 IMLOD, FITLD Eric The image reading code did not consider that NAXIS could be 1. Fixed it to make sure the number of Y pixels is at least 1. These tasks had trouble with a 1-D IMAGE attached to SDSS FITS images. Moved nowhere - patch? 13889. July 10, 2012 UVCOP Eric Changed it to extend the copy times of tables by UVCOPPRM(7) minutes with 0.0 -> 15. Moved nowhere. 13890. July 13, 2012 RFLAG, WIPER Eric RFLAG: improved progress message while flagging WIPER: added CATFIX so writing of FG table will not fail Moved nowhere. 13891. July 13, 2012 SNIFS Eric Fixed picking up of PSUM and added option to time average amplitudes in a scalar fashion. Moved nowhere. 13892. July 17, 2012 WIPER, REFLG, FGDIF Eric WIPER: Changed to allow the center of the fat eraser to leave the data area (and hence do just a little flagging at the edges of the plot). FGDIF: Limit the flag comparisons to the number of Stokes actually in the data ('1111' = '1100' when there are only 2 Stokes). Improved the difference reporting. REFLG: Changed to limit work to those flags that flag all of the Stokes now in the data set. REFLG: and FGDIF help files - changed the discussion of CPARM(1) to clarify things hopefully - in some cases there is no perfect answer. Moved nowhere. 13893. July 17, 2012 TBAVG Eric Put in a prohibition for auto-correlation data. Further comments in help file. Moved nowhere. 13894. July 18, 2012 ATLOD Eric It did computations of how much to increase the file and the current file size all in INTEGER*4. That overflows. Changed things to double precision for large files. Moved nowhere. 13895. July 18, 2012 UVAVG Eric Rearranged code - Intel in the LNX64 version was computing the average time differently for some reason then the computation of the times in the list which is sorted. This produced occasionally times out of order and times that were slightly different than the ones without optimization on Intel 64 and with optimization on GNU 64 and Intel 32. Moved nowhere. 13896. July 24, 2012 TBAVG Eric Corrected the exclusion of auto-correlation data. When an AC point occurred the rest of the IO buffer was ignored. Moved nowhere. 13897. July 25, 2012 RFLAG Eric Adjusted the handling of FG table versions. New flags are written to OUTFGVER directly unless OUTFGVER=FLAGVER. Old flags from FLAGVER are copied to OUTFGVER if and only if OUTFGVER represents a new table. If FLAGVER=OUTFGVER (after application of defaults), the task uses a temporary FG table filled with the old flags and any new ones and then replaces the old FG ver with the temporary one. Moved nowhere. 13898. July 25, 2012 CODECIML Eric New verb to change between X, Y in decimal degrees and COORDINA in sexagesimal. Changed POPSDAT, NEWPARMS, AU7B, and new help files for CODECIML and DECIMAL (an adverb to give the direction of the conversion). Moved nowhere. 13899. July 25, 2012 STARS Eric Added adverb DECIMAL to this task to specify that the RA/Dec coorinates in the input text file are decimal degrees rather than sexigesimal. Moved nowhere. 13900. July 26, 2012 MAXCHA Eric Changed PUVD.INC to change the parameters MAXCHA and MAXCIF from 65536 to 132096 (2*65536 + 1024). The EVLA hit the old limit! Changed ZPBUFSZ.INC accordingly. Corrected UVGET which had a computation (MVIS*(ECHANS-BCHANS+1)) / (ECHAN-BCHAN+1) where MVIS is the length of the record in values. For 2 IFs and 32768 channels, MVIS = 65536 and (ECHANS-BCHANS+1) is 32768 and the numerator overflows. I wonder how many more of these there are. Moved nowhere - massive compilation/link job. 13901. July 26, 2012 VLBAEOPS Amy I introduced a bug with my June 21 change of VLBAEOPS, fixed. Moved nowhere 13902. July 27, 2012 IMVAL, QIMVAL Eric Added adverb DOCENTER to IMVAL and QIMVAL. If DOCENTER > 0, the verbs find the image intensity and coordinate at the nearest pixel to PIXXY. If <= 0, a cubic interpolation is done (as always before). Updated DOCENTER help file as well as IMVAL and QIMVAL help files. Changed AU9 to implement the function. Moved nowhere. 13903. July 27, 2012 FITLD, CLCOR Eric In CLCOR, added code in SOUMOD to compute the apparent coordinates over again to make sure they are right and to report the change if it exceeds 0.01 degrees in RA or Dec. In FITLD, in the FITS-IDI section, changed SO2SU to recompute the apparent coordinates in all cases and report those with a change > 0.01 degrees. Previously, it recomputed only if the apparent coords were identical to the epoch coords. It did not even recompute when the apparent coords were 0.0! Moved nowhere. 13904. July 30, 2012 U2CAT, M2CAT Eric NEW VERBS: U2CAT and M2CAT are like UCAT and MCAT but with IN2DISK. Changed POPSDAT.HLP, NEWPARMS.001, AU3A.FOR and made new help files for U2CAT and M2CAT. Moved nowhere. 13905. July 30, 2012 IMAGR Eric With heavy editing, the average frequency of an image may not be the center channel. Changed IMAGR to compute this average frequency and store it in the Clean output objects (fields). Changed U2IDES in UVUTIL and IMGATT in IMAGE to detect this keyword if present and to override the default frequency in the header. Moved nowhere. 13906. July 30, 2012 IMEAN Eric Re-instated the use of BLC and TRC in finding the initial (robust) guess at the rms. DOINVERS is also supported in that rouine now. This is needed since the rms may be a function of position as in beam-corrected images. Moved nowhere. 13907. August 1, 2012 HGEOM Eric HGEOM used the wrong version of the image header when calling table copy routines. If any tables were copied, the resulting header had magic blanks for the min and max. Moved nowhere. 13908. August 2, 2012 M*CAT, U*CAT Eric Made NEW VERBS: M2CAT, M3CAT, M4CAT, MOCAT, U2CAT, U3CAT, U4CAT, UOCAT and put them all in a new verb subroutine AU4A. Changed POPSDAT, NEWPARMS to list these verbs, AIPS, AIPSB, AIPSC to call AU4A. Changed AU3A to support M2CAT and U2CAT since they were released in this routine but to tell users to run COMPRESS to update their vocabulary. More new help files and correct old MCAT, M2CAT, UCAT, and U2CAT helps for the sort limit (now 5120) and wording. Moved nowhere. 13909. August 2, 2012 GETVERS, QGETVERS Eric NEW VERBS: GETVERS finds the maximum version number of a specified extension file typ. QGETVERS does the same quietly. New adverb MAXVERS also. Moved nowhere. 13910. August 2, 2012 IMAGR Eric de Sade's law: I wanted to set the frequency of the header properly for averaged channels, but neglected the spectral line case in which that frequency then got changed with every channel. Changed IMAGR to do this new keyword only for SUM case, not LINE case. Changed QCLEAN to pick up the keyword if it exists for use in the CG table. Moved nowhere. 13911. August 3, 2012 KA-band Models Amy/Eric Put Amy's models for 3C48, 3C138, 3C147, 3C286 into the system. Moved nowhere. 13912. August 6, 2012 EVAUV Eric Changed adverb DOHIST to DOPLOT and added APARM to control the Real versus Imaginary plots. Changed the last 2 plots to be contours at logarithmic steps of a 2-D histogram image with real and imaginary on the x and y axes. APARM controls the range of values in the image, the size of the image in pixels, and any smoothing applied before the contour plot. Moved nowhere. 13913. August 6, 2012 IMAGR Eric Added messages to the code that finds the average frequency to tell users when there is flagging how many channels out of how many were used and the resulting average frequency. Moved nowhere. 13914. August 7, 2012 RM Eric RM failed to get the TABFQ call sequence change done May 25. Corrected a format error in the HI file writing that caused 64-bit Linux with the Intel compiler to abort. Moved nowhere. 13915. August 8, 2012 CLCOR Eric The subroutine SOUMOD picked up the equinox as CATBLK(KREPO) which gave it a terrible value - should be CATR(KREPO). This was left over from when the CATBLK was a call sequence argument and tricks had to be played. Moved nowhere. 13916. August 8, 2012 IMEAN Eric Added DOMODEL adverb. Gaussian is plotted if DOMODEL > 0. Not plotted otherwise. Moved nowhere. 13917. August 8, 2012 NX tables Eric The routine GETNX used to build the list of time intervals in the input NX table which is then used to guide scan boundaries in the output table had a bug (wrong variable name). This affected only data sets with multiple sub-arrays. Moved nowhere. 13918. August 8, 2012 FRING Eric Added DOAPPLY adverb to control whether the solution is appled to write an output data set or not. This only happens for single-source files. Corrected behavior when UVGET ('INIT' returns error code -1. The data set must be closed in that case and the error code preserved for the routines one level up. Added automatic NX table generation to the output file. Added a call to CATFIX to guarantee that the input file is in an appropriate status to have the SN table added to it. The history file writing routine wrote IN2NAME etc to the HI file when SMODEL > zero - it should be when SMODEL is zero. Moved nowhere. 13919. August 9, 2012 SNPLT Eric Changed DO3COL to have a value = 2 (> 1.5 actually) to mean colorize by source number, It makes a color pattern based solely on those sources that actually occur in the table, source list, time range, etc. If that list consists of source numbers 3 4 8 and 23, there will be 4 regularly spaced colors from red to blue. Moved nowhere. 13920. August 10, 2012 SLICE Eric Added DOKEEP which says to skip interpolation when that is possible (slice along X or Y axis). It already does this on the special slice along Z axis. Moved nowhere. 13921. August 10, 2012 AFARS Eric Added options to control how the maximum amplitude and its position are estimated fromthe pixel values surrounding the peak in each RM row. The default is an exact parabolic fit to 3 points, with the previous 5-point least-squares parabolic fit and no fit at all as other options. Moved nowhere. 13922. August 10, 2012 LISTR Eric Added DPARM(10) to set the value of a print of 1 in LIST, GAIN, and MATX forms. DPARM(10) <= 0 gives what has always (recently) been done. Moved nowhere. 13923. August 13, 2012 TARS Eric Changed the output formats (amp, real, imag multiplied by 10^6, more suitable number of digits and blanks). Changed the output text file to checked for errors and handle them without thousands of messages. Changed the output text file to allow it to be concatenated. Changed task to print to the message file only when OUTFILE is blank. Moved nowhere. 13924. August 13, 2012 PCLOD Eric PCLOD uses the damn KEYIN routine and so needs an enormous buffer in case there are lots of values to be read. Changed 100000 to 20000000. Fixed the help file: frequency and pulse-cal are spelled without the last character. Moved nowhere. 13925. August 13, 2012 TAPLT Eric Added RPARM, KEYSTR, DOINVERS to allow selection of which rows to plot. Added comments in the Explain file about SY table plotting. Moved nowhere. 13926. August 14, 2012 COPIXEL Eric Verb COPIXEL, subroutine AU7B, was not careful on output about coordinates between -1 and 0 degrees declination. Changed it to set COORDINA of 4, 5, and 6 negative, not just COORDINA(4). Changed help file of COORDINA to warn about typing -0, m, d. Upgraded that help file to include all verbs, procs, tasks now using the adverb. Moved nowhere. 13927. August 14, 2012 TARS Eric Put in a call to JTRIM to turn TAB characters in blanks and trim off unprintable junk from the end of lines when reading whatever garbage the user has prepared. Moved nowhere. 13928. August 15, 2012 TAB characters Eric All text reads should be checked for TAB characters except those of outside-world, fixed format text files. In general a call to JTRIM will replace TABs with single blanks and drop all characters after the first (other) unprintable character. Changed in this way: ZLWIO ZTXREN AU5F AU7 AU7B GNETSP K2TEXT OFMDIR WRIBOX GETOFM COOKNS COOKNX HLPA HLPB SHOPA SHOPB SHOPH SETSP CHKFC CPYRT DRCHK PRTAC SAD AFILE BOXES BPCOR CLCOR DFCOR FACES FARS FILLM FIXBX RLDIF SETAN SETFC SHADO STARS TBIN TBOUT TFILE UVLIN UVSIM VLOG CCEDT INDXR UVCON IMAGR BPERR GREYS IMRMS SERCH CONFI GAL PLOTR ICON Also did: DRCHK Changed the spelling of the ACD.. name - two slashes are needed to get one in fact. PBEAM Tried to change it to use free-format parsing for the data and, when the header card fixed format read has an error, with that card. I have no data to test this. Moved nowhere. 13929. August 17, 2012 IMAGR Eric Use of FQTOL with the frequency-dependent options has turned up some errors. Fixed IMAGR Lost the value of FQTOL from the work object if that object (file) was found to pre-exist and need replacing. The keyword had to be in the main UVDATA object so that it would reappear in the work object in this case. VISDFT Two places used the incomong NCHAN argument rather than the max of the number of channels in any model. That made the model get erased in part by data and resulted in NaNs possibly in Gaussian modeling but probably got wrong results in point-source modeling also. In addition, a summary message used the number of channels in model #1 rather than model #I to get the highest channel number and so listed more than it should in some cases. Moved nowhere. 13930. August 20, 2012 UVFIX Eric Changed LPSEC to add the latest leap second. Changed UVFIX to count the number of failures from FRQUPD and report the first 4 and the total count. FRQUPD is supposed to find the source, time, antenna in the CL or FO table and returns -1 if it does not. UVFIX used to report this as if an I/O error occurred and then turn off any looking for Doppler offsets. Moved nowhere. 13931. August 20, 2012 CD matrix Eric Coordinates are often transferred from IRAF et al with a CD matrix which can contain (and frequenctly does contain) skew. Changed the PCHDR routine which tries to make sense of this to take the smaller terms less seriously. Axis increments will be set by the bigger terms CD(1,1) and CD(2,2) usually, but CD(1,2) and CD(2,1) when the rotation is near 90 degrees. Moved nowhere. 13932. August 20, 2012 MORIF Eric Added the option to set BCHAN and ECHAN so that one can get a number of channels that divides into N equal parts. This forced changes in how a variety of things are handled. It also turned up errors in the handling of BIF in BP and similar tables. The BP tables now come out right and the data too. Moved nowhere. 13933. August 21, 2012 IMAGR Eric In the presence of subarrays with SUBARRAY=0 but with other adverbs (notably TIMERA set) IMAGR dies badly. The root cause is the failure to return a negative error code when UVGET ('INIT'..) returns -1 even though this return means that files are not open and there are no data. Changed: UVDATA S.R. OUVOPN to issue a message about no data selected and return the -1 error code. UVUTIL S.R.s SCNAVG, SCNHAY, NXTAVG changed to return the -1 if it arises rather than to ignore it. IMAGR When adverb SUBARRAY is zero, make a list of subarray numbers that have data and use only those. Found and fixed an error - the maximum baseline used to set the averaging when doing baseline-length time averaging was determined from the first subarray only. It looped but always looked at the first (when SUBARRAY=0). Moved nowhere. 13934. August 23, 2012 IMAGR Eric Changed OUVOPN in UVDATA to close the UV file when UVGET ('INIT'..) returns -1. The file is closed already on error returns > 0, so make this consistent. IMAGR had to be changed to do OCLOSE only when OOPEN was successful - the error in OCLOSE caused it to issue messages and die otherwise. Moved nowhere. 13935. August 23, 2012 IMSTAT Eric Improved formats in AU6D for IMSTAT and TVSTAT. Moved nowhere. 13936. August 23, 2012 Model subtraction Eric The automatic change of CMETHOD ' ' to 'IMAG' when there are no CC files was broken in SETGDS by the change that allowed there to be no CC files in some but not all facets. Fixed it although an error code is returned whn the user explicitly asked for CC models. Moved nowhere. 13937. August 23, 2012 CC2IM Eric NEW TASK: converts a CC file to an image in JY/PIXEL for point or Gaussian Clean components. Moved nowhere. 13938. August 23, 2102 IMEAN, IMSTAT Eric Added code to IMEAN and IMSTAT/TVSTAT (subroutine AU6D) to report the total flux from images in JY/PIXEL. Moved nowhere. 13939. August 24, 2012 install.pl Eric Added comment about editing $SYSLOCAL/UPDCONFIG for the e-mail address which is usually wrong in two places. Moved nowhere. 13940. August 24, 2012 PRTAB Eric Added the ability to do F format displays of reals and doubles as well as the exponential form. The default (NDIG=0) is now F format. It tries to represent the smallest non zero (with 2 digits) as well as the largest value, within the character limits of the exponential formats previously used. It can be more efficient however depending on the data. Moved nowhere. 13941. August 24, 2012 IMAGR Eric Found an error that I am suprised had not appeared previously. In QUVUTIL, subroutine OUVIMG it made reference to BEAM(LFIELD) but that does not exist for all LFIELD when ONEBEAM is true. This could have produced an error return with no properly printed error. Changed the handling of the message suppression in a number of places throughout QUVUTIL. Moved nowhere. 13942. August 24, 2012 CPASS Eric Added the spectral index correction adverbs and capability. As BPASS, it knows the main calibrators' spectral parameters and can be told those for others. Moved nowhere. 13943. August 27, 2012 TYAPL Eric Added OPTYPE to allow a request that the gains be set from the SY table using the post-detection gain rather than Pdif and Psum. Weights will not be computed in this approach. Also added the ability to do the gain operation with TY or SY (Psum and Pdif or Pgain) by writing a new CL table rather than with copying the data. This cannot do weights and will not correctly handle changes that occur anywhere other than between scans. It is a lot faster however. Moved nowhere. 13944. August 27, 2012 SETFC Eric Added adverb OPTYPE to specify 'USEF' to have the task write F type field cards (RA/DEC shifts) rather than C type (RA/Dec) field cards. Moved nowhere. 13945. August 28, 2012 install.pl Eric Added a close parenthesis that I forgot on August 24. Moved nowhere. 13946. August 28, 2012 MEMO 117 Eric Added a remark in the SU table section that it is required in the presence of a SOURCE random parameter and usually omitted when there is not SOURCE random parameter. Also change log chapter too. Moved to ftp site. 13947. August 29, 2012 PRTAB Eric Messed with the Fm.n formatting which had issues. It should be nicer now. Moved nowhere. 13948. August 29, 2012 FITLD Eric In the IDI section, the print format for sky frequencies did not allow for >= 100 GHz. Moved nowhere. 13949. August 29, 2012 EVAUV Eric Added adverbs DOOUTPUT, OUTNAME, OUTSEQ, and OUTDISK to help and Fortran. Added capability to catalog the output files as UV files rather than temporary scratch files. Added writing of history and copying of table files to each. Changed DOPLOT so that > 0 does the real versus imaginary plots of the residual and gain files. DOPLOT > 1.5 does (first) the mean/rms versus radius and image histogram plots. Added messages explaining the real vs imaginary plots better. Moved nowhere. 13950. August 30, 2012 DTSUM Eric Raised the limit on the number of scans and number of sources from 1000 each to 10000 each. Moved nowhere. 13951. August 31, 2012 INDXR Eric Tried to improve the warning messages when some IF is not consistent with the band selected. Moved nowhere. 13952. September 3, 2012 INDXR Eric Added messages like those in FILLM about the zenith opacities used. Moved nowhere. 13953. September 3, 2012 TARS Eric Found error: the Clean algorithm failed to subtract the contribution of the last iteration so the restored plus residual or residual only outputs were incorrect. The task also ignored the Q and U values in the data rows when a model was given. Changed it to add the model components to the data from the rows that define the frequencies to be used. Also changed it to read a fourth column of weights. Added a new OPCODE to control which Clean algorithm is used and added one called CMPL. The new Clean does a pattern match of the inner portion of the complex beam with the RM spectrum modified by all possible phase shifts (in increments of 0.1 degrees). That shift with the largest sum of the product of the real part of the beam with the real part of the shifted row plus the imaginary part of the beam with the imaginary part of the shifted row determines the location of the peak in RM with the amplitude of the sum and the phase shift setting the Clean component amplitude and phase. Added code to fit the width of the real part of the RMTF to give the width of the Clean restoring Gaussian (as a default). Changed all floating-point computations to double precision - it makes a difference when working with noise-free problems at least. Moved nowhere. 13954. September 3, 2012 FARS Eric Found error: the Clean algorithm failed to subtract the contribution of the last iteration so the restored plus residual particularly or residual only outputs were incorrect. Added fitting of a Clean beam to the real part of the RMTF. Moved nowhere. 13955. September 4, 2012 UVLOD, FITLD Eric Added adverb DOKEEP to UVLOD and FITLD. Fully flagged visibility records are kept only if DOKEEP > 0. Moved nowhere. 13956. September 7, 2012 FARS Eric Dropped OUTCLASS and OUT2NAME, OUT2CLASS, OUT2SEQ, and OUT2DISK adverbs from the inputs. The task forces the outclass for the two output images and uses the same OUTNAME, OUTSEQ, and OUTDISK for both. Changed Fortran for the change in adverbs. The old code used OUT2SEQ and OUT2DISK for the first image along with OUTNAME and OUTCLASS and vice versa for the second output image. Now both use OUTNAME, OUTSEQ, and OUTDISK, with the class forced to the previous defaults. Got rid of erroneous STOKES axis in the output headers. Moved nowhere. 13957. September 7, 2012 AFARS Eric Dropped OUTCLASS and OUT2NAME, OUT2CLASS, OUT2SEQ, and OUT2DISK adverbs from the inputs. The task forces the outclass for the (now) three output images and uses the same OUTNAME, OUTSEQ, and OUTDISK for all. Changed OPCODE to have 'CMPL' request real and imaginary out, otherwise amplitude and phase are written. Changed Fortran for the change in adverbs and to write out 3 images at all times - RM plus either amplitude and phase or real and imaginary. All but the phase attempt to use a fit position for the maximum between pixels around the maximum pixel. The old code used OUT2DISK for the RM image but OUTNAME, OUTCLASS, and OUT2SEQ and vice versa for the second output image. The code wrote some abnormal string on top of the STOKES axis in the output images and that axis should no longer show up. Got rid of erroneous STOKES axis in the output headers. Moved nowhere. 13958. September 7, 2012 MODIM Eric NEW TASK: Makes model image cubes in I, Q, and U either starting with suitable input cubes or creating new ones. Model components have values of I, Q, U at 1 GHz plus spectral index, rotation measure, position, and widths (major, minor, pa) of 6 different types - point, Gaussian, uniform disk, uniform rectangle, sphere, and exponential. Up to 9999 components may be read in via a text file. Moved nowhere. 13959. September 12, 2012 TARS Eric Correct to handle files with no weight correctly (rather than setting the weights to 0). Moved nowhere. 13960. September 12, 2012 CLCOP Eric Added OPTYPE 'AVER' to average a list of IFs and then replace a list of IFs with that average. This will deal with an IF adversely affected by line emission, RFI, etc. Fixed spelling in SYCOP help file. Found error message in SNCOP that used the wrong format number. Moved nowhere. 13961. September 12, 2012 CLCOR Eric Improved the description of the INFILE option for the ANTP function in the help file. Moved nowhere. 13962. September 14, 2012 QUXTR Eric NEW TASK: This task extracts Q,U spectra from a pair of Q,U cubes to generate text files for input to TARS. Moved nowhere. 13963. September 14, 2012 TARPL Eric NEW TASK: This task plots rotation measure spectra from TARS output text files. It can plot multiple lines/symbols each representing a block of TARS output contained in the TARS output text file. Moved nowhere. 13964. September 17, 2012 SETJY Eric Corrected the handling of RADIO velocities in OPTYPE 'VCAL' - they were all wrong. A similar attempt to use a cute trick involving a simple sign flip led to incorrect computation of the LSRVEL in other OPTYPEs when the frequency reference pixel and the velocity reference pixel differ. Moved nowhere. 13965. September 17, 2012 IMEAN Eric Added TRIANGLE as a return adverb. It contains in TRIANGLE(i) the flux level exceeded by only i per cent of the data samples. This will let one pick off flux levels in procs to set PIXRANGE or other flux-dependent variables. Note that this only tells about the flux up to 50%. Moved nowhere. 13966. September 21, 2012 Clean beam plots Eric Changed PLTBEM again to make it define line types rather than live with the one it inherits. At the end it restores the line type it started with. Line type 1 is used for the box and line type 2 is used for the beam and any cross-hatching. Was changed on September 16 to output a comment indicating the start of the beam plot. Moved nowhere. 13967. September 22, 2012 LWPLA Eric (1) Added a new page type - landscape rotated the other way (to +90 rather than -90). (2) In positioning character strings, LWPLA had assumed that all characters are the same length. That is only true of Courier and not of the default font. Changed the rtoutine that checks that the character string is legitimate to count the actual length by assuming that small characters are only 2/3 the size ('.-, :;()'). Moved nowhere. 13968. September 22, 2012 EVLA.GAINS Bryan A corrected EVLA.GAINS file has been installed in $AIPSIONS. Moved nowhere. 13969. September 22, 2012 WINDOW Eric The subroutine that interprets BLC and TRC dropped BLC to the integer below BLC(i)+0.01 but raised TRC with TRC(i)+0.89. That should be 0.99. Moved nowhere. 13970. September 24, 2012 plotting Eric Changed LABINI and CLAB1 to use capitalized lower case axis labels with longer strings in some cases (ARC SEC). Will affect output of plots to LWPLA but not to TV. Moved nowhere. 13971. September 25, 2012 LWPLA Eric A change in the handling of strings to look for characters smaller than the average led to the code not handling strings that it thought had unprintable characters. They really did not but the subroutine testing for them was inadequate. Changed to let JTRIM do that and rearranged the code to avoid the issue. Moved nowhere. 13972. September 26, 2012 PCCOR Eric Changed message when the number of tones in less than two. For some reason it gave an elaborate message when 2 * Nif * Npol > 16. Dropped that in favor of the simple message. Moved nowhere. 13973. September 26, 2012 FLATN Eric FLATN did not defend itself against the observed RA and Dec being 0.0. That made correcting for primary beam dicey. Put in set it to the ref values. Moved nowhere. 13974. September 26, 2012 FLATN Eric The inscribed ellipse that set weight to 0.0001 outside the ellipse was done **after** the primary beam which tried to set the weight to 0.0 outside its limits. Fixed the order of the weighting so that inscribed ellipse is first, then primary beam, and then radial downweighting (which is rarely recommended in mosaicing). Moved nowhere. 13975. September 27, 2012 SLICE, SLFIT Eric Fixed SLICE to allow non-integer BLC and TRC when DOKEEP is false. Fixed SLFIT to take absolute value of LTYPE and to convert values <= 1 to 3. It failed otherwise. Moved nowhere. 13976. October 1, 2012 PRTAB Eric The change to F format did not deal with columns with real values that are too small for the allowed F formats. Changed code to detect this and use exponential formats for those columns. Moved nowhere. 13977. October 1, 2012 RM Eric Fixed some HI file formatting problems. Moved nowhere. 13978. October 3, 2102 Misc Eric RLDIF Did not write history file when DOAPPLY was false. Now it will write it with the DOAPPLY value at the end. TARS Made the print format even wider to encompass large values as well as small. POSSM Modified some formats related to channels although all should have been okay at present. VPLOT Modified a few formats related to channels - one was rather wrong (I4) while most of the rest were functional. The plot header used I8 but then only the lower 4 characters - fixed to use 5. TARPL There was a typo causing line plots to be omitted. Moved nowhere. 13979. October 3, 2012 documentation files Eric Update: 1. $HLPFIL/HLPIT.LIS the file used to do TAB completions 2. Apropos text file ($HLPFIL/LSAPROPO.HLP) did not need changing - putbck does an excellent job. 3. About files updated: ADVERB ANALYSIS AP CALIBRAT CATALOG COORDINA EDITING EXT-APPL FITS GENERAL IMAGING INTERACT MODELING ONED OOP PLOT POLARIZA PROCEDUR SPECTRAL TABLE TAPE TASK TV UTILITY UV VERB VLA Moved nowhere. 13980. October 4, 2012 SETJY Eric The computation of velocities at the reference pixel given the velocity at some other pixel was grossly in error since 17 September (a factopr of the channel width in Hz!). However, the computation was in error for optical convention for forever (probably). A full non-linear computation must be done Vref = Vc + C (1/nuref - 1/nuc) where c is the channel at which the velocity is known. A linear formula was used after a correct computation of the velocity slope at c. Moved nowhere. 13981. October 4, 2012 FRING Eric The change in FRING to allow 1/N subdivisions of the full set of channels into N solutions broke the mode of multi-band/single-band represented by APARM(5)=2. I used an old version (2009) of FRING to inform me of the intentions of this mode and now there appears to be a sufficient set of If-Then-ElseIf-Then-Else clauses. Moved nowhere. 13982. October 5, 2012 GCPLT Eric NEW TASK: Plots selected contents of the antenna gain curve test file used by INDXR and FILLM in AIPS and by other software packages. Moved nowhere. 13983. October 5, 2012 LWPLA Eric Corrected error - the number of characters in a string to be plotted was altered by the code and that new number used to move the buffer pointer. Changed to remember the initial number which is needed to move the buffer pointer. This does mean that some task wrote a string with trailing blanks. Moved nowhere. 13984. October 8, 2012 IMAGR Eric Both UVUTIL and IMAGR were changing the sort order to '??' when there is rotation. That is not necessary with BT or TB for which the sort order is not lost. The BLVER is getting carried along so UVGET was unhappy with DOBL true and sort order not T*. SELINI did not initialize DOBL so it was kept from a previous data set and tested in UVGET before it could be properly set. UVGET used DOBL before GAININ could test for BLVER and set it. Changed it to test the image header for a BL table and set BLVER=-1 if there is none. Then set DOBL = BLVER.GE.0 before the time order and other tests. Moved nowhere. 13985. October 8, 2012 CookBook Eric Changed: 1 Updated the count of lines, version name, etc 3 M2CAT, U2CAT et al, IM2HEAD, Q2HEADER, et al. double precision pseudo-AP 4 SNFLG (A&P), setjy (spectral), SNREF, Use RLDIF not CLCOR, self-cal before DOPOL application 5 OOCAL, SCMAP additions, IMAGR average channels, SNFLG self-cal before DOPOL 6 CHARMULT, TVHUI (hue transfer func, wedge output image), SNPLT color by source 7 TARS, TARPL, QUXTR; SLICE with FQID, 3rd axis, no interp option nd more text-file info; Extend modeling to incl GAL/CUBIT ref, OMFIT, DTSIM, UVHOL, MODIM 9 SNPLT color by source 12 NAMEGET, GETVERS, QGETVERS 13. Updated 27 ABOUT lists E LISTR (SY), OOCAL, SETJY (VCAL), FRING/RLDLY quarters, LINIMAGE. SNFLG (A&P), MORIF, RFLAG freq changes, INDXR EVLA gains and OBIT warning, PRTSY, TYAPL (PGN) 0 updated I updated J updated Recipe Used every one of the unused in various chapters. Moved nowhere. 13986. October 11, 2012 MORIF, VBGLU Eric A call to determine whether a column was in a list had the wrong length for the list. This caused it to look in the next list and find things it should - but only with some compilers. Affect GC tables badly on Intel 64 bit. Moved nowhere. 13987. October 12, 2012 SETJY Eric Changed to add the Perley/Butler 2013 values for 3C286 and 3C295. This shifts all previous systems 1 higher in the APARM(2) adverb. Moved nowhere. 13988. October 12, 2012 DEFCOLOR Eric Added proc to POPSDAT to set default PLCOLORS - most useful when Tv is not running. Edited help files for LWPLA, PLCOLORS, and TVCOLORS and made a new help for DEFCOLOR. Moved nowhere. 13989. October 12, 2012 SOUSP Eric Added explanatory message ahead of printing fit values. Corrected the positioning of the strings inside the plot giving the values. They were based on TVs and not proper plot file concepts. Moved nowhere. 13990. October 15, 2012 editing Eric Changed QEDIUTIL to have "SWITCH ALL IF" be a 3 position switch (when there are >2 IFs and all are in ram). The positions are 1 IF (the current one), a range of IFs (the current one may not even be included), or all IFs. This required modest changes to the help files EDITA, EDITR, SNEDT, HLPEDICL, HLPEDISN, HLPEDISS, HLPEDITY, HLPEDITS, HLPEDITY, and HLPEDIUV. Moved nowhere. 13991. October 15, 2012 OBIT verbs Eric Changed AUO to make sure the "line printer" does not appear to be open when it has not been opened. Changed content of error report when ZSHCMD (the operating system) reports an error. Fixed CookBook chapter E about DOWAIT in BDF2AIPS and the table of contents to make the last page not truely a hanger-on. Moved nowhere. 13992. October 16, 2012 PLOTR Eric Fixed the line type (TV color) for strings. Fixed a bad subscript on a test for symbol type - got bad values for data types greater than a few. Moved nowhere. 13993. October 16, 2012 UVRFI Eric The task attempts to do TB sorted data but did not allocate enough memory (by a factor of number of samples in the interval). Changed to allocate that much memory and to hint in the help file that TB data might well work. Moved nowhere. 13994. October 16-17, 2012 SETJY Eric Corrected history strings (and message) for the new set of calibration numbers. Then fixed it to get the correct "previous" values which are 2010 for 3C286 and 3C295 (and 3C196 sort of) but 1999 for 3C48, 3C147, and 3C138. Rearranged messages to show what is being done. Moved nowhere. 13995. October 18, 2012 KEYIN Eric Again! It did not check for length of strings at all the right places so a bad input file could cause abrupt aborts for bad formats. Added JTRIM on the input text file line - they now have to obey modern rules. Note that this is the AIPS version of KEYIN which had strayed far from the original CIT version and that version has also since the split evolved. But that version is still written with Fortran 2 rules. Tasks ANCAL, ANTAB, DTSIM, FETCH, M3TAR, MK3IN, PCLOD, SETAN, USUBA, UVFLG, VLBIN, INDXR, LDGPS, BLING, OMFIT, APCAL were re-linked. Moved nowhere. 13996. October 18, 2012 POSSM Eric Fixed labeling of multiple plots so that the horizontal axis gets labels only when the plot is at the bottom of a column. They were ok on the TV in non-fixed scale plotting, but not elsewhere where characters are by default larger. They overlapped things all the time when fixed scale plots were done. Moved nowhere. 13997. October 24, 2012 TVFLG, SPFLG Eric When the IO buffers got bigger, TVFLG's pointer set did not adjust properly. Bad things happened on CLIP INTERACTIVE - pieces of the master grid got messed up. Both TVFLG and SPFLG required modernizing the list of display types in the subroutine that does LIST FLAGS. TVFLG also messed up the NEXT IF/ST option by assuming there were 2 Stokes without checklng. This also made for bad addressing. Moved nowhere. 13998. October 25, 2012 KEYIN Eric Changed it to look at the length of the input key strings and compare strings only to that number of characters. PCLOD was failing. Moved nowhere. 13999. October 26, 2012 KEYIN Eric Further change - when a parsed string is too long it will now report it but not count it as an "error". It will ignore the long symbol and the rest of the line but not return an error code at the end of the execution (and not quit after 10 long symbols). With the number of bad KEYIN-data files I have seen recently this could be pretty bad. Moved nowhere. 14000. October 26, 2012 EXTLIST Eric AU8A needed updating for changes, some rather old, in plot task adverb lists. Fixed BPLOT, POSSM, ANBPL, CAPLT, CLPLT, FRMAP, FRPLT, UVHGM, UVPLT, and VPLOT. Reviewing all tasks in the list turned up TAPLT, SNPLT, ISPEC, RSPEC, and PLOTR. Fixed the number of adverbs list in PLGET (AU2A). Moved nowhere. 14001. October 26, 2012 Slices Eric The SL file format changed in an improper way (no proper version control) so modern programs read old SL files badly. Fixed AU8A EXTLIST - to test record 2 for old versus new form and interpret the models appropriately (new has physical values as well as pixel values of fit) TVSLIN Changed to test record 2 for old type and set the FQID values appropriately. Expanded units handling. TKSLIN Changed to test record 2 for old type and set the FQID values appropriately. Expanded units handling. MV2C06SL $DOCTXT: upgraded the documentation incl notes about older formats. SL2PL Changed to test record 2 for old type and set the FQID values appropriately. Expanded units handling. SLFIT Changed to test record 2 for old type and set the FQID values appropriately. Expanded units handling. Moved nowhere. 14002. October 31, 2012 EXTLIST Eric Added tasks CCNTR SERCH, EVAUV, DELZN, ELINT, FGPLT, GAL, SNIFS, SOUSP, and WETHR to AU8A for EXTLIST. Changed the tasks themselves to use the new plot number codes and changed the precursor comments in GINIT. Also changed: RFLAG Fill in IF defaults, fix UVRANGE units for plot files (and fix them back). CCNTR Fix defaults, deal with CLEV and PLEV both zero, reset multiplier when ST file is plotted. EVAUV Add a 10-value array at end of adverbs for plot file type and X and Y ranges. DELZN Filled in defaults in several adverbs. SOUSP Added some more extra parameters to give plot type, fit flux, rms ELINT Filled in defaults for some adverbs GAL Filled in defaults for DPARM FGPLT Filled in defaults. AU2A Added the new tasks for PLGET GINIT Updated the precursor comments. Moved nowhere. 14003. October 31, 2012 history Eric Changed CALHIS to compare CALSOU and SOURCS. If they are equal, then the HI cards call the adverb CALSOUR, otherwise it is called SOURCES. Upgraded the writing of history files to use CALHIS in tasks FRING, RLDLY, PCAL, BPASS, CPASS, CALIB, KRING, and LPCAL. Moved nowhere. 14004. October 31, 2012 UVCOP, TYAPL Eric Changed the maximum number of flags applying to s single time to 600000 in these two tasks (only). All others are 60000. Moved nowhere. 14005. November 1, 2012 EVLA.GAINS Brian Butler New EVLA gains files - retains frequency column but has gone to one frequency per band. The accuracy of the measurements was not felt to be good enough to justify multiple frequencies. Certain unreliable antennas were down-weighted in the solution or omitted. The curves were re-normalized to 50 degrees elevation. Moved nowhere - MNJ prep done at 8:45am however. 14006. November 1, 2012 EXTDEST Eric Added DOCONFRM adverb with value -2 to allow RUN and batch jobs to delete CL and HI tables version 1. It will ask interactive inputs for permission if DOCONFRM >= -1 and will skip it for non-interactive. Moved nowhere. 14007. November 19, 2012 POSSM Eric Corrected some of the bad effects from the last change to the labeling. Added a change to the labeling of the panels with IF numbers and polarizations. The old method ("IF nn (RR)") is used when the total number of panels in the X direction is less than 13. The new method, a simple IF number with the polarization above whenever it changes, is used when there are more panels (IFs, polarizations, NPLOTS in the X direction). A subtle shift of the number occurs when the number of panels exceeds 48 as well. Moved nowhere. 14008. November 19, 2012 BSCAN Eric NEW TASK. It examines the data for a selected source (calibrator usually) in a UV data set, trying to find the best scan to use for various calibrations. It first tries to find scans containing all antennas and if it finds some it returns an ordered list starting with the one with the most baselines and samples. The time range and number of antennas are returned for each selected scan. If no scan has all antennas, it looks for scans containing all but one. It returns the best of those and another scan containing the missing antenna (if any) along with the number of the antenna missing in the first returned scan. Moved nowhere. 14009. November 25, 2012 VPLOT, Calibration Eric Subroutines SOUFIL and SELINI assumed that there could be no more than 50 antennas requested even though the array is dimensioned MAXANT=90. VPLOT made that mistake a variety of places also. Moved nowhere. 14010. November 26, 2012 DOCENTER Eric Changed the default value of DOCENTER to 0 in POPSDAT and AU2A. This is better with IMVAL and okay for all the rest. Moved nowhere. 14011. November 26, 2012 MAXFIT Eric Changed AU9 and the MAXFIT help file to add output adverb SHIFT to give the shift needed to put a source on top of a pixel if one re-images. Note that the quirks of Clean may force you to iterate 2 or even more times depending on how accurate you wish to be. Moved nowhere. 14012. November 27, 2012 RLDLY Eric When doing the solution in dividing the IFs in halves or quarters, the task wrote over correct answers in half number 2 and quarters 2, 3, and 4 with values from the first gropup of IFs. I have a clear memory of fixing this previously, but apparently the fix did not make it into the system at that time. This bug has presumable been present since March 19. Moved nowhere. 14013. November 28, 2012 IMLOD, FITLD Eric Changed these tasks and subroutines PCHDR and PCCARD to also trap PVi_j FITS cards and to report the presence of the PV cards if not understood and to switch -SIN to -NCP projection when the PVs specify the right values. This enables WSRT images with modern headers to be read correctly (e.g. the WENSS postabge-stamp server). Rerquired changes too to CALRD, PRTTP, UVLOD, and AIPS subroutine FWRITE. Moved nowhere. 14014. November 28, 2012 control files Eric Updated the tab completion file HLPIT.LIS in $HLPFIL. Also the ABOUT files ZZCALIBRAT, ZZPLOT, ZZPROCEDUR, ZZTASK, ZZTV, ZZUV, ZZVLA, and ZZVLBI. Redid Chapter 13 of the CookBook (LaCook13.tex) and the web versions. Moved nowhere. 14015. November 28, 2012 POSSM Eric Changed it to use the new IF labeling on NPLOTS=0 also. Changed it to set the flux scaling on the first panel on a page and use that power of 10 for all plots on that page. This led to the discovery that the minimum tick interval was 1.E-5 and it really should be 1.E-6. Changed TICINC to implement this and TICSTR and ITICS to react to the new meaning of the variable usually called ITRY. Lots to link. Moved nowhere. 14016. December 3, 2012 EDITA, et al Eric The option to give a range of IFs to be flagged made it most places in the code but not quite everywhere - the output FG table did bif-bif rather than bif-eif where bif,eif is the IF range requested to be flagged. Killed one old line from QEDIUTIL. Moved nowhere. 14017. December 5, 2012 FRING, KRING, RLDLY Eric Added BIF, EIF adverbs to these three tasks, but they are used only when doing all IFs together as one, halves, or quarters. Added these adverbs to the tasks - all data are still used but the data in excluded IFs is flagged locally when read. RLDLY - corrected to keep track of overall weights of antennas in case that an antenna lacks data in some IF. This was necessary to handle the quarters and halves when one might be excluded. KRING - added H and Q modes to the N mode. This required using other variables to hold the working values of phase, delay, rate, etc and fixing up the data expansion/contraction routine. Changed the reference frequency used for delay to be the same for all sections of the spectrum. *********************************************************************** *********************************************************************** Changes while 31DEC12 is NEW *********************************************************************** *********************************************************************** 14018. December 5, 2012 UVCOP Eric The PDSEL routine used in copying PD tables from one UV data set to another had a bad variable causing it to zero the results. Moved to 31DEC12 this date. 14019. December 6, 2012 FARS, AFARS Eric FARS now gets rid of the STOKES axis on output - but this led one of the number of points on an axis to be 0 (an axis not actually used). AFARS did not test for this and so looped over nothing at all producing a constant, blanked image. Fixed this in both tasks. Moved to 31DEC12 this date. 14022. December 12, 2012 RM help Eric Cleaned up the help file and fixed the "limited to 4 frequencies" statement to the correct limit - 65536. Moved to 31DEC12 this date. 14027. December 17, 2012 FRING, RLDLY Eric Changed both to do N groups of IFs so long as N divides evenly into the total number of IFs. Moved to 31DEC12 this date. 14028. December 20, 2012 FRING, RLDLY Eric Corrected dimension of parameter list - it was adequate for up to 60 antennas but looked like it would work for MAXANT (90). Corrected the phase reference frequency to first IF of the group whenever IFs are grouped. Changed dimension of the variable to save the fits of groups so that it could handle more than 4 groups. Moved to 31DEC12 this date. 14029. December 20, 2012 VLBAEOPS, VLBACALA Amy Removed a bug from VLBAEOPS which was introduced with my June change. In June I removed the INFILE but there was still an if statement in VLBAEOPS that was based on the INFILE. VLBACALA change the SOLINT in ACCOR because it was set to 2 which is too long and averaged over scans, changed to -0.5. Also changed the clip parameters in SNSMO because 2Gbps data seems to need more leeway. Moved to 31DEC12 this date. 14030. December 21, 2012 FRING, RLDLY Eric Added test for DOEVLA = NumIF and forbid BIF and EIF in that case too. Moved to 31DEC12 this date. 14033. December 28, 2012 FRING Eric When it solves for multi-band delay and then blanks solutions for low SNR, it did not blank the MB delay. Fixed the routine which finds the MB delay to do this. Moved to 31DEC12 this date. 14034. December 28, 2012 CONVL Eric Fixed output header to declare no blanked pixels and then reset it if it makes blanked pixels. Moved to 31DEC12 this date. 14035. December 28, 2012 VTESS, UTESS Eric Changed to find the beam center for all images rather than assume that the beam center is NX/2+1,NY/2+1 for those that do not need a primary beam correction. Should be no harm... Moved to 31dEC12 this date. 14038. December 31, 2012 IMAGR Eric Changed QUVUTIL - OUVIMG was summing up the total ram needed for the requested images and, with less intelligence, for the requested beams. This sum could overflow a 32-bit integer. Changed the summing variable to LONGINT and changed the places where it was used to start by dividing by 1024 rather than doing that later. Moved to 31DEC12 this date. 14039. January 1, 2013 IMAGR Eric In QCLEAN, found 2 places that referenced IFIELD rather than the correct loop variable LFIELD when seeking the next facet to Clean in OVERLAP 2. Moved this correction only to 31DEC12. ----------------------- patches -------------------------------- 14050. January 16, 2013 smoothing after BP Eric The bndpass application was performed only for channels BCHAN through ECHAN, not the wider range required if smoothing were to be done after the bandpass cal. Fixed DATBND to do the wider range so that smoothing after the BP rather than just before can be done properly. Normally bandpasses are normalized so this was not so obvious. Moved to patch this date. 14054. January 22, 2013 POSSM Eric Corrected: (1) It did not know about "formal I" Stokes and so looped over non-existent Stokes values. (2) It did not initialize values for the upper plot in terms of scaling and axis prefix/label. I probably introduced that bug when trying to keep scaling constant on each page. (3) Changed it to specify the 2 antennas it needs when it reads the data each time. Previously it was reading/calibrating all data and then keeping only the needed baseline. Moved corrections only to patch 5 Feb. 14058. January 29, 2013 TICINC Eric The subroutine that computes tick increments was changed in late November to allow even finer tick intervals. The change was not properly carried to a piece of code computing round-off parameters, leading to excessive roundoffs and thus offsets in the plotted coordinates. Two minutes of time was a common such offset (0.5 degrees). Moved to patch this date. 14062. February 4, 2013 POSSM Eric The subroutine to plot BD tables ended with a call to TABBP rather than TABBD to clos the file. This worked sometimes but aborted under a variety of conditions. Moved to patch Feb 5. 14064. February 5, 2013 FITLD Eric Changed FITLD to leave DIGICOR set to the user's value independent of the name of the array. If that name is not VLBA a warning message is generated if the correlator is not DIFX or VLBA. Altered the wording of FITLD help file and the DIGICOR help file. Moved to patch this date. 14067. February 13, 2013 FITLD Eric The recent change to control digital corrections for any and all correlators also enabled delay corrections - basically the CQ table. Not all FITS-IDI files have the necessary info to do this. Changed the program to issue all the error messages but to keep going when the CQ table fails (setting DELCOR false after a failure). Moved to patch 18-Feb. 14078. February 19, 2013 CL2HF Eric A COMMON in the task was named ANINI causing an addressing failure when the subroutine ANINI was called. Renamed the COMMON. Moved to patch this date. 14085. March 1, 2013 PRTAB Eric With really large floating point numbers (e.g. frequencies) the floating point formatting could generate an illegal format. There is a simple fix. Moved to patch this date. 14087. April 4, 2013 COMB Eric Scaled the BPARM(1) and (2) fluxes in POLC by the apparent beam factor which could be extreme. Led to POLC doing nothing. Fixed it to keep Clean beam area factors at 1.0 for those images not being read. Moved to patch this date. 14092. April 5, 2013 AFARS Eric The correction to make it handle headers without having 0 points on an axis left it trying to have way more than 3 on the 3rd axis - when it knew it did not want that. Moved to patch this date. 14099. April 11, 2013 FITLD Eric The include DSOV.INC used to handle SO tables from FITS IDI declared a buffer to be 31+11*NIF in a comment and 200 in the code. This is 10 words short for 16 IFs which are common in VLB and many many words short for 90 IFs which AIPS supports. Fixed to the correct size = 34 + 11*MAXIF + a little. This caused issues when compiled with the Intel compiler at least in 64 bits, but not for all data sets (EQUINOX ended up -1 rather than 2000 with a bad error message). Moved to patch this date. 14110. May 3, 2013 COMB Eric Found a strange scaling happening when combining an image with JY/BEAM with one of some other units. Moved to patch this date. 14126. May 21, 2013 FITLD Eric The subroutine SO2SU had a format with more arguments than format translation codes. Fixed it into two lines so it will work correctly on all machines and not overflow the line. Moved to patch this date. 14154. June 17, 2013 BPASS Eric The code extended scans by 4 seconds in each direction while the comment said 2 seconds. Changed this to 0.4 sec since new scans can start quickly if the pointing position does not move. This caused data in the later scan to be thought to occur in the earlier scan and so to make a BP solution with a tiny amount of data in this artificially overlapped time range. That lousy solution would then get serious weight wrt the solutions with the intended integration times. Moved tp patch this date. 14163. June 26, 2013 SETJY Eric The code used the Perley 2010 coefficients for Perley 2013 because the code was copied and not corrected. It reported the Perley 2013 coefficients however. Moved to patch this date. 14171. July 2, 2013 IMAGR Eric IMAGR computed the average frequency by finding the average channel number and then interpolating in the list of frequencies. This fails if there is more than one Stokes (which that subroutine allows) or if the frequencies are not in a monotonic sequence. Moved to patch this date. 14175. July 7, 2013 IMFIT, JMFIT Eric When the Gaussian peak was negative, it was possible for the error bars to be negative also. Put absolute value on all error bars. This issue caused DOPOS and DOWIDTH to be returned as negative values so that the next fit held them constant. Moved fix only to patch. 14188. July 17, 2013 TRANS Eric The TRAROW subroutine that swaps rows did bad things when one wanted to reverse an axis - i.e. TRANSC='-312' which is a common request. The data were not reversed but were written to later in the output scratch file and so went off the end of that file. Moved to patch this date. 14189. July 25, 2013 PRTAB Eric June 10: The row selection (RPARM adverb) for integer-valued columns set the selection value to an integer. This prevents setting the selction value to say 2.5 and the accuracy value to 0.6 to look at rows with either 2 or 3 in the specific column. Changed it to keep the user-specified value as floating-point. jULY 25: Changed the scaling routine to read up to 50000 records. It read only 5000 before and that could end up finding no valid values in some of these new tables. Adjusted handling of the columns some including when no max.min found in the 50000 and placement of the 'INDE'. Moved to patch this date. 14192. July 29, 2013 SU table Eric The SU table has 21 columns (MAXSUC in PUVD.INC) and most places honor that. Found two subroutines GETSOU and SOUFIL and 7 tasks FILLM, FIXAL, HAFIX, TRUEP, UVFIX, FRMAP, and POSSM that did not. Put fixes only in UVFIX and POSSN which are different in 31DEC13. Moved to patch this date with large COMLNK file. 14206. August 16, 2013 SPFLG Eric Gridding counters can overflow (> 2 billion) so it died rather finishing - except that it ignored the error code and got in deep trouble. Counters -> double precision, error codes get checked. Moved to patch this date. 14213. August 23, 2013 PRTAB Eric Dropped the last digit of 2 E formats used when doing the F format display with columns requiring E format (e.g. delays). Moved to patch. 14234. October 21, 2013 SETJY, CVEL Eric Subroutine CVLDOP used by the above tasks called a low level routine with a hollerith argument when a character string was expected. It got away with this on some compilers but not the 32-bit Intel compiler. SETJY (OPCODE 'VCAL') and CVEL probbly got wrong answers on this architecture (date = 0/0/0 rather than correct). Error made March 2012. SETJY got the OPTYPE ' ' radio velocity wrong because it omitted SYSVEL. Added that offset. Moved to 31DEC12 this date. 14240. October 24, 2013 ISPEC, etc. Eric Subroutine FRQGET declared the number of columns in the FQ table to be 5. It should be MAXFQC from PUVD.INC which is now 6. caused executions with an FQID axis to fail in RSPEC, ISPEC, BLSUM, and SLICE. Moved to patch this date.