;----------------------------------------------------------------------- ; Copyright (C) 2007-2009 ; 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 31DEC08 is TST *********************************************************************** *********************************************************************** 12511. December 2, 2007 New version Eric Put new version string in GETRLS, INSTEP1, install.pl, update.pl, ZXHLP2.PL, UPDCONTROL, AIPSPATH.SH, AIPSPATH.CSH. Moved nowhere. 12512. December 2, 2007 Accounting Eric Created DSUMIO.INC and put it in ZMIO, ZFIO, ZMSGDK, ACOUNT to sum read/write counts and bytes. The accounting will report the IO in Mbytes. Moved nowhere. 12513. December 5, 2007 RECAT Eric Added code to the stand-alone program RECAT to detect that it is being run as a task and to forbid that. Moved to NEW this date. 12514. December 7, 2007 VLARUN Lorant Added a check for long (>12 char) source names, typical when planets are FILLMed without CPARM(2)>=16, or for mozaic fields with the same name (and occasional long IRAS/CXO etc entries). VLARUN will warn that it won't run and allow to do a quick fix using an included procedure to shorten the names automatically. Moved nowhere. 12515. December 10, 2007 IMEAN, IMST Eric Changed IMEAN to add DOINV adverb. Changed AU6D (IMSTAT verb) to match the handling of circular apertures in floating point rather than integer cells and AU8A (verb EXTLIST) for new adverb in IMEAN. Moved nowhere. 12516. December 10, 2007 install.pl Eric MACINT was missing from the list of extra architectures for which diretories must be created. It was also missing from the MAC case issue when making a link file named AIPS as well as the one named aips. Moved fixes to NEW this date. 12517. December 11, 2007 I/O accounting Eric Changed DSUMIO.INC, ZFIO, ZMIO, ZMSGDK, ACOUNT to count ZFIO and ZMIO I/O counts separately. If DBGAIP > 0, ACOUNT will report full details at the end. DBGAIP is set by pseudoverb SETD and any value < 10 will be meaningless as far as debug is concerned. Moved nowhere. 12518. December 19, 2007 RMSD Leonia The number of RMS used at the multi iterational calculating of the final RMS has been fixed to 3. Now it is under control of input parameter SCAL3. The help file is updated. Moved nowhere. 12519. December 25, 2007 VLBAEOPS Eric Cleaned up typing of help file. Lines were too long causing them to be truncated for the user displays, etc. Moved to 31DEC07 this date. 12520. December 28, 2007 AIPSLetter Eric Released LET07B.{TEX,PS,PDF} for 31DEC07 final revision. Moved nowhere ($AIPSPUBL is in all versions). 12521. December 28, 2007 CookBook Eric Changed $SYSUNIX/AIPSWC to include more areas and count PDF to skip. Changed chapters: 1 Updated line and task counts, shipping info, mailing address, aips++ info, etc. 3 Comments re message server, Web and street addresses, REBYTE mention. 4 FILLM (corr coef), TYSMO/TYAPL, TAPPE, OUTFGVER, CLIPM -> CLIP, WIPER, VLANT, FITAB changed, VLACALIB does models 5 Numerous wording changes - esp CLIP usage different, EDITR has FLAGVER and OUTFGVER, IMAGR has two new menu options, new OOSUB mentioned. Weights from FILLM are good not poor as stated. WIPER more general now with baselines known. 6 Numerous wording changes - esp TVLOD FUNCTYPEs, CORER calibration added, COPIXEL error return, IMEAN doinver. 7 IMCENTER, IMEAN, reword some 8 Wording changes, CUBIT, TAPPE, UVLSF 9 DBCON does NOT fix antenna numbering (fixed several errors on this point). Numerous wording/typo fixes. UVCOP now does NX tables. VLBA verb in MkIII section. TYSMO added. LISTR corrected. OUTFGVER with UVFLG. Drop SLIME. Add WEIGHTIT in CALIB self-cal section. 10 Wording improvements, time order for flagging, OUTFGVER 11 Updated wording, FITAB 12 Tab completion, COPIXEL, FUDGE, move data to new computer section incl REBYTE 13 Updated every list to be sure A OUTFGVER in UVFLG, add major section on VLARUN., use model in VLACALIB B Tape -> archive disk files, FILLM parms now default, lots more rewrites incl SETFC. C CLIP, OUTFGVER, UVCOP (NX table), numerous minor wording changes F Dropped 9-track tape section, FITAB added some and 31DEC07 change in it, fixed size of CL table Moved nowhere ($AIPSPUBL is in all versions). 12522. December 29, 2007 CookBook Eric Changed chapters: LaCook*.tex in $DOCTXT, COOK*.PS in $AIPSPUBL. Copies in ftp site. 0 Updated D Add SNFLG, correct INDXR remark, emphasize models, VLANT, correct editing of SN tables (EDITA does not, CLCOR does CL), other re-wordings. I Updated J Updated Recipe chapter rearranged and reorganized as needed. $AIPSPUBL/COOKBOOK.PDF replaced as well as the full html version. Moved nowhere ($AIPSPUBL is in all versions). 12523. December 31, 2007 LISTR Eric Made FREQID apply to gain listings too, but allow all freq-id adverbs <= 0 to mean do all. moved nowhere. 12524. January 2, 2008 FILLM Eric Correct time computation in FLMPO - no longer needs to subtract half the integration time. Changed averaging to average all parameters written to TY, PO, OF, WX, and OT files and changed those routines to write the average values. Changed way averaging decides when is enough so that an integration time or an elapsed time exceeding the requested ends the integration. Previously it was only elapsed time and so could give, for example, 5-sample integrations when the requested time was exactly 4 times the sample interval. Changed default table sampling so that it will come out as the data sampling. Changed $APLUNIX/ZTPOP2.C to try 12 times with delays of 60 seconds to open the on-line disk. Previously it tried 100 times with a delay of 100 seconds! It also did not quit when it detected a request for on-line with a version not capable of such although it would report that every 100 seconds for 9900 seconds. Changed MCSUM to check if the current, already begun, integration has any valid data at the start of the next sample to be included. If not, then the integration is restarted. This led to short integrations with incorrect average times previously. Changed CL table writing to use average elevation and antenna delays and OR'ed OTT bit. It was possible for CL table writing to be called with AVGIAT not yet averaged so time could be very wrong. Corrected this Moved nowhere. 12525. January 2, 2008 ZCPU Eric CLK_TCK is not easily defined in leopard - add a fix to the $APLMACIN version. Moved nowhere. 12526. January 2, 2008 AIPS Eric Updated the copyright. Moved nowhere. 12527. January 2, 2008 IMLOD, FITLD Eric The default reference pixel is 0.0 not the center of the image. Changed only the IMLOD portion of FITLD. Moved nowhere. 12528. January 3, 2008 RMSD Leonia Add an option using circle instead of the box surrounding the pixel to evaluate RMS. Moved nowhere. 12529. January 4, 2008 IMLOD, FITLD, UVLOD Eric Added defenses against strings with null bytes in tables and elsewhere. Changed also RWTAB and R3DTAB. Moved nowhere. 12530. January 4, 2008 UVDIF Eric Changed it to ignore differences in random parameters so long as the essentials all work okay. Moved nowhere. 12531. January 7, 2008 XAS, START_AIPS Wes/Eric The new Mac operating system "leopard" has a clever scheme in which $DISPLAY is set to a named pipe of for "/tmp/gobbldegook" which confused aips significantly. Made small changes to XAS.SHR (init.c) and START_AIPS and START_TVSERVERS. Moved nowhere. 12532. January 9, 2008 VLANT Eric Moved the - I hope - completed VLA station corrections file for 2007 to $AIPSTARS and saw to a dummy 2008 file being created in the appropriate mysterious places. Moved nowhere, but 31dEC07 gets it automatically. 12533. January 9, 2008 REBYTE Eric The table translation made a hash of tables with long rows - IM tables from VLBI and potentially BP tables for significant numbers (~ 120 or so) of channels times IFs. Moved to 31DEC07 this date. 12534. January 9, 2008 FILLM Eric Changed 1. FILLM was able to miss changes of mode or number of spectral channels if they occurred at file boundaries. Added additional parameters to catch them. 2. Added option v (13'th bit) to CPARM(2) which says to write an OF table in any case. Set IF table parameter WANTOF to true is shadow flagging is suppressed or < 25 meters, if Tsys variable flagging is suppressed or if warning level flagging set higher than 3 (CPARM(3)). 3. Add parameters MCSTAT being the array status and AVSTAT being the logical OR of the MCSTATs. 4. Changed format of OF table, changing DOFV.INC, OFINI, TABOF and the calling areas in FILLM. Changed it to write the OF table only when a flag or status is non zero. Perhaps we should write the OF table all the time. Moved nowhere. 12535. January 10, 2008 PRTOF Eric NEW TASK: prints the contents of an OF table. Help could probably use a bit more explain. Moved nowhere. 12536. January 10, 2008 VLBAUTIL Amy Changed the default CLINT to 0.25, seemed more rational for the default CLINT to be at the low end of the range that should probably be used rather than the high end (1 minute). Also changed INDXR input that set the max time gap (CPAMR(1)) to 1 whenever INDXR is run. The default in INDXR is 10 minutes which seems long. Changed the following help files to reflect the CLINT default change: VLBAFIX.HLP, VLBAFQS.HLP, VLBALOAD.HLP, VLBASRT.HLP and VLBASUBS.HLP. Moved nowhere. 12537. January 11, 2008 PRTOF Eric Enhance the correctness and information content of the help file and the header comments in the Fortran. Moved nowhere. 12538. January 15, 2007 START_AIPS START_TVSERVERS Eric Solaris could not handle the grammar we put in for Macs. Changed to skip some statements when Solaris. Moved nowhere. 12539. January 16, 2008 UVDIF Eric Added APARM(6) to allow for comparing fully flagged data and changed code to ignore fully flagged data unless APARM(6) is at least 1. Moved nowhere. 12540. January 16, 2008 OFLAG Eric New task: applies OF table selectively to write a FG table. It can copy an existing table and append the new flags. Moved nowhere. 12541. January 17, 2008 TABED Eric Implemented CLIP, DELE, UFLG options and cleaned up the code typing a whole lot. Moved nowhere. 12542. January 18, 2008 FITLD Eric Corrected errors I put in the call sequence to TABWR so that weather tables can be properly interpreted. Included use of the number of parameters to handle weather tables missing WVR and ION values. Added more OPCODE options to allow the correlator tables to be retained and/or the VT table. Moved WR fix to patch, rest nowhere. 12543. January 19, 2008 SNEDT, EDITA, EDITR Eric Changed QEDIUTIL, the EDIT class routine, to handle flagged table rows properly (rather than qutting for error), to get a proper ANTMAX for TY tables even though they do not have a NO_ANT keyword, and to perform "restore area" with the same complicated logic as flag area (it was restoring all times in the box ignoring the amplitude window). Moved nowhere. 12544. January 21, 2008 WIPER Eric Added options FLAG BASELIN and UNFLAG BASEL - changes task to interactive with slow return to aips since we have to read the baseline numbers from the terminal. Changed the number of baselines remembered to 10 although only 2+ can be displayed. If a baseline is flagged that is known at each point where it occurs even if that does not flag the displayed point (because of other valid data there). Changed help and real-time help. Moved nowhere. 12545. January 23, 2008 SNEDT et al Eric Forgot to COMRPL QEDIUTIL so the changes of Jan 19 did not get made until today. Moved nowhere. 12546. January 24, 2008 APCAL Amy Added a couple checks of the WX table in case it is corrupt. The FITLD bug that was recently fixed corrupted the WX table so that no rows were read from the WX table, for this I put in an error message. I also check if any temperatures are read in for an antenna and if not, print a message for that. Moved nowhere. 12547. January 24, 2008 VPLOT Eric Create new adverb AVGCHAN, changing POPSDAT, NEWPARMS, and AU2A and writing a new help file. Added adverbs AVGCHAN and CROWDED to VPLOT. The former says whether to treat spectral channels individually or as a single vector average and the latter controls whether the IFs and channels appear all together in each panel or in separate panels. Moved nowhere. 12548. January 25, 2008 TVFLG Eric Added the adverb options NCHAVG and CHINC. Required a change to the call sequence and functions in GRIDTB and a new useful subroutine UNFRMT. The latter is REFRMT except with an added symbol to say 0 blanks desired at this point (used for a range of channels such as 11-13). Lots of minor changes throughout TVFLG for these options. UVIMG needed a change to handle the new GRIDTB call sequence. Moved nowhere. 12459. January 25, 2008 VPLOT Eric Corrected to draw colored connecting lines when requested. Also found a few more places where the multi-channel option was not fully implemented. Moved nowhere. 12550. January 27-28, 2008 VPLOT Eric Changed it to try to avoid phase plots that scale -180 to 180 when something like 100 to 200 would do. 28th - added adjust phase on fixed plot where min >= 0. Moved nowhere. 12551. January 28, 2008 UVPLT, WIPER Eric Changed: 1. Make phase plots try to avoid the -180 to 180 where possible. 2. The self-scale within a fixed range was not implemented. 3. Found binned plots (UVPLT only) did not average phase as a vector. Changed to allow 4096 bins and to do vector binning for phase. Moved nowhere. 12552. January 28, 2008 BPLOT Eric Changed to avoid any phase wrap issues if possible. Moved nowhere. 12553. January 30, 2008 EDITA, EDITR Eric Changed QEDIUTIL to be more aware of which antennas have data rather than just those that are allowed to have data. Changed it to choose the better phase plot range and convert the data to that range. Added routine to attempt to put phases in the user-specifed range if possible. Moved nowhere. 12554. January 30, 2008 VLBATECR, VLATECR Amy There was a small error in VLBATECR and VLATECR in calculating how many files to download to cover the entire timerange of the experiments. Only projects observed before 2003 with day numbers larger than 305 would have been effected. Moved nowhere. 12555. February 1, 2008 MFPRT Eric Changed: 1. Added print or sorted row number with new adverb FORMAT to suppress it if desired. 2. Added OPTYPE 'USER' to allow the user to select rows with BOX using appropriate formats for the columns. 3. Added OPTYPE 'LIDE' to do LIST with deconvolved widths instead of fit ones. 4. Added OPTYPEs LINE and CONT to force the default style printing to be that for multi- and single-channel images respectively. 5. Changed to symbolic column numbers from include PMFC. 6. Changed line display to omit velocities when they cannot be computed. 7. Omitted code for unused time columns. 8. Corrected the line flux cutoff to use total flux column (like the others) rather than peak brightnes. Moved nowhere. 12556. February 1, 2008 TVFLG Eric Found bug in GRIDTB introduced when I added the NCHAVG and CHINC options. It messed up Stokes QU and RLLR. Moved nowhere. 12557. February 4, 2008 MFPRT Eric Corrected scaling of X and Y offsets when DOHMS=1- - they were left in degrees. Added "column" 40 to USER - as the computed velocity from the header and plane column (1). Moved nowhere. 12558. February 4, 2008 RMSD Leonia The histogram concept for calculation of RMS has been added. The new input adverb OPTYPE is added to switch the old concept and the histogram concept. The histogram is built at the range of the input data +-1.5sigma, near the maxinum position. The histogram has 16 intervals. Non linear least square method is used to fit Gaussian into the histogram. The initial parameters for this fitting: SIGMA, MAXPOS are calculated using the old concept of the RMS calculation. The initial value for the Gaussian maximum is estimated as maximum of the histogram. The three initial parameters of the Gaussian are actually calculated only for the first pixel of each row. The relevant values of the previous pixel are used as initial Gaussian parameters for the other pixels. Moved nowhere. 12559. February 4, 2008 FILLM Eric Changed to take advantage of the new bandcode available in the subarray data area. Now, two frequencies from the same receiver will be treated as the same band even if one lies in the formal frequency of another. Moved nowhere. 12560. February 4, 2008 TV memories Eric It would seem that more TV memories would be useful. Changing to 8 memory planes. Changed: TVDEVICE TVMOVIE AU6 GETOFM TVFIDL COLORC ILNCLR SPFLG TVFLG IBLED to copy the first NLEVS values into repeated parts of the OFM up to 8. Tasks TVHXF and TVHLD are build to do 256 levels only and zero the OFM above that. Raised the buffer sizes in them so that things will not die for that reason. MUCH MORE is coming. Moved nowhere. 12561. February 9, 2008 RMSD Leonia The last RMSD version has not worked with the special Frazer's data which have many zeros at the each row begin. So the task codes were modified: 1. The task started calculating SIGMA with the first pixel at each row. Now it starts with the pixel where SIGMA.NE.0 2. The double precision version of the least square routine is used instead of previously used single precision one. 3. Several conditions added to exclude the failure inside of of the least square routine. Moved nowhere. 12562. February 12, 2008 RMSD Leonia The double precision version of the Least square routine has been attached to the RMSD to control "matrix is inconsistent..". Now the solution for SIGMA is zeroed if it happens. That allows to circumwent bad solutions which can occur at the case of bad data (more zeros for example). The least square solution is considered bad under the following new condition: IF(SIGMA.LE.0 .OR. MAXIMU.LE.2). The previos condition IF(SIGMA.LE.1.E-7) may create a problem if the data consist mainly of the very small numbers. So the new version can more succesfully fight with the very special data: many zeros, histogram is two far from a Gaussian.... Moved nowhere. 12563. February 25, 2008 APCAL Amy Format error in a print statement. Caused APCAL to croak on certain computers, but always wrote the wrong thing in the history file. Moved nowhere. 12564. March 3, 2008 TV memory Eric Changed TV memory in XAS to allow 16 memories and image intensities to 8191 and outputs from the LUT of 2046. A serious attempt has been made (using new opcodes) to allow older releases of code to talk to 31DEC08 XAS. The reverse will not work. Changes to XAS include: 1. To make NGREY a local variable and NGREYMAX=16 a defined thing that controls the size of the image catalog and various other arrays. Ngrey may be reset with .Xdefaults. 2. catalog.c - changed error in find operation in which the 1st character was tested twice rather than testing the 2nd character 3. colors.c - changed lut read/write to 16 bit, saving old routines under new names using 256 values. Made new name copies of ofm read/write to use 1024 values rather than FNINTENS. cmap_split also changed to use 16-bit data and to have old 8-bit clone under new name. 4. image.c - Cloned old image read/write routines and made new ones under the old names to handle 16-bit pixel data. Changed Interogate to return the desired new intensity values and made clone of old one with new name to return old values. Made clones with new name for fill and vector routines with revised using 16-bit data. Changed both to call Init routines if the type changes. 5. init.c - NValue, NColour now 8192 not 256. Show user max intensity and number gray planes. Image memory now short int in allocation. New name for plane_data when must be 8 bit. OFM setting done more generally, LUT no longer assumes out = in. Colour table not needed for 24-bit and now too small so comment out. Made new opcode to clone the INIT function so that we init with 4 Ngrey for old and 16 for new with different default LUT and OFM scalings. 6. screen.c - plane_short where using 8-bit plane data. 7. xas.c - add to case statement all "old" functions now under new names. Correct return bytes. 8. xas.h - NINTENS now 2046 (note 2048*16 gives I/O touble), FNINTENS now explicitly NGREYMAX*NINTENS, buffers bigger, plane data short int not unsigned char, new opcode names for old functions and new opcode values for old names, declare new subroutines. Z routines: ZSSSO2 Raise work buffer size to 32768 ZSSSX2 Raise work buffer size to 32768 ZSSSC2 Raise work buffer size to 32768 $YSS routines: YCNECT Changed code from 25 to 28, to 16-bit data YFILL Changed code from 23 to 27, to 16-bit data YIMGIO Changed codes from 21 and 22 to 19 and 20, changed to 16-bit data YINIT Changed code from 15 to 29 to invoke new InitXAS. YLUT Changed codes from 41 and 42 to 39 and 40, changed to 16-bit data. YOFM Changed to codes 37 and 38 and raised size of local buffer to 32768. YSPLIT Changed code to 48 from 47 and data from 8-bit to 16-bit Moved nowhere. 12565. March 3, 2008 TV enhancements Eric Application code changes: NEWPARMS New adverbs GR2CHAN and TV2CHAN to specify a 2nd graphics and 2nd grey-scale channel POPSDAT New adverbs GR2CHAN and TV2CHAN, correct TVON et al. references in procedures. TVLEVS default -> 0.0 AIPS Add TVOPEN/TVCLOS to prime the TV parameter common in start up. AU1A Added TVCHAN, TV2CHAN, GRCHAN, GR2CHAN, TVLEVS, TVCORN, and TVXY as system adverbs with locally controlled max values. AU2 As AU1A PTVC.INC Parameterize LUT and OFM size and max number TV and graphics channels DTVC.INC Add parameters TVSCGX, TVSCGY for graphics YTVCIN $YGEN: init new parameters YTVCIN $YSS: init new parameters YSCROL $YSS: set correct parameter for graphics YZOOM $YSS: use correct parameter for graphics AU5A To load images in a more general fashion for TVROAM, to ignore SETROAM, and to make the TV channels used be 1 - N where N is the number of frames up to 16. Drop DECBIT calls. TVROAM To allow 16 memories max and to stay in sensible order (previously went in a circle in 2x2). ROMODE now 100 * Nx + Ny. TVWIND To be more general in analyzing roam requirements and to set ICHAN to a bit pattern of low order bits ROMODE now 100*Nx+Ny. AU2A TVLEVS default now 0. TVDEVICE Use graphics scroll, not TVSCRX/Y; set YOFM data more generally [loop to (OFMINP+1)/NLEVS], insure through PTVC that the OFM buffer is adequate and use it in OFM function calls everywhere. Changed TVDFUN to default TVCHAN using all 4 quadrants. TVALUE Use graphics scroll, not TVSCRX/Y GRBOXS Use graphics scroll, not TVSCRX/Y GRSLIC Use graphics scroll, not TVSCRX/Y TVFLG Use graphics scroll, not TVSCRX/Y SPFLG Use graphics scroll, not TVSCRX/Y WIPER Use graphics scroll, not TVSCRX/Y TVCPS Use TV or graphics scroll as needed; set OFM and LUT buffer sizes with PTVC, drop DECBIT silly usage. Drop GRCHAN adverb and use graphics which are on. TVMOVI Bigger local OFM buffer using PTVC, general copy form AU6 As TVMOVI. Change meaning of TVCHAN by dropping DECBIT and rearranging code. TVLEVS defaults to LUTOUT+1 for <10 and old value 256. Changed handling of TVCHAN 0 to check all 4 quadrants for channels that are on. TVFIDL As TVMOVI AU5E Bigger local buffer to manage OFM subroutines. TVLEVS defaults to LUTOUT+1 for <10 and old value 256. ROMODE now 100*Nx plus Ny. COLORC General copy to fill OFM, buffer from outside ILNCLR Made a general copy TVHXF raised buffer size for OFM TVHLD raised buffer size for OFM IBLED general copy form for OFM TVMON Set buffer size with PTVC TVDIC Set buffer sizes with PTVC, drop silly DECBIT usage, drop GRCHAN to use TVLIMG. TVPSUD Fixed precursor re size of buffer, changed call to COLORH COLORH Corrected size of local OFM buffer, dropped unused call argument buffer SPFLG Copy form for OFM, argument to TVPSUD TVFLG Copy form for OFM, argument to TVPSUD GETOFM General copy form to build up OFM, local buffer of right size using PTVC - changed call sequence! KNTR GETOFM call seq GREYS GETOFM call seq PCNTR GETOFM call seq LWPLA GETOFM call seq TVPL GETOFM call made no sense and should have done nothing - removed it. Also removed DECBIT call and changed handling of TVCHAN to try to accomodate 3-color if it occurs. AU5 Drop DECBIT, change TVON et al to bit pattern on channel from stack. Default TVCHAN for TVCLEAR to check all quadrants. AU5B Add GR2CHAN to control background color in TVANOT. Drop DECBIT and some mod's by 10 on GRCHAN. AU5D Dropped DECBIT on TVCHAN setting. AU5F Add GR2CHAN to eliminate decimal coding. AU6A Add GR2CHAN and TV2CHAN to eliminate decimal coding; TVLEVS defaults to LUTOUT+1 for <10 and old value 256. Use PTVC to insure LUTBUF big enough. AU6C Drop DECBIT in TVFIDDLE. HIENH Drop DECBIT, set bit patterns and local values, PTVC to set LUT size HILUT PTVC to set LUT size GINIT Drop DECBIT. BLANK Drop DECBIT, compute channels simply. BLSUM Drop DECBIT, compute channels simply. PLAYR Add TV2CHAN rather than use decimal coding. TVRGB Raise buffer sizes for lUT TVHUI Raise buffer sizes for lUT DECBIT Dropped from system Moved nowhere. 12566. March 3, 2008 TV changes Eric Help files: DELBOX Add GR2CHAN DFILEBOX Add GR2CHAN GRBLINK Add GR2CHAN instead of decimal coding GRCHAN No use for decimal coding any more GR2CHAN New adverb GROFF Describe binary mask immediate argument or single value GRCHAN GRON Describe binary mask immediate argument or single value GRCHAN OFFHUINT Add TV2CHAN. PLAYR Add TV2CHAN. SETROAM Changed to show as obsolete. TVANOT Add GR2CHAN for backgrounding. TVBLINK Add TV2CHAN for 2nd channel TVCHAN No use for decimal coding any more TV2CHAN New adverb TVCLEAR To clear all channels that are on if TVCH=0. No decimal coding. TVCPS Drop GRCHAN adverb, add remark that graphics channels that are on will be displayed. TVDIC Drop GRCHAN adverb, add remark that graphics channels that are on will be displayed. TVHUEINT Add TV2CHAN. TVMBLINK Add TV2CHAN for 2nd channel TVOFF Describe binary mask immediate argument or single value TVCHAN TVON Describe binary mask immediate argument or single value TVCHAN TVROAM Drop TVCHAN, change expl ROMODE and channels. TVSPLIT Add TV2CHAN instead of decimal coding XAS Fix wording about 24-bit versus 8-bit pseduoColor visuals, add nPlanes option. Simply drop decimal coding or adjust TVLEVS from help file: COTVLOD OFFPSEUD OFFROAM OFFSCROL OFFTRAN OFMCONT TVALL TVFIDDLE TVILINE TVLEVS TVLINE TVLOD TVLUT TVMLUT TVPHLAME TVPSEUDO TVSCROL TVTRANSF GRCLEAR BLANK GREYS PCNTR PFPL1 FRMAP Moved nowhere. 12567. March 3, 2008 CONVL Leonia The cross correlation of the two given images has been added under control of OPCODE=IMAC. The convolution of the two images is calculated if OPCODE=IMAG (as it was previously). Nothing is changed to add the cross correlation of the given image with GAUSSIAN, because convollution and cross correlation should be coincided at this case. (GAUSSIAN is a symmetric function.) Moved nowhere. 12568. March 3, 2008 more TV eric Changed in testing: 1. AIPS, AU1A, AU2 - the TV opens too slowly sometimes so moved the TVOPEN in AIPS to after the user number and added if NGRAPH <= 0 TVOPENs in AU1A and AU2. All 3 are in message suppression state to hide failures. 2. G3VCOL used 255 rather than MAXINT! 3. GETOFM - change to return number values, TVMLOU from TV, whatever from text G3SCAL - add argument to give dimension of OFM (TVMLOU normally) KNTR, GREYS, PCNTR, LWPLA - change ROFM etc dimension, add argument to G3SCAL amd GETOFM 4. AU5E, AU6, AU6C - changed TVLEVS to be > LUTOUT only. Changed help for OFFPSEUD, OFMCONT, TVALL, TVFIDDLE, TVLEVS, TVPHLAME, TVPSEUD. 5. Fell into my own trap with TVALL - fixed immediate arg to TVON two places in POPSDAT. 6. Typo in TVFLG too. 7. Upgraded XAS.SHR and ZSSSO2 to match on buffer list. Moved nowhere. 12569. March 4, 2008 More TV Eric Did not remember to COMRPL 2 of the changed subroutines. I thought I had done all, but seemingly not. Also DGPH.INC has a parameter to keep track of channels that are on and it needed to go to at least 24 from 16. GINIT inits it too. Moved nowhere. 12570. March 4, 2008 model division Eric The correct weights after model division are multiplied by amplitude squared, not amplitude. The gridded modeling was corrected years ago, but the DFT was not until today. Fixed QPTDIV, QGADIV, and QSPDIV. Moved to 31DEC07 patch March 5. 12571. March 4, 2008 more TV Eric $YGEN versions of YCURSE and YCUCOR made reference to the channel scroll and were not fixed to use the new variable for graphics channel scroll. This messed up interactive routines some places, esp WIPER. Moved nowhere. 12572. March 5, 2008 FILLM Eric FILLM has been scaling and weighting cross-hand data incorrectly in those cases in which the antenna order in the on-line system is reversed from the antenna order implied by the numbers painted on the sides of the telescope. The reversal took place before the scaling and weighting and it must take place after since we cannot change the antenna numbers themselves. Note that this affects ModComp-era data when converted to correlation coefficients and post-Modcomp data when converted to visibilities. It affects weights of Modcomp-era data when not converted to correlation coefficients. The affects on solar data cross-hands are more extensive since they are converted to correlation coefficients. Note that this bug was found by George Moellenbrock in trying to understand problems in casa. Moved to 31DEC07 patch this date. 12573. March 5, 2008 CONVL Leonia The new option 'IMAC' at CONVL (#12567) requires array XTEMP. The dimension of this array was 1000 which is not enough for the image of bigger size (1024x1024 for example). So I have changed the dimension to the maximum allowed by AIPS: MAXIMG. Moved nowhere. 12574. March 5, 2008 FACSET Eric Added messages to FACSET about source model usage in the presence of an SU table but 0 flux for the source. One cannot use GETJY after setting gains in such a case. Notes about this have been added to the help files of SETJY and GETJY. Moved nowhere. 12575. March 5, 2008 GREYS Eric GREYS limited the FUNCTYPE to omit L2, N2. Moved nowhere. 12576. March 6, 2008 WETHR Eric Brian Butler requested an new opcode : Dew point depression defined as Temp - DewPt. It has been added and a puzzle re antennas clarified in the help file (VLA WX tables do not have antennas). Moved nowhere. 12577. March 6, 2008 more TV Eric Changed $APLUNIX routines ZSSSO2, ZSSSL2, ZSSSX2, and ZSSSC2 to use the aips message system for error messages rather than puting them out to stderr. This way, they can be suppressed at level 7 and can be saved in the message file. Moved nowhere. 12578. March 7, 2008 More TV! Eric 1. XAS: The output buffering was not large enough to read back OFMs. The data read routine returned a misleading value - 0 or -1 are all that is allowed. The parameters of tthe TV common were being handled in 16 Bits - but TVLIMG now requires 24 - so change to 32, making an old version subroutine for saving 16-bit parameters. 2. YDTVC: change to be 32 bit for parameters and change the save opcode. 3. TVCPS: change the number of memories parameter to 16. 4. SPFLG and TVFLG: parameterize the LUT size. The CLIP INTERACTIVE function went nuts when the LUTs were too small. Moved nowhere. 12579. March 11, 2008 misc TV Eric Corrected the messages from ZSSSX2 - some messages were not errors and became like error messages. Moved nowhere. 12580. March 11, 2008 LISTR Eric Added two gain cenversions EFST = effective system temperature and SEFD = system equivalent flux density. Both are just SN or CL amplitude gains squared and scaled. Moved nowhere. 12581. March 12, 2008 more TV Eric Typo in AU5 meant that TVCHAN=1 did not clear properly. This led to oddities in TVALL for channel 1 only. Moved nowhere. 12582. March 13, 2008 further TV Eric TVFIND: Changed to specify what sort of image it is looking for - otherwise when asking for wedges it confuses users. IAXIS1: Did not set string lengths correctly, so there was too much background. Moved nowhere. 12583. March 13, 2008 LISTR Eric A READ status was not put on the file for the new OPTYPEs and therefore also not removed. Strangely, sometimes a READ status did appear and remain after LISTR was done. Since this makes no sense, the correction to handle the new OPTYPEs like GAIN may or may not do something useful. Moved nowhere. 12584. March 14, 2008 CONVL Leonia The recentely put CONVL (#12567, #12573) does not work with very large array (OPTYPE='LIMAC') when more then one pass is required. So I have fixed it changing WIN(2,4) for mirror reflect at Y axis. Moved nowhere. 12585. March 14, 2008 APCAL Amy Added an error if no entries where written to the SN table. Moved nowhere. 12586. March 19, 2008 improve I/O count Eric Imaging in the case of large numbers of facets and large UV data set sizes has become completely I/O bound. I have begun a process of looking at the sources of that I/O. One obvious fix to make is to reduce the usage of UVPROT which finds the maximum rotated UV value for 3D imaging. It is called in ALGSTB but the rotation matrix is known only for one facet at a time, so UVPROT is called once for each facet in such tasks as CALIB and UVSUB. Changed: DGDS.INC Add array UMATS(3,3,MAXFLD), the uv rotation matrix. UVPROT Change call seq adding MFIELD (# fields), DISKO, and CNOOUT (output uv data set to be written later) and changing UMAT to UMATS. Inside it computes the max U over all fields if necessary (header keyword MAXABSU <= 0 or absent) and writes that keyword to the input and future output files. ALGSTB Change call seq to UVPROT UVGRID Change call seq to UVPROT - make sure out = in UVPROT should never be called UVGRTB Change call seq to UVPROT - make sure out = in UVPROT should never be called SETGDS Change call seq: add main data set disk, cno. Find UV rotation from main header, compute UV rotation matrix for each field in UMATS. Fix it to handle transposed RA/Dec. Changed to read for MAXABSU and change it to negative or 0 to force one read through the data by UVPROT. Tasks getting new call sequence to SETGDS are: CLPLT, IBLED, VPLOT, BLCAL, BPASS, CALIB, CPASS, FRING, KRING, LPCAL, PCAL, UVADC, UVCON, UVSUB. SETDO3 Changed to record which data set was used and to compute UMATS and to notice if it changed. It resets MAXABSU if the file changes and UMATS changes. QUVUTIL Changed call seq to SETDO3 OOSUB Error allowed EIF to be larger than present in the data and led to loops over non-existent channels. Moved nowhere. 12587. March 19, 2008 VLARUN Lorant Noticed that sometimes the ampcal model was not loaded - fixed 12588. March 20, 2008 imaging Eric BLCAL, BPASS, CALIB, CPASS, FRING, KRING call SETGDS twice - fix the 2nd call sequence. Fix SETGDS not to force UVPROT to do a read if UMATS does not change (like SETDO3). UVPROT changed to handle output disk from the SCR common (DISKO <= 0) which CALIB etc produce. Forced quick MNJ prep. 12589. March 24, 2008 rethink pseudo AP Eric Changed PAPC.INC Comment all old parameters, new one PSAPMN set to 5120 kwords for minimum pseudo ap size DAPC.INC Changed APCORE dimension from whatever PAPC give (5 million usually) to 2, all that is needed for dynamic memory addressing. Saves 20 Mbytes. ZDCHIN $APLGEN: Drop PAPC, default KAPWRD = 132*1024 kWords, read disk and if KAP2WD (value 69) > 0 or KAPWRD (value 11, was always 64) < 1024, then reset values to defaults. Added ZADRSZ call and set limit to AP to 1536*1024 (6 Gbytes) for 64-bit addressing machines. QINIT $QPSAP: change to make allocated size max (min (I, KAPWRD), PSAPMN) Also cleaned up left-over comments QCLNSZ $QPSAP: drop PAPC.INC, fix comments, calling will need to do error tests. FILAIP Use KAPWRD, drop PAPC and its parameters FILAI2 Use KAPWRD, drop PAPC and its parameters SETPAR Drop secondary AP parameter, reword KAPWRD parameter and set limits using ZADRSZ (1.5 or 6 Gbytes) SETSP As SETPAR POPSDAT Add verb SETMAXAP NEWPARMS Add verb SETMAXAP AUC Code new verb SETMAXAP: telss current values and allowed range or updates the SP table. SETMAXAP New help file for new verb. ALGSTB Improve test on adequate pseudo-AP memory after QINIT call. ALGSUB Fixed typo'd message, require desired AP memory. GR3DUV Dropped local limits, added tougher test GRIDUV Dropped local limits, added tougher test VISDFT Dropped a limit, added a missing error test VSFDFT Dropped a limit, added a missing error test VSCDFT Dropped a limit, added a missing error test, removed an erroneous use of the old KAPWRD parameters QCLEAN Found error: used APSIZE from DAPC common for the current AP size - fixed to PSAPNW. Improved QINIT error handling, dropped limits. SDIMG Changed what were supposed to be AP limits from MAX to MIN and dropped first one. Dropped erroneous reference to KAPWRD, improved final error message on QINIT. CXCLN Made the QINIT error test check that we got requested memory. Add error tests after QINIT call (not present previously): CONV CORR GRDCOR GRDSUB IMGMEM UVUNIF BPASS CVEL SDVEL CONPL Drop limits on AP size prior to call to QINIT which does that now and add or improve error tests: MAKMAP OLDMAP UVFRID UVGRTB UVMDIV UVSORT QARRAY QUVUTIL UVWAIT CONVL FFT IM2UV IMERG UVCON RSTOR APCLN APGS APVC MX SDCLN STEER UVMAP VTESS UTESS Moved nowhere. 12590. March 25, 2008 finish ap business Eric Found residual uses of PAPC.INC: DAPC.INC Drop PAPC include. SETSP Drop PAPC include. SETPAR Drop PAPC include. KRING Drop PAPC include: APSIZE used for no discernible reason other than perhaps a guess as to machine size. Changed to KAPWRD * 1024. QINIT $PSAP: added PAPC include since uses PSAPMN. Left PAPC include in new verb in AUC where PSAPMN is also used. Moved nowhere. 12591. March 25, 2008 PEELR Eric Added SOLMODE as an adverb to the procedure and changed the procedure to set solmode to 'A&P' unless it is 'P'. Added cautions to the help file. Moved nowhere. 12592. March 25, 2008 SPFLG, TVFLG Eric GRIDTB in $APLNOT and GRIDTC in SPFLG stopped there from being phases on autocorrelations. There should be none on RR and LL of course, but RL and LR can have phase. Moved nowhere. 12593. March 26, 2008 UVFIX Eric Rearranged the clock error comments to make the VLA correction less obvious since 0 is now usually correct. Moved nowhere. 12594. March 26, 2008 SNFLG Eric The dynamic memory allocation depended on a guess of the number of times in the data set - a guess that could be too low. Added code to count the times from the SN/CL table rather than guess. Also put the time array into dynamic memory so there are no limits now for large SN tables. Moved nowhere. 12595. March 26, 2008 UNFRMT Eric Improved the precursor comments which did not previously distinguish this routine from REFRMT. Moved nowhere. 12596. March 26, 2008 POSSM Eric Added tests for SOLINT not 0.0 when UVGET returns -1 meaning no data found. In the case of multiple time intervals, an error here should be trapped and looping to the next interval allowed. Moved nowhere. 12597. March 27, 2008 SETSLICE Eric Changed SETSLICE help file to include ONED as a category. Changed TVSLICE help file to mention the renaming of TVSLICE to SETSLICE (in 31DEC00!). Moved nowhere. 12598. March 27, 2008 FREESPAC Eric $APLGEN/ZFREE.FOR had a bug affecting those disks that have assigned user numbers. It caused the time-destroy limit to be shown as 9 days rather than whatever was correct for that disk. Moved nowhere. 12599. March 28, 2008 VPLOT Eric The switch of phase from -180 to 180 plots to 0 to 360 when that reduces the range did not work for AMP plus PH in a single run. Moved nowhere. 12600. March 31, 2008 gridded modeling Eric Change things to allow gridded subtraction without use of scratch files when enough AP memory is available: QGRDCC Changed so that GRD is both in and out CCSGRD Changed the write APIO call to use the new output address of QGRDCC UVMSUB Comment out the GRDSET call - leave to GRDSUB. GRDSUB Change the computation of the amount of AP memory required to set a minimum and a desired. If the amount is >= desired, then call new routines CCSMEM and ALGMEM rather than CCSGRD and ALGSTB or ALGSUB. Put the GRDSET call in GRDSUB rather than CCSMEM New routine: grids/FFTs the Clean components keeping full grid in AP memory. ALGMEM New routine: ALGSTB but with CC grid in ram not APIO and assume all fits well. UVPREP New OPCODE = 'REED' -. open for read but do not worry about MAXBLN. Moved nowhere. 12601. April 2, 2008 gridded subtraction Eric Looking ahead to the hope that we can do more than one grid at a time in some cases, the recomputation of UVW for 3D and rotation and the setting of the data to be all at U >= 0 will need to be done in the AP to avoid moving buffers back and forth a lot (with compression possible!). Changed: UVPREP Forgot to issue the initial read on REED opcode. ALGMEM Cleaned up code setting buffer size, move UVW changes into AP and conjugate there too. GRDSUB Raise the needed memory by one word per record for the conjugate pointer QPRJM $QPSAP version of PRJMUL QUCONJ $QPSAP: conjugates or undoes conjugation of full UV data buffer requiring U >= 0. Moved nowhere. 12602. April 4, 2008 TI2HA Eric When the cal system was added, we omitted a UVGET ('CLOS' which should appear in the loop over subarray. This left LUN 28 open so that the history file could not be written. Moved nowhere. 12603. April 6, 2008 CONVL Leonia Solution for the maximum position at the output data is added. The subroutine POL2DI fits the two dimensional polinom (including the XY term) by least square method in the output data. The selected box (IMSIZE) of the output data is used. Then the found coefficients of the polinom are recalculated in the X0,Y0 position of the maximum (and the value of the maximum itself). Moved nowhere. 12604. April 8, 2008 TV verbs Eric The new code and some older stuff was not careful to close the TV when errors occurred. Then the TV could not be accessed again. Fixed error branches in AU5F, AU6A, AU6D, and a message in AU6E. Moved nowhere. 12605. April 9, 2008 CALIB Eric Changed default of CPARM(1) to -100 degrees for an elevation cutoff for determining the average amplitude gain. Otherwise, for data where "time" is actually hour angle, the "elevations" for good solutions could be negative. CPARM(1), when set and when mean gain will be constrained, functions as it should. Changed the DOFLAG write-up in the help file to be a bit more self-evident. Moved nowhere. 12606. April 18, 2008 IMAGR Eric Changed UVWAIT to do summing with double precision vaiables. Summing large numbers of small numbers can lose accuracy once the sum gets to 10 million or so - the small numbers no longer change the sum even if there are still a large number of small numbers to go. Perhaps this explaains why some uniform weighting schemes seem to improve noise over natural! Moved nowhere. 12607. April 21, 2008 misc, EDITR Eric Changed QUVUTIL and QEDIUTIL to remove references to APCORE that were purely useless comments - I'm tired of greps turning these up. Also dropped unused include DAPC.INC from IMAGR, WFCLN, and FRING. FRING did need DAPM.INC. Changed QEDIUTIL to match SOURCE to the sources in the SU table. It lost contact with the source name/number when a single source was loaded from a multi-source file. This led to the name of source 1 appearing in the editing displays and source number 1 appearing in the flag table when one-source was requested even when the source number was really something else. Moved nowhere. 12608. April 24, 2008 CONVL Leonia Changed default for IMSIZE: the size of the box determined the data used at fitting of the parabola. It was 9, 9. Now it is 5, 5. The English at the output message is improved. Moved nowhere. 12609. April 28, 2008 Imaging Eric Try to reduce the amount of IO by making gridded models have more than one model available in the AP at a time. Find order of facets matters, order of computation in gridding matters. We could fix that with R*8 but then the computation would go a bit slower and would have half the available words in the AP. QGRDCC Fix precursor comments - they were not changed as intended. DMOD.INC New include for multi-facet and/or multi-channel model parameters MODFIT New subroutine to see how many channels and/or facets can be made to fit in the max allowed AP. Prefers N channel groups with all facets, else will do N channels w 1 facet if necessary or all channels with M facets preferably. GRDSUB If # models = 0, set up the DMOD common to do as many as possible (assumes 1 spectral channel group only - see PBUTIL below). Loops doing as many facets at a time as possible. Changed call seq to CCSGRD and CCSMEM and ALGMEM. CCSGRD Drop channel argument which was unused. CCSMEM Change call seq - add NCHAN. Set up DMOD if not already set to contain only 1 model as given by the subroutine args. Do a loop over the models setting up all the grids. ALGMEM Change call seq dropping field, channel, and nchan. Precompute projection matrices for all fields in DMOD lists, loop over models within the loop over data computing and reloading AP parameter memory as needed (when field or freq changes between models). QCLEAN Change CLNUV1 (overlap < 2, do3d true) to do all cleans then UVSUBM for all planes at once. Previously did the UVSUBM inside the loop 1 plane at a time. PBUTIL PBFSUB: get more info MODCCEND, MODMETH, LREC to allow decisions about components to subtract, list frequency groups rather than doing on fly, for gridded CC call MODFIT, create groups of CC tables (potentially many for each facet) in DMOD.INC and let OUMSUB do the work, then clean up. Leave the one channel at a time code in case it is needed. PBFDIV: as PBFSUB but must do all facets even if it means fewer channels in a pass. PBFCCT: remember the spectral index for the CCs where possible to avoid having to reread the images all the time. UVMDIV Change to use DMOD to determine the channels for the call to UVDOUT. Moved nowhere. 12610. April 29, 2008 DFT modeling Eric VISDFT found the model type from the first facet, first component and then ignored any model types not matching without comment. Changed it to check all facets and to use Gaussian or Sphere for point models if there are also Gaussian or Sphere models. It will complain if both Gaussians and Spheres occur, but points can be either with 0 width. It takes more space and computing, but this is better than ignoring most of the flux! Moved to 31DEC07 this date. 12611. April 29, 2008 FILLM Eric Changed test on pointing and tilt data to issue a suitable message at the start of such data (either omit or include). Moved nowhere. 12612. May 1, 2008 Color plots Eric G3VCOL used MAXINT to scale the requested colors, but that parameter is not set when the TV is not actually used. This means that color plot files were all black, which is not entirely useful. The corresponding point in TVPL remained a hard-coded 255.0 which caused TVPL plots to be 1/32 as bright as they should be. The former error appeared on March 3 of this year. Moved nowhere. 12613. May 2, 2008 TVPL Eric TVPL has an array to keep track of what planes are on. It was dimensioned 16 which no longer works - there are 16 grey-scales plus 8 grpahics. Fixed it to 32. This caused it not to turn on graphics planes that it was using. Moved nowhere. 12614. May 5, 2008 imaging Eric Changed things to allow DFT subtraction/division to work on more than one spectral channel (group) at a time. This will help in IMAGR and OOSUB when applying channel-dependent corrections for primary beam and spectral index. Changed: GRDMEM Make frequency handling variables all double precision as much as possible to avoid optimizers reducing intended precision ALGMEM As GRDMEM MODFIT Changed call sequence to allow input and output subtraction methods. Now calls QMTYP so must move to $QSUB. Has method for DFT as well as GRID. GRDSUB Changed the call to MODFIT forcing GRID. PBUTIL PBFSUB: changed to allow DFT and GRID possibilities in call to MODFIT, to force the resulting choice, to order the models in MODFIT with facet number in the inner loop altough the computations are with frequency in the inner loop, to use new call seq for PBFCCT: New call sequence to include range of CCs to be subtracted/divided this time. It now computes the corrected CCs only for that range although it must make a full CC table for each version. CCSMEM If it sets MODMAX it must reset it to 0. VISDFT To use pre-existing DMOD.INC parameters or set them locally, to determine groups of facets with the same frequency and handle them together. to use MODCCB since NSUBG(field#) gets reset the first time that field is accessed. Thus it helps to have facet number vary more rapidly than channel number. Moved nowhere. 12615. May 5, 2008 VLBAUTIL Amy Replaced all minmatch adverbs, verbs and task names. Moved nowhere. 12616. May 7, 2008 FITLD Eric If the first record of a new scan contained an antenna not in the previous scan, subroutine GET3DT would write CL entries for that antenna for the previous scan. Furthermore, if an antenna only occurred in the first record of a scan, then CL entries for that antenna and scan would not appear. The former led to error messages from the software correlator that did not provide useless IM table entries for the antenna that was not used. Moved nowhere. 12617. May 20, 2008 MODFIT Eric Apparently this subroutine can cinfuse optimizing compilers. Added it to OPTIMIZE.LIS as a no-optimize. UVSUB went nuts although it seemed to work sometimes as well. Most confusing. Moved nowhere. 12618. May 22, 2008 SPLIT, SPLAT Eric Added automatic creation of NX tables to these tasks. SPLAT really requires it when making a multi-source output. But an NX table in a single-source file can help CALIB do a better job of averaging within scans. Moved nowhere. 12619. May 23, 2008 PEELR Eric The addition of the SOLMODE option to the procedure was made with incorrect POPS grammar: I used != when <> is correct. This made the proc fail when attempting to loop over peeled facets, even when there was only one. Error made March 25. Moved nowhere. 12620. May 29, 2008 UVPLT Eric Code got rearranged apparently so that the multiplier that set u,v,w to the correct value was reset to 1.0 before application. The uvw should now be a proper function of channel and IF. Moved nowhere. 12621. May 29, 2008 fix Alias Eric The VLA using the old correlator has problems on EVLA-EVLA baselines for narrow spectral bandpasses. The continuum of the other sideband is aliased into the observing band with a channel-dependent amplitude and opposite phase. A new task FIXAL has been written to read a data set containing continuum, bandpass-quality calibrators to determine this pattern, which appears to be the same for all antenna pairs, polarizations, and IFs. FIXAL then applies the correction to the primary data set, determining the phase of the continuum by fitting the pattern to those channels not containing spectral-line signals and then removing the aliased signal. A procedure FXALIAS has also been written. This flags EVLA-EVLA baselines, runs BPASS, unflags EVLA-EVLA baselines, runs SPLAT, runs UVCOP to separate the bandpass calibrators toa second data set, and finally runs FIXAL. This procedure will function for simple cases and offers a pattern for cases in which the user wishes less standard BPASS parameters or for the taking the alias pattern parameters from a different data set. Both FXALIAS and FIXAL are marked in the helps as experimental for good reason. Moved nowhere. 12622. June 4, 2008 CLCOR Leonia The variable source position correction has be added at OPCODE= 'ANTP' or 'ANTC'. The variability can be added as drifts of RA, DEC at CLCORPRM(8,9) or as a table of shifts given at INFILE. Moved nowhere. 12623. June 6, 2008 DELZN, CLCOR Leonia DELZN has been created with the output file, using the number concept for the antennas. When this file is used for CLCOR (OPCODE=ATMO) with another UV data, the numbering of the antennas can be different, So I changed both DELZN and CLCOR using name instead of number for the antenna identification at this file. Moved nowhere. 12624. June 07, 2008 IRING, ISPEC Eric Changed to give it more plot controls: choice of symbol type, size of symbol, error bar plotting, connecting of symbols, rescale and relable X axis (e.g. to go to pc rather than arc sec). Added text output option (for PLOTR) and fixed plot type from 1 (misc) to a newly reserved 33. Changed: POPSDAT.HLP, NEWPARMS.001 to define new adverbs INTEXT, OUTTEXT, DATAIN, DATAOUT. Wrote help files for them. Changed GINIT comment to allow plot type 33 for IRING and brought comment up to date. Changed AU2A and AU8A to support PLGET for IRING type 33. Added displays of flux sums and number non-blanked plot points to the print out (or message file) in ISPEC. Moved nowhere. 12625. June 7, 2008 XGAUS Eric There was a TVOPEN that could be called when the TV was already open. Changed it to open the TV or Tek once and for all at the start. Corrected code that could erase the screen and then plot only a changed guess (if GRCHAN > 0). Added info to help file. Moved nowhere. 12626. June 11, 2008 IMAGR Eric The gridded subtraction could get into situations in which too little memory was allocated to the AP for large images. Changed MODFIT to add more overhead to the number words. Moved nowhere. 12627. June 12, 2008 EDITA Eric EDITA did not handle the UV-data FG table well when doing editing based on TY table values - BCHAN and ECHAN were 0 and this confused the interpretation of the values in the FG table. Fixed QEDIUTIL to set BCHAN to 1 and ECHAN to UV-data max. Also fixed code in TY and SN editing to avoid an initial plot using an antenna with no data. This was avoided at later stages, but not at the start up. Moved nowhere. 12628. June 12, 2008 SAD, IMFIT, JMFIT Eric Put an upper limit of 360 degrees on the position angle error although in fact it is indeterminate when the major and minor axes are equal. The display of 10^6 gave ***. Moved nowhere. 12629. June 18, 2008 APCAL Amy Fixed bug I introduced when I added some checks to the WX table in January. Moved nowhere. 12630. June 19, 2008 VLBAARCH Lorant Took out the '='-sign in the output FITS-file naming so that the NGAS archive doesn't stumble over this special character. Moved nowhere. 12631. June 19, 2008 VLARUN Lorant When renaming the input variables not all error check messages got updated with these new names - hope all are correct now. Moved nowhere. 12632. June 19, 2008 FILLM Eric The setting of the CORRCOEF keyword was all messed up. It should only care about the setting of the ISSUN and RICKS options rather than whether a correction needs to be made. Moved to patch this date. 12633. June 20, 2008 install.pl Eric Added commands to insure the group ownership and permissions on the SPACE files. Moved nowhere. 12634. June 23, 2008 VLANT Eric Have it try curl first for Macs, then wget. Moved nowhere. 12635. June 23, 2008 FITLD Eric Added recognition of keyword CORRELAT in main FITS header and passed that string on to BFAC subroutine. For DifX correlator, the SATUR parameter wants to be 1.0 rather than the computed value used for the VLBA correlator. A note about this was added to the help file. Moved nowhere. 12636. June 23, 2008 IMFIT, JMFIT Eric The GAUSPS outputs were ignored for user display of position angle. They are our best guess for Gaussian parameters in celestial coordinate space and were used everywhere else. SAD displays the position angle correctly. Fixed the one display line in *MFIT that displayed the pixel position angle with the celetial major and monor axes. Moved nowhere. 12637. June 23, 2008 TIORD Eric NEW TASK: simply tests the time ordering of a data set. Tells you at what vis number data are not in increasing time order. It examines the entire data set and prints up to 100 of the failures. Includes the calibration adverbs in case the time order error arises after application of those adverbs - a very rare case I would expect. Moved nowhere. 12638. June 25, 2008 IMAGR Eric Changed UVWAIT to add a SUMWTIN keyword to the UV data header when doing the weighting. Changed MAKMAP and IMGMEM to carry this from the UV file to the image file. Moved nowhere. 12639. June 30, 2008 AIPSLetter Eric $AIPSPUBL/LET08A.{PDF,PS,TEX} put bcak with the latest AIPSLetter content - also mailed today and put on web. Moved lots of places. 12640. June 30, 2008 IMAGR et al. Eric Changed SETDO3 to recognize UVDISK=0 to mean a scratch file - SCMAP has a scratch file rather than a fully cataloged file which confused it no end. Moved nowhere. 12641. June 30, 2008 UVHGM Eric It does fit Gaussians - the inputs section on DOALL was telling an old, now incorrect thing. Moved nowhere. 12642. July 1, 2008 FITTP, FITAB Eric Added FORMAT 4 to allow for 8-bit integer images which may be needed to feed some other picture-oriented software. Moved nowhere. 12643. July 2, 2008 SETFC Eric Changed its decision making re maximum radius. If the zenith angle is large, reduce allowed phase error by cos (za). Also reduce allowed phase error if the average abs(w) > Wmax/4 or sqrt(average W^2) > Wmax/3. Display reason and result. Changed default phase error to 45 degrees from 60. Moved nowhere. 12644. July 4, 2008 APCLN, VTESS, etc Eric The TV routines in APCLN, APGS, APVC, MX, SDCLN, STEER, UTESS, UVMAP, and VTESS did not make sure that the desired TV channel was on and the the others are off. Wrote TVSET to off all channels, on 1 channel, zero and YCINIT that channel and fixed these tasks to call it. Moved nowhere. 12645. July 4, 2008 CC files on UV data Eric Apparently CC files can be found on UV data sets. Changed PRTCC to notice that it is UV data and to set the print scaling from the CC content rather than using the image header for guidance. Changed SETGDS and GRDAT to check for UV data and take defensive action. Changed VPLOT - the shift of 360 for phase plotting was not included in the model phase plot. Moved nowhere. 12646. July 6, 2008 TVFLG Eric Changed TVFLG to set the number of channels in the work file and ECHAN and CHINC more carefully. It got in trouble previously with not setting CHINC to a high number when 1 channel is the result of the averaging. It also had a round-off issue in doing I/O to the grid file if ECHAN was too large at that point. Changed GRIDTB to set its own ECHAN and check that it agrees with the input. Found error in restart with existing work file as well (ref pixel rather than invrement used one place). Moved nowhere. 12647. July 6, 2008 IMAGR Eric Changed IMAGR.HLP and QCLEAN.FOR to have the option controlled by IMAGRP(8) amd (9) function slightly differently. This option deletes weak isolated CCs (or perhaps all negative regions). When the Clean would normally end due to FLUX or NITER et al., this option is run and the Clean is now restarted. The next time it would normally end, the filtering is done but the Clean then finishes. If one has filtered several times during exwecution, then this forced restart is not done. Moved nowhere. 12648. July 7, 2008 RMSD Leonia Apparentely RMSD has not worked on small-memory computers because the size of required memory was too big. I diminished this size without any damage to the task. The history file is modified to include OPTYPE and FLUX. Moved nowhere. 12649. July 8, 2009 FILLM Eric/George M As far back as our records go (12/1990), FILLM has confused which IF goes with which in applying on-line flags for modes 2BC, 2CD, 4, PA, and PB. This means some data were flagged that should not have been and some were left unflagged that should have been flagged. Note that the new OF table with PRTOF and OFLAG would have pointed out this error had we looked at appropriate data sets. Moved to 31DEC07 patch this date. 12650. July 8, 2008 CONVL Eric The change of April 6 introduced an erroneous variable into the COMMON for the input parameters before the end of the list of input parameters. Adverbs BPA, FACTOR, DOBLANK, and BADDISK were therefore lost. Cleaned up help file also - the inputs section is not the help section. Moved nowhere. 12651. July 9, 2008 SPLAT Eric If SPLAT is run with APARM(7)=0 but only 1 source to be written it goes to a great deal of trouble to insure that a multi-source output file is written with the source number unchanged. This tripped up RNXUPD which makes an NX table on the fly and which uses source 1 when there is no source in the calibrated data. Made a local copy of RNXUPD with a default source number argument for SPLAT alone. Moved nowhere. 12652. July 9, 2008 UVFLG Eric Changed FLAGUP call sequence to return the actual number of FG records written or re-written. Changed UVFLG, OFLAG, IBLED, TVFLG, and SPFLG for the changed call sequence and to report the number records written. Changed FLAGUP to support REAS and WILD opcodes where WILD is REAS with wild cards/ Changed FLGSTK to accept WILD as an opcode. Changed UVFLG help and Fortran to accept REAS and WILD and to make UFLG rather restrictive. Unless you set adverbs to "I do not care" values, the FG table record and the adverb must match. None of the dropping flags if there is some overlap. Of course REAS and WILD only examine reason and all other adverbs are ignored. Moved nowhere. 12653. July 11, 2008 CLCOR.HLP Leonia My recent change at DELZN/CLCOR (#12623, June 6) was not shown at help file. I do it now. Moved nowhere. 12654. July 12, 2008 RMSD.FOR Leonia Following the request of Frazer Owen, another small add to the history file has been made. Moved nowhere. 12655. July 14, 2008 CLCOR Leonia Recently I have modifyed DELZN/CLCOR changing antenna numbers to antenna names at the out(in)put file. This change can solve the dismatch of the antenna numbers at the DELZN's data and CLCOR's data if they are different. On the other hand it can create a problem when the file created earlier by DELZN (or by something yet) has antenna number concept, but CLCOR accept only antenna names concept. So I have modifyed CLCOR with "ATMO" to make CLCOR accept the both formats. Moved nowhere. 12656. July 14, 2008 VPLOT Eric Added the phase plot fix to the place where phases are computed for models at computed U,V,W (NMAPS < 0). Movedd nowhere. 12657. July 15, 2008 UVFLG Eric My changes made UFLG not the reverse of FLAG when BIF/EIF were defaulted. Fixed this. Added WILD to inputs, clarified defaults of IF and chan. Moved nowhere. 12658. July 15, 2008 FXALIAS Eric BPVER = 0 and apply flags (FLAGVER=0) in SPLAT. Spell out all names and use = where it was being assumed. Correct in2seq and in2disk in fixal setup. Moved nowhere. 12659. July 15, 2008 DELZN Leonia Previously the fitted model was plotted vs of the recalculated SN table data. This was an only option. Now the direct SN table data can be plotted vs recalculated model. This option is added under control of APARM(10). The ploting itself is improved: Scaling and labeling of each plot is done for each individual plot. Previously it was identical for all plots. Moved nowhere. 12660. July 15, 2008 EDITR, EDITA Eric Changed QEDIUTIL: 1. EDSLST - get the SOURCE adverb from the UV data set and if the first is not blank and the second blank (or = 1st), use that rather than the 8-character "OBJECT" name to search the SU table for the source number. Also used for the SLIST(1) default. 2. EDFCAP - use FLGANT(2) with AREA since it is not explicit about other baselines (FLAG ABOVE looks at all baselines to current antenna when one-antenna and marks each separately) like TIME, TIMERANG, and POINT. 3. EDITxx - added code to detect pre-existing FC table and start with a REDO FLAGS command rather than show the data without the FC table applied. Moved nowhere. 12661. July 16, 2008 CALIB Eric ACCOR, BPASS, CALIB, and CPASS all use the index table but had no defense against trying to read past the end of information. Added defense against that error - which time inaccuracies can cause to occur - by testing on the known length and setting an appropriate scan end time when we have read all there is. Added messages to CALIB when there is a high failure rate to suggest WEIGHTIT=1, 'L1R', etc. Moved nowhere. 12662. July 18, 2008 UVFLG Eric Changed UVFLG and FLAGUP to allow STOKES='----' to mean don't care and SOURCE='-' to mean don't care. Changed help and added a hint about TAFLG if the UFLG is too demanding. Added simple example to TAFLG's help too since it is a confusing task. Moved nowhere. 12663. July 18, 2008 SDVEL Eric Put in a check of dates between the data header and the antenna file. The antenna date is used for this shift and could be wrong if the AN file has been copied. CATH(KHDOB) was used where a string variable should have been. Moved nowhere. 12664. July 18, 2008 CALIB Eric CALIB had an algorithm that attempted to avoid short integrations within a scan by breaking up the scan into equal intervals. With a tiny round out as well, this led to very unfortunate alignments between the fixed intervals and the regularly occurring data samples. Changed the algorithm to use a local, scan-dependent SOLINT but to run the intervals from the data samples not from a regular grid of times. This should get us outputs that have equal integration time almost all the time. This has become important due to the new presence of NX tables even on sihngle-source files. Moved nowhere. 12665. July 18, 2008 FITLD Eric Code contained a computation of file size for expansion in AIPS bytes that overflowed on a 4 Gbyte file. Fixed the computation to have a wider dynamic range - the actual integer is in kbyte units. Moved nowhere. 12666. July 22, 2008 VLBAARCH Amy Fixed the archiving script so it would not croak if it encountered calibration tables (TY, GC & PC) with no rows. Moved nowhere. 12667. July 24, 2008 VLANT Eric The change for macs to do curl first and then wget forgot that curl uses -o and wget uses -O for output. Fixed I hope. Moved nowhere. 12668. July 25, 2008 RMSD Eric Discovered a REAL variable being used as a DOUBLE, causing an important variable to be trashed on Intel Macs when optimized. In the process, cleaned up the code some, repaired an error in BLAN optype and improved the process of binning data for the histogram fit. It should now be faster. Moved nowhere. 12669. July 28, 2008 CALIB Eric Fixed the time adjustment to be ).05 of the interval or 0.1sec whichever is smaller. Otherwise intervals of 0.1sec became infinite loops. Moved nowhere. 12670. July 28, 2008 FITLD Eric The UV table reader set the time range for the table update routines. If there was > 1 UV table in a file, the update would be done only on the time ramge of the second table. The first would be ignored. GETMC and GETDEL had a gotcha - they tested if a local saved variable = the number of MC/IM records. If not they set up that hash table. But this does not cover the presence of a new MC/IM table with exactly the same number of records as the last. Put the record as an argument so that the calling task can force a re-hash. Changed FITLD, INDXH, INDXR, and TABLEUTIL for this. Moved nowhere. 12671. July 28, 2008 LISTR Eric Rick was wrong! It seems clear that the VLA's *continuum* correlator efficiency is 0.89, not 0.805, as advertised in various documents. (This is due to the oversampling by a factor of two that is effectively done -- the 'all products' formulation that Ken and Durga implemented back in 1998 effectively improved our sensitivity by 10%). So -- when LISTR computes the SEFD or effective Tsys, we need to use different conversion factors: SEFD = 228 * G2 was 206 EFST = 40.5 * G2 was 36.8 These are not the right factors for spectral line and they are a function of bandwidth. This may be fixed later when they are known. Moved nowhere. 12672. July 29, 2008 RMSD Leonia Previously, RMSD returned zero for rms, if the histogram solution failed. Now the robust solution is used at such case, if it is available. If not then robust solution is determined and then used. Moved nowhere. 12673. July 29, 2008 APCAL Amy Removed spurious message for people using weather file rather than WX table. Moved nowhere. 12674. July 29, 2008 LISTR Eric Rick was wrong again - the coefficients are returned to their original values. Moved nowhere. 12675. July 29, 2008 aliasing Eric Changed: BPASS to offer the option of amplitude only BP function FIXAL to reduce the adverbs users should not use and add options to not average across baseline, IF, polarization, and time, to allow a full complex aliasing function, and to allow Hanning smoothing of the aliasing function. FXALIAS to set BPASSP(4) to -1 (amplitude only), BPASSP(9) to 1 (interpolate flagged), and to set bchan=echan=0 in SPLAT. Added DPARM to FXALIAS help file and made appropriate changes to BPASS and FIXAL help files. Moved nowhere. 12676. July 30, 2008 imaging Eric GRDSUB did not handle the old case - needed when FFT of image - and did not call GRDAT so parameters were not known. Added another ELSE for image so that size parameters are set. Relink a lot. Moved nowhere. 12677. August 4, 2008 SHIFT Eric I took a look at the usage of SHIFT and it is a mess. 1. Changed help files for FACES, SETFC, HORUS, IM2UV to make it clear that SHIFT is in asec at the reference position. The help files simply did not say. 2. FRPLT applied the phase correction with a strange sign (as if coordinates had been shifted but they had not) and did the shift as a pure Delta RA. 3. UVLIN, UVLSF, UVLSD shifted with delta RA, DEC coordinates rather than arcsec at the source. Fixed help and Fortran. 4. UVLSD, UVLSF, UVLIN ignored IF in computing the frequency to shift phases. Added CHNDAT call and use results. 5. FRPLT also ignored the IF in doing the shift. It invoked FOFF and FINC elsewhere but never set them or used the result. Added CHNDAT call which enables the frequency to be used in shift and fixed the shift code. 6. DFTPL help file implied SHIFT was in coordinates, but the code multipled SHIFT(1) times U and SHIFT(2) times V directly! It did have correct frequency handling however. Fixed the coordinate handling. Moved nowhere. 12678. August 4, 2008 UVLIN, UVBAS Eric Besides the SHIFT corrections, added cal adverbs and changed input data reading accordingly. Moved nowhere. 12679. August 4, 2008 TBAVG Eric Changed: Help Add SHIFT adverb, make clear the absence of flagging, calibration, etc. Fortran Drop excess junk and clean up TBAVER. First sample in each integration was not counted in average time. Added code to get the FOFFs and to shift the phases before summing. Changed averaging to weighted. Dropped copying of numerous tables which are no longer meaningful or, in the case of NX, correct. Moved nowhere. 12680. August 4, 2008 help files Eric Added note to help files pointing out that the task does not apply FG or cal tables: UJOIN, DFTPL. Moved nowhere. 12681. August 4, 2008 AIPS Eric Removed FPOS from the list of system adverbs in AU1A and AU2. It is an offset in arcsec as well as a position in pixels. Moved nowhere. 12682. August 4, 2008 UVMOD, PHSRF Eric Added calibration and flagging. PHSRF never worked properly for multiple polarizations or IFs - there was an addressing problem with the answer from AVGCHN. That has been corrected as has the weight computation. Moved nowhere. 12683. August 4, 2008 help files Eric Explain lack of calibration and flagging in the help files for COHER, CVEL, FIXWT, FRCAL, SPECR, UVADC, BASRM, BSMAP, MSORT, OOSRT, UVSRT, UVSUB, OOSUB. Moved nowhere. 12684. August 5, 2008 UV2MS Eric Added cal/flag adverbs so that single-source files can have their FG and SN tables applied before going to the multi-source output. Moved nowhere. 12685. August 5, 2008 SNSMO Eric It tied itself in knots when the input SN table was empty and otherwise did not handle error conditions well. Changed to make empty input be a serious error and one hopes to have full error accounting and reporting throughout. Moved nowhere. 12686. August 5, 2008 CALIB Eric Added AVGIF variable to the flagging routine so that it will know to set the IFs to BIF, EIF in the flag record. Moved nowhere. 12687. August 6, 2008 APCAL Amy APCAL aborted with no (interesting) message when it encountered a certain type of corrupt WX file. Fixed so now it dies with more grace. Moved nowhere. 12688. August 12, 2008 FITLD Eric FITLD (1) Correct recognized random parameter types UU---SIN etc along with all historical ones. (2) Fix processing of BC table to recognize actual dimensioning of the data. DBCV.INC Drop NOPOLZ, change data from (4,maxif) to (4*maxif) since (nstokd,nband) is how it is processed. Reduce number keys. BCINI Drop NO_POL keyword. BCIO Drop processing of NO_POL keyword BCTAB Drop comments re NOPOLZ, change dimensioning info on the BL corrections. Moved nowhere. 12689. August 13, 2008 FITLD Eric Corrected usage of deleted parameter and mis-spelled parameter in the BANDPASS table conversion. Corrected the IFFREQs written to the BP table - they were all 0.0D0 and should be the BANDFR minus the reference freq. Moved nowhere. 12690. August 13, 2008 GPOS, FPOS Eric System adverbs need to have standard usages - both FPPOS and GPOS are used as absolute pixels, relative pixels, and arc seconds in different places. Removed both from the system adverbs list. Moved to 31DEC07 this date. 12691. August 16, 2008 FILLM Eric DOUVCOMP was made true on value 0.0 - non=standard and not desirable. Help filed said this, so fixed it too. Moved nowhere 12692. August 26, 2008 UVPLT Leonia The new option to plot out UV depending on the W range is added under controll of APARM. UVPLT has added until now the mirrow set of UV. The new version of UVPLT can plot only one given (APARM(3)) half. The new option can be usefull in particular when investigating the effect of the W term on imaging. Moved nowhere 12693. August 28, 2008 TACOP Amy The error statement that was generated when output table exists printed the input table number rather than the output table number. Fixed. Moved nowhere 12694. August 28, 2008 VLARUN Lorant Noticed that VLARUN never checked on PHACAL value other than '*', ie, never caught if it wasn't specified at all. Fixed. Moved nowhere 12695. August 29, 2008 VLARUN Lorant Been programming perl too much lately - used && instead of & in IF. Moved nowhere. 12696. September 2, 2008 UVCOP Eric The copying of IM tables depends on IMSEL which depends on the Flatters' routines IMINIT and TABIM. The latter had an incorrect subscript so that the PDELAY for the 2nd polarization was read from the GDELAY 2nd polarization column. IMSEL is called with all copying tasks so this will be a common problem. I can find no place in AIPS that uses the PDELAY parameter however. Moved nowhere. 12697. September 3, 2008 FACSET Eric FACSET sums the CC flux over all facets except for standard calibrators. In that case, it was summing only the flux within the central radius of the source but it allowed that radius to be placed at the center of each facet. Fixed code so that it looks only at facet 1 in the standard calibrator case and says as much at the end (when there is > 1 facet). Moved nowhere. 12698. September 3, 2008 CALIB Eric CALIB rounds times outwar when looking at scans from the index table. This has always given trouble so I have removed all round outs. Now we wait to see what goes wrong this way. Moved nowhere. 12699. September 5, 2008 CCEDT Eric Changed empty output CC file to a warning with actually writing out the CC file instead of a quit now error. Moved nowhere. 12700. September 5, 2008 FACSET Eric FACSET confused the subscript order in the alternative names and so used the wrong radius for names for the cal sources. Moved to patch Sept 10. 12701. September 5, 2008 IMAGR Eric Changed WINDF to allow a code to state that no boxes are desired for a field: nnnn 0 0 0 0 Added words to BOXFILE and IMAGR help files for this. Moved nowhere. 12702. September 5, 2005 modeling Eric CCSMEM did not handle the case in which a CC file has no valid components (either empty or all flagged) correctly. It regarded this as an error and bailed out causing TABED modified CC files to fail to model correctly in CALIB et al. This only becomes a problem when we were trying to do many fields at once which is a relatively new capability. Moved nowhere. 12703. September 8, 2008 FITLD Eric The attempt to add INFILE to the history file was foiled in 2 of 3 places by the use of an incorrect variable name. Moved nowhere. 12704. September 8, 2008 CALIB Eric Suggestions from Olaf Wucknitz (Uni-Bonn): CLBPA: Error branch kept solution from propogating to 2nd polarization when polarizations averaged. CALIB: Count of bad solutions used pol 2 for counting twice, ignoring pol 1, a typo CALIB: The output routine to the SN table messed with the weight of the reference antenna, forcing it to be at least SNRMIN+1. This messes up counting of bad solutions and passing failed solutions as 1,0. Dropped the special treatment of the ref antenna and the change in weights. Moved nowhere. 12705. September 8, 2008 FIXBX Eric Changed it to ignore totally all boxes in BOXFILE even if INFILE contains no boxes for a facet in BOXFILE. The output will be given a default box (inscribed circle) if no other boxes appear. Moved nowhere. 12706. September 8, 2008 LPCAL Eric NCOMP is used in a very non-standard way in this task. Fixed the Help file to say the correct thing. Moved nowhere. 12707. September 10, 2008 FITLD Eric Changed to take CORRELAT more seriously, checking value against array name and turning off VLBA corrections when CORRELAT is not ' ', 'VLBA', or 'DIFX'. Then CORVER can be used with other software. Moved nowhere. 12708. September 11, 2008 UVPLT.FOR Leonia The recent modification of UVPLT includes condition .EQ.42 to carry out the plot of U vs V (or V vs U) under control of APARM. The product 6x7 is equal 42. But the product 14x3 is equal 42 also. This fact may corrupt the plot of hour angle(14) vs UV dist(3). So I have changed the condition. Moved nowhere. 12709. September 12, 2008 FILLM Eric The new code to show receiver ID was used before it was set and so, once set, it never changed from an unwanted band. Added setting code in MCWANT which must translate some parameters before doing the full header translation to see if the full translation is desired. Moved nowhere. 12710. September 17, 2008 VLBAUTIL, VLBATECR, VLBACPOL Amy VLBACPOL - fixed the 'FIT FOR BASELINE' to refer to correct baseline. VLBATECR - added 'zcat' to the possible uncompress commands to use on IONEX files. Some recently encountered operating systems no longer have 'uncompress'. Moved nowhere. 12711. September 17, 2008 COMB Eric Added DOHIST to allow suppression of the history file copyiing. DOHIST=-2 => copy the 1st HI file. DOHIST=-3 => no copy of HI file but add COMB HI and comments about no copy. Moved nowhere. 12712. September 23, 2008 IMAGR Eric Develop the ability to grid multiple facets at one time and then take advantage of that in IMAGR. Changed UTILITY Added OBFSIZ to return the size of the disk file attached to an object. Used to decide if the UV work file is large or if the OS will hold it in memory. BEMFIT New: fits elliptical Gaussian to beam image known in imaging commons. Called by IMGMEM, needed to make call before extended resolution "beam" writes over the normal tapered beam. QNGRD New: A version of Q1GRD but it must move data from protected area to work area, do the 3x3 matrix multiply on u,v,w to rotate to the new facet phase center, check/fix for rotated U negative, and check/blank when rotated u,v outside image as well as the taper, freq adjust etc done in Q1GRD. Q1FIN Cleaned up typing while correcting error. It tested for the symmetric row being below the address of the X convolving function rather than below the start of the grid. Fortunately, this test never has anything to do. IMGMEM New call sequence to give a list of fields to image this time including their tapers. Now loops over field to prepare gridding parameters, then calls GRDMEM to grid all request fields, then loops over fields to do FFT and grid correct and beam fit if appropriate. GRDMEM New call sequence to pass in a list of fields and tapers and send bak lists of GRID addresses, max Y row used, and sum of weights. Now loops over fields to init parameters and set grid addresses. Then loops through data filling data buffers into the AP and, when needed, looping over fields to grid the data now in AP to the fields. Finally, loops over fields with zero spacing and finishing operations. GRCMEM New call sequence to pass in the scaling parameter rather than depend on common variable BEMMAX(field). GRDFIT New gridding routine reuires 2 copies of data and it was noted that data come in blocks of the buffer size rather than simply the logical record size when moved to the AP by GRDMEM. QUVUTIL OUVIMG: change call sequence meaning to allow IFIELD to be optionally a forced list of fields in addittion to a single number or 0 => all. Sum total memory needed as well as max for 1 image. Decide FAST (=> IMGMEM) based on largest single image but ask for AP memory enough to do everything. Call GRDFIT and sort only if not FAST. Break the DOBEAM and not DOBEAM segments into not FAST (MAKMAP essentially unchanged) and FAST segments. In the latter, make a list of the fields that will fit at the moment and call IMGMEM with that list. Manually loop for the next list if any. In the image FAST section, break it into 2 parts one of which does consecutive field numbers and the other of which supports the forced list. QCLEAN CLNUV2: Put in code at start to decide how many facets (MXLIST) to image each time we look for the next strongest. Large work file => want more, big images prevent more. Use list returned by CLBSTF - if the top of the list has not been imaged lately, image as much of the list as is not current and then call CLBSTF to make a new list of apparently most desirable facets. If the top of the list is current, then clean that facet. CLBSTF: changed to make an ordered list of apparently desirable fields rather than returning only one with a clue about the second. CLORDR: new to insert a field in the ordered list when appropriate. Moved nowhere. 12713. September 23, 2008 model Amy/Eric Putbck Amy's new model : 3C147 at C band. Moved nowhere. 12714. September 24, 2008 IMAGR Eric Fixed QCLEAN to remember the actual score for the top file after re-imaging and if the apparent top score for the next group is close, use the already re-imaged facet rather than image more. This option has been present under control of IMAGRPRM(20) for some time, but has been lost for 24 hours. Moved nowhere. 12715. September 24, 2008 SIZEFILE Eric New verb SIZEFILE and adverb FSIZE added to POPSDAT.HLP and NEWPARMS.001. New help files written. Changed AU3 to implement and added comments to SETMAXAP.HLP. Returns file size in Mbytes, interpreted sometimes by BCHAN,ECHAN,BIF,EIF, NCHAV, CHINC. Moved nowhere. 12716. September 26, 2008 GRDMEM Eric One too many words were addressed in the zeroing of the AP memory - apparently okay under GNU and Intel on Linux boxes but not okay under Max OS/X with the Intel compiler. Moved nowhere. 12717. September 26, 2008 DBCON Eric Added code to copy the first file's keywords to the output. Also added code to refuse to concatenate two data sets with opposite CORRCOEF settings. Moved nowhere. 12718. September 29, 2008 IMAGR Eric Corrected: GRDMEM It used LROW as a synonym for NY so LROW must be of the dimension MXIMAG and subscripted. Messed up when not all images are the same size. QCLEAN Did not offer the TV option to look at any field when there are > 64. QCLEAN Remade all images when one tried to force a field. TVUTIL Used the image header in COMMON which is not guaranteed to be correct for the current image when drawing the inscribed circle in OTVBOX. Moved nowhere. 12719. September 25-30, 2008 header keywords Eric Most tasks fail to copy the header keywords to the output file sometimes for partial good reason. Wrote new routine KEYPCP to copy all keywords except for specified list with a default list of ISCALE, IZERO, ACTNOISE, and ACTMEAN. Added KEYCOP to tasks BSROT DIFUV DSORC DSTOK REAMP REMAG SAD SDLSF SDMOD TRANS UVCMP UVPOL ACFIT AVER BLAVG BLOAT BSMAP CLINV CORER DAYFX DESCM DFQID DIFRL DQUAL FIXAL FLGIT FXTIM FXVLA HAFIX HGEOM LTESS MATCH MSORT MULTI PADIM PGEOM PHNEG PHSRF REGRD SBCOR SDCAL SKYVE SPLAT SPLIT SWPOL TBAVG TI2HA UJOIN UVAVG UVBAS UVDEC UVDGP UVDI1 UVFIX UVGLU UVIMG UVLIN UVLSD UVLSF UVMTH UVNOU UVSRT UVWAX VBCAL VBGLU VBMRG VPFLG WTMOD XTRAN CALIB CVEL FRING HORUS SDVEL UVADC UVSUB MX UVMAP Added KEYPCP with default list only to AHIST COMB HISEQ MCUBE SCLIM SPIXR SQASH SUMIM TAFFY WTSUM XMOM XSMTH XSUM BDEPO DCONV IMFIT IMLIN IMMOD JMFIT LAYER MATHS MOMNT MWFLT NINER POLCO RGBMP RM RMSD SHADW SMOTH STESS STFUN SUMSQ UVHIM CONVL FFT IM2UV IMERG RSTOR APCLN APGS APVC SDCLN STEER UTESS VTESS BLANK BSCLN SERCH XGAUS GAL Added KEYCOP to OHCOPY in HISTORY.FOR which will cover all of OOP codes. Moved nowhere. 12720. October 1, 2008 UVMOD Eric For some reason the HI write and table copy came before the UV subroutine. The actual number of vis written did not then make it to the header. The calibration read routines survived this, but not more normal routines. Moved nowhere. 12721. October 1, 2008 TYAPL Eric Added a default for INVERS = IN2VERS = 0. moved nowhere. 12722. October 2, 2008 UVCOP Eric Added option (UVCOPPRM(6)) to flag the TY tables with the flag table used on the UV data. Moved nowhere. 12723. October 3, 2008 preparation Eric Changed QUVUTIL, FACSET, VSCDFT, CCSMEM, VSFDFT, GRDSUB, VISDFT, and GRDCRM to use ABS(CCVER(i)) rather than CCVER directly. This is in preparation for the sign of CCVER to convey the fact that it is a temporary CC file. Moved nowhere. 12724. October 3, 2008 3C48 X-band Eric/Amy There have been reports of trouble with the 3C48 X-band model previously released. It was based primarily on an 8-hour Craig Walker run from ~1991. It has been replaced with a model based primarily on other data sets which works better for most data although not Walker's. This mystery is not understood. Moved nowhere. 12725. October 7, 2008 CCRES Eric Added FACTOR adverb to control scaling the residual image. Added default scaling of the residual image to correct for "beam area". Added OPCODES 'ADDP' and 'S+AP' to put the components back as 1-pixel points. Added code to test the actual scaling to Jy/beam - the beam area for largish beams is about what theory would say (1.1331 * bmaj * bmin in pixels) but for small beam, the fact that one samples only at a few points makes the actual integral of the Gaussian something else. This made of order 18% difference with a bmaj and bmin intended to make 1.1331*bmaj*bmin be 1.0. With the '...P' change output units to JY/PIXEL'. Moved nowhere. 12726. October 8, 2008 UVCOP Eric Extended the UVCOPPRM(6) option to flag tables tp a bit pattern with the 1 bit for TY and the 2 bit for SN and added SNFSEL and SNFLG subroutines to UVCOP. Moved nowhere. 12727. October 9, 2008 APCLN Eric A typo caused APCLN not to link edit - why did this not appear previously? Moved nowhere. 12728. October 9, 2008 modeling with images Eric Modeling with images works fine if the images are suitable except that subimages failed. Created two new routines: IMGSIZ figures out a power of 2 image size that will allow the reference pixel to be at the usual center SUBGET a simplified version of PLNGET that does the placing of a subimage within the larger area of zeros. Changed routines: FFTIM to call SUBGET rather than PLNGET SETGDS to call IMGSIZ to set FLDSZ GRDAT to call IMGSIZ to set FLDSZ GRDSET to call IMGSIZ to set the size of scratch files Moved nowhere. 12729. October 13, 2008 VLANT Eric Removed FREQID adverbs and made task do all FREQIDs at once. Otherwise the AN table gets changed Number of FREQID times. Moved nowhere. 12730. October 13, 2008 Create new adverbs CALIN, INLIST, DATA2IN in POPSDAT and NEWPARMS. Changed which adverbs are used by tasks to reduce the name conflicts: ANCAL INFILE -> CALIN ANTAB INFILE -> CALIN APCAL INFILE -> CALIN ATMCA OUTFILE -> OUTTEXT BLSUM OUTFILE -> OUTTEXT BOXES INFILE -> INLIST BPCOR INFILE -> CALIN BPERR INFILE -> INTEXT CANDY INFILE -> DATAIN CONPL OUTFILE -> OUTTEXT CPYRT INFILE -> INTEXT DFCOR INFILE -> CALIN DTCHK OUTFILE -> OUTTEXT EXTAB OUTFILE -> OUTTEXT FACES INFILE -> INLIST FETCH INFILE -> DATAIN FILLM INFILE -> DATAIN, IN2FILE -> CALIN FIT2A INFILE -> DATAIN, OUTFILE -> DATAOUT FITAB OUTFILE -> DATAOUT FITDISK OUTFILE -> DATAOUT FITLD INFILE -> DATAIN FITTP OUTFILE -> DATAOUT FRMAP OUTFILE -> OUTTEXT FRPLT OUTFILE -> OUTTEXT GAL OUTFILE -> OUTTEXT, INFILE -> INTEXT GSCAT INFILE -> DATAIN HINOTE INFILE -> INTEXT HITEXT OUTFILE -> OUTTEXT IMEAN OUTFILE -> OUTTEXT IMLIN INFILE -> INTEXT IMLOD INFILE -> DATAIN IMRMS OUTFILE -> OUTTEXT IMTXT OUTFILE -> DATAOUT INDXR IN2FILE -> CALIN MOMFT OUTFILE -> OUTTEXT OTFBS INFILE -> DATAIN OTFIN INFILE -> DATAIN OTFUV INFILE -> DATAIN, IN2FILE -> DATA2IN PBEAM OUTFILE -> OUTTEXT PCLOD INFILE -> CALIN PLOTR INFILE -> INTEXT POSSM OUTFILE -> OUTTEXT PRTTP INFILE -> DATAIN READISK INFILE -> DATAIN SERCH OUTFILE -> OUTTEXT SETAN INFILE -> INTEXT SETFC INFILE -> INLIST SHADO OUTFILE -> OUTTEXT SLCOL OUTFILE -> OUTTEXT SLICE OUTFILE -> OUTTEXT STARS INFILE -> INTEXT TBIN INFILE -> INTEXT TBOUT OUTFILE -> OUTTEXT TCOPY INFILE -> DATAIN, OUTFILE -> DATAOUT TPHEAD INFILE -> DATAIN UVCRS OUTFILE -> OUTTEXT UVFLG INFILE -> INTEXT UVLIN INFILE -> INTEXT UVLOD INFILE -> DATAIN VLBAARCH INFILE -> DATAIN, OUTFILE -> DATAOUT VLBAIT INFILE -> DATAIN VLBALOAD INFILE -> DATAIN VLOG INFILE -> CALIN WRTDISK OUTFILE -> DATAOUT XTRAN INFILE -> INTEXT Changed source code for new adverb names: AU4 for TPHEAD AU7B for HINOTE, HITEXT Changed $RUNSYS procedures: DDTLOAD DATAIN for FITLD, DATAOUT for FITTP FITDISK DATAOUT for FITTP HYB OUTTEXT for IMEAN OTFPROC DATAIN, DAT2IN for OTFUV VLACLOAD DATAIN for FITLD, DATAOUT for FITTP VLALLOAD DATAIN for FITLD, DATAOUT for FITTP VLARUN INLIST for SETFC VLBAARCH DATAIN for FITLD, DATAOUT for FITAB VLBAIT DATAIN for FITLD, OUTTEXT for POSSM VLBAUTIL change 1 comment re DATAIN for FITLD WRTPROCS DATAIN for INFILE, DATAOUT for OUTFILE, spell adverbs completely Y2KLOAD DATAIN for FITLD, DATAOUT for FITTP Changed descriptive help files: ASTROMET DATAIN for FITLD, INTEXT for UVFLG, CALIN for PCLOD, OUTTEXT for BPASS Changed allowed DOHIST range in all relevant help files to allow the -3 used in COMB. Moved nowhere. 12731. October 14, 2008 SPFLG Eric Corrected bug setting the pix-ranges for display modes > 4 - including the 3 difference modes. Raised the buffer size which should speed up the initial I/O and also allow larger scan lengths. When scan length is lowered due to memory limitations, add message. Fixed TSKHLP to first try for exact match in the names of the functions and then do the min match. DISPLAYY PHASE was coming out when DISPLAY PHASE DIFF was desired. Moved nowhere. 12732. October 15, 2008 adverbs Eric Corrected a typo in NEWPARMS.001 and added to it and POPSDAT adverbs for ATLOD from ATNF - NSKIP, NIF, IFMAP, FREQSEL, IFSEL, XYPHASE. Moved nowhere. 12733. October 15, 2008 Nasmyth telescope mounts Eric Richard Dodson has supplied code modifications to allow AIPS to support a variety of new mount types: 0 alt-az, 1 equitorial, 2 orbiting, 3 X-Y, 4 right-handed Nasmyth, 5 left-handed Nasmyth (0,1,2 are old codes) 6 prime focus/folded cassegrain Changed CLCOR ANAXIS allows mount >= 3 to do as mount=0 DFCOR as CLCOR DTSIM Exclusion of orbiting antenna mount types needs to NE.2 now rather than LT.2. Extra computation for Nasmyth and X-Y mounts in s.r. ORIENT. PRTAN Add to list of mount types SNPLT Add code for position angle by more mount types in s.r. ANTGEO APCAL Orbiting antenna selectioon is .EQ.2 not .GT.1 TABAN add to pre-cursor comments PARANG Major changes: add code to call SOUELV for Nasmyth and to compute extra for mount types 3,4,5,6. Moved nowhere. 12734. October 17, 2008 UVPLT, POSSM Eric Corrected log amplitude plot in UVPLT to be LOG10 not natural log. Added CODETYPE = 'LA&P' and 'LAMP' to POSSM to plot BPARM(5) * LOG10 (amplitude). Corrected bug that caused the imaginary plot in R&I to have no plus signs. Moved nowhere. 12735. October 17, 2008 GAL Eric Added to the OUTTEXT output when CPARM(6) not zero and OUTTEXT not blank. It will list both ends of the major axis and the rotation curve: radius (arcsec), velocity, and rms of velocity (km/s). Moved nowhere. 12736. October 20, 2008 PARANG Eric Put in a corrected version of pARANG from Richard Dodson. This one is simpler and leaves out the unused type 6. Link edit all direct calls and all DOPOL tasks. Moved nowhere. 12737. October 20, 2008 FILLM Eric The change of September 12 (#12709) had unintended consequences. The band changed to the new band before the present scan was finished. The only known consequence was to the last record in the CL and TY tables. The latter would matter only for 4 or P band and would cause a different system temperature to be chosen. The CL table however would have the atmospheric opacity and antenna gain correction for the next band rather than the present band. This could make a substantial (e.g. 5 %) difference in the gains. Use of INDXR to make a new CL table erases this problem. Changed the code to find the new band only for deciding about keeping the data or not but leaving the band ID unchanged until the data are fully absorbed. Moved nowhere. 12738. October 21, 2008 IRING Eric Moved the setting of XRNG for plotting to after the filling in of defaults for CPARM and created axis units for pixel-oriented plots. Moved nowhere. 12739. October 28, 2008 AIPS Eric Created new verbs ASIN and ACOS. Changed POPSDAT, POPSYM, NEWPARMS, KWICK.FOR and created ACOS and ASIN help files. moved nowhere. 12740. October 29, 2008 string adverbs Eric Changed POPSDAT, NEWPARMS to make new adverb FITOUT Created DATA2IN, CALIN, INLIST, FITOUT help files with full listing of tasks/verbs that use them. Updated INFILE, IN2FILE, OUTFILE, OUTPRINT, INTEXT, OUTTEXT, DATAIN, and DATAOUT help files with full lists of who uses them. Changed usage in help files for: TPHEAD INFILE -> DATAIN (AU4 was already changed) IMFIT OUTPRINT -> FITOUT JMFIT OUTPRINT -> FITOUT SAD OUTPRINT -> FITOUT Moved nowhere. 12741. October 30, 2008 UVCMP Eric Dropped OPTYPE from the adverb list - it will now always switch between the compression states. Moved nowhere. 12742. October 30, 2008 IMAGR Eric Changed name multi-resolution to multi-scale in help files IMAGR, IMAGRPRM, and BOXFILE. Moved nowhere. 12743. October 30, 2008 lists Eric 1. Updated TAB completion list $HLPFIL/HLPIT.LIS 2. Changed SHOPH.FOR to put out a directly useful message. 3. Changed ABOUT files: ADVERB ANALYSIS CALIBRAT EDITING FITS GENERAL HARDCOPY IMAGING MODELING OBSOLETE ONED OOP PARAFORM PLOT POPS PROCEDUR SINGLEDI SPECTRAL TABLE TASK TV UV VERB VLA VLBI 4. Fix categories in TIORD, UVFIL, CANDY help files 5. Cleaned up LSAPROPO - it was okay but had numerous places where the same line was repeated. Moved nowhere. 12744. November 3, 2008 CookBook Eric 2: A few notes re XAS changes 3: INFILE, OUTFILE changed; more text file name adverbs; FILESIZE & SETMAXAP 4: INFILE, IN2FILE changed; UVFLG codes WILD/REAS; UVCOP flag TY before TYSMO; WIPER falgs baseline; TVFLG channel averaging; FILLM OF tables (OFLAG, PRTOF) 5: INFILE changed; reword a little re IMAGR option to remove weak/isolated components; multi-scale not multi-resolution; adjust one SHIFT wording; reference CCRES in units mis-match discussion 6: INFILE changed, XAS changes -> TVON etc, TVCPS, TVHUEINT 7: CCRES changes units of residual, CONVL 8: INFILE changed 9: INFILE, OUTFILE changed some places; add REAS/WILD to UVFLG; VPLOT 10: INFILE, IN2FILE changed 12: Added ACOS and ASIN; INFILE, OUTFILE changed for FITS 13: Updated all 20 changed ABOUT lists B: INFILE -> DATAIN, UVFLG, archive web site, mention multi-scale C: INFILE changed to INTEXT, CALIN; Recipe: rearranged only LaCook.PL: update version Moved nowhere. 12745. November 10, 2008 VBLAUTIL, VLBATECR Amy Fixed a bug I added with my September change to VLBATECR. It was only triggered when downloading multiple files. Moved nowhere. 12746. November 17, 2008 VTESS Eric Raised format size for number iterations. Moved nowhere. 12747. November 17, 2008 IMAGR Eric Changed Fortran in IMAGR.FOR to allow DO3D if necessary to do ATNF with multi-scale or large numbers of facets with a warning. Moved nowhere. 12748. November 18, 2008 FILLM Eric The subroutine that is used when CPARM(4) is not 0 or 25.0 had an error in computing sqrt (u^2 + v^2) due to a misplaced close parenthesis. abs(u) + v^2 was examined instead. This matters since apparently the on-line system has been setting the shadow-flag bit wrongly and that bit is used when CPARM(4)= 0 or 25.0. Clarified this in the help file. Moved to 31DEC07 this date.. 12749. November 20, 2008 INDXR Eric Changed limit in INFILE option from 50 to 5000 breaks. Moved nowhere. 12750. November 24, 2008 LOCIT Eric Raised max points from 400 to 50000 for unreasonable people. Added test and die for overflow. Added a flag for when no OT table is present to read SN table anyway. Moved nowhere. 12751. November 26, 2008 DOUVCOMP Eric Changed POPSDAT.HLP and AU2A (verb DEFAULT) to have the default value for DOUVC be -1 (false). Moved nowhere. 12752. December 7, 2008 TABSRT Eric The sort of BP tables in particular is inefficient since there are a lot of data values in each record. Added to TABSRT a call to a new subroutine BTBSRT which does a "brute force" sort. It makes a list of the 2 key values and record number, sorts that, and then does a gather read to the output file. Moved nowhere. *********************************************************************** *********************************************************************** Changes while 31DEC08 is NEW *********************************************************************** *********************************************************************** 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. 12758. 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. 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. 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. 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 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. 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 drop IMGSIZ call and set FLDSZ to max of its current value and the image's size. SETGDS to defer the setting of FLDSZ until the full set of images are examinined and the MODEL 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 spexcial 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. 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, 2008 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. 12787. January 6, 2009 CALIB Eric This business at the end of scans is still getting us. Changed the latest changed code so that if 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. 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. ============== now patches =============================== 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. 12803. February 4, 2009 TYAPL Eric It applied the FG table since the SELINI default is FGVER=0. Moved to patch this date. 12824. April 1, 2009 OFMs Eric The files in $AIPSOFM havingvarious 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. Moved to patch this date. 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. 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. 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. 12863. June 16, 2009 IMAGR Eric 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 10 % and try a second time to avoid a sort. Moved to patch June 19. 12905. 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. 12923. August 18, 2009 FITLD Eric When rearranging the frequencies in reading IDI-format data, FITLD did not correct the ANtenna file frequency. This is used by UVFIX and so is important. Moved idea from 31DEC09 this date. 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 fix from 31DEC09 this date. 13021. November 30, 2009 ANFIX8 Eric Created a RUN file ANCHECK 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.