;----------------------------------------------------------------------- ; Copyright (C) 2008-2010 ; 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 31DEC09 is TST *********************************************************************** *********************************************************************** 12753. December 7, 2008 version Eric Changed 31DEC08 to 31DEC09 in: GETRLS.FOR INSTEP1 install.pl update.pl ZXHLP2.PL UPDCONTROL (comment) AIPSPATH.* Changed 16384 to 32768 for MAXCHA and MAXCIF, did corresponding change in ZPBUFSZ plus doubled to extra buffer size as well (BP tables need that). Moved nowhere. 12754. December 7, 2008 misc Eric Fixed $SYSLINUX/LDOPTS.SH to set STRIP=FALSE (now not a known option). Moved to 31DEC08 this date. 12755. December 8, 2008 ACCOR Eric Added code to trap the no data found case and set errors. Moved to 31DEC08 this date. 12756. December 8, 2008 DFTPL Eric Due to a typo, DFTPL plotted data only from spectral channel 1 including averaging channel 1 from BCHAN to ECHAN times. Moved to 31DEC08 this date. 12757. December 10, 2008 VLBAUTIL/VLBAPCOR Amy If INVERS was set in VLBAPCOR it was used when CLCAL was run as part the procedure. This was a bug and INVERS is now set to 0 for CLCAL. Moved to 31DEC08 this date. 12756. December 10, 2008 3C147_X Amy/Eric Installed the last standard image: 3C147 at X band reduced and imaged all by Amy. Moved nowhere (all versions get it automatically). 12759. December 10, 2008 PCCOR Amy/Eric Changed OPTIMIZE.LIS to have PCCOR not optimized. The Intel compiler gets 1 antenna wrong when optimized. Moved to 31DEC08 this date. 12760. December 10, 2008 TIMDEST, RENUMBER Eric Changed AU3A to inhibit TIMDEST (although I kept the messy code since it also does ALLDEST) and RENUMB to allow expansion of the CA file so that SLOT numbers greater than the current maximum are acceptable (no more than 25000 greater than the current max however). Moved nowhere - these do not go to new. 12761. December 11, 2008 AIPSASSN.*SH Eric Add site ALVOARN to special codes for VLA. Moved nowhere. 12762. December 11, 2008 ACCOR, CLCAL Eric Changed ACCOR to delete the new SN table when there is an error, such as no data found. Changed SNAPP to return error -1 on empty SN table or empty output SN table. Changed CLCAL to trap this error return and to quit promptly when no SN data were found. Moved to 31DEC08 this date. 12763. December 11, 2008 JTRIM Eric The decision to leave TAB characters in strings made 1/94 was wrong. Changed JTRIM to look more mdern and to convert TAB characters to blanks. Thus a TAB does not terminate a string like the other non-printing characters, but it no longer remains in the string. GMRT is sending source names with TABs and possibly other garbage characters and POPS cannot help the user with this. Linked FITLD, UVLOD, IMLOD - it will propogate further with time. Moved nowhere. 12764. December 11, 2008 POLRXY Eric A decision is made: the pole position will be in arc seconds not meters. The former is what is tabulated by the people who measure such things. The actual units may be told apart: if sqrt (POLRX^2 + POLRY^2) <= 0.6 then they are in arc seconds. Changed: SETAN help and fortran comments about units ANTINI changed comments and changed code to convert parameters in meters into arc seconds (divide by 30.8) Moved nowhere. 12765. December 12, 2008 UVFIX, HAFIX Leonia Investigation of the polar motion from 1979 till 2008 shows: 1. ROMAX/ROMIN<30.8; RO=SQRT (POLRX^2 + POLRY^2); 2. RO < 0.6 arcsec; 30.8 is coefficient of recalculation meters to arcsec. This fact allows to determine the units of POLRXY based on the value of POLRXY. The following codes are added to UVFIX and HAFIX IF (RO .GT. 0.6) THEN UNITS of POLRXY are meters; recalculate POLRXY to arcsec ELSE UNITS are arcses END IF Moved to 31DEC08 this date (where it is more useful). 12766. December 12, 2008 BPASS Eric One pair of arrays were dimensioned MCHAN=2048! Phil code! When dimensions MAXCHA they work better with Widar. Moved to 31DEC08 this date. 12767. December 12, 2008 FILLM Eric Changed: 1. For revisions >29 (post ModComp) the shadowing will be handled by FILLM rather than by using the erroneous on-line flag bit. 2. The array configuration will now be computed as best as possible by FILLM. The on-line data set depends on operator intervention and so is sometimes incorrect. This only appears in the FILLM history cards. Moved to 31DEC08 this date. 12768. December 12, 2008 TVCPS Eric Something got changed in the software that prints PostScript at the AOC - it claims to care more about following some sort of standards. Added some junk and took out the %%BeginProlog comment and now it seems to work. Magical thinking moved to 31DEC08 this date. 12769. December 14, 2008 UVDEC Eric Alternate ref pixel not adjusted. Moved to 31DEC08 this date. 12770. December 15, 2008 ZMIO Eric Changed max I/O request to 16 megawords. VBGLU was now asking for 1.5 megawords per half buffer. Fixed error message format too. Moved nowhere. 12771. December 18, 2008 TABSRT Eric The new BTBSRT can call ICSORT with a serious number of records. Changed the untested internal limit of 1024 to a tested limit of 20480 in ICSORT. Moved to 31DEC08 this date. 12772. December 18, 2008 TVCPS Eric 8 More commands were found to be out of order in a formal PostScript sense. The EndComments and EndProlog must come after the BeginSetup .. EndSetup that specifies fonts used. Moved to 31DEC08 this date. 12773. December 19, 2008 CALIB Eric Data were included in the average if the time <= the begin time + solint - they should not be on = so changed that to time <= begin time + solint - 0.001s. Moved to 31DEC08 this date. 12774. December 24, 2008 UVCOP Eric When ANTEN = 2,0; BASELIN=8,0, it copied baseline 2-8 as expected but also copied 2-2 if DOACOR. Droped the code that allowed this. If you want that result, set BASELINE=2,8. Moved nowhere. 12775. December 24, 2008 UVDIF Eric The CLIP only subroutine had an erroneous subscript KK rather than K which caused it not to find bad amplitudes although it reported them in the max difference over max flux value. Moved to 31DEC08 this date. 12776. December 24, 2008 imaging problem Eric The change that allowed subimages to be padded out and used for modeling has caused nasty problems with non-3D imaging (i.e. de Sade's principle). Changed: GRDAT to call IMGSIZ to set FLDSZ only on DO3D true SETGDS to defer the setting of FLDSZ until the full set of images are examinined and the DO3D value determined. Saves the IMGSIZ outputs and the image size values in a temp array for this. Must move to 31DEC08 but this should delay a final freeze! 12777. December 26, 2008 VBGLU Eric The task had 3 main problems: 1. When disk I/O had to be restarted because the required UV record preceded the ones in RAM, the task checked a few of the upcoming records to find the minimum needed input record number. Thus the first record number in the new buffer could be one that preceded the currently required one. Unfortunately, the pointer was set to the first in the new buffer always. This led to wrong baseline data being included in the output record. Raised the search region and made sure it did not search over empty records and made the pointer aim at the desired record. 2. On AT tables, the task asks for a SOURCE column that is not present in the table and so quit on AT tables. Added a special error test for this, allowing it to procede normally since the copy can handle a missing selection column. 3. The handling of character-data columns was incorrect. So far as I can tell, the error only damaged CQ tables. Corrected all the length handling for character-valued columns of 1 or more dimensions (includes AT, CT, and SU tables). Moved to 31DEC08 this date. 12778. December 26, 2008 ACFIT Eric Was setting CLVER not CLUSE from adverb GAINUSE and so always applied the highest CL table. Moved to 31DEC08 this date. 12779. December 29, 2008 AIPSLetter Eric Put back the TeX, ps, and pdf forms of the 31Dec AIPSLetter. Moved to all versions automatically. 12780. December 30, 2008 FILLM Eric Changed the shadowing again: if < 0, no shadow flags; if = 0, use system bit for versions 20-29, and >=34 or use 25m for versions <20 and 30-33; if > 0, use specified value with shadowing computed in FILLM. Added an extra test on VLAOBS to help online FILLM stop when requested and the observation code has changed to one that is not desired. Moved nowhere. 12781. January 1, 2009 VLA.ant.2008 Eric Install the 2008 baselines file under its new name in $AIPSTARS and fix the 2007 file to have the copyleft. Moved to all versions automatically. 12782. January 1, 2009 CALIB Eric The business of dividing a scan up somehow into intervals when the actual integration time and the user's SOLINT may be an exact match or not is simply a mess. The existing code where I add a -DEPS could cause an infinite loop if the scan integration time is zero which is legal although not usually correct. It could also cause it to do an extra solution on the last sample of a scan. Changed it to fix the integration end to be > scan end when it was computed to be exactly the scan end to avoid both problems. Deleted a bunch of old code that has been commented out for a while. Moved to 31DEC08 this date. 12783. January 2, 2009 VLBAUTIL (VLBACPOL) Amy Several minmatchs still existed in VLBACPOL. Replaced. Moved to 31DEC08 this date. 12784. January 5, 2009 UVW scaling, IMAGR Eric The calibration routines do not correct UVW for a change in the header frequency even though they do issue a new header frequency. Checked over all the code for errors and was relieved to discover that UVCOP, SPLIT, and SPLAT among others did this right. IMAGR attempted to do it right too but a recent (2008) change to GRDMEM made some error: BASAVG $APLNOT subroutine that appears not to be used - changed it to scale output UVW GRDMEM $QNOT: the UVW scaling for the first channel was used before it was computed - so first call to GRDMEM would make an error but later calls would have it okay. Affects IMAGR, SCMAP, SCIMG. DECOR Already did scaling UVLIN The TRUEF used for phase shifting should have used UVFREQ since it is based on the input FOFFs rather than FREQ which is the BIF/FQID selected new frequency. Add UVW scaling but after function call since that deals with input UVW correctly. UVLSF As UVLIN UVLSD As UVLIN UVMOD As UVLIN Simple addition of UVW scaling: AVSPC DIFRL FUDGE PHSRF TI2HA TYAPL UVAVG UVBAS Moved to 31DEC08 this date. 12785. January 5, 2009 FLOPM Eric Added OPTYPE = 'IFTO' to flop the IFs as well as the channels. In this case, many IF-dependent tables are not copied but some are: SU, CL, FG. Moved nowhere. 12786. January 5, 2009 PCCOR Leonia Add the print out, if the number of tones per IF/POLZN less than 2, including the case of 2*NIF*NPOLZN .GT. 16 (number of required PC detectors exceeds the number of available ones (16)). Moved nowhere 12787. January 6, 2009 CALIB Eric This business at the end of scans is still getting us. Changed the latest changed code so that is the current interval is within 60% of SOLINT of the end of the scan, the end time of the interval is made significantly larger than the end of the scan. (Note that making it 3 milliseconds larger than a scan in day 3 was for some reason not enough.) Moved to 31DEC08 this date. 12788. January 8, 2009 CALIB Eric Yet again! The NX table contains center times and intervals which are in single precision floating point. It is thus able to indicate an end time that is actually less than the real end time by more than a few milliseconds. Added an error margin of 0.2 seconds which will still be inadequate if the times get into several days or more. The handling of NX records with subarray and freqid that do not match the current ones was incorrect. An NX table can contain overlapping scans with different subarrays and/or freqids. Changed the NX reading code to accept only matching scans and dropped the tests on data matching NX record subarrays/freqids since that would cause pre-mature reading of a new NX record in the case of overlaps. The data from UVGET will always be only the desired subarray and freqid and therefore does not need examination. Added a long section to the help for SOLINT. Moved to 31DEC08 this date. 12789. January 8, 2009 FRING Eric FRING does much the same as CALIB in trying to break up intervals in scans. I fixed the handling of freqids and subarrays in it but left the old form of time management in which the intervals start at regular times without regard for when the data occur and times are expanded outwards a good bit. This is more likely to be desirable in FRING then in CALIB's self-cal modes. Moved nowhere. 12790. January 9, 2009 CALIB Eric The subarray got lost so the SN table did not have a subarray value and so the next subarray deleted the previous one. Moved to 31DEC08 this date. 12791. January 14, 2009 RUN help Eric The help file for RUN was not sufficiently clear. For VERSION blank, the file $RUNFIL/name.usr is sought followed by $RUNSYS/name.001. For version = 'xxx', the files $xxx/name.usr and then $xxx/name.001 are sought. Moved nowhere. 12792. January 14, 2009 SPLIT, SPLAT Eric Fixed to close NX tables when there is a problem with a source so that they can be written for the next source. Did not deal with a bad output file (e.g. no data found) when there are also good output files. If there are no good output files, then the problem is handled. Moved nowhere. 12793. January 16, 2009 SNPLT Eric Added DOBLANK option to plot location of blanked solutions. It will use a different line type on GRCHAN=0, DO3COL<=0 and so be a different color on the TV. Moved nowhere. 12794. January 19, 2009 VBGLU Eric Added code to control the sort order in the output header. It will almost certainly end up 'T' although 'TB' is possible for 2 fully sorted data sets with all matching visibilities. Moved nowhere. 12795. January 26, 2009 ISPEC, BLSUM Eric Add the concept of a slice of an area in XY over the Z pixel. ISPEC add code to save spectrum as a SLice file optionally SL file format changed to tell is FLUX or brightness Add adverb to help. BLSUM add code to save spectrum as a SLice file optionally SL file format changed to tell is FLUX or brightness Add adverb to help. SLBINI add whole 2nd type to handle slice on axis 3 SL2PL pass on BLC/TRC(3) to image catalog (misc words) also trap FLUX units from slice file PFPL2 force TRC(3-7) = BLC(3-7) TKSLIN save BLC/TRC(3) in image header, flux units TVSLIN save BLC/TRC(3) in image header, flux units SLOCIN pick up BLC/TRC(3) from image header for SLBINI LABINI If 3rd axis involved in slice coordinates add it to centered on string AU6E pick up BLC/TRC(3) from image header for SLBINI AU9A change TKPOS and TKXY to return correct values for the spectral coordinate and central values for XY in the case of the new type slices SLFIT add code to handle spectral case on input and output. Add output adverbs (a la IMFIT) to return the answers. Control flux units. Save the physical units answers and units strings in the SL file. Changed help to adverb returned adverbs. SLICE Don't write all the inputs into the SL file record 2 and set OPTYPE there to AVER. AU8A Tell about fit models using newly saved parameters and the rms for SLices. MV2C06SL Revise format description Moved nowhere. 12796. January 27, 2009 CALIB Eric The fix for solution interval getting set to zero was not made for the first record in the NX table only for all later ones - so of course that bit us. Moved to patch this date. 12797. January 28, 2009 64-bit Linux Eric Create a new architecture - Linux compiled for 64-bit machines using the latest Intel compiler on rishi. AREAS.CSH Add APLLNX64, INXLNX64, SYSLNX64 AREAS.SH Add APLLNX64, INXLNX64, SYSLNX64 AIPSPATH.CSH Add LNX64 case for setting library path AIPSPATH.SH Add LNX64 case for setting library path LIBR Add LNX64 $ARCH for standard ar methods START_AIPS Add LNX64 in group using "host" 2 places TVALT Add LNX64 case in search for X server ZLPCL2 Add LNX64 to group using "lpr" ZLASCL Add LNX64 to group using "lpr" STARTPMON Add LNX64 to group that is not system V LINK Add LNX64 with LINUX to options reversed group AIPSWC Add APLLNX64, SYSLNX64 in word counts MAKE.MNJ Add LNX64 to list of $ARCH MAKE.BMNJ Add LNX64 to list of $ARCH UPDUPDATE Make cases to fix SYS${ARCH} for AXLINUX, LINUXPPC UPDCONFIG Add RISHI for AOC site UPDREMOVE Add LNX64 case for TSTFLG UPDERROR Add LNX64 to group with -s grammar for "mail" UPDCONTROL Add LNX64 to group with -s grammar for "mail" AIPSUPD Add kluge to copy LIBR areas to 2nd set install.pl Add LNX64 case in places, detect linux 64. Moved contents of $SYSLINUX and $APLLINUX to $SYSLNX64 and $APLLNX64 and then changed: ZDCHI2 UNIX, Linux 64-bit in strings FDEFAULT Fix COMPILER and ifort options LDOPTS Fix COMPILER and ifort options CCOPTS Fix COMPILER and icc options LIBR.DAT Change: Z routines go to $LIBR/APLSUB, link edit lists contain only $LIBR/APLSUB once rather than 2 each of APLSUB and APLLNX64 COMRPL.ADD APLLNX64 INCS INCLNX64 Moved nowhere. 12798. January 29, 2009 LNX64 Eric Further changes: INSTEP2 was fooled by $LIBR/APLSUB/SUBLIB existing before APLGEN was to compile. Removed test on SUBLIB existing and do only check on are all files compiled in list. AIPSUPD remove kluge copying LIBR files - solved by having Z and APLSUB in same library AIPS updated copyleft XAS.SHR Update Makefile for LNX64 Moved nowhere. 12799. January 29-30, 2009 TABSRT, LNX64 Eric OPTIMIZE.LIS Put ZSTRTA in level 0 for LNX64 - it worked on primate and hamal, infinite loop on pilabo! ICSORT Length 20480 was not enough for some FITLD cases - used to use 1024 untested! Now 204800! Moved nowhere. 12800. January 31, 2009 EDITA Eric The FLAG ABOVE and BELOW commands behaved oddly - when the flag all polarizations was turned on they found high/low points in both polarizations even when only one was displayed. Changed them to behave as FLAG AREA which controls the finding outside of EDFLAG. Moved nowhere. 12801. January 31, 2009 ZDCHIN Eric Changed the default max ap to 1 Gbyte for 64-bit machines. It is too dangerous to let the default get way large. Moved nowhere. 12802. February 2, 2009 TYAPL Eric Help file initial description was for TYSMO. Moved nowhere. 12803. February 4, 2009 TYAPL Eric It applied the FG table since the SELINI default is FGVER=0. Moved to patch this date. 12804. February 23, 2009 QINIT Eric Fixed the error message after ZMEMRY failure to report Kilo words with an adequate format. Moved nowhere. 12805. February 23, 2009 DELAY Eric New verb to delay DETIME seconds. Moved nowhere. 12806. February 24, 2009 DIFUV Eric Added OPTYPE = 'DIV' option to make gains rather than a difference of 2 UV data sets. Moved nowhere. 12807. February 25, 2009 USERID adverb Eric The USERID adverb is now obsolete for most applications since it is not required to separate multiple users inside a single catalog file. That old construct was forbidden a couple of years ago in install.pl, FILAIP etc. since not all tasks and verbs supported it even though they should have. Did the verbs first fixing Fortran: AU2A AU3 AU5 AU5A AU5D AU6D AU6E AU7 AU7A AU7B AU8 AU8A AU9 AU9B AUT and help files: ACTNOISE ADDBEAM ALTDEF ALTSWTCH ANTNUM AXDEFINE CATALOG CELGAL CHKNAME CLRNAME CLRSTAT COODEFIN COPIXEL COTVLOD COWINDOW EGETHEAD EGETNAME EPOSWTCH EXTLIST GET2NAME GET3NAME GET4NAME GETHEAD GETNAME GETONAME GETTHEAD HINOTE HITEXT IMDIST IMHEADER IMSTAT IMVAL MAXFIT MAXTAB MCAT PCAT PLGET PRTHI PUTHEAD PUTTHEAD PUTVALUE QHEADER QIMVAL RENAME RESCALE SCANTIME STALIN T1VERB TABGET TABPUT TKAGUESS TKAMODEL TKARESID TKASLICE TKGUESS TKMODEL TKRESID TKSLICE TVAGUESS TVALL TVAMODEL TVARESID TVASLICE TVCUBE TVDIST TVFLUX TVGUESS TVLOD TVMAXFIT TVMODEL TVMOVIE TVNAME TVRESID TVROAM TVSLICE UCAT ZAP Moved nowhere. 12808. February 25, 2009 EDITR Eric Changed the limit on number of sources from 100 to 10000 in QEDIUTIL. It was making bad source numbers in the output FG table and not showing the source name(s) when it should. Moved nowhere. 12809. February 25, 2009 USERID in tasks Eric USERID is needed in DISKU, PRTAC, and MOVE. Dropped it from $APLPGM tasks: COMB CORFQ MCUBE MF2ST PBCOR PRTAB PRTCC PRTIM PRTPL QMSPL SAD SLFIT SQASH STFND SUBIM SUMIM TAFLG TKPL TRANS UVDIF Moved nowhere. 12810. February 26, 2009 USERID Eric Plot files have a problem with dropping an adverb, especially at the beginning with both EXTLIST and PLGET. So retain USERID adverb but mark it unused in the help files and set it to NLUSER in the code. Tasks done this way: CCNTR CNTR IMEAN ISPEC PCNTR PROFL SL2PL TAPLT ANBPL BPLOT CLPLT DFTPL FRMAP FRPLT GAL IRING LOCIT PFPL1 PFPL2 PFPL3 PLCUB PLOTR PLROW POSSM UVHGM UVPLT VPLOT CONPL Y area tasks that can simply drop USERID (several did not use it anyway!) are LWPLA TVHUI TVPL TVRGB IBLED IMLHS SPFLG TVFLG TVHLD WIPER Q tasks had USERID dropped: BPASS CPASS Moved nowhere. 12811. February 27, 2009 CALIB, SNPLT Eric Changed CLBPA to count up the occurrences of unblanked data for each antenna/IF/polarization. Blanked data are then not counted as failed solutions either when the full solution fails or when it is just the individual antenna that "failed". Changed SNPLT to plot, when DOBLANK > 0, fully failed SN/CL table records. Previously this option plotted the failures only when some solutions in the record were good. Moved nowhere. 12812. February 27, 2009 Drop USERID Eric Dropped USERID from $APGOOP tasks CCEDT CCGAU CCSEL CL2HF HF2SV HFPRT MBDLY POLSN TBDIF TBSUB TBTSK Moved nowhere. 12813. March 2, 2009 Drop USERID Eric Dropped USERID from $APGNOT tasks - note that quite a few of these did not actually use the adverb that they pretended to support. ACFIT BDEPO BLWUP BPSMO CANPL CCFND CCMRG DCONV GSTAR IMFIT JMFIT LGEOM LISTR MATHS MEDI MFPRT MOMFT MOMNT OBEDT PADIM PGEOM POLCO PRTAN PRTSD PRTUV REGRD RGBMP RLDIF RM SHOUV SKYVE SLCOL SLICE SMOTH STARS STFUN STRAN SUMSQ TAMRG TASRT TBIN TBOUT TXPL UVFND UVHIM UVHOL UVIMG UVPRM UVPRT UVSEN WARP XTRAN Moved nowhere. 12814. March 2, 2009 CL2HF Eric The GNU compiler aborts optimizing this task. Changed OPTIMIZE.LIS to add LNX64 to LINUX OS in downgrading this task. Moved nowhere. 12815. March 3, 2009 UVMOD Eric Changed it to offer the option to read components from a text file, up to 9999 of them. This will allow for more general models including all types and spectral index. Moved nowhere. 12816. March 3, 2009 USERID Eric USERID remained erroneously in CCSEL and WIPER help files. Moved nowhere. 12817. March 3, 2009 AVSPC Eric Plugged a hole in which a user could ask for AVOPTION 'SUBS' with CHANNEL > number of input channels. This set the number of output channels to zero and proceeded along merrily. Fixed to go to mode ' ' with 1 output channel and a message. Moved nowhere. 12818. March 4, 2009 IMAGR Eric Add IMAGRPRM(18) to limit the number of facets that will be imaged together when trying to find the next facet to Clean. The defaults which depend on work-file size, image size, max AP size (as estimate of RAM) may not be optimal and may cost rather than save time. Moved nowhere. 12819. March 4, 2009 AN Table Eric The AT table format is specified to have pol cal values dimensioned (Npcal, Nif). Corrected DATV.INC and FITLD which reversed these subscripts. The AN table is specified to have pol cal also dimensioned (Npcal, Nif) where Npcal is always 2 (or 0). Numerous tasks got this wrong and ANTINI requires the creating task to provide these values. Fortunately, PCAL and friends fixed this bad dimensioning but it would be good to get it right. Changed call seq to OANINI in TABLEUTIL so that the pol cal will have the right dimension. FXPOL fix to find max IF number out and force that into new AN table INDXR, LDGPS, OBTAB, DTSUM - use dummy variable for numif UVSIM Change Npcal from 3 to 2, there is 1 IF UVFIL Change Npcal from 3 to 2, there is 1 IF UVCON Change Npcal from 3 to 2, there is 1 IF BSMOD Change Npcal from 3 to 2, there is 1 IF SETAN Pick up Nif from header, change Npcal from 4 to 2, change dimension from 3 to 2*Nif OTFBS Change Npcal from 3 to 2, there is 1 IF OTFUV Change Npcal from 3 to 2, there is 1 IF SDTUV Change Npcal from 1 to 2, pick up NUMIF to set correct pcal column dimensions IM2UV Fix col dimension from 3 to 2, Npcal already 2 UVLOD In DUMPAN and ANTREC, simply set Npcal 2, Nif 2 for AN file (VLA-only formats), in ANTAB and UVFEXT do better with picking up NIF and setting Npcal to 2 etc FITLD Pick up NIF and use it in setting up ANTINI calls. VLBIN Pick up NIF and use it in setting up ANTINI call. SPLIT Zero the polcal values in the output antenna, using the correct column dimension, not 2 SPLAT as SPLIT. Moved nowhere. 12820. March 11, 2009 DFTPL Eric Changed the time averaging to include up to but not T1 + BPARM(2). Added some hints to the help file. Moved nowhere. 12821. March 11, 2009 LWPLA Eric Two places did a draw of a full image background - on an init line when a background is requested and on an init color grey-scale when no background is requested. Both used the coordinates of the bounding box, but only the first one came ahead of the scaling command. Fixed the second to scale the bounding box so that it would make a full area box. Moved nowhere. 12822. March 12, 2009 XBASL Eric The task had no defaults for BOX and so never found any valid pixels to allow for a fit. It also did not set a subscript when not writing images of the fit parameters, causing an address to get trashed and an abort. Fixed these and made the setting of fitting windows friendlier. Added some EXPLAIN to the help file and clarifications about NBOXES and BOX. Moved nowhere. 12823. March 19, 2009 UVCON Leonia The sign of W term contribution to the phase has been changed to match with the UVMOD and IMAGR. Moved nowhere. 12824. March 26, 2009 VLANT Eric Persistent reports of aborts on MACINT laptops running the latest operating system have been received. We have been unable to replicate this on any of our Macs including a laptop run from home. But changed OPTIMIZE.LIS to compile VLANT with no optimization on MACINTs to see if that helps. Moved nowhere. 12825. March 27, 2009 UVLOD, FITLD Eric Changed DFUV and VFUV includes to add SORTORD as a synonym for SORTORDR and adjusted the code in UVLOD and the UVLOD section of FITLD for the option. Moved nowhere. 12826. April 1-2, 2009 OFMs Eric The files in $AIPSOFM having various OFM coloring schemes needed to be updated for the new depth of a plane (0-2045) from the old depth (0-255). Did this with AIPS itself on some and with a stand alone program for others. Fixed those with abrupt changes by hand also. Put them back as *.000 to allow 31DEC07 to read the old ones. Changed OFMDIR and OFMGET to handle the new names and possible format revisions for the larger intensity numbers which can go to 16*2046-1. New OFMs are CIRCLE and CIRCLE2 and GREY100W. Reworked OFMS are: BLACK BPHLAME BPPHLAME COLOUR16 COLOUR8 COLWED16 COLWED8 COMB2152 DEMO3 EYEBW FOTOBW GCPHLAME GREY01W GREY02W GREY04A GREY04B GREY04C GREY04D GREY04W GREY08A GREY08B GREY08C GREY08W GREY128W GREY16A GREY16B GREY16W GREY256 GREY32A GREY32W GREY64W GREYLOGA GREYLOGB GREYLOGC GREYLOGD GYPHLAME RAINBOW ROB1 ROB2 ROB3 ROB4 RPPHLAME RYPHLAME STEP10 STEP64 STEP8 STEPS8 WHITE ZERO Moved to patch this date. 12827. April 6, 2009 UVCOP Eric Doubled the max number of flags applying to one time to 120000. Moved nowhere. 12828. April 8, 2009 SPFLG, TVFLG Eric Subtle errors can arise when making the master grid which cause time-averaged displays to omit data. The source number may not appear in the master grid if the planes gridded in that pass through the data had no valid samples at the time of the row even though other passes did have valid samples. This is rather more likely in SPFLG than TVFLG. Changed SPFLG and TVFLG to pass the array of source numbers which are known to the grid subroutines GRIDTB and GRIDTC (latter in SPFLG) which then pass them to GTBWRT. Changed UVIMG to provide an array for source numbers and call GRIDTB with the new call sequence although in that case they get filled with -1 as before. Changed limits from 16384 times to 32768 times. Moved nowhere. 12829. April 10, 2009 LISTR Eric Made displays subtly nicer: 1. Help file to document phase self-scale better (DPARM(7)), add DPARM(8) to LIST each baseline twice, document new usage of ANTENNAS/BASELINE where non-existent telescopes are not included. 2. Fortran: add subroutine GETANS replacing GETNAN to get list of those antennas used in the SUBARRAY (incl subarray 0 for GAIN) and max antenna number. GETANS includes only antenna numbers that occur in the antenna file(s) and only if their names are not blank and do not include the string OUT. Only antenna numbers returned by GETANS are displayed even if others occur in adverbs ANTENNAS or BASELINE. 3. Changed format of column titles to I2.2 for antenna numbers and made titles align with right hand edge of columns. 4. Added decimal places to header scaling display which could say 1000 = 0.000 Jy with a wide enough format. 5. Fixed column listing to admit that a baseline could occur twice on the same page. This was allowed rather generally in LIST previously and now is controlled by DPARM(8). 6. Fixed gain header display to account for self-scaling in phase. 7. Dropped odd difference in spacing for DOCRT true in gain displays that forced interactive gain displays to use > 1 pass. Moved nowhere. 12830. April 10, 2009 FITLD Eric A pointer that was not reinitialized caused the image header to be copied from a saved header when creating a UV file from a FITAB-written tape on the 2nd and later data sets being loaded in a single execution of FITLD. Init the pointer in the right place. I think the only really bad result of this was the loss of the prior name and sort order information. Moved nowhere. 12831. April 11, 2009 CMODEL Eric Changed handling of CMODEL, adding it to a number of tasks. SETGDS Changed to suppress CC error messages if input MODEL=0 and return no error code in that case while setting MODEL=2 for image. FACSET Remove messages about type of image except when normal Clean or dirty. CALIB Change messages before first FACSET call and made default MODEL = 0. LPCAL Add CMETHOD and CMODEL with now normal defaults. Modified help re CPARM(2). PCAL Add CMETHOD and CMODEL with now normal defaults. UVCON Changed default to 0 for MODEL, added messages. UVSUB Changed default to 0 for MODEL, added messages. Following tasks: added adverb CMODEL with default 0 to allow either type (CC preferred). Add messages about model type ahead of first FACSET call. BLCAL BPASS CPASS FRING KRING The plot tasks CLPLT, IBLED, and VPLOT use CCs only and UVADC is also geared solely for CC models. Moved nowhere. 12832. April 21-22, 2009 IMAGR Eric Found error in QCLEAN with FORCE A FIELD option that caused all images to be remade rather than just the requested one. Found error that caused scratch disk files to be way too large for gridded subtraction with DO3D false. Added FFTIM to GRDSET to set size directly from image for FFTIM false and to use IMSIZ which uses the reference pixel too for images. Changed GRDSUB for this call seq change. Moved to patch April 22. 12833. April 25, 2009 Imaging Eric Leonia figured out a different mathematical approach to multiple facets that allow the facets to be computed all on the tangent plane at the UV phase stoping point. This is about like our old DO3DIM false, except that the samples are gridded at corrected values of u and v, with the correction depending on w and the direction cosines to the center of the facet. I have installed this as an improved DO3D false. This means that u,v depend on facet which makes the assumptions in UVGRID and UVGRTB incorrect. Removed from the system: MX and HORUS (leaving both help files to point at IMAGR and to continue to provide their EXPLAIN information). Aldo dropped subroutines OLDMAP, UVGRID, UVGRTB, GRIDUV, UBTBGD, and ALGSUB all of which had trouble with the 3D concepts which are now in DO3D false as well. Other changes made: IMAGR Dropped forcing DO3D when multi-scale and/or > 64 facets. Number of beams always same as number facets. Changed to compute max U and V via matrix for both values of DO3D with false calling P2DMAT. Added adverb ONEBEAM to select a single beam for multiple facets (per resolution). IMAGR Help: added ONEBEAM, changed discussion of OVERLAP, mention more accurate DO3DIM false. P2DMAT New subroutine, computes L and M of facet center and makes appropriate UMAT to get shifted U,V,W from input U,V,W. PMAT return = identity matrix Added rotation to both. PRJMAT Fixed RFILL call to DFILL. CLEAN Get NUMRES and ONEBEAM so as to make only necessary beams QCLEAN Add object IMAGE(1) keyword ONEBEAM - make one facet at a time vs all facets at once depend on it and the choice of which beam in the beam histogram and beam to AP load routines. (note need to do beams 1, 1+nfield, 1+2*nfield, etc for multiple resolutions). QUVUTIL Add object IMAGE(1) keyword ONEBEAM and set number of beams and pointer to beam number accordingly (note need to do beams 1, 1+nfield, 1+2*nfield, etc for multiple resolutions). SETGDS Call P2DMAT if DO3D false - loop to redo UMATS computation if DO3DIM ends up not the same as on the first time through and UMATS needed. Use subimage sizing code only on input model 3. Subimage code never chosen even when CCs missing. SETDO3 As SETGDS. UVPROT Do proj multiply in all cases, drop DO3DIM from call sequence DGDS.INC Put parameter ONEBEM in also - use in MAKMAP MAKMAP Drop call to GRIDUV, use only GR3DUV GRDMEM add call to P2DMAT to make non-3D UMAT instead of simple identity matrix, use PMAT to rotate x,y,z GR3DUV add call to P2DMAT on non 3D and use UMAT on all 3D to adjust u,v, use PMAT to rotate x,y,z GRDSUB Changed call seq adding MODEL, dropped ALGSUB possibility, changed call seq to GRDSET and FFTIM GRDSET Change call seq again, passing MODEL and using model 3 to call IMGSIZ ALGMEM Do matrix multiply on uvw always, with P2DMAT for non-3D ALGSTB Change UVPROT call seq, do matrix always using P2DMAT for non-3D, save u,v,w in local array and restore from that rather than from re-rotating which introduces numerical issues. FACSET Allow model 3 == model 2 UVMSUB New call seq GRDSUB FFTIM Add MODEL to call, model 3 calls SUBGET, else PLNGET. VISDFT Do matrix multiply on uvw always, with P2DMAT for non-3D VSCDFT As VISDFT VSFDFT As VISDFT UVCON Fix local VISDFT as general one. SCIMG Add ONEBEAM adverb, replace the DO3DIM use with ONEBEM. MX Inputs/Help file cleaned out with explanatory text, but EXPLAIN left. HORUS Inputs/Help file cleaned out with explanatory text, but EXPLAIN left. UVSUB Incl help, changed CMODEL code: 'SUBI' => model 3 POPSDAT New adverb ONEBEAM NEWPARMS New adverb ONEBEAM ONEBEAM New help file DO3DIMAG Change help file Moved nowhere. 12834. April 26, 2009 Additional imaging Eric Changed: CLEAN Fix handling of NUMRES which is an optional keyword. IMAGR Beam size when one beam should start at MAXNX/NY SCIMG Be more careful about user's IMSIZE. Pass CLEANO down to local SCUV2S to get shifts from CLEANO and then use them as IMAGR does to find max u,v (omits points that fall of the grid for any facet). Tasks that use CMODEL: BLCAL, BPASS, CALIB, CPASS, FRING, KRING, LPCAL, PCAL, and UVSUB. Moved nowhere. 12835. April 27, 2009 DBCON Eric Changed KEYCOP call to KEYPCP to block MAXBLINE and MAXABSU keywords from being copied into the output data set. Moved to patch this date. 12836. April 29, 2009 CALIB.HLP Amy Added section on how to get and use flux calibrator models. Moved to patch this date. 12837. April 30, 2009 MCUBE Eric Put images on output if the expected plane matches the computed plane but issue a message unless they are within 0.02 pixel of an integer - was 0.2 which is way too far off. Moved nowhere. 12838. May 2, 2009 TV roam Eric The TVROAM function leads to problems with the TV display. Some functions can handled a roamed screen (split and scrolled) and many cannot. Added protection against the bad things that the latter encounter/cause while adding a verb that will allow the "final" roamed image to be captured as a simple TV image instead. Corrected the others to handle scrolled images and where possible roamed ones. Changed: POPSDAT.HLP Add ROAMOFF verb, delete OFFROAM procedure NEWPARMS.001 Add ROAMOFF verb GRPOLY Work in presence of scroll (graphics to scroll like grey scale and allow scroll correction, plus special tricking of YCUCOR) GRBOXS Work in presence of scroll GRBOXD Work in presence of scroll GRSLIC Work in presence of scroll TVFIND Test for different images on in different quadrants and return error. Suggest ROAMOFF if it appears to be roam mode. This will block all the verbs/tasks that do not work correctly on a roam-state TV. AU5A Wrote new verb ROAMOFF. Corrected TVROAM handling of 1xN and Nx1 roams - used bad windows - and changed it to zero the image planes before writing. Block TVLOD on roamed-state TV. YSCROL $YSS: Commented out desirable code and put in code that recognizes that XAS has only 1 scroll. XAS.SHR Changed screen.c to use upleft_x,y[0] for the window corners (zoom/scroll) commenting out the attempt which cannot work of allowing separate scrolls for separate channels (now no longer used except in TVSCROL. TVROAM.HLP Adjusted description of what works, mention ROAMOFF REROAM.HLP Altered TVSCROL.HLP Add info about XAS does all together ROAMOFF.HLP New modest help file. OERROR Added TV ROAM MODE error number 77. AU6E Test TVLIMG and quit with 77 all opcodes. AU6D Test TVLIMG and quit with 77 on TVSTAT AU6C Test TVLIMG and quit with 77 on TVFIDDLE AU6A Test TVLIMG and quit with 77 on all but TVTRA, OFFTR, GRBLINK AU6 Test TVLIMG and quit with 77 on all but TVPS, TVPHLAME AU5E Test TVLIMG and scroll and quit with 77 on those that use the graphics AU5 Block TVON, TV3COLOR with error 77, code to find TVNAME when roamed OFMGAM Shift lettering by TVSCRX/Y for scroll. IENHNS Shift lettering for scroll - watch edge trouble TVALUE Shift lettering for scroll - watch edge trouble. Dealt also with odd problem that had XAS thinking graphics channel 2 went off when it should not have done so - happened only in roamed images. Moved nowhere. 12839. May 5, 2009 CLCOR, RLDIF Eric Changed CLCOR Fortran to allow OPCODE='POLR' when no valid PCAL solution is found in the AN file. CLCOR Help to describe the POLR and PHAS opcodes more correctly - both can do a range of 20 IFs. RLDIF Changed to average spectral channels before averaging into the main arrays. Adjust sigma of vector average of matrix halves by square root of the number of baselines contributing to the average. RLDIF Help changed to drop adverbs not used from the help and to mention RLCOR along with CLCOR. PCAL Added CPARM(2) > 0 to have the source models in the SU table updated. This was done always by PCAL and never by LPCAL. RLCOR NEW TASK: applies the POLR correction directly to a uv data set (either single- or multi-source) Moved nowhere. 12840. May 5, 2009 DO3D false Eric A Taylor expansion of the terms in Leonia's math used for DO3D false has given a correction to the fomulae already implemented - a division by sqrt (1-L^2-M^2). Changed P2DMAT and relinked everything. Moved nowhere. 12841. May 6, 2009 image models Eric Both SETGDS and SETDO3 try to force UVPROT to do its thing at least once - but both assumed that MAXABSU in the header was always positive. Bad things happened when it was not. Fixed to put -abs(maxabsu) back in header to do this forcing. Moved nowhere. 12842. May 7, 2009 FRING Eric Changed the FRNSRC error message which used to correctly say that there was not enough memory. Now the limitation is in how large an FFT we are willing to do (currently 32768). Changed the limit to use MAXIMG parameter which reflects the FFT limit and fixed a work buffer to also use MAXIMG for sizing and in the error tests. Moved nowhere. 12843. May 8, 2009 FITLD Eric The attempt to allow DOCONCAT > 0 for normal UV binary tables from FITAB was doomed to failure. At the point where the concatenation must be determined, nothing is now known about the structure of the UV data (e.g. number polarizations, spectral channels, IFs, random parameters, etc). This the concatenation cannot be accepted. Fixed the help file to say that DOCONCAT applies only to FITS-IDI binary tables from the VLBA correlator and fixed the code that attempted to allow binary tables to concatenate. Note that this used to work when FITAB filled the main header with info about the uv data. Now that that was deemed inadvisable for moderately good reasons, it no longer works. Moved nowhere. 12844. May 14, 2009 PRTAN Eric PRTAN confused the maximum row number in the AN table with the maximum antenna number and on occasion the row number with the antenna number. Also fixed it to handle flagged rows in an AN table. Moved nowhere. 12845. May 14, 2009 BPASET Eric Correct missing +1 in a test for valid sums - invalid ones would always be caught anyway - so this correction is just a nicety. Moved nowhere. 12846. May 15, 2009 IMAGR Eric The restoration of CCs from one image to other overlapped ones uses the pseudo AP memory directly. A test for whether it fit check against the size in kilowords ratther than words - a mere 87000 overlapped components would cause it to do bad things when in fact it should have been 87 million CCs. Fixed QCLEAN Moved to patch May 18. 12847. May 18, 2009 SABOX Eric NEW TASK: finds Clean boxes from a collection of facet images. It uses SAD's island discovery routines and then outputs rectangular and/or circular boxes encompassing them. A single flux cutoff = APARM(1) * rms(i) is used in each facet i; the rms is found from the header or from a robust computation using the image values. Moved nowhere. 12848. May 20, 2009 VPLOT Leonia VPLOT has had an option to report the max/min value inside of the time averaging interval (SOLINT). This option provides the vertical lines near the average point with bottom/top edge corresonded to min/max of averaged points. This option may be very useful as a diagnostic of possible errors. Unfortunately this option has not worked for a long time (or never). So I returned this option to the life. Moved nowhere. 12849. May 22, 2009 VPLOT Leonia 1. The option plotting min/max (see the previos change) does not work at the case of the self scaling if the phase is plotted. Fixed! 2. The symbol used to plot the data is forced to '+' if min/max is plotted. Moved nowhere. 12850. May 22, 2009 AIPS Memo 113 Eric Put back the new memo on Faceted imaging in AIPS by Leonia and myself and noted it on the web pages. Moved nowhere. 12851. May 22, 2009 OOP TV device Eric 1. Discovered that the OOP TV device routines made allowance for only 8 TV channels. Changed dimension to 16 in TVDEVICE, TVUTIL, QCLEAN, QEDIUTIL, SCIMG, SCMAP, PLAYR. 2. PLAYR: dimensioned the Clean boxes wrongly, causing an abort. They need to be 4*nfield, maxboxes. 3. SABOX: reordered adverbs a little, added APARM(7) to limit the size of boxes to stop 1-pixel ones normally unless very strong. 4. TVDEVICE: new routines TVDCAT to read/write image catalog, TVDROM to do roaming on already loaded images. Changed order of routines to alphabetical, TVDFUN to apply CHANS for FIDD with default those that are on (as done already for TRAN). 5. TVUTIL: New routines OTVROM to do a load and roam, OTVRRM to do a re-roam on already loaded, OTVROF to do a roam-off screen capture. Changed OTVFID to set CHANS properly, OTVBOX to pick up print level for GRBOXS. 6. GRBOXS, GRBOXD, DRBOXS, LMBOXS: Change to have an X and Y radius for circular boxes. This allows TXINC and TYINC to be different while ending up with circles in actual image pixels. Moved nowhere. 12852. May 22, 2009 FILIT Eric New task: allows FILEBOX and DFILEBOX (effectively) on a collection of facet images including allowing for images that are larger than the TV screen. The user roams in each and then a roam-off operation captures the current display for rebox and delbox. A re-roam allows the rest of the image to be processed as well. A TV menu allows selection of previous or next facet and the usual zoo of enhancement and curvalue functions. Moved nowhere. 12853. May 27, 2009 FLATN Eric Changed INTWIN in INTERPLATE.FOR to return the blc/trc of the input image as well as the location in the output image. Then changed FLATN to impose a weight of 0.0001 outside the inscribed ellipse in the input image. Fixed the WEIGHTIT option to work from the center of the image not the reference pixel (which could be way outside if DO3D false). Changed how EDGSKP is used to inhibit the inscribed circle if desired. Moved nowhere. 12854. May 27, 2009 NMAPS help files Eric Added comments about modeling with multi-scale images to the help files of BLCAL, BLING, BPASS, CALIB, CLPLT, CPASS, FRING, IBLED, KRING, LPCAL, OOSUB, PCAL, UVADC, UVSUB, VPLOT. UVCON should have these too but it is checked out now. Moved nowhere. 12855. May 27, 2009 UVCON Leonia UVCON has supported simulation of even distributed frequencies. Now I have added simulation of group (IFs) frequencies located not necessary even. CPARM deliveres number of groups(IFs) and set of begin frequencies of the groups. APARM(8,9) delivers increment inside of the group and number of frequencies inside of the groups. Moved nowhere. 12856. June 4, 2009 UVCON Leonia The limit of number of antennas has been increased up to 5000. Moved nowhere. 12857. June 7, 2009 CLCOR Leonia Add the new OPCODE='IONO' to correct the delay at inosphere. This opcode is similar to 'ATMO' (correction for atmosphere), but has 2 differences: 1. opposite sign of the phase-delay correction; 2. different mapping function, because the ionosphere is located at the 'narrow' ring at the high height. The new opcode was ordered by Mark Read. Moved nowhere. 12858. June 9, 2009 CLCOR Leonia The change of the phase sign was forgotten (for STOK=1) at last modification of CLCOR. Moved nowhere. 12859. June 9, 2009 CLCOR Leonia The logic of calculation of DPDLAT for IONO has been wrong. Moved nowhere. 12860. June 14, 2009 IMAGR Eric Created a new option to generate Clean boxes automagically. Changed: POPSDAT To make new array IM2PARM(40) NEWPARMS Ditto GRBOXS To take ellipses as BLC(1,2) < 0 = radii in x,y TVUTIL To pass the inscribed circle as an ellipse to TVDBOX when the image size is not equal on x and y. WRIBOX To write out facets with no boxes as # 0 0 0 0. IMAGR HELP: add IM2PARM and describe the use of parms 1 through 6 for the autoboxing. Fixed GRCHAN info and upgraded the multi-scale explain with paper reference and discussion of IMAGRPRM(11). IMAGR FOR: add IM2PARM, set the defaults for parms(1-6) and pass on as AUTOBOX. Fix GRCHAN = 1, 2, or 3 to 5 for outlining the inscribed ellipse (1, 2, and 3 are erased by the other graphics - menu, label, and Clean boxes, resp.). QCLEAN Pick up AUTOBOX incl with TELL, find max in the inscribed ellipse in each facet when autoboxing and use it when the max in Clean boxes is small or zero. These ellipse maxima are set to zero when the overall rms in the ellipses gets small enough that the autobox is unlikely to add boxes - otherwise it would keep trying for a facet that would never get a box incompasssing its current non-boxed maximum. Call CLABOX everywhere that CLNTV is called with the same choice of fields. In OVERLAP 2, change to all fields every 30 major cycles - otherwise facets would never get boxes even though they had sources unless those sources were really strong. Add subroutines CLABOX, CLABXW, ISLAND, ADDPK, MERGPK to do the island finding and fitting much like SABOX. Corrected error: the CLRMAX routine used VERY STRANGE windows to find the max/min that were way outside the Clean boxes they were supposed to be checking. Moved nowhere. 12861. June 15, 2009 IMAGR Eric QCLEAN: A TVSTAT array needed to be (16) now - FORCE SDI CLEAN caused it to be filled and overwrite the TVNAME. IMAGRPRM: Help file needed updating. IM2PARM: Help file needed creating. Moved nowhere. 12862. June 16, 2009 SNFLG Eric One array was not big enough to include a double word - this caused the number of keywords to be overwritten on the 64-bit Linux Intel load modules and led to an abort. Made the array bigger. Moved nowhere. 12863. June 16, 2009 IMAGR Eric IMAGR Fortran: changed the default number of Clean boxes (FLDSIZE) to 0 if auto-boxing is selected. QUVUTIL Raise the needed memory returned by GRDFIT by 10 kwords - previously it truncated that needed value so that the QINIT call did not get enough and a sort was almost always "needed". QCLEAN Raise the return requested size from GRDFIT by 5 % and try a second time to avoid a sort. Add another round of trial as well before resorting to sorting. Made necessary by the fact that the AP is used for data too and so with long records may need more if the AP is bigger. Also fixed the auto-boxing so that if the number of islands found is close to the limit it raises the island selection level and tries again up to 4 times. Otherwise it was finding islands only in the lower portion of the image and that did not meet the island peak versus peak residual criterion. TVROAM Help file: comment at start still said 4 planes - it is now up to 16 IMAGR Help file: limit on autoboxes now 50. Default Clean box now a null box when auto-boxing is selected. IM2PARM Help file: limit on autoboxes now 50 per cycle. Moved nowhere. 12864. June 17, 2009 FRING Eric The routine that figured out the frequency increment and checked for regularity was very confused. It asserted that all frequenciies are greater tha FREQS(1) but that other frequencies FREQS(M)-FREQS(n) with M > N might have either sign! That would be a problem I suspect and the first statement is not true. Fixed the routine to get the abs value of the frequency increment and then fix the sign at the end. Later places in the code assume that FREQS(M)-FREQS(N) has the same sign as DF. Fixed the default delay search which used the frequency increment to use the abs value of the increment. Delay search range < 0 => no delay search. Moved nowhere. 12865. June 20, 2009 PRTAN Eric Added code to find casa's EVLA array station names and convert to the ones AIPS is used to. This allows printing of the Y-shaped map of antennas. Moved nowhere. 12866. June 20, 2009 SABOX Eric Raised maximum number islands to 50000. Moved nowhere. 12867. June 20, 2009 FLAGR, FINDR Eric Changed max number of sources to 5000 from 100. Moved nowhere. 12868. June 23, 2009 IMERG Eric Fixed a typo in the history routine that caused it to report the name of the low res image as that of the high-res image. Class, seq, etc were correct in the report. The code and the history file were okay. Moved nowhere. 12869. June 23, 2009 SPCAL Eric Messed up the creation of a scratch AN table to update and so in the end deleted the highest AN table which in many cases would be the only AN table. Moved nowhere. 12870. June 23, 2009 IMAGR Eric QCLEAN did not realize that the low level ISLAND finder could return an island as at (0,0,0,0) meaning it was deleted from the list. As a consequence it accessed IMAGR(0,0) which is an addressing exception with dynamic memory. Added test to skip such islands. Moved nowhere. 12871. June 24, 2009 SABOX Eric The ISLAND subroutine occasionally returns nulled out windows since they merged with another window. Fixed the code following ISLAND to test for and skip such windows. Moved nowhere. 12872. June 24, 2009 IMAGR Eric Fixed QCLEAN to do AUTOBX defaults correctly in TELL. Changed IMAGR to allow shrinking of windows by no more than 1 pixel. Fixed IMAGR and IM2PARM help files to show the shrinking (expansion by -1 and if < -1 -> +1 for IM2PARM(5)). Also fixed a gotcha in the autoboxing in which a tiny island would not be erased from the list of possible boxes and so no weaker but larger islands would be included. Moved nowhere. 12873. June 25, 2009 AIPS MEMO 114 Eric Put back aips memo 114 on FITS-IDI. Put both color and black and white versions of the PDF and PS files. Moved nowhere. 12874. June 25, 2009 FILIT Eric Add IM2PARM adverb, AUTOBOX, AUTOBOX PARMS, and RESET BOXES operations. The auto-box is similar to that in IMAGR except that the "residual" image is the full image with pixels outside the inscribed circle and all current boxes blanked out. So the "peak residual" equivalent is the peak value not currently in a box. Added to HLPFILIT as well as HELP and Fortran. Re-ordered the menu to separate EXIT from the more commonly used menu items. Changed limits to allow up to 200 boxes per execution of autobox and to allow IM2PARM(6)=-1 to mean no inscribed ellipse. TVUTIL was changed to make a "draw boxes and return" option not tied to drawing the inscribed circle. The BOXONLY keyword controls this and is used by FILIT when boxes are updated by the RESET BOXES or the AUTOBOX function. TVMENU was changed to deal with blank menu items properly. Moved nowhere. 12875. June 25, 2009 IMAGR Eric Changed the filtering option so that the Clean is restarted after the "final" filtering only if IMAGRPRM(9) < 0. The absolute value of IMAGRPRM(9) is used as the radius. Changed IMAGR.HLP, IMAGRPRM.HLP, IMAGR.FOR, QCLEAN.FOR. The CLNTEL, CLNUV1, and CLNUV2 routines needed the new option. In OVERLAP < 2 but ONEBEAM false, the forced Clean resumption could cause a complete meltdow in the residual images had no pixels strong enough to Clean. Changed CLRHIS to set FINISH and CLNUV1 to trap returns of -1 and -2 as normal exit requests. Moved nowhere. 12876. June 26, 2009 WIPER Eric Changed menu shape to put ABORT and EXIT above all others items with a blank in between. This should reduce inadvertent ABORTs and EXITs. Fixed the scaling of the numbers shown (CURVALUE equivalent) during flagging. Wrote $APLSUB/REMETS.FOR to convert metric scaling prefixes into scaling multipliers to enable this. Moved nowhere. 12877. June 26, 2009 ISLAND.FOR Eric Pulled the routines ISLAND, ADDPK, MERGPK from QCLEAN, SABOX, and FILIT to a new file $APLSUB/ISLAND.FOR. The versions of these in SAD itself are rather different. Moved nowhere. 12878. June 26, 2009 QEDIUTIL Eric Added a blank in each menu between ABORT and EXIT and the rest of the menu. The top of column 1 is also often used so moving them to the top would not help with the accidental selection issue. Affects EDITR, EDITA, SNEDT, SCMAP, SCIMG. Moved nowhere. 12879. June 27, 2009 BPASS, BPLOT Eric BPASS did not change the BP output table sort order even though it knows it to be 1,5 for a new table and knows that it is indeterminate on adding to an old table. Fixed it to set both correct orders for output. BPLOT handled the default for EIF incorrectly. When it was equal to BIF it was changed to the max. That should have been, when it is less than BIF. Added 0.5 character to separate the left hand axis text from the numbers. Moved nowhere. 12880. June 29, 2009 AIPS Memo 115 Eric Put in memo on auto-boxing. Moved to web site this date. 12881. June 29, 2009 AIPSLetter Jun30 Eric Put in AIPSLetter for June 30. Moved to web site this date. 12882. June 29, 2009 UVFND Eric Added option to CLIP on phase using APARM(5-8). Moved nowhere. 12883. June 30, 2009 EDITR Eric When requesting ONE ANTENNA to be flagged with FLAG BELOW and FLAG ABOVE, the flag commands were correctly interpreted by the displays, but not when the FC table was translated into the main FG table. For those two OPs, the baseline involved was the only baseline flagged. The change of 31-January to QEDIUTIL meant that the translation at this level had to change as well. Moved nowhere. 12884. June 30, 2009 IMAGR Eric My fix of the forced filtering neglected to clear the error code and so led to a failure when all was in fact well. Fixed QCLEAN. Moved nowhere. 12885. July 1, 2009 IMAGR Eric Changed IMAGR.FOR and QUVUTIL.FOR to determine the rms time in a data set as well as the average. The zenith and parallactic angle are written to the output images only if that rms < 15 minutes or so. Moved nowhere. 12886. July 1, 2009 CLPLT Eric Corrected the setup for PNTPLT which confused where the corners were for X and Y and so drew a diagonal line. Added FACTOR to control the size of the plotted symbol. Moved nowhere. 12887. July 3, 2009 CLPLT, VPLOT Eric Changed subroutine $APLNOT BASLIN.FOR to take the antenna locations as an array and the number of such locations as an argument. Changed VPLOT for this - it needs only 2 antennas and so dropped the subscripts on U, V, W to 2 (for 2 alternate descriptions) and all subscripts on other parameters. Fixed the orbit routine in VPLOT to not loop from 1 to 1 and not have subscripts on U,V,W. Fixed the sign of BY for non-vlbi modeling in CLPLT. Changed CLPLT for BASLIN call seq. Changed CLPLT to plot models at the sampled U,V,W if NMAPS>0 and at computed U,V,W if NMAPS<0. Wrote Subroutine DEFRMT to remove all blanks from a string, replacing some special character with blanks if desired. Changed VPLOT, CLPLT to use DEFRMT for the station name and number displays. Moved nowhere. 12888. July 3, 2009 CAPLT Eric NEW TASK: plots closure amplitude with models like CLPLT but with 4 baselines: [(1-2) * (3-4)] / [(1-3) * (2-4)]. Moved nowhere. 12889. July 7, 2009 CLPLT, CAPLT Eric Added the option (SOLINT < 0) to average the closure phases or amplitudes in time rather than the visibility data. Removed the scalar averaging code that was never invoked and cleaned up CAPLT's help file - "phase" was mentioned several places rather than amplitude. Moved nowhere. 12890. July 9, 2009 MCUBE, FQUBE Eric FQUBE is a NEW TASK to build frquency cubes with an FQID axis. It can take a mix of planes and cubes or even hyper-cubes with FREQ or FQID axes. Changed MCUBE to write only the rows of the FQ table that are meaningful. Changed MCUBE help to mention FQUBE. Moved nowhere. 12891. July 10, 2009 UVMOD Eric The V polarization was put in backwards due to the sign swapping being done on polarization 1 rather than 2. Added option for INLIST to give polarizations for all components. This means that it reads the INLIST file if it is not blank and NGAUSS>1. Moved nowhere. 12892. July 10, 2009 IMAGR Eric IMAGR with auto-boxing turned on defaults to no Clean boxes at all. There is then some trouble catching things gracefully if the auto-boxing doesn't make some boxes. Changed. IMAGR Tests on ONEBEAM did not take into account the case where NFIELD=1, causing the beam not to be made. Set ONEBEAM true when only 1 field. CLEAN Tests on ONEBEAM did not take into account the case where NFIELD=1, causing the beam not to be made. QCLEAN Added test and messages for OVERLAP=1 w ONEBEAM false to detect that no field was Cleaned and exit gracefully. It checks if the reason is a lack of Clean boxes. OVERLAP 2 closed down well but silently, so I added messages there too. OVERLAP < 2 with ONEBEAM true closed gracefully anyway. Moved nowhere. 12893. July 13, 2009 calibration Eric DATCAL had a lazy bit of code that decided whether to do delay and rate cal based on the sum of the 4 delays and rates involved. If they summed to zero, nothing was done although the individual values were non zero. Changed to do delay-rate if any are non zero. Moved nowhere. 12894. July 13, 2009 IMAGR Eric QCLEAN had a gotcha waiting for the change where I forced ONEBEAM true when imaging 1 facet. CLBSHV got called with facet 0 and did not make the beam # be 1. Added option: IM2PARM(7) to control whether channel N+1 begins with the boxes of channel N or only with the initial boxes set up via BOXFILE, CLBOX, etc. The default is to pass them along when not auto-boxing (as before) but to reset them when auto-boxing. You may choose either mode explicitly. Moved nowhere. 12895. July 14, 2009 EVLA Eric The EVLA through the Widar correlator uses the Earth's center as the array center, the hour angles are measured from Greenwich, and the sign of the BY antenna location is reversed. Changed: BPASET Detect that EVLA array name is not the VLA and so the BY sign is not changed PBUTIL Make EVLA be a known array name for PBCAL call ONLY APPROXIMATE AND SHOULD BE FIXED. PBCAL Make EVLA equivalent to VLA for the moment - THIS IS ACCOR Comment on BY sign that EVLA included with VLBI - this seems not to be actually used. CLCOR Test for EVLA array name and control CLCORPRM(4) and (7) accordingly. CLCOR help - update meaning of these parms DFCOR Test for EVLA array name and control CLCORPRM(4) and (7) accordingly - tests made better like in CLCOR. DFCOR help - update with EVLA added HAFIX Add EVLA as a VLBI like array and fix message usage. HAFIX help - clarify array type PCCOR Comment on BY sign that EVLA included with VLBI - this seems not to be actually used. TI2HA help - add warning about hour angles for EVLA being wrt Greenwich rather than VLA site as before UVCRS Comment on BY sign that EVLA included with VLBI - the Y terms are used in this task. UVFIX Add EVLA as a VLBI like array and fix message usage. UVFIX help - clarify array type VLANT Added test on ANAME with error quit if not VLA or EVLA. Change all data file names if EVLA from VLA to EVLA. Swap sign of BY corrections for phase and then back for fixing the AN file. VLANT help - emphasize for VLA and untested and unprepared for EVLA. CVEL Added EVLA test which makes VLA logical false. The BY phase change done only on VLA true. CVEL Help - add comment EVLA is APARM(7) 0. VPLOT Changed handling of DOARRAY - the user's value was completely overridden. Now it is zero if the array center is at the center of the Earth except VLA and ATCA where it is 1. The user of merlin or other arrays may set the correct value for his telescope. This only matters for NMAPS < 0. VPLOT Help - correct the DOARRAY explanation CAPLT Help - correct BPARM(10) explanation (as VPLOT) CLPLT Help - correct BPARM(10) explanation (as VPLOT) Moved nowhere. 12896. July 15, 2009 MNJ Eric Increased the timeouts in the UPDRSYNC procedure to 8 minutes for SYSTEM and LIBR and 16 minutes for LOAD. Moved nowhere. 12897. July 17, 2009 IMAGR Eric Changed QCLEAN in the relatively new process of finding the max pixel within the inscribed ellipse. It was finding single pixels above a limit that were inadequate to produce an auto-box but enough to send the Clean to the facet in a near infinite loop. Changed it to find the max of 2 consecutive pixels looking only within rows. This parameter was zeroed when the max in any facet was < 1.5 * rms * IM@PARM(3). The 1.5 is not needed for protection against bright single pixels so was reduced to 1.05. In CLNUV2, reset this max in the ellipse for the facet just Cleaned to 0 - it was re-imaging the just Cleaned facet and then finding that it did not want it almost always. Now it re-images only for good reasons. Moved nowhere. 12898. July 20, 2009 FILIT Eric Changed output routine to put a no-boxes box for facets that end up with no boxes. Moved nowhere. 12899. July 20, 2009 AIPSC Eric The parameter checker on GO printed only the first 2 characters of the offending adverb. Corrected 5A2 to A in the format statement. Moved nowhere. 12900. July 20, 2009 UVFLG Eric Added the option to flag antennas for shadowing and baselines for being too short. It uses the existing BASLIN subroutine which is thought to be inaccurate but appears adequate for this purpose. Moved nowhere. 12901. July 24, 2009 IMAGR Eric The formatting of box and facet numbers in the history file was I2! Worked fine for small numbers of facets and boxes but not for larger numbers. Fixed it to be adaptive to the size needed. Moved nowhere. 12902. July 24, 2009 MODVF Eric Code did not allow for comments in INFILE, did not allow for fewer than 4 numbers in INFILE lines (zeroing the remainder), and did not initialize ERRNUM which for some reason was not zero in some Mac computers. These oversights have been corrected. Moved nowhere. 12903. July 24, 2009 GAL, others Eric Changed GAL to initialize ERRNUM in parsing the input text file. Changed OFMCON, TVPSUD, CONVL, PRTAC, DRCHK to drop the DERR.INC include that was never used. Moved nowhere. 12904. July 24, 2009 ATLOD Chris Phillips/Eric Chris has sent us the latest revision of ATLOD to load ATNF data into aips - called 5.3k. I have put it in as a single Fortran routine plus Z routines ZATIO and ZATCVT. I have put versions of those in $APLUNIX (really Linux byte swapped), $APLSOL (Solaris), and $APLMACPPP (solaris with tape removed). All needed some work to correct the worst non-standard problems including mis-aligned commons, function statements not recognized as such by the PP preprocessor, "type *" statements and the like. The Fortran does not look like AIPS Fortran but we have to trust that it works. Moved nowhere. 12905. July 25, 2009 FARS Leonia The new task FARS to serve the Faraday rotation synthesis. FARS read the two input image cubes. The first cube should correspond to the Q-linear polarization component. The second cube should correspond to the U-linear polarization component. The main purpose of the current release of the task is to get the Faraday rotation free (lambda=0) brightness and the Faraday depth of the media on the way from the source to the observer. The multi frequency observations are required to get this purpose. The main ideas of the algorithm are described by M.A.Brentjens and A.G.de Bruyn at: Astronomy&Astrophysics, 441, 1217-1228(2005). FARS carries out the Fourier transform along the frequency (lambda square) axis for each pixel at the image plane. This Fourier transform converts the observed complex brightness as a function of the lambda^2 to the output complex function of the Faraday depth. The real part of this output is recorded into the first output file and the imaginary part of this output is recorded in the second output file. The amplitude of this output or Rotation Measure Transfer Function (RMTF) can be recorded instead under control of input parameters. The eight parameters of the found maximum athe Fourier transform can be recorded at the output instead of the full Fourier transform. This option serves to the most typical case: one source and one rotating media. Moved nowhere. 12906. July 26, 2009 install.pl Eric Changed SCREEN_1 - when it did not find a tarball it did not hop to SCREEN_1_RETRY which it should have done. Rearranged a few } to have it do that. Moved to patch this date. 12907. July 26, 2009 CAPLT Eric The code to manage a specified TRIANGLE was all messed up. Fixed it to test a labeling complaint. Also fixed an incorrect INCLUDE in CLPLT. Moved nowhere. 12908. July 27, 2009 ATLOD Eric Changed direct call to gmtime - a standard C function - issued from Fortran to a use of ZXDATE. Dropped cvs lines from ATLOD and ZATIO and ZATCVT since they were causing out MNJ prep to update the files daily. Moved nowhere. 12909. July 27, 2009 CLPLT, CAPLT Eric Changed to trim the list of possible closure triangles/quadrangles to a list of actual ones. Then the plot routines will know when the last plot is about to be done and enable the X-axis labeling. Checks number plots per page and will do one less if that uses the same number of pages. Moved nowhere. 12910. July 28, 2009 ATLOD Eric Appended the ZATIO and ZATCVT routines to ATLOD.FOR and removed them from $APLUNIX, A$APLSOL, and $APLMACPP. Changed the ZATCVT individual routines to do the right thing depending on the BYTFLP parameter known inside aips. Changed the I/O routines to drop the direct access Fortram I/O which breaks the Intel compilers and replace it with TAPIO which is portable. However, TAPIO and $APLUNIX/ZTPMID had to be changed to allow 2560-byte records as well as 2880-byte records. Moved nowhere. 12911. July 29, 2009 TVFLG Eric Fixed the declaration of the big buffer. The IO buffers got bigger and so the buffer had to grow and the two ways it is used changed relative size. Added an extra IO buffer for safety. It was failing when more than 6 spectral channels were gridded. Moved nowhere. 12912. July 30, 2009 CLINV Eric The call to KEYCOP used the wrong catalog number for the output file. Moved nowhere. 12913. August 2, 2009 CALIB Eric Fixed SETGDS - it can loop a second time through the model catalog and did not fix the class name 3rd character back from 1 when loop over 1000. Fixed DMOD.INC which limited models to 1000 rather than 4096 (should be mAXFLD). Moved nowhere. 12914. August 3, 2009 BPASS Eric BPASS keeps a table of all entries in the BP table which is used to manage the writing and re-writing of the BP records. That table was limited to 2000. Changed to 5000 which allows 10 second averages for 12 hours. Moved nowhere. 12915. August 4, 2009 CHKFC Eric The number of adverbs to GTPARM was wrong - led to abort on Solaris. Moved nowhere. 12916. August 5, 2009 UVMOD Eric Fixed formats in history writing for spectral index. The I4.3 format for the component number was missing. Moved nowhere. 12917. August 7, 2009 SPLIT, SPLAT Eric Clean up on failure - clear status and delete empty UV file. Moved nowhere. 12918. August 10, 2009 FARS Leonia The new version of FARS includes a simple version of CLEAN of the complex image to find a solution for multi complex components at the Faraday rotation spectrum. The relevant change of the help file is added. Moved nowhere. 12919. August 10, 2009 SAD, IMFIT, JMFIT Eric SAD applied the rotation to the position angle with the wrong sign in a number of places that end up in the MF table and in a display location also. The latter arose only when the minor axis of the deconvolution was 0.0 (i.e. failed). IMFIT and JMFIT had a similar issue but only for when the deconvolution failed fully. Moved nowhere. 12920. August 11, 2009 UVMOD Eric Fixed the HI recording of widths to correct units and added FWHM to comments. Moved nowhere. 12921. August 13, 2009 DISP, DDISP units Eric Changed the units description of DISP and DDISP in CALINI, CMINI, IMINI, IMINIT, MCINI, CMTAB, IMTAB, MCTAB, TABCAL, and TABIM. They are sec per meter squared and sec/sec per meter squared, rather than declaring them at some "fiducial" wavelength. Moved nowhere. 12922. August 13, 2009 UVMOD Eric The HI file had the subscripts of ZEROSP in the old order and so displayed the peak values wrongly. Moved nowhere. 12923. August 19, 2009 STUFFR Eric Changed procedure file to (1) force outname=inname still for the UBAVG step and (2) convert all variables to fully spelled out rather than min-match. The former caused a failure in the last UVSRT when INNAME not equal OUTNAME. The latter was a time bomb waiting to go off when the current min-matches fail due to AIPS changes or to the users' own local procedures. Moved nowhere. 12924. August 20, 2009 FREESPACE Eric Charlottesville has purchased a disk so large that its size in kbytes overflows a 32-bit integer. Changed ZFREE and 11 versions of ZFRE2 (incl UNIX, SOL, MACPPC, MACINT) to use Mbytes instead and to round off the internal computations rather than truncate. Moved nowhere. 12925. August 20, 2009 ZCREAT Eric Added call to ZFRE2 in ZCREAT to test the requested disk space before the time-consuming actual creation. This should help with the very large files which are coming. Moved nowhere. 12926. August 21, 2009 Autoboxing Eric Changed the decision about whether a box should be a circle or a rectangle by adding the requirement that circles be fairly small. Minimum diameters must be <= 8 pixels to become circles - large areas lose to much in the corners which may well contain needed emission. Changed QCLEAN (IMAGR), SABOX, and FILIT. moved nowhere. 12927. August 24, 2009 CONVL.HLP Leonia Add 'IMAC' into the list of OPCODEs on the inputs page. moved nowhere. 12928. August ??, 2009 Box overlapping: IMAGR, FILIT Eric In OVERLAP<2 mode, IMAGR will over-Clean when two facets have a box on the same source. Wrote two subroutines: BOXCHK to make a list of overlapping boxes and BOXFIX which will delete the smaller of the two overlapping boxes if > 1 fiducial point overlaps. Fiducial points are the center and corners of rectangles and the center and points every 45 degrees around the circumference of circles. Changed QCLEAN to call this routine when OVERLAP<2 and added it as an option to the TV menu to appear only for multi-facet images. Changed also IMAGR and HLPCLEAN help files. Placed BOXFIX in QCLEAN and BOXCHK in IMAGEUTIL. Added menu option to FILIT to CHECK BOXES. Adjusted also FILIT and HLPFILIT help files. BOXFIX needed recoding for FILIT since boxes are ordered differently there - made a local version. Moved nowhere. 12929. August 25, 2009 IMAGR Eric Changed QCLEAN to restore components to all resolutions in multi-scale Clean. Previously, the components were only restored to the images with the highest resolution. Small-diameter components are convolved with the Clean beam of the image to which they are being restored. Extended components are added to images at finer resolutions with their actual size (the Clean beam size of the images at that resolution). Found the auto-boxing did not work well for multi-scale: it needed finding the peak value and rms in each resolution separately and then applying the appropriate peak to the auto-boxing process. Restoring the images at all resolutions will allow FILIT and SABOX to be applied to NFIELD * NGAUSS facets fully. Changed IMAGR help file to mention the full restoration - it was silent on the subject previously. Checked for spelling and changed some in the MX explain section. Moved nowhere. 12930. August 25, 2009 IMAGR, FILIT Eric Changed QCLEAN and FILIT routines that delete overlapped boxes. The which box is larger test requires that one be larger by more than 10% or they are viewed as equal. The selection between equal sizes tests which one's center is further away from the center of its facet. Significantly larger boxes are kept and among equals, ones closer to facet centers are kept. If all inequalities fail, the box in the lower numbered facet is kept. Moved nowhere. 12931. August 26, 2009 more with boxes Eric Changed: QCLEAN Added an additional call to BOXCHK after the box compression and only at that point report any apparent overlaps remaining including 1-point Changed to tell everything about overlaps only when invoked by the TV option or when SETDEBUG is > 0. IMAGEUTIL Changed BOXCHK to be silent on one-point overlaps for any value of MSGSUP. FILIT As QCLEAN SABOX Help: added NGAUSS adverb and APARM(8) to suppress the checking of overlapped boxes between facets SABOX Added new adverb. Reordered the code to build up a full list of boxes for all facets, then optionally do a boxfix including the extra pass to report any remaining overlaps including 1-point ones, then write out the boxes to the OBOXFILE. Moved nowhere. 12932. August 26, 2009 FILEZAP Eric NEW VERB to zap external files from inside AIPS. Changed AU3, POPSDAT, NEWPARMS, and made new help file. Moved nowhere. 12933. August 27, 2009 IMAGR Eric Created new adverb OVRSWTCH which will allow switching between OVERLAP >= 2 mode and OVERLAP=1 mode. Wrote help file for it and for ONEBEAM which got overlooked before. Changed POPSDAT and NEWPARMS. Changed IMAGR and its help to include OVRSWTCH. Changed QCLEAN to compute the maximum in the inscribed ellipse always rather than only when autoboxing. In QCLEAN, changed CLNUV to call CLNUV1 after CLNUV2 if appropriate and changed CLNUV2 to pick up OVRSWTCH and watch for the level getting low enough and then switching OVERLAP and possibly also ONEBEAM, re-making all images and then exiting for CLNUV1 to take over. OVERLAP=2 is faster in my tests by quite a bit but many of my facets are pretty empty. Moved nowhere. 12934. August 28, 2009 IMAGR Eric IMAGR was Cleaning each facet below the flux level and then re-making the residual images. At that point, some facets would pop slightly above the flux cutoff causing a whole major cycle to be done (or maybe even several) for a small number of weak components until all the bumps were - almost accidentally - reduced below the magic FLUX cutoff. Changed the tests in QCLEAN where all facets have just been imaged - in the CLNUV upper routine at the beginning (i.e. on restarts) and in CLNUV1 after each major imaging - to look at 1.05*FLUX. The Clean of each facet still quits at 1.0*FLUX. OVERLAP 2 mode has another way to turn facets off when they drop below and essentially never knows the correct residual for all facets so no changes were made in it. Re-worded a bit in help file. Moved nowhere. 12935. August 30, 2009 IMAGR Eric Finding and keeping the max in the inscribed ellipse (but not in boxes necessarily) is not a good idea for non-autoboxing cases. Changed QCLEAN to find these since we need the overall maximum. but then zero the by-field list. It was picking up fields to Clean that had high fluxes outside boxes and then having trouble with Clean. Moved nowhere. 12936. September 20, 2009 IMAGR Eric There was an error recovery section that talked about an infinite loop and said switching to OVERLAP=1. But it did not really do that - at least after the OVERLAP 1 and >= 2 parts were made into separate subroutines. Fixed that recovery to make the switch like that triggered by OVRSWTCH. Of course, it should never need such a recovery. Moved nowhere. 12937. September 20, 2009 AN files subroutines Eric Changed ANTINI to force the antenna locations to be in a right hand system, avoiding all the if-then-else guess work. Took advantage of this change to add 3 keywords and two columns. The keywords define the handedness (XYZHAND - signature of this change), TFRAME (terrestrial frame for antennas) and NO_IF (number IFs defines size of 3 columns). The new columns are antenna diameter in meters and the FWHM of the single-dish beam for each IF but scaled to 1 GHz. These may be - when filled in - of assistance in imaging. Changed: PUVD.INC MAXANC parameter 12 -> 14 for 2 new columns. DANT.INC Added XYZHAN, TFRAME, ANTNIF for ANTINI call, added DIAMAN, FWHMAN for TABAN call DANS.INC Added ANDIAM, ANFWHM for full antenna common ANINI Revised version of ANTINI to handle 3 more call arguments for 3 more keywords and 2 more columns. The argument of number of values is no longer used on input. TABAN Revised to include two more columns which may or may not be there. ANREFM New routine to reformat AN table - check and change sign By and/or add the 2 columns. RMEXT Standardized and moved from $APLNOT to $APLSUB to allow all this AN stuff to stay at $APLSUB level. ANFIND New routine to recognize handedness by the By sign and array and station names. ANTINI Changed call sequence - now just calls ANREFM and then ANINI. THIS IS STILL THE ROUTINE TO USE. ANSAME Added 2 copies of new parameters, changed call sequences, checked that new parameters match ANTDAT Dropped the return argument VLB which was not used by DBCON (the only task to use this). Simplifies the code a lot. Fixed call sequences. TABCT Drop erroneous ANTINI from comment COPTAB Change AN file reference to call ANSEL in a loop. ANSEL New subroutine to copy AN files correct header frequency and number of IFs. GETANT Change 2 call sequences, pick up new parameters in DANS commons. Change handling of ant coordinates: do array center correcion for VLA and GMRT correctly and stop flipping BY for VLBI. GETTIM Drop VLA argument - was only for error messages (used by CVEL and SDVEL) TABLEUTIL Changed call seq to OANINI and OTABAN and their calls to ANTINI and TABAn, RESP. BASLIN Changed call sequence to drop array type BPASET Changed antenna location computation to rotate local relative station locations by the array longitude (as GETANT) Simply change 2 call sequences: AU3B ANMAXA GETNAN PDRGET PDRSET Moved nowhere. 12938. September 20, 2009 AN files - tasks Eric Tasks affected by the above changes: ACCOR Remove much of ANTDAT (no need for antenna locations which were done wrongly), fix NANT to be max antenna number not rows in antenna table. AVSPC Change COPTAB to do IF selection etc like COPTAB library subroutine. CLCOR Changed call sequences, changed ANTMOD to use the normal arguments for TABINI/TABIO rather than pseudonyms. Dropped the open for READ and close - ANTINI sets all the keywords for existing files even on WRIT. Dropped CLCORP(4,7) for ANTP, ANTC, SSLO CLCOR Help - dropped CLCORP parameters DBCON Changed all ANTINI/TABAN call sequences, added the new variables as dimensioned local variables for the two input files. DFCOR As CLCOR including help HAFIX As UVFIX HAFIX Help: as UVFIX OBEDT Replaced local declarations with DANT.INC, changed call sequences. PRTAN Extend testing to see if antenna diameter and beams are same and are not zero. Change display accordingly. Fixed call sequences of course. SETAN Change call seq, allow new header values in KEYIN section including XYZHAN and TFRAME, drop excess re-write of the AN file header keywords. SETAN Help: Add header keywords SPLAT Change to copy AN tables via new ANSEL updating the ref freq in subarray 1 only then use new call seq to zero the pol cals if needed. Change to use standard variable names and to fix all AN tables not just that for subarray 1. SPLIT As SPLAT UJOIN Change to use ANSEL rather than direct copy. UVCOP Change to call ANSEL (must have separate inver, outver arguments to support UVCOP UVCRS Change call sequences, get ant coordinates simply as copy of station coordinates but pass ARRLON to the modeling and add it in to the ATAN2 (BY, BX), change back to LH in crossing subroutine. UVFIX Drop much of APARM(1) and ISVLB. Need to keep special VLA and GMRT things. Set BY = -BY for all antennas, change call seq ANTINI and TABAN. UVFIX Help: changed APARM(1) which is only now needed to force use of the array center adverbs. VBGLU Add AN tables as special table type to be translated. Fix finding of row parameters for non-time tables including AN. Force translation of N tables near the start (with GETNAN). VLANT Fix call sequences and change to use standard variable names. CVEL Changed call to GETTIM, drop most VLA business, always change BY for low-level doppler routines. Correct handling of VLA/GMRT - the X/Y is wrt the array center not Greenwich for them so simple addition does not work (as in GETANT). CVEL Help file: APARM(7) not used now SDVEL As CVEL. UVCON Changed call seq, dropped NUMV, setup variables better and dropped the BY reversal for VLBA. CAPLT Fixed ANTINI, TABAN, and BASLIN call seq; dropped BPARM(10). Make XB, YB, ZB be simple copies of the station coordinates - no need to rotate anywhere since that is covered in BASLIN and modeling. Changed help also CLPLT As CAPLT. FRMAP Drop the DOVLA stuff, force into left-handed, compute ARRLON (in LH) and pass along. Fix call sequences. LOCIT Fix call sequences; drop the if-then-else on array name, compute ARRLON correctly, fixing signs (hopefully) VPLOT Dropped DOARRAY from help and Fortran, changed ANTINI, TABAN, BASLIN call seq, XB etc now simple STNXYZ. Changed help. Simple call sequence changes plus stop setting column sizes for ANTINI (called NUMV normally), init variables better: ATLOD BSMOD DTSIM FILLM FILLR FITLD M3TAR MK3IN OTFBS OTFUV SDTUV SPCAL UVFIL UVLOD UVSIM VLBIN IM2UV LPCAL PCAL Simple call sequence changes: AVER BLAVG BOXES CLCAL FACES LISTR MATCH PCCOR RLCOR SETFC UVAVG UVDI1 VBMRG ANBPL FGPLT FRPLT IBLED POSSM TVFLG DTSUM FXPOL INDXR OBTAB (DTSUM, FXPOL, INDXR needed dimensions of POLCA/B corrected) Moved nowhere. 12939. September 20, 2009 FILLM Eric Changed to add half an integration time to the specified stop time test so that the ultimate test is on the center of the integration rather than the end. Moved nowhere. 12940. September 20, 2009 FITLD Eric The AN file was being writen with the frequency in the AG table, even if the frequency order was swapped inside FITLD (for IDI data). Changed to replace this with the ADJRFQ which is used for the header (at least when there is one subarray). Moved nowhere. 12941. September 20, 2009 FO tables Eric The frequency offset may be pulled out of the CL table for receipt from other packages or submission to them. It should be of use to CVEL and UVFIX as well (perhaps when no CL table is available) and to INDXR to make a full CL table. Therefore, created FOINI and TABFO to initialize and read/write them. Created CL2FO to copy a CL table directly to an FO table selecting on source and only if there are non-zero Doppler offsets for that source. Changed SPLIT and SPLAT to call CL2FO when run on multi-source input files making single-source output files. Changed FRQUPD to test on subarray and freqid and to use FO tables if present rather than CL tables. This involves a changed call sequence. Improved its handling of IO to store all antennas for the current time as they are encountered - it was rereading the CL file many times until the antennas were found. Moved nowhere. 12942. September 20, 2009 CVEL Eric Changed it to use the correct SUBARRAY in DSHIFT. It was using subarray 1 for antenna positions. Correct call sequence to FRQUPD. Add test on subarray so that only the correct subarray is passed into DSHIFT. Moved nowhere. 12943. September 20, 2009 UVFIX Eric UVFIX uses the AN file frequencies since for subarrays > 1 the actual frequency might be different from that of the initial data set (DBCON allows this for non-line data). Changed it to use the header frequency for subarray 1 since that is almost certain to be correct. Added FRQUPD to find Doppler tracked frequency offsets when there is one FQID, data are in TB order, and there is a CL or FO table. The fully correct frequency can then be used for the phase shifting. Moved nowhere. 12944. September 21, 2009 CVEL Eric Version put back lacked some Fortran changes that were required to make it work at all. moved nowhere. 12945. September 22, 2009 antini changes Eric BPASET had a bad call sequence for ANTINI. Moved nowhere. 12946. September 22, 2009 DATCAL Eric Added code to fix amplitudes whenever there is a delay to be fixed. Moved nowhere. 12947. September 22, 2009 CVEL Eric 1. Found where LUN 49 already closed in FTAB cam from and dropped the TABIO close. 2. Added code to create an NX table on the fly, dropping the NX table copy and deletion in the existing code. Moved nowhere. 12948. September 23, 2009 CVEL Eric Changed CVEL to read the frequency offsets correctly from the CL or FO table and to update them on output. This means that CVEL run on the output of CVEL should do nothing unless told to do a further change in velocity. Corrected FRQUPD in $APLNOT since it did not select on antenna as it should. Moved nowhere. 12949. September 23, 2009 TABAN Eric TABAN only copied 12 of the 14 pointers and so messed up the polarization terms and possibly the Bx one as well. Moved nowhere. 12950. September 24, 2009 TVFLG, SPFLG Eric Added the option to set the transfer function in displaying amplitude like things. LINear, LOG, SQRT, LOG2 are offerred in a cycle. Changed the HLP files TVFLG, HLPTVFLG, SPFLG, and HLPSPFLG to reflect the new option. Moved nowhere. 12951. September 25, 2009 CLCOR Eric The ANTP and ANTC phase corrections are made by computing the delay based on the antenna location ignoring the array center. For the VLA this was fine - each antenna was at a comparatively small distance. But for the EVLA each antenna is at a distance of about the Earth's radius making many vastly large phase corrections in an antenna basis. These corrections are not stored accurately enough to later compute the much smaller phase differences with any accuracy at all. Changed CLCOR to find and use a reference antenna location to keep the relative antenna lengths much smaller. Moved nowhere. 12952. September 29, 2009 TACOP Eric Drpped call to TABCOP and added a new TBLCOP that uses the TABINI/TABIO routines. This allows for deleting flagged rows. Changed also TACOP.HLP adding adverb DOFLAG to control this option. Moved nowhere. 12953. September 29, 2009 SNFLG Eric Added OPTYPE = 'AMP' to the task. It looks at the robust mean and rms of the SN amplitudes as a function of source, antenna, IF, and polarization and writes flags to a flag table to flag the UV data at bad times. Moved nowhere. 12954. October 1, 2009 IMAGR Eric Added yet another option: IM2PARM(11) and (12) allow for baseline-length averaging of the data on the fly as they are copied from the input data set to the work file. This does have an impact on data weighting and will reduce the emphasis on short spacing data in VLA data sets. But it also reduces the size of the work file which will speed the whole process. The speed up should be linear in work-file size except for a great jump in speed if the work file is reduced to a size that fits in ram. Moved nowhere. 12955. October 2, 2009 UVFLG Eric Reworded the inputs section on elevation flagging and added a sentence ("normal values would be..") to the help. Moved nowhere. 12956. October 2, 2009 POSSM Eric Found one place where the NPLOTS > 0 code was not changed when I added the log amplitude plotting. As a consequence it tried to plot non-existent upper plots in some cases. Changed it to plot log amplitude even for auto-correlations. Fixed the plotting of plus signs and lines with FACTOR in the NPLOTS > 0 code to match that described in the help and implemented in the NPLOTS = 0 code. Moved nowhere. 12957. October 5, 2009 SPLAT Eric Fixed writing of NX table - the source numbers were being set all 0 when a multi-source output file was written with time averaging. The random parameters must be in the buffer before the call to RNXUPD. The special version of RNXUPD which handles the special case of one source being written to a multi-source file is still needed. Moved nowhere. 12958. October 6, 2009 PRTUV Eric The old code here will no longer do. Changed it to used fixed scale by default; set DPARM(1) < 0 to get self scaling. Added routine to use the start time and the NX table to get the correct offset into the file. Now it will be sufficiently fast for the Widar era - or at least the eraly part of that era. Minor corrections in SNFLG. Moved nowhere. 12959. October 9, 2009 IMAGR, PRTUV Eric IMAGR: Forgot to increase the pointer in the baseline-dependent averaging when writing intermediate results. Worked fine with 1 channel, not with > 1. PRTUV: Added test for BPRINT > NVIS. Moved nowhere. 12960. October 11, 2009 GETJY Eric Changed the methods of averaging to robust forms and added reporting of numbers of samples possible, used, and excessively bad. The flux error bar is still the formal error (divided by the square root of number of antennas times number of polarizations actaully used in the average) but is rported both over the sample used in the robust average and the full sample. Added a bunch of explanation of this to the help file. Also fixed IM2PARM help file to mention baseline based averaging. Moved nowhere. 12961. October 12, 2009 GETANT Eric This widely used subroutine was not careful when asking the AN file for the POLTYPE keyword. It supressed error messages when the keyword is not present but filled the string even when there was an error. Added error test and set the string to non blank only when the keyword is read properly. Moved nowhere, many relinks required. Moved nowhere. 12962. October 12, 2009 ANREFM Eric The reformat routine lost the "non-standard" keyword POLTYPE when reformting AN tables. The PCAL created word is needed for later pol cal. Moved nowhere. 12963. October 12, 2009 UVDEC Eric Had a silly error - when setting the alternate reference pixel it did KRARP+JLOCF but there is only one KRARP no +JLOCF. This put a "shift" in the Y direction. Moved nowhere. 12964. October 12, 2009 SNPLT Eric Changed the minimum allowed plot range to allow more detailed plots. Delay and rate things get 2 more orders of magnitude whilst most others get 1 more order of magnitude. Moved nowhere. 12965. October 12, 2009 GETJY Eric Numerous errors: (1) Left small buffer used for debug rather than the dynamic memory that was supposed to be there. (2) The robust averaging must allow for an rms of 0.0 (1 sample) and so test on LE not LT. It was finding no valid data when there was 1 sample. (3) The overall rms should be wrt the robust average not its own average. Corrected computation. Moved several times to force early MNJ preps. 12966. October 13, 2009 FLGIT Eric Work on the messages some more to try to be clearer about what they mean. Added prominent message that no flags were written or no new data flagging occurred and drop the lengthy report in those cases. Moved nowhere. 12967. October 13, 2009 LOCIT Eric Changed to avoid trying to open an over-the-top table once that open or an OTT I/O has failed. Otherwise, the error messages came with every antenna. Moved nowhere. 12968. October 14, 2009 TABSOU Eric Change CALCOD = 'NONE' to blank on input and output. Moved nowhere. 12969. October 14, 2009 IMSTAT, IMEAN Eric Modified AU6D and IMEAN a little to not count pure zero pixels in the rms/mean and to report how many of them there were. Played a little with declarations - there is weak evidence of compiler problems perhaps linked to double precision variables (which are needed and used here). The simple mean/rms of IMEAN should produce the same answer as IMSTAT always. Moved nowhere. 12970. October 15, 2009 FILIT Eric Changed FILIT and the help and real-time help to 1. Add option ENTER FACET to type in a fect number to look at next. 2. Add option DEL ALL BOXES to set the number of boxes in the current facet to zero. 3. To fix the logic about BOXFILEs - it did not read the BOXFILE if the OBOXFILE was the same (and so did not know any initial boxes) and I suspect messed up the output as well. Changed it to refuse to allow OBOXFILE to be the same as BOXFILE and will add an O (not zero) to the end if OBOXFILE if needed to force a new name. Fixed the logic about starting with an INCLASS other than ICL001. It now sets up all facets independent of ICLnnn but starts with a display of nnn. Before, if one started with nnn then the boxes for facets 1 - (nnn-1) were lost. Moved nowhere. 12971. October 21, 2009 AIPSMEM110 Amy Revised DELZN memo, some of the recomended APARMs were incorrect. Also updated the part about editing CL tables. Moved nowhere. 12972. October 21, 2009 FRING Eric The output data set is - by default - averaged over all spectral channels, IF by IF. The output FQ table and header did not reflect the resultant change in frequency increment and reference pixel. Moved nowhere. 12973. October 21, 2009 UVUTIL Eric Had an error in UVFRQS when the BCHAN, ECHAN, and number of channels were inconsistent. This should never happen but Fraser provoked the similar IF test. The fix up for frequency used the number of IFs rather than frequencies. Moved nowhere. 12974. October 22, 2009 WIPER, UVPLT Eric Added adverb NCHAV to go with CHINC to control what is plotted. Averaging channels gets better S/N. Moved nowhere. 12975. October 22, 2009 FRPLT Eric 1. Added option DOACOR. This means that DOXCOR becomes false when adding up all included baselines in one plot (NPLOTS=0) when DOACOR is true. 2. Added option APARM(10) to convert the vis by either normalizing the amplitude or making the phase zero while retaing the amplitude. This option does not apply to autocorrelation data. 3. Corrected local FILANT to have both DOACOR and DOXCOR - before it used only an uninitialized variable for DOACOR and made DOXCOR = .NOT.DOACOR which is NOT what is desired. 4. Fixed it to reset NPLOTS > 9 to 9 rather than quit. 5. Changed setup to UNDISK to ask for the 2 desired antennas only on each pass - should speed things a lot. Moved nowhere. 12976. October 23, 2009 FRPLT Eric The task did not take the plot as fixed scale when the limits for an unused data type were not set. Otherwise the scaling does seem to work and the fixed scale test was corrected. Moved nowhere. 12977. October 23, 2009 CALIB Eric Fixed formating of file names for seqence numbers > 999. This is pretty widespread in AIPS. Moved nowhere. 12978. October 26, 2009 IMAGR Eric Changed IMAGR.FOR to fill in the name parameters for the work file when IN2SEQ is specified and then check if the work file already exists. If so, it deletes it unless ALLOKAY >= 2. We do this since changes in parameters such as BIF, EIF, NCHAV, CHINC, etc will change the shape of the work file and this serious change escaped the notice of the task until bad things happened. Changed the help file to mention this and to add more warning about beam size when OUTSEQ is set. A change from NITER=0 to NITER > 0 changes the size of the beam. Moved nowhere. 12979. October 26, 2009 INDXR Eric Add comments to the help file about making a new CL table with no initial calibration if that cal has already been applied. Moved nowhere. 12980. October 26, 2009 SPLAT Eric Changed COPTAB to make a CL table 1 with (1,0) for gain and 0 for delay and rate and fix it to do CL tables only on multi-source output. A special version of CLSEL called CLNULL is appended to do this. Moved nowhere. 12981. October 26, 2009 CLIP Eric Added APARM(7) option to flag n channels on each side of a bad channel. Moved nowhere. 12982. October 27, 2009 UVAVG Eric The dynamic memory buffer must be 512 words or more since it is used in ANTINI as well as for the data. In BT order the buffer was sometimes too small, causing ZEXPND or other problems. Moved nowhere. 12983. October 27, 2009 FRING Eric Changed it to make 2 duplicate samples for the fitter when only 1 time is found. The fitter is happy with that. Moved nowhere. 12984. October 29, 2009 FRING Eric Added DPARM(9) to the help and cleaned out some Ketan "jokes". Changed the Fortran to trap DPARM(9) > 0 to mean that one wants to suppress the fitting of rate entirely rather than simply zeroing the fit rate after the fact. What happens is the data are accumulated in a number of times as always. But when it comes time to call the fitting subroutines, the data are averaged together fully and then duplicated. The times are set to +- a small amount and the fitting routines which demand a rate fit are given data that absolutely have 0 rate. The data buffering is still managed in such a way as to allow the subinterval processing (SOLSUB and SOLMIN controlled functions). Moved nowhere. 12985. October 29, 2009 FRING Eric Added tests for data falling betwen scans. Raised the time correction to 10 milliseconds and put in a way to skip data that falls outside the apparent scan boundaries. Moved nowhere. 12986. October 30, 2009 CCRES Eric Found an array that was not dimensioned. This caused the task to fail under the Intel 64-bit compilation. Every where else seems to work so no need to patch. Moved nowhere. 12987. November 2, 2009 VPLOT Eric Correct and amplify the SYMBOL information in the inputs and help. 0 => a cross or no symbol depending on FACTOR. Moved nowhere. 12988. November 2, 2009 APCAL Eric Corrected silly repetitions in the code stumbled upon by Craig. Moved nowhere. 12989. November 2, 2009 UVWAIT Eric The request for pseudo-ap space could get truncated and then it would tell you to do XY sorting when that should not be required. Fixed the truncation by rounding well up. Moved nowhere. 12990. November 2, 2009 VBCAL Eric Old code limited number of antennas to 28. Fixed it to use MAXANT and look a bit more modern. Moved nowhere. 12991. November 3, 2009 VBCAL Eric Corrected the mistakes I put in yesterday in the do all baselines mode and added a DOWEIGHT option to scale weights too. Moved nowhere. 12992. November 4, 2009 SPLAT Eric The code used random parameter pointers freely not realizing that the input and output random parameters may be different. Changed to trap the (SPLAT-intended) output ones in separate variables since UVGET ('INIT'...) will over-write the output ones with ones that UVGET thinks should be used. This appeared in a casa-delivered file where FREQSEL preceded SOURCE. FREQSEL is dropped and so the offset numbers changed. The code itself had all sorts of potential for trouble that would rarely occur in practise. Dropped message about new CL table. Moved nowhere. 12993. November 5, 2009 CALIB Eric Changed inputs and task to use ICHANSEL rather than BCHAN to ECHAN. Set BCHAN and ECHAN to the maximum range in ICHANSEL so as to limit the work file I/O. Fixed HI writing to include ICHANSEL (it ignored BCHAN and ECHAN!). Moved nowhere. 12994. November 6, 2009 POLTYPE Eric This keyword is a pain. SPLIT and SPLAT attempted to copy the AN file and then fix this to blank - but the copy did not include the keyword. SPLIT and SPLAT took this seriously and quit. Changed ANSEL to copy the keyword when present (supressing any errors). Changed call sequence to ANSEL to include DOPOL and then clear the POLTYPE keyword and cal parameters when DOPOL is applied. Changed call sequence to ANSEL in UVCOP, UJOIN, and COPTAB. Changed also in SPLIT and SPLAT and dropped all the AN table fix up now done in ANSEL. Found FITLD was prepared to copy this keyword from the AT file and left out the call to write it in - which I have now added. Moved nowhere. 12995. November 7, 2009 SNCOP Eric NEW TASK: averages selected IFs in each record of an SN table and writes those averages in all IFs in a matching output SN table. This could be used to propagate good IFs to those damaged by RFI or to take e.g. a 1 IF file's SN table and propagate it to a 4 IF data set. Moved nowhere. 12996. November 10, 2009 ZFRE2 Eric Corrected a very long-standing C error in most versions of ZFRE2.C. A new gnu compiler on ubuntu seems suddenly to take exception to the error which did not bother numerous previous compilers. It will take a while for this error to be corrected in existing load modules. moved nowhere. 12997. November 11, 2009 BPASS, CPASS Eric BPASS, CPASS, and PHSRF all picekd up the ICHANSEL adverb badly, leading to IFs > 1 averaging all channels! The other ICHANSEL tasks did this correctly. Moved to patch this date. 12998. November 12, 2009 Image extensions Eric IMAGE extensions are a standard FITS construct that is used occasionally. Until now AIPS has ignored them. Changed: PRTTP Changed the reading code to recognize IMAGE extensions, to allow them BITPIX=-32, and to display the contents sensibly. PRTTP Help - added mention of IMAGE extensions. IMLOD Changed internal call sequences to pass additional information internally (to allow > 1 output image). Corrected too small buffer error when creating dummy image and changed to delete dummy image if it contains no useful information and real images had been written from IMAGE extensions. In extension reading routine, added case for IMAGE extensions and wrote a new routine to read the header and data and make additional output images. IMLOD Help - change opening remarks to mention IMAGE extensions. FITLD As IMLOD more or less to change the IMLOD section to handle IMAGE extensions. Note that IMAGE extensions will only work from IMAGE main data or empty main data with BITPIX not 8 or NAXIS not 0. FITLD Help - small note re IMAGE extensions Moved nowhere. 12999. November 12, 2009 FITLD Eric Changes in FITS IDI cause: 1. Change FITLD to pick up and use EQUINOX and WEIGHTYP from the UV data table if present. WEIGHTYP controls what the data weights mean. Correct VIS_SCAL usage for weights of normal type. 2. Changed DAGV.INC, AGINI, AGTAB to add a DIAMETER column. No error if missing. Changed FITLD to use the diameter returned by AGTAB (0 if absent). 3. Changed DATV.INC, ATINI, ATTAB to add a BEAMFWHM column. No error if missing. Changed FITLD to use the fwhm returned by ATTAB (0 if absent). 4. Changed FITLD local include DWRV.INC, local subroutines WRINI and WRTAB to allow a preciptation column. AIPS will ignore this for now. Moved nowhere. 13000. November 12, 2009 FILLM Eric The AN file creation forgot to init the FRAME keyword. Decided that 'VLA' best describes the not very well defined frame used by the VLA (as opposed to EVLA). Moved nowhere. 13001. November 12, 2009 DIFUV Eric The weight computation could get bad results when weights were 0.0 - leading to NaNs on some machines and aborts on others. Cleaned up the logic. Moved nowhere. 13002. November 12, 2009 SUFIX Eric NEW TASK: accepts a list of time ranges and changes the source number in those ranges to a new number and then adds the new source to the source table. (A FUDGE clone.) This may be needed when the on-line systems or off-line software lose a source break, running one source's data into another. Moved nowhere. 13003. November 13, 2009 BLCAL Eric Add ICHANSEL to control averaging over spectral channel. Moved nowhere. 13004. November 13, 2009 ISPEC Eric When the axis was reversed, the SL file did not convey the fact. CHanged SLBINI and ISPEC to use TRC(3) < BLC(3) as well as TRC(3) > BLC(3). Moved nowhere. 13005. November 17, 2009 GETANT, TABAN Eric GETANT did not set up the polarization calibration parameters correctly. The loop counter was changed to NIF and the counter used for the Pol cal values stayed 0. As a result the ellipticity was set 0 and the ori parameter set to the ellipticity of IF 1 for all IFs! This error appeared with the September 20 changes. TABAN used the size of column 9 to tell it the number of IFs. After the format change this is a bad idea since 0 is normal. Changed it to use the new header parameter. Moved nowhere - lots of link edits. 13006. November 17, 2009 AVSPC Eric The local copy of COPTAB was overlooked when changing the ANSEL call sequence. Led to an abort when copying the tables. Moved nowhere. 13007. November 19, 2009 BL cal Eric The table for BL cal was large enough for the VLA up to 8 or so IFs. Had to change PUVD.INC to make a larger buffer and so recompile a lot and link even more. Moved nowhere. 13008. November 20, 2009 TACOP Eric The portion of the task that copies the keywords, selection strings etc copied up to and including the first data record. If there is no data record, this could cause a bad failure. The error message reporting that failure also used the wrong format number and so caused a second failure. Moved nowhere. 13009. November 20, 2009 INDXR Eric Tone down the warning about making NX table for single-source file and note that no CL table will be made. Moved nowhere. 13010. November 20, 2009 CALIB Eric Make NX table for output file. Moved nowhere. 13011. November 20, 2009 VBGLU Eric (1) Dropped requirement that all data sets have NX tables. The requirement for them disappeared in 1995. (2) Make NX table for output file. (3) Moved the test for there being no extension files of the given type. Otherwise it tried to do CT tables with error messages when there were none. (4) Fixed call to UVTIME which required the DUVH include parms to be set up for each file - they pointed to the output which was completely wrong. Moved nowhere. 13012. November 23, 2009 CLCOR Eric The POLR opcode adjusts the pol angles in the AN file. It got the number of IFs from a risky approach using a column dimension and the column number was changed Sept 20. Fixed to use the number of IFs which is now kept in the table. Moved nowhere. 13013. November 23, 2009 POLCO Eric The KEYCOP routine did not have the call arguments prepared yet. Moved it a few lines where disk numbers will be known. Moved nowhere. 13014. November 23, 2009 MODFIT Eric The subroutine MODFIT that makes a list of models which will fit in doing a subtraction stopped listing at the first facet with no CCs. This caused tasks to do more passes through the data than necessary when some of the facets have no CCs but other higher numbered ones do have CCs. Moved nowhere. 13015. November 24, 2009 BLCAL spectral Eric Write NEW TASK BLCHN to solve the BLCAL problem as a function of channel and write out a corrected data base. Required new table type and BDINI and TABBD subroutines to support it. Changed POSSM and BPLOT to offer options to plot the BD table. Moved nowhere. 13016. November 24, 2009 info files Eric Updated the TAB completion file $HLPFIL/HLPIT.LIS. Updated the apropos file $HLPFIL/LSAPROPO.HLP Updated the ABOUT files: ZZADVERB.HLP ZZANALYSIS.HLP ZZAP.HLP ZZCALIBRAT.HLP ZZFITS.HLP ZZGENERAL.HLP ZZIMAGING.HLP ZZINTERACT.HLP ZZOBSOLETE.HLP ZZOOP.HLP ZZPARAFORM.HLP ZZPLOT.HLP ZZPOLARIZA.HLP ZZPOPS.HLP ZZPROCEDUR.HLP ZZSPECTRAL.HLP ZZTABLE.HLP ZZTAPE.HLP ZZTASK.HLP ZZTV-APPL.HLP ZZTV.HLP ZZUV.HLP ZZVERB.HLP ZZVLBI.HLP Moved nowhere. 13017. November 24, 2009 SETSP Eric Changed limit - it allowed a change of speed only to 100 although it claimed to allow 500. Now it does. Moved nowhere. 13018. November 25, 2009 FITLD Eric FITLD had a second place where the antenna Y coordinate was forced negative. This is on antennas that did not occur in the first IDI files being read but did occur on later ones. Moved nowhere. 13019. November 26, 2009 MAPIT Eric Deleted RUN file and 4 help files (actually moved to $APLCONTR) that depended on MX for imaging. Moved nowhere. 13020. November 27, 2009 TVTRANSF Eric Fixed AU6A to use all roam channels when the screen is in roam mode. Mentioned this in the help files for TVTRANSF, TVROAM, and REROAM. Moved nowhere. 13021. November 30, 2009 ANCHECK Eric Created a RUN file to check antenna By positions in the antenna file. The recent bug found in FITLD means that a few antennas in the archive will have erroneous By values. This proc is to check on them. A different proc ANFIX8 is put in 31DEC08 to allow VLBA archive people to change back to the 31DEC08 version. Moved to 31DEC08 as described above. 13022. December 1, 2009 VLOG.HLP, ANTAB.HLP Amy Mentioned in help files that ANTAB cannot handle mixed modes. Instructed people with mixed modes to seperate *cal.vlba file into single mode files before running VLOG and/or ANTAB. Moved nowhere. *********************************************************************** *********************************************************************** Changes while 31DEC09 is NEW *********************************************************************** *********************************************************************** 13025. December 3, 2009 CookBook Eric 1 : Updated numbers, aips++ remarks, misc wordings 2 : Drop TIMDEST reference, misc minor wordings 3 : RECAT, AN file format, drop HORUS, MX, TIMDEST, WHATSNEW, add FILEZAP, correct VLA archive web address 4 : 3C147 model, RLCOR, ATLOD now AIPS, SNFLG flags data, GETJY robust average 5 : IMAGR (DO3D false, autobox, IM2P(7), restart after filter, OVRSWTCH, baseline-length dependent averaging) MX/HORUS, SABOX, FILIT, UVFND phase, FQUBE 6 : ISPEC/BLSUM make SL files, REROAM amd ROAMOFF, FILIT, CAPLT, UVPLT NCHAVG option 7 : ISPEC, BLSUM slices; UVMOD text file, FQUBE, FARS 8 : ISPEC, BLSUM slices; MX, HORUS gone; IMAGR autobox/carry boxes over between channels; FQUBE 9 : SNPLT DOBLANK, RLCOR, FITLD read pre-processed archive data but without DOCONCAT, VPLOT max/min, CLCOR 'IONO', CAPLT, FRING DPARM(9) no rates, SNCOP for copying SN tables, BLCAL ICHANSEL and BLCHN 11 : Drop TIMDEST several places 12 : Only private CA, FILEZAP, FQUBE, DELAY, several misc. 13 : Updated all changed ABOUT files A : Drop MX, HORUS, WFCLN, MAPIT proc; reword VLACALIB. B : Wording corrections I : update J : update Recipe: rearranged only LaCook.PL revised to correct version(s). Moved PS to shared TEXT, $DOCTXT/LaCook* to 31DEC09 and 31DEC10. 13026. December 3, 2009 ANCHECK Eric Effelsberg is EB not EF - corrected ANFIND.FOR and ANCHECK.001 in NEW and TST and ANFIX8 in the patch page for 31DEC08 (OLD). Both EF and EB are now recognized per Craig. Moved as described above. 13028. December 7, 2009 SPLIT, SPLAT Eric SPLIT and SPLAT assumed that the header value of the alternate reference pixel was correct even for multi-source files. This is not necessarily the case and should always be the frequency reference pixel. Changed usage. Moved from 31DEC10 this date. 13029. December 8, 2009 ATLOD Eric Messed up one of the 2 calls to TABAN. Moved to 31DEC09 this date. 13030. December 10, 2009 TIMDEST Eric Removed messages that mentioned TIMDEST - it no longer is available to users in PRTMSG, AIPSC (added block), and ZDCHIN. Moved to 31DEC09. 13031. Dcember 10, 2009 COMB Eric Added code to copy tables to the output except for CC tables. Moved to 31DEC09. 13032. December 11, 2009 install.pl Eric Added code to trap the current list of hosts that are found in a pre-existing HOSTS.LIST file and the to reread that file to make sure that those hosts have the current $ARCH and $SITE that install.pl assumes. This should stop mysterious errors which arise from a mis-match of the old and new values of these critcal strings. Moved to 31DEC09 December 29. 13037. December 15, 2009 Memo 114 Eric Put back a revision of the memo - numerous minor errors plus BITPIX, 3 types of weights, more care with EQUINOX and EPOCH, antenna diameter and full-width at half-maximum, units corrected for dispersion, etc. Corrected web pages. Moved nowhere (all versions get it). 13039. December 18, 2009 SMOOTH Eric SMOSP made an erroneous reference to CATBLK (output header) rather than CATUV (input header). In IBLED it got a very wrong set of increments through the data - so IBLED messed up badly if SMOOTH was set. Most tasks have not messed up the CATBLK enough to matter after the INIT call to UVGET but before the READ calls. Discovered that spectral smoothing was occuring before delay calibration application in the before bandpass smooths. This is a very bad idea but not important for small delays. Moved to 31DEC09 this date - these have been with us a long time. 13041. December 23, 2009 BLCAL Eric The help file referred to writing an output file. BLCAL does not do this. Dropped the remark and cleaned up the Explain typing. Fixed history file writing in the Fortran since it left out many of the calibration parameters and was coded oddly. Moved to 31DEC08 this date. 13042. December 23, 2009 BLCHN Eric Dropped remark in help file - output is written whether or not there is a division by the source model. Changed Fortran to compute spectral rms and honor the print option. Changed help and Fortran to offer a post-solution spectral smoothing identical to the usual smooth. Changed Fortran to write all calibration parameters to the history file and also cleaned up the rest of the HI writing. Moved to 31DEC08 this date. 13044. December 25, 2009 TVFLG Eric Tried to read the number of records in an antenna file equal to the highest antenna number in that file. That is wrong. Changed the length finding routine to issue a warning if some possible antennas have no position info and so have length 0. Moved to 31DEC09 this date. 13045. December 25, 2009 SETJY Eric It issued a message about available models for each IF and had messages about some of the bands saying models may be available. Changed to issue only one message per source and make all certain that models are available. Moved to 31DEC09 this date. 13046. December 25, 2009 ISPEC Eric 1. Fixed return quickly to return quickly on DOCRT 0. 2. Fixed setting of status so it will clear at end of making plot file. Moved to 31DEC09 this date. 13047. December 25, 2009 IMEAN Eric Counted histograms as real*4 - limit of accuracy is 2 or 4 million after which count ceases. Changed to double precision. Moved to 31DEC09 this date. 13048. December 25, 2009 MCUBE Eric Found some code that limited the axis on which we are building to 2048. Fixed. Moved to 31DEC09 this date. 13049. December 29, 2009 CALIB, FRING Eric The handling of the index table, dividing the scan into equal increments ignored the start and stop times. Fixed to change the scan length and start and stop times to honor the overall start and stop times (from TIMERANG). Moved to 31DEC09 this date. 13057. January 5, 2010 IMSTAT, IMEAN Eric Images exceeding 2^31 - 1 pixels (e.g. 400x400x16112 in my case) will be upon us soon. Changed counters in AU6D and IMEAN to double precision float and modified formats. Moved to patch January 18. 13070. January 18, 2009 DOBAND Eric Corrected BPASET, restoring ISVLA variable. DATBND uses it to decide between all channels and the central 3/4 and we do not want to use the outer channels except for crazy VLBIers who insist on it. Moved to patch this date. 13076. January 25, 2009 FITLD Eric Olaf Wucknitz (Bonn Uni) reported another typo causing the polarization type stored in the AN table to contain garbage. Moved to patch this date. 13084. February 5, 2009 IMEAN, IMSTAT Eric The format printing the number of pixels has to be corrected for a previous format that shifted the decimal place. The number printed was a factor of 10 off. Moved to patch this date. 13085. February 10, 2010 Delay correction Eric Amplitude should not be corrected for a delay error unless we know how much averaging was done after the initial correlation. Commenting out the current correction. Moved to patch this date. 13086. February 11, 2010 APCAL Eric The read routine used the loop counter as a row number argument to TABWX. But that routine increments the row counter so only every other row was read and occasionally it would try to read off the end. Moved to patch this date. 13092. March 1, 2010 CVEL Eric My recent code contained the ability to go into an infinite loop when minor roundoff errors with times arose. Added a slight extra bit to the time and added other checks to avoid the infinite loop. BUT - in the course of testing, I found that with single-source files, the source name was null which prevented the message showing the shift or lack thereof from being completed properly. With single-source files having an index (NX) table, the task only shifted the first scan! This fact was obscured by the error in the source name. Corrected. Moved to patch this date. The lack of shift is old and maybe needs to be in the NRAO newsletter. 13098. March 4, 2010 UVFLG Amy Flagging on shadowing was not working because UVFLG was not calculating the baseline length correctly. This was traced to the fact that ARRTYP had not been taken out of the call sequence for the subroutine BASLIN. Eric said he took that variable out of the BASLIN last Sept., so shadowing in UVFLG has been broken since then. Moved to patch March 5. 13100. March 5, 2010 FITLD Eric RDATE is not a required keyword. Fixed tests on it to check only non-blank values. Moved to 31DEC09 since it alarms EVN users. 13111. March 16, 2010 self defense in modeling Eric Changed UVPROT to drop an odd line that set the max in U to the max rotated U. Added line to set the max rotated U to the max input U if the rotation appears to make a short maximum (less than 0.9 max input u). This alone will solve UVCON. Moved UVPROT to patch March 17. 13114. March 17, 2010 SPLAT, SPLIT again Eric The alternate reference pixel was computed ignoring the fact that the reference pixel was corrected for BCHAN by UVGET. Moved to patch this date. 13124. April 5, 2010 CVEL Eric The logical variable ISVLBA got lost last September 20 and so VLBA shifts were based on an antenna position rather than the center of the Earth. The VLBA correlators rotate everything to 0,0,0 to make life easy on CVEL among other things. For other correlators we have to guess to which antenna the velocity was shifted inside the correlator. That guess is antenna 1 except for EVN for which it is antenna 2. NOte that any renumbering of antennas makes a mess of this. Moved to 31DEC09 this date. 13129. April 8, 2010 INDXR Eric Discovered that the AGAINS array in memory allowed only 8 bands whilst the lower read routines returned band numbers 1-10. This led to troubles for P band (using Q values) and Q band for antennas that did not have any values of their own. Also discovered that the band ID routine would never return KA band but used Q instead so KA had the same troubles as Q and used Q band values. Moved to 31DEC09 this date. 13140. April 21, 2010 BPASS Eric The call sequence to FRQUPD was not corrected last year in this task. Has no VLB person run BPASS? Moved to 31DEC09 this date. 13142. April 21, 2010 UVFLG Amy Fixed flagging for shadowing. The hour angle was being misscalulated because the station longetude was being used instead of the array longitude. Also added a 10 second the end time used for the shadowing. Moved to 31DEC09 23 June, 2010. 13206. July 8, 2010 ANREFM troubles Eric DBCON with DOARRAY = 1 could mess up the input headers when the AN file needed reformating. Changed: RMEXT Use LUN=118 not LUN=28. ANFIND Add 'MWA' as a known array ANREFM Do not use the provided CATBLK since it may have been revised on us. Get a safe one from CATIO for creating corrected AN files. Moved to 31DEC09 July 9. 13207. July 8, 2010 copy BP tables Eric Changed BPFSEL and local copy of it in UVDEC. They were deleting partly flagged (one polarization, not other) BP table records. In fact one must keep even fully flagged solutions or the application routines will behave differently after a copy. Moved to 31DEC09 July 9.