;----------------------------------------------------------------------- ; Copyright (C) 2014-2016 ; 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 31DEC15 is TST *********************************************************************** *********************************************************************** 14498. December 2, 2014 Version update Eric Changed $APLSUB/GETRLS.FOR. $INSUNIX/INSTEP1, $SYSPERL/install.pl, $SYSPERL/update.pl, $SYSPERL/ZXHLP2.PL, $SYSPERL/LaCook.PL, $UPDUNIX/UPDCONTROL, $SYSUNIX/AIPSPATH.CSH, $SYSUNIX/AIPSPATH.SH to make 31DEC15 instead of 31DEC14 - also fix dates in install.pl. Moved nowhere. 14499. December 4, 2014 SAD, TVSAD Eric Changed: SAD: Subtract Gaussian over wider area, TVSAD suggested that the previous setting really did not go far enough. TVSAD: 1. Add CURVALUE option to both menus. 2. Have CURVALUE run when setting the Gaussian parameters. 3. Allow up to 8 (parameterized) Gaussians in any one island (only under user command). Also TVSAD and HLPTVSAD help files edited. Moved to 31DEC14 this date. 14500. December 6, 2014 TVSAD Eric Corrected the menu placement which should not expand if the character size expands (and it was too far in in any case). Moved to 31DEC14 this date. 14501. December 11, 2014 Misc Eric Changed 1. KWIKHD to understand XX/YY polarization 2. SAD to subtract components over a much wider range of rows 3. TVSAD as SAD Moved to 31DEC14 this date. 14502. December 11, 2014 AIPS Memo 119 Eric AIPS Memo 119, describing the use of the new TVSAD task, has been put in $AIPSPUBL in PS (very large), PS.gz, and PDF forms. $AIPSPUBL is in all releases. 14503. December 11, 2014 UVSUB Eric removed simple prohibition to allow XX/YY to be modeled with I models. Moved to 31DEC14 this date. 14504. December 11, 2014 CENTFQ Eric Changed this subroutine, called by numerous tasks, to adjust the FQ table for centering the frequency reference channel. The routine did not consider that the column numbers and pointers might differ from the input to the output files. But if an old FQ with no bandcode comes in a new one with blank bandcode will go out. So sepearte numbers/pointers for the two files are required. Moved to 31DEC14 this date. 14505. December 12, 2014 UVSUB Eric The desire to divide the RL and LR correlations by the I model leads to the addition of OPCODE 'DIV4' to UVSUB and its help. Also change: DGDS.INC Add logical DOALL4 GETCTL Now can have 4 factors and set NSTOK to 4 if DOALL4 is true and the model is IPOL. ALGMEM FACT2 now dimensioned 4 and all 4 need scaling ALGSTB FACT2 now dimensioned 4 and all 4 need scaling VISDFT FACT2 now dimension 4, scale all 4, in one mode set all 4 based on DOALL4, reorder writing to AP memory since subtract will never have DOALL4 true and divide does not care about "keep data" VSCDFT Dimension FACT2 as 4 since calls GETCTL VSFDFT Dimension FACT2 as 4 since calls GETCTL UVCON Dimension FACT2 as 4 and reorder loading of AP memory EVAUV Dimension FACT2 as 4 and reorder loading of AP memory UVSUB Recognize DIV4 as both DIVIDE and DOALL4 true. UVSUB Help: add OPCODE 'DIV4' Moved nowhere. 14506. December 12, 2014 CLIP Eric The type of clip level to use for V was compromised by a typo. Should work now. Moved to 31DEC14 this date. 14507. December 15, 2014 HUINT Eric Add remarks about HUINT to help files for TVHUI and TVHUEINT. Also corrected typo in AIPS Memo 119. Moved to 31DEC14 this date. 14508. December 15, 2014 INPUTS Eric Setting the page full suppression at DOPRINT=-2 hits too many other uses - so changed it to <= -9. Changed AU1A and help files for INPUTS and INP. Moved to 31DEC14 this date. 14509. December 18, 2014 CORFQ Eric Changed it so that FACTOR < 0 causes the uvw to be scaled but not the header and antenna-table header frequencies. This can be used to undo the unfortunate re-scaling of holography "u,v,w"'s in almost all AIPS tasks. Moved to 31DEC14 this date. 14510. December 18, 2014 UVHOL Eric Added adverb DPARM to specify the max weight, amplitude, and u,v for the display. The self-scale routine remains present if DPARM(1) < 0 but is skipped unless requested. Corrected the display of channel and frequency to use the correct average channel and frequency computed with the average channel and BIF. Moved to 31DEC14 this date. 14511. December 23, 2014 TVHLD Eric NEW TASK: The old tasks TVHXF and TVHLD have been removed. They depended on hardware functions of the IIS Model 70 (and perhaps 75) which are not available in XAS and other TV devices. The new task, written in modern AIPS style, attempts to offer greater functionality than the old TVHLD. It loads the TV with a "histogram equalized" version of the input image. An interactive menu allows you to control the PIXRANGE used in the equalization and select various functions to be performed on the normal histogram before it is used to do the equalization. Some of these functions enhance the higher levels of the image relative to the lower levels more than others. The usual image display/enhancement options are also offered. When done, you may write out the equalized image or you may exit without writing an image. Also created new help files TVHLD and HLPTVHLD (real-time help for menu). Moved nowhere. 14512. December 26, 2014 AIPSLetter Eric Put back LET14B.{TEX,PS,PDF} in $AIPSPUBL. Moved to all versions automatically. 14513. December 30, 2014 SNPLT Eric Test on outside plot range not on correct plot types (tested REAL) and type 24, not the PC phase. Moved to 31DEC14 this date. 14514. December 30, 2014 SNFIT Eric NEW TASK: Drift (or driven) scans through a source can produce in the SN-table gains a plot of the primary beam. Made a version of SNPLT that can fit a 2nd or 4th order polynomial to such gains, solving for the time at the peak, the width, and the function. Moved nowhere. 14515. Decmber 30, 2014 EVLA.ant.2014 Eric Copied the on-line EVLA antenna position correction file for 2014 to $AIPSTARS and created a new 2015 on-line file for use for new antenna location corrections. Moved to all versions automatically. 14516. January 2, 2015 DOFARS Eric Changed the help file to emphasize things about BLC and TRC. Values 1,2,4,5,6,7 are applied in TRANS, BLC(3), TRC(3) are applied in the FARS step. Dropped OUT2NAME et al from inputs and help and DOFARS.001 - they were left over from earlier versions of FARS. Moved to 31DEC14 this date. 14517. January 8, 2015 DOFARS Eric I messed up the proc (close quotes issue). Moved to 31DEC14 this date. 14518. January 13, 2015 RMFIT Eric Task needs to copy FQ table (if any) to output residual images but not the model fit images. Moved to 31DEC14 this date. 14519. January 13, 2015 VLBAUTIL/VLBATECR Amy When VLBATECR calculated which files it needed to download it did not handle observations that went over the new year correctly. Fixed. Moved to 31DEC14 Jan 19. 14520. January 15, 2015 HOLOG Eric/Rick NEW TASK (sort of) - a simplified version of HOLGR which i will leave alone. Rick Perley developed the simplifications and I have worked over the code. The most significant changes were to make it play well with the UVHOL and PANEL program that precede and follow the HOLGR step. The DFT routine was changed to use double precision internally which changed its answers at the 1 part in 1000-10000 level. Moved nowhere. 14521. January 15, 2015 PANEL Eric/Rick NEW TASK: Rick's panel.f developed initially by Mkie Kestevan was converted into an AIPS task. The pgplot plots were replaced with AIPS plots. The code was restructured to suitable subroutines. The unformatted file inputs were replaced with reading the AIPS images from the AIPS files directly and reading text files for the panel layout (if not EVLA which is available without a text file). Blanking under the feed legs and subreflector is now handled with a blanking image (the V_AMP image from HOLOG). The task gets some of the parameter values (e.g. frequency, mapsize, # pointings) from the input image header and offers controls over which plots are desired and allows suppresion of portions of the plots (grey scale, contours, panel layout). The fitting routine was enhanced to take the masking into account and to avoid singular matrices when a panel is fully masked. Moved nowhere. 14522. January 16, 2015 VLBARUN Amy Fixed two small gotchas in VLBARUN. If a user starts aips with "notv" then then default TVCHAN=1 is invalid, this caused KNTR to die which then caused VLBARUN to die; changed TVCHAN=0. Second if the command "convert" does not exist on system and OUTFILE is set (i.e., the user wants output plots in a html file) VLBARUN would run, but no plots would be produced and the html file would be useless. Changed VLBARUN to check if OUTFILE was set that "convert" exists, and not to run if it does not. Also undated HLP file to say that "convert" is required for those features. Moved to 31DEC14 Jan 19. 14523. January 16, 2015 FTFLG Eric Some vestiges of the way SPFLG handles baselines remained in the code and they were not correct for FTFLG. It led the output flag table to be incorrect when one or all baselines to one antenna were specified in the ANTENNAS and BASELINE adverb. Moved to 31DEC14 this date. 14524. January 19, 2015 HOLGR Eric Changed Fortran to do the DFT in double precision, to put out a FREQ axis in the output images, to output a keyword for use by PANEL, and to read the frequency from the input file. Moved nowhere. 14525. January 20, 2015 BPASS Eric One array was dimensioned 500 and should have been MXBASE. ALMA with 36 antennas breaks this badly. The history routine listing antennas also had a typo bug and was not general and so failed in this case. Moved to 31DEC14 this date. 14526. January 21, 2015 IMAGR Eric Changed: IMAGR When DOTV is true at the beginning, DOWAIT is now assumed. If OBOXFILE is blank at the beginning, a temporary OBOXFILE='HOME:OBOXtempYYYYMMDD.HHMMSS' is created and used. If the final OBOXFILE is this file, then it will be deleted inthe history routine at the end. QCLEAN In CLNTV, when TVBOX would delete a boxes, ask for permission. If not YES or yes, do REBOX instead. In CLNTEL, trap a change in OBOXFILE and handle it properly, copying the old one to the new name and, if the old one is the temp file, deleting it. The expansion with date and time is allowed in the TELL function too. IMAGR help file - add info on OBOXFILE and TVBOX. OBOXFILE Help file - added comments about IMAGR options. Moved nowhere. 14527. January 22, 2015 TVHLD Eric The writing of the top label did not take the actual window into account properly. Its position must pay attention to the window and since they change, the graphics plane needs to be zeroed in advance. Changed to have a subroutine write the top label line and to call that routine in the load function (where it was before) and also ahead of the menu function. This should trap many of the window size/shape changes promptly. Discovered a buffer used for OFM that was only 8192 words, just 1/4 of the required size. Changed TVSAD to make its comparable buffer be equivalenced to adequate buffers. Moved nowhere. 14528. January 26, 2015 INPUTS Eric Setting the page full suppression at DOPRINT=-2 hits too many other uses - so changed it to <= -9. Changed AU1A and help files for INPUTS and INP. Somehow this did not get into 31DEC15 (except CHANGE.DOC) but it is in 31DEC14. Moved to 31DEC14 15 DECEMBER. 14529. January 26, 2015 INPUTS, GO Eric Changed AU1A and AU2 to let the inputs section of the help file set the lower limit of "system" adverbs (disk, tape, TV, etc). In this way, tasks, varbs and procedure that require INDISK > 0 can enforce it. Moved nowhere. 14530. January 26, 2015 DOFARS Eric DOFARS does not use OUTCLASS so remove it from the help and run files. It does require INDISK > 0 and IN2DISK > 0, so put that requirement in the procedure and in the inputs and help for DOFARS. Moved nowhere. 14531. January 26, 2015 SCIMG, SCMAP Eric Changed SCIMG and SCMAP to support TV image labeling, more general OBOXFILE usage (see IMAGR) including stars and changing with TELL, and correcting the setting of the output UV data sort order. Required changing QCLEAN so that the name of the internal WRIBOX does not conflict with the $APLSUB version used by these tasks (no UN-boxes). Changed the self-call TVBOX command to request permission to delete pre-existing boxes and both to allow interactive text input in the Clean section. Moved nowhere. 14532. January 27, 2015 REFREQ Eric Removed this dangerous RUN file and its help. It changed the offsets in the FQ table so that the first IF offset was not zero so that it could move the frequency reference pixel without have to copy and change the u,v,w values. There are places in AIPS where this would lead to bad values of u,v,w being used since not all tasks read the details of the FQ table. Task CENTR and adverb FQCENTER in many other tasks deals with this problem correctly. Moved nowhere. 14533. January 29, 2015 MORIF Eric Added option to combine IFs as well as subdivide them. There was a bug in the handling of BP-like tables that made a mess of the results. Moved bug fix to 31DEC14 this date. 14534. February 3, 2015 NOIFS Eric The SY tables were copied with all IFs. They need to be selected properly. Moved nowhere. 14535. February 3, 2015 TVHLD Eric TVHLD did not put the depth in the image catalog header, meaning TVWINDOWs off of the image were bad. It also used the number of pixels in the working image as the number in the full image, causing it to fail when BLC and/or TRC were specified. Moved nowhere. 14536. February 4, 2015 UVFRE Eric NEW TASK: An HGEOM for frequencies in UV data sets. Regrids the data of one data set onto the frequency structure of another (as best it can). Updated INTPARM help file which was all wrong for current usage. Moved nowhere. 14537. February 5, 2015 consecutive seq numbers Eric Changed 1. MCUBE help file to point out that some of the sequence numbers from INSEQ through IN2SEQ are allowed to be missing. 2. FQUBE Fortran and help to allow the same. 3. STUFFR RUN file and help to allow missing ones. This was rather tricky due to all the intermediate files that get created and then destroyed. Verbs NAMEGET and CHKNAME were essential in this. Moved nowhere. 14538. February 10, 2015 UVPLT Eric Changed the print out of the binned values to include the number of samples going into each bin. Moved nowhere. 14539. February 10, 2015 BPASS help Eric Add phrasing to point out that division by a model is turned off when channel 0 division is done. Moved nowhere. 14540. February 11, 2015 TYSMO Eric The clip parameters (APARMs) were not applied unless the MWF cutoff parameters (CPARM(6-9)) were set. Corrected this to apply the APARMs with or without the CPARMs. Moved to 31DEC14 this date. 14541. February 11, 2015 FRING Eric ALMA produces data with one IF having negative frequency increment and the other with positiv frequency increment. FRING required a couple more absolute value calls in handling this. Moved nowhere. 14542. February 12, 2015 UVCON Eric The inputs section stated that APARM(10) = 0 would become 1 and allow no blockage at all. In fact, the help said the truth, in which 1 means a fully blocked aperture is allowed. Moved nowhere. 14543. February 16, 2015 table format updates Eric The failure to update table formats is not a disaster for some of the updates (AN, SU, FG, BL, BP) but is for SN and CL. Read only file systems are more likely to have old formats and cannot be updated, so changed ANTINI to not die on ANREFM error, updated message SOUINI to not die on SUREFM error, updated message TABSOU to check that RAOBS and DECOBS have a valid column BPASET to issue message but not die on BPREFM error GAININ to issue message but not die on BLREFM error, will die on SNREFM, CLREFM errors TABBL to check the FREQID has a valid column Moved nowhere. 14544. February 17, 2015 UVFIT Eric Added messages about whether it converges or not and did an extra JTRIM on the two text file names (some ZTXOPN issues were reported). Moved nowhere. 14545. February 18, 2015 VLOG Amy VLOG has frequency ranges hardwired, I widened then to the maximum allowed by VLBA recievers. I also added 3mm to the set of frequencies. Moved nowhere. 14546. February 18, 2015 XAS Eric Character writing into grey-scale memories used a background of 1 which means that the characters could disappear with some LUTs. Changed it to zero in XAS and in $YGEN/YCHRW. Moved nowhere. 14547. February 19, 2015 POPS Eric On occasion for no known reason, LTSTOR returns a pointer to a data area already used. This seems to happen when concatenation is being done. Changed CONCAT to drop the saving of the string data to memory which has already been done by the call to LTSTOR anyway. It was that saving that was causing the failures. Moved nowhere. 14548. February 19, 2015 TVMOVIE, TVCUBE Eric Added the adverb DOALL to these verbs to have it loop over all planes on the data "cube". DOALL false loops only over the first axis (> 2) with more than one pixel. Labeling in DOALL true (when there are more axes > 2 with > 1 pixel) is a simple plane number. Changed both help files and AU5D. Moved nowhere. 14549. February 19, 2015 UVIMG Eric UVIMG has been re-written to make images with any axes, not just those of the UV data set sort order. It can also make an image of the count in each pixel. The images may be done with a count of 1 in the nearest pixel, or the standard interpolation functions may be used over a range of pixels around each sample. NCHAV spectral channels may be gridded into each plane, planes are separated by CHINC spectral channels, and there may be multiple IFs, producing a 4-dimensional "cube" (hence the above TVMOVIE change). The image of data parameters may be an interpolated image or a convolved image. Moved nowhere. 14550. February 21, 2015 bandpass Eric The low level, VLB routine DETRAT was still in left-hand antenna coordinates. This was known to the bandpass application routine BPASET but not elsewhere, particularly BPASS. Therefore BPASS made solutions which were not good when applied to the calibration data at least for high-frequency, narrow-channel data. Changed DETRAT to be right handed and BPASET to go along with this. Changed some vestigial bits in BPASS that had no effect. Moved to 31DEC14 this date. 14551. February 21, 2015 GC tables Eric Changed PGCV.INC to reduce the excessive allowance of entries per row of tabulated gains from 200 to 120 and to correct the buffer size which was a factor of 2 too small (did not allow for 2 polarizations I guess). Changed ANTAB to read the KEYIN file twice, the first time to determine the type of GAIN entries and their size and the second to build the TY, GC, et al. tables. Moved nowhere. 14552. February 23, 2015 PANEL Eric Rick gave me the wrong values for the VLA panel radii. He has now "corrected" them (I hope). Moved nowhere. 14553. February 23, 2015 TVHLD Eric The output routine build the output header including the increments which were supposed to apply only to the display. Fixed it to use 1.0. Moved nowhere. 14554. February 25, 2015 SNSMO Eric The VLMB mode uses the average of the phases over all IFs after they have had the phase due to delay removed. They should all be the same at that point, but what is removed is the smoothed multi-band delay, not the unsmoothed one which created the phases to begin with. So they are not the same and the relative phasing between IFs can be damaged. This is especially so since PHASES are what the task smoothed (after trying to avoid phase wraps). So Changed SNSMO help file to add info in the inputs for choices of SAMPTYPE and SMOTYPE. Added statement that BIF/EIF are ignored in VLMB and the new VLDE mode. And added a VLDE mode with lengthy descriptions of it and an improved description of VLMB. SNSMO Fortran file to set BIF/EIF to the full range in VLMB and VLDE modes. Changed to set NPIECE=NUMIF except for VLMB and VLDE. Changed call to smooth amp/phase to loop from BIF to EIF by NPIF where NPIF is number of IFs in the group (# groups or pieces = max of 1 and FRING's APARM(5)-1). The smooth amp/phase now uses multi-band or single-band delay depending on SMOTYPE to correct the phase using all IFs (VLMB) or only the first valid IF in the IF range (VLDE) before smoothing. After smoothing, it applies the smoothed phase plus an appropriate delay correction to each IF in the group. It now smooths real and imaginary separately and computes phases from there. Moved nowhere. 14555. February 25, 2015 IMAGR Eric QCLEAN turned of interaction "ok" when OVERLAP < 2, perhaps once for good reason, but no longer. Leave that flag as the main programs set it. Moved nowhere (31DEC15 bug only). 14556. February 27, 2015 DOVLAMP Amy Procedure to produce amplitude calibration for phased-VLA data used in VLBI. The procedure uses VLAMP and opptionally CALIB (if flux calibrator is present) to produce an ANTAB style file that can be loaded with ANTAB. The procedure attempts to be as automated as possible. Moved nowhere. 14557. February 27, 2015 TVLAYOUT, KNTR, GREYS Eric Made holography panel layout plots more widely available. Changed GREYS Fortran and help to allow INFILE='VLA' or 'VLBA' to mean use standard values of the layout rather than require text files. Added INFILE option to KNTR Fortran and help to plot layouts on each panel of a single or multi-panel plot like GREYS (so can also do "bolts"). Changed POPSDAT.HLP, NEWPARMS.001, AU5B, and made a new help file and new Fortran routine LAYOUT to create verb TVLAYOUT which overlays a layout plot on a TV image. Moved nowhere. 14558. February 28, 2015 Obit things Eric Discovered bad usage in many ZDCHI2.FOR versions. Fixed LINUX, LNX64, MACPPC and LINUXPPC versions to have SYSTYP='UNIX' and SYSVER=$ARCH. Changed $AIPSUB/AUO.FOR to leave off the --pid=$! argument to tail for SYSVER='MAC...'. Moved nowhere. 14559. March 2, 2015 calibration Eric CSLGET fetches the current calibration from the SN/CL table data. It was using time uncertainties of 50 milliseconds in the initial search and 400 milliseconds in setting the overall range (which determines when new data are sought). Since data may now be taken at 100 milli-second intervals and need calibration on that level, this is crazy. Changed both "slop" values to 5 milliseconds. Should go to patch after much testing. 14560. March 6, 2015 VLAMP (GETCDS) Amy VLAMP had a subroutine (CHKBND) in it that was converted to an external existing subroutine with more parameters and the same name. The internal subroutine was never removed and caused VLAMP to die. GETCDS - just removed all the print statments Eric put in to help me debug the VLAMP problem, no real change. Moved nowhere. 14561. March 18, 2015 debug lines Eric Removed debug lines from INDXR, PANEL, QCLEAN, QEDIUTIL, SPLAT, TVFLG, UVCOP, UVIMG. Moved nowhere. 14562. March 24, 2015 OOP sort Eric OOSRT uses the pseudo-AP for data sorting as does IMAGR and other OOP tasks. The lowest level routines in this sort were broken: QOSORT Arguments are DOUBLE PRECISION and LONGINT, call sequence to QPERMA was messed up (to undo errors in QPERMA I suspect) QPERMA Arguments were confused causing completely wrong data rearrangement (if one depended on the comments at the top of the routine). Also Arguments are DOUBLE PRECISION and LONGINT. OOSRT The OUVSRT routine will copy any SU and AN tables. The table copying at the end needs to omit these two (plus FQ) from the final copy to prevent error conditions being raised. Moved nowhere. 14563. March 24, 2015 LISTR Eric Changed the LIST option to include autocorrelations when DOACOR is true. Added separate scaling at 1/100 for the auto amplitudes. Moved nowhere. 14564. March 24, 2015 UV format Eric The BASELINE random parameter limits the number of antennas to 255 [it is 256 * ant1 + ant2 + (isub-1)/100.0]. Changed to allow a different set of random parameters SUBARRAY, ANTENNA1, ANTENNA2. BASELINE will be converted to the new set as the data are read in and will be put back to the old BASELINE as they are written out, but only when the maximum antenna number <= 255. Changed includes and subroutines: DUVH.INC Added ILOCSA, ILOCA1, ILOCA2. DUV3.INC Removed unused include. UVPGET Added search for the new parameters, changed parameter missing complaint to require either BASELINE or all 3 of the new ones. AVERT Added baseline extraction, changed BL code to be 32768 * IANT1 + IANT2 DATFLG Baseline pickup, FLGBAS to base 32768 DATGET Baseline pickup 4 places. DGETSD Baseline pickup GRIDTB Baseline pickup NXTFLG FLGBAS to base 32768. RNXUPD Subarray pickup RUVFIL Determine if need switch to antenna1/2 and set up random parameters RUVTAB Apply baseline -> antenna1/2, subarray UVDESC UVDPNT returns 3 new arguments UVPHAS UVDPNT call seq, baseline pickup UVUTIL several UVDPNT call seq, 2 baseline pickups SRTIN On baseline sort, use ILOCB if present, otherwise compute a base 4096 baseline number from the two antenna numbers using QVSMA. QEDIUTIL Baseline pickup 3 places, UVDPNT call seq 2 places, plus direct fetch of antenna pointers QUVUTIL One call to UVDPNT Changed tasks to alter 1 antenna/baseline pickuo BSROT NANS REAMP REWGT SDLSF SDMOD UVPOL ACCOR ACLIP ACSCL ALVAR BASRM BDAPL BSCAN BSFIX BSMAP BSMOD CLIP DQUAL DTCHK FGDIF FIXRL FLGIT FUDGE FXTIM HA2TI INDXH MULTI PCCOR PHNEG PRTSD REFLG RLCOR RLDIF SBCOR SHOUV SPCAL SPMOD SWPOL TRUEP UV2TB UVDEC UVDGP UVFIT UVIMG UVLIN UVLSF UVLSD UVMLN UVMOD UVNOU UVWAX WTMOD BLCAL KRING LPCAL ANBPL CAPLT CLPLT FRMAP FRPLT POSSM TVFLG Changed tasks in more ways: DIFUV Differentiate the random pointers between the 2 data sets. Antenna pickup separately, drop the averaging of the random parameters (was not correct when scaling uvw etc.) RESEQ 1 antenna pickup and 1 antenna putup UVDIF Add more pointers, allow the 2 to be of different baseline format, numerous antenna pickups changed ACFIT 2 baseline pickup, 1 baseline putup AVER 2 baseline pickups, 2 baseline putups, fancier test for baseline changed BLAVG 2 baseline pickups, 1 baseline putup BPWAY 3 baseline pickups CORER 2 baseline pickup DAYFX drop baseline pickup DBCON subarray adjustment change DECOR baseline pickup and 1 baseline putup DTSIM use ANTENNA1, ANTENNA2, SUBARRAY, not BASELINE FITAB Decide whether to convert to BASELINE if Nant<256 and implement that conversion FITLD Added conversion of standard UV FITS to the new form and conversion of IDI to the new form. FITTP Decide whether to convert to BASELINE if Nant<256 and implement that conversion FIXAL 2 baseline pickups FXVLA 1 baseline and 1 array pickup HAFIX 2 baseline pickups and 1 array pickup LISTR 2 baseline pickups M3TAR to create file with ANTENNA1/2, SUBARRAY instead of BASELINE MATCH 1 baseline pickup, 2 baseline putups MK3IN to create file with ANTENNA1/2, SUBARRAY instead of BASELINE MSORT Special code for ANTENNA1/2, then compute sort key baseline base 4096 (rather than 256) PHSRF 2 baseline pickups PRTSY Changed FLGBAS comment. PRTUV Added new parameters to 2 parameter lists, changed antenna/subarray pickup 2 places REWAY 6 baseline pickups SDCAL subarray set to 1 in new way SPLAT Changed FLGBAS comments. Baseline pickups several places including messy one looking for baseline to do time averaging. SPLIT Subarray drop code adjusted TBAVG 1 antenna pickup, 1 antenna putup TI2HA 1 antenna pickup, 1 antenna putup TIORD baseline to antenna pickup, save baseline base 4096 and pick that up in error message TYAPL Changed FLGBAS to base 32768 in comments and local NXTFLG. Changed 2 antenna pickups. USUBA Changed to use include DUVH rather than pass numerous arguments and to do 2 pickups and putups UVAVG 1 baseline pickup plus encode antennas appropriately in 2 places rather than use ILOCB UVCOP Changed FLGBAS to base 32768 in comments and local NXTFLG. Changed antenna pickup in subarray reassignment and in flagging. UVCRS 2 pickups UVDI1 3 baseline pickups plus encode antennas appropriately in 2 places rather than use ILOCB UVFIL to use the new random parameters (old shown in comment) and 1 putup UVFIX 2 baseline pickups and 1 array pickup UVFLG 2 baseline pickups UVFND 3 baseline pickups UVGLU 2 baseline codes base 4096 UVHIM 2 baseline pickups UVHOL 3 baseline pickups UVLOD Added conversion of standard UV FITS to the new form and also UV table form to the new internal one. UVMTH 2 baseline pickups UVPRM 2 baseline pickups UVPRT 2 baseline pickups UVRFI 1 baseline pickup plus encode antennas appropriately in 2 places rather than use ILOCB UVSIM to use SUBARRAY, ANTENNA1/2 instead of baseline UVSRT to compute base 4096 baseline if needed or use ILOCB VBCAL 1 baseline pickup, 1 array pickup VBGLU drop unused pointers, baseline pickup to use base 256 but mixed baseline types still will not work so added test and error termination. VBMRG baseline comparison not has to set logical for testing VLAMP FLGBAS to base 32768 (mostly not used). VLBIN to use SUBARRAY, ANTENNA1/2 instead of baseline and the one baseline putup COHER UVDPNT call sequence, 1 antenna pickup DTSUM fetch new pointers, 1 baseline pickup, display total integration times FINDR 3 UVDPNT calls, 3 baseline pickups FIXWT 3 UVDPNT calls, 3 baseline pickups FLAGR 5 UVDPNT calls and 5 baseline pickups INDXR Move baseline pickup above work routine, 1 UVDPNT call MAPBM 1 UVDPNT call, 2 antenna pickups OBTAB 1 UVDPNT call, typing SPECR 1 UVDPNT call, use its results rather than getting the pointers over again in 1 baseline pickup BLCHN 2 baseline pickups BPASS baseline encoding and 6 baseline pickups CALIB baseline encoding and 1 pickup CPASS 4 baseline pickups CVEL 4 pickups, FLGBAS goes to base 32768 DEFLG 2 pickups FRING Changed BL code for call to AVERT to base 32768. baseline encoding and 2 pickups IM2UV Use antenna1/2, subarray rather than baseline PCAL 2 pickups PCVEL FLGBAS to base 32768, 4 pixkups RLDLY Changed BL code for call to AVERT to base 32768. baseline encoding and 3 pickups SDVEL FLGBAS to base 32768, 2 pickups UVCON Changed to do ANTENNA1/2,SUBARRAY rather than BASELINE BLING UVDPNT call sequence, 1 pickup OMFIT UVDPNT call sequence, 1 pickup FTFLG 2 pickups IBLED 2 pickups RFLAG Changed FLGBAS to base 32768 in comments and local NXTFLG. Changed antenna pickup 4 places. SPFLG 3 pickups UVHGM 2 pickups, plus logic on whether there is BASELINE or not. Also changed the meaning of the baseline axis to do a proper computation to N*(N+1)/2 in each subarray appending them to each other. Previously it plotted the baseline code and overlapped subarrays. UVPLT 4 pickups VPLOT 3 pickups WIPER 4 pickups Moved nowhere. 14565. March 24, 2015 SN table Eric Add dispersion and delta dispersion columns to SN table. Changed: PUVD.INC MAXSNC becomes 26 (add 4 columns) PSNTAB.INC Add SNRDS1/2, SNRDD1/2 SNINI Add the 4 columns with units, dimensions (1/pol), no complaint if missing or no values TABSN Add the 4 columns when pointers set, DISP(2), DDISP(2) argumnets added after MBDELY, default 0. TABLEUTIL Changed call seq of OTABSN and the call therein to TABSN CLCOP Added DISP/DDISP to polarization switching as well as the 4 calls to TABSN SNDUP Does not use TABSN, but ignoreded MB delay. Added MB delay and the dispersions (if present). SWPOL Switch the DISP/DDISP for SN as well as CL, correct typo in CL MBDLYs, TABSN call seq BLAPP Put DISP and DDISP everywhere that MBDELY occurs (lots of places but never computed) Changed TABSN call with no use of DISP, DDISP except to init as zero SNAMBG SNREFM SNRESC SNSEL UVUTIL ACCOR ACFIT ACSCL ATMCA CLINV DTCHK PCCOR SNCOP SNP2D SPLAT UVCOP UVCRS CL2HF MBDLY CALIB FRING RLCAL OMFIT APCAL DELZN ELINT Moved nowhere. 14566. March 26, 2015 DOVLAMP Amy Added check on the tables if there is an input file rather loading from disk. Also checked on the number of visibilites for the selected calibrator, if there are less than 10 then don't bother with additional calibtration. Moved nowhere. 14567. March 26, 2015 More BASELINE Eric References to the BASELINE random parameter were hiding: LISRAN Add new random parms, none are averaged UVUTIL Make 4 places find alternatives to BASELINE and then use them as needed QEDIUTIL Correct fetching of alternative pointers RFI Make 1 place find alternatives to BASELINE and then use them as needed BLING Try alternatives for BASELINE before dying because the task could not set an unused parameter (!) IMAGR BASELINE random parameter was fetched, add the new ones and then use them as an alternative. OMFIT Make 1 place find alternatives to BASELINE and then use them as needed Moved nowhere. 14568. March 27, 2015 FITLD Eric A short section of essential code did not make it from UVLOD into the simple-FITS sections of FITLD causing bad things to happen. UVLOD worked. Moved nowhere. 14569. March 27, 2015 PGNV.INC Amy FITLD was not loading the gain table correctly because the include file PGNV.INC had not been relinked when MAXTBGC in PGCV.INC was changed. This meant that FITLD thought there were a possible 200 gain entries per IF when there were in fact 120. Moved nowhere. 14570. March 30, 2015 VLAMP Amy Changed reference to ANCAL at the top of the output file to ANTAB. Also moved the "BAND=" part of the gain curve definition a comment in the output file. Moved nowhere. 14571. April 2, 2015 STACK Eric New task to do image stacking - ignores coordinates, does weighted average (or median) of a matching set of images either INSEQ through IN2SEQ or planes of a cube. Moved nowhere. 14572. April 8, 2015 VLBARUN Amy Didn't work when loading a file from disk. Had to restructure a bit because it tried to "check" the file before it was loaded. Moved nowhere. 14573. April 8, 2015 ANTNUM (VLBAUTIL) Amy ANTNUM would set INVERS to 1 when run. It tried not to do this, by setting VBA_VERS to the INVERS, but ANTNUM called MAXTAB and MAXTAB also set VBA_VERS to the INVERS. Changed VBA_VERS to VBA_VERS2. Moved nowhere. 14574. April 20, 2015 DBCON Eric The format change is much more complex than what I did. The number of pointers and their names needed to be increased and special code has to appear to change the data when the 2 files have different versions of the BASELINE or ANTENNAn parameters. Moved nowhere. 14575. April 20, 2015 EDITR Eric This task did not handle the new format of random parameters properly. Fixed QEDIUTIL to use baseline numbers in the old form even with data of the new form. Limits in formats keep the task working properly only with MAXANT < 100. Fixed error message to give proper baseline values rather than the coded ones. Moved nowhere. 14576. April 20, 2015 RFLAG Eric The call to plot results after flagging was missing an argument. So DOPLOT = -12 et al. did not work although a good flag table was made. Moved nowhere. 14577. April 21, 2015 RFLAG Eric Corrected typo that caused RFLAG to ignore baseline-based flags as it was reading data. Moved nowhere. 14578. April 21, 2015 START_AIPS, TVSERVERS Josh Malone/Eric Changed ^/tmp to ^/{private/)*/tmp in both scripts to account for the new usage in Mac OS 10.10.3 (Yosemite). Moved to 15DEC14 24 April. 14579. April 21, 2015 TVFLG, SPFLG, FTFLG Eric The new option to allow pieces produced code which would try to load a window larger than the actual work file when one switched back to one piece. It rounded up by the smoothing interval. Added a MIN of the actual work file number of rows to each of the 3 tasks on the ENTER SMOOTH TIME option. Moved nowhere. 14580. April 22, 2015 DBCON Eric One more place that needed special care for baseline adding. Moved nowhere. 14581. April 22, 2015 CALIB help Eric Changed old reference from CPARM(2) to NORMALIZ in inputs section. Moved nowhere. 14582. April 23, 2015 UVINIT (!!!!) Eric UV data sets are always supposed to be larger by 4 kilobytes than they actually have to be. But we have encountered some that are exactly the right size for some reason. In that case, if the data set is small enough to fit in the first IO buffer, there should be no read into the second buffer. The loop thatstarted the first 1-2 reads skipped the second if the count was < 0, but my reading of the code said that the count would never be < 0 for the first buffer but could be zero. Changed the LT to a LE. Moved to patch and 31DEC13 (for Huib). 14583. April 23, 2015 START_AIPS, _TVSERVERS Josh Malone Corrected the grammar and typing of the change of 2 days ago. With the correction Mac yosemite is friendly again to simple aips commands with no TV= options. Moved to 15DEC14 24 April. 14584. April 24, 2015 AIPSMEM117 Eric Added UV format and SN table format changes to this memo. Moved nowhere. 14585. April 24, 2015 SNINI Eric Did not block message about 1 of the new parameters. Moved nowhere. 14586. April 28, 2015 UVAVG, ANFIND Eric Changed ANFIND to recognize LOFAR including the UK station which should have the other By sign. LOFAR seems to be using a strange antenna coordinate with a small but non-zero array center and the Earth radius-like station coordinates. They are using a ???? handedness which is in fact left handed. UVAVG had a buffer too small for 54 antennas. Changed it to handle MAXANT. Moved nowhere. 14587. April 29, 2015 CLCOR Eric Fix spelling of milli. Add some more statements to bring HA between +- PI. For EOPS, changed handling of CLCORP to allow the user to actually request day offset 0. Added info in help file for how to do this and why you would want to. Moved nowhere. 14588. April 29, 2015 FITLD Eric Fixes for use of REFDATE: 1. Trap that RDATE occurs in a table and update the value to that of the user. Times in such files were already updated. 2. Change AG2AN to receive the user REFDATE and to enforce that for RDATE and for the GST at time 0 in the AN file. 3. Change SO2SU to include the day offset in the JD to which we precess to get the apparent coordinates. Moved nowhere. 14589. April 29, 2015 TECOR Eric The JPL files for TEC include a value at the end of the day for midnight. The file for the next day includes the same time, but the data are different. Presumably, the later file was made with information from later in the 2nd day, informing the model for 0 hours. TECOR previously skipped the first record of each JPL file afterthe first. Now it will replace the last record of the previous file with the first of the current file. For data in, e.g. March 27, you will now get the same answers if you started with the file for the 27th or with multiple files starting before the 27th. Moved nowhere. 14590. May 1, 2015 OMFIT Eric Removed a multiplication by sqrt(2) initially suggested by Bientenholz on the Uncertainty in the parameters. Tests suggest that it is not correct. Worked on the help file to make the help section match the inputs section better and to say truth about the code. Moved nowhere. 14591. May 1, 2015 COMB Eric When it has no good idea what to say about the output Stokes, COMB used to say -1 (RR). Now it will say +1, -1, or -5 after looking at the input Stokes values. Moved nowhere. 14592. May 5, 2015 OBIT jobs Eric Changed AUO to make some strings longer. When executing an OBIT task and tailing the logfile, 200 characters were not enough. Moved nowhere. 14593. May 9, 2015 BPASS Eric To save work, the task did not divide by the fluxes in the source table. BPMOD was called only if using a model and not dividing by channel 0. It should also be called when doing no bandpass normalization and also when the division by channel 0 is only in phase. Moved nowhere. 14594. May 11, 2015 IMAGR Eric Changed: UVMTYP to have the same bias in favor of DFT as MODFIT GRDSUB to loop over models if the full set does not fit - the fix to UVMTYP should prevent this but it is safer this way. Call seq of CCSMEM and ALGMEM change. CCSMEM Call sequence to have range of models to do. ALGMEM Call sequence to have range of models to do. QCLEAN TVBOX procedes to delet box is there is only 1 without questioning the user. Moved nowhere. 14595. May 12, 2015 BPASS help Eric Under DOSCALE, clarify that the spectral index correction is done within an IF, not between IFs. Moved nowhere. 14596. May 13, 2015 IMAGR Eric The change to QCLEAN did not work as intended. It went into REBOX when there was 1 box. Fixed it to delete single boxes without a question. Added message when doing REBOX instead of TVBOX on a TVBOX request. Moved nowhere. 14597. May 14, 2015 BPASS Eric The logical DOMOD meant less than I thought, so the logic in calling BPMOD was essentially correct. Fixed the "fix" of 14593 plus a bug that caused the task to keep going when no data were found. Moved nowhere. 14598. May 19, 2015 Cal routines Eric GAININ Computed wavelength as if the reference pixel in frequency is always 1. It is now usually Nchan/2+1 making problems with dispersion and polarization calibrations. DATCAL Required both antennas to have non-zero dispersion to apply the correction. This leaves out all baselines to the reference antenna. Moved to 31DEC14 this date. 14599. May 20, 2015 FITLD Eric Changed handling of DIGICOR. 1. It used to lie and claim to do corrections but then if the array was not 'VLBA', not actually do them. Fixed the messages tell truth. 2. Changed DIGICOR in help and Fortran to have values 3 and 4 mean do the corrections even if not VLBA. Values 1 and 2 do the corrections only if VLBA (and 0 -> 1). 3. Fixed saturation correction to be done only on CORRELAT 'VLBA' (or ' ' but array VLBA). Moved nowhere. 14600. May 20, 2015 XAS Eric Test for Mac's odd DISPLAY variables dir /tmp but not /private/tmp which is now used. Added the latter so that shared memory will be used on Macs. Moved to 31DEC14 this date. 14601. May 25, 2015 UVMOD Eric The adverb QUAL was used twice in this task for different purposes. Changed the non-standard use to adverb WTUV. Moved nowhere. 14602 May 28, 2015 UVUTIL Eric A vis buffer was parameterized at 30000 to allow 4 * Npol * Nif * Mant * Mant. This would only allow 4 IFs for the VLA! Changed it to 1000000 which is too small for 90 antennas and 90 IFs but is much better. COMLNKed $QPGOOP and $APGOOP. Moved nowhere. 14603. June 15,2015 Dispersion Eric Changed: PUVD.INC XCLRSZ had 13*MAXIF - should have been 15*MAXCIF. MBDLY Added an OPTYPE to fit single-band delays with a single multi-band delay plus dispersion. Output contains dispersion columns with values plus corrected phases and SBdelays. CALREF Changed MBKOLS adding dispersions and changed call to MBDREF. MBDREF Changed call sequence adding a buffer, changed column pointers to (7) pointing at dispersion, and added pickup, smooth, interpolation of dispersion values. SN2CL Changed the buffer sizes - changed 1 or 2 * 1024 to 1 or 2 * XCLRSZ (1024 is too small for SN, yet say CL). Added pointers to dispersion and code to interpolate and copy dispersions (like MBdelay but simpler). CLCAL Made the scratch memory larger for everything and added 2 more parts so that the task could actually smooth multi-band delay and dispersion as advertised. Added pointers for dispersion in the polynomial fit routine and calls to fit the dispersions. In the smoothing subroutine, added the buffers for dispesion and MBdelay and did the smoothing in appropriate cases. Changed buffer size 1024 to XCLRSZ in several places (SNFCL does not need a double SN buffer like SN2CL). SNSMO help file: added DELCORR adverbs to set clip level for dispersion and showed other MBdelay parameters are also used for dispersion. SNSMO Added adverb to clip dispersion, added calls to smooth/clip, re-reference, and finally smooth dispersion. Treated it like MBdelay. FRING Changed call sequence to MBDREF. Added DISP to KEYS column pointers and fixed up handling of column pointers (all in FRNADJ). Added routines like in MBDLY to fit the SB delays for MBdelay and dispersion and correct the SB delays and phases for dispersion, writing the full result to the SN table. FRING help file: Add APARM(10) to fit MBdelay plus dispersion (from SB delays). Moved nowhere. 14604 June 16, 2015 Dispersion Eric More upgrades: MBDREF Corrected a pointer so it will re-reference dispersion as it was supposed to do. KRING Help: Add adverb DOIFS to control solutions in pieces. THE FORTRAN HAD THIS ADVERB BUT THE HELP FILE DID NOT. THIS HAS TO GO TO A PATCH SOMEHOW. Show mode T in inputs for dispersion. KRING (a) Changed call sequence to TABSN -- it was missed somehow! Added real dispersion values. (b) Changed call sequence to MBDREF. Added DISP to KEYS column pointers and fixed up handling of column pointers (all in KRNADJ). (c) Dropped requiremet of multi-source. NX tables are now present on all files. (d) Changed labeling and displayed units of dispersion. (e) If not solving for rates, do the FRING trick of averaging all data in a SOLINT and then making a second copy at a "different" time. Speed the task a lot! (f) Change conversion routine between SN table units and solution parameters a lot. It did phases badly when doing delays so that the resulting phases, while flat, were not zero. It took the dispersion values and adjusted delays - no longer needed since dispersion can go into the SN file. Also added documentation about the call sequence here. (g) In the fitting routine, corrected phase computation to use the frequency difference rather than a full frequency. Moved nowhere. 14605. June 20, 2015 SOUSP Eric Added SNVER and INVERS to inputs. Task will adjust gains in this range of SN table versions if the user chooses to replace the SU table fluxes with the fit fluxes. Took the SN table adjustment routine from GETJY and changed the old fluxes, after writing to the SU table, into the ratio of new to old for scaling the gains. Changed $INC/PSNTAB.INC to have format version 11 for the new columns and added SNREFM call in the SN adjustment routine. Changed GETJY to have this extra SNREFM, but it probably does not need it since SNREFM was called earlier for all versions. Moved nowhere. 14606. June 22, 2015 POSSM Eric Added code to make STOKES='HALF' (and RRLL and XXYY) be more forgiving if there is only one parallel hand polarization. Moved nowhere. 14607. June 22, 2015 IMAGR Eric The code that did the baseline-length time averaging copied too few spectral channels worth of data to the output buffer in the final dump of all remaining baselines (by the number of random parameters). Moved to 31DEC14 this date. 14608. June 23, 2015 MOVE Eric Changed from ZFIO (1 kilobyte per I/O) to a large buffer with ZMIO up to 64 Megabytes per I/O. This makes a ten-times speed up in some cases. Moved nowhere. 14609. June 24, 2015 CAPLT, CLPLT Eric There were some logic difficulties when the earliest scan did not have data for all baselines of a triangle. Moved nowhere. 14610. June 25, 2015 PIPEAIPS Lorant Included PIPEAIPS as clone of current DOOSRO to advertise it with a more general name for data reduction of connected element interferometers. Maybe more people will find it in the distribution and make use of it. Moved nowhere. 14611. June 26, 2015 BLING Eric Corrected handling of times in NX tables by adding 0.02 sec to the length of the scan. Also changed the NX handling to avoid changing the source number when one is past the last calibrator scan. Changed it to force starting over in the NX table at each scan. Moved nowhere. 14612. June 30, 2015 AIPSLetter Eric Put in $AIPSPUBL/LET15A.{TEX,PS,PDF} the June 30, 2015 AIPSLetter. AIPSPUBL goes everywhere. 14613. July 1, 2015 REVERSN Eric NEW VERB added to POPSDAT, NEWPARMS plus AU3 and a new help file. This verb looks on disk for all possible extension files and updates the header. Moved nowhere. 14614. July 2, 2015 PRTMSG Eric Chamged PRTMSG (Fortran) to ask the user to confirm jobs that will possibly print > 600 lines to the actual line printer. Added comment to the help file also. No confirmation required for batch jobs, text file outputs, etc. Moved nowhere. 14615. July 2, 2015 FLOPM Eric The low level routine UVGET can alter the frequency of the data when the frequency offset of IF 1 is not zero. But it does not scale u,v,w for this alteration! Fixed FLOPM to grab the initial frequency and the final frequency and scale appropriately. Moved nowhere. 14616. July 7, 2015 UVLOD, FITLD Eric These tasks read index tables written by FITTP and FITAB. When reading the former it is possible to have UV records discarded if they are completely flagged (under control of DOKEEP). Added subroutine NXCHCK to both tasks to check if there is an index table and if so does the peak vis record number match the peak of the vis file just written. If not, the NX table is recomputed (or created new if no input table was provided). This covers for the lack of NX tables from CASA and for odd behaviors when records are omitted. Moved nowhere. 14617. July 8, 2015 UVLOD, FITLD Eric Pulled NXCHCK out as a separate $APLNOT subroutine. Changed both and RUVTAB to honor DOKEEP with UV binary tables, report the number of records omitted, and compress the UV file accordingly. Previously, this operation was done on random groups format only. Moved nowhere. 14618. July 8, 2015 PRTHI Eric Changed AU7 to ask for confirmation on line-printer jobs potentially greater than 500 lines. Note added to PRTHI help file. Moved nowhere. 14619. July 9, 2015 UVHOL Eric Change WTUV in HOLG from unlimited (if positive) to 1.0 if outside the range 0.1 to 10. Fix help file to admit to the allowed scaling in HOLG and to include a paragraph to explain why one will want to do scaling whenever the IF came from a copy of the UV data with IF selection. Moved nowhere. 14620. July 9, 2015 LISTR Eric Added the reference channel value in header of Frequency Table listing. Corrected the format of that listing. Moved nowhere. 14621. July 10, 2015 TV3COLOR Eric Changed AU5 and the help to have this verb use TVCHAN to specify the first of the 3 channels/colors. Useful for the next item in particular. Moved nowhere. 14622. July 10, 2015 EDIT Class Eric When editing with CROWDED true, it is useful to try to differentiate the various IFs and polarizations with color. The option to do so has been added to the EDIT class with some 3-color support n the TV device class. Changed: EDITA Fortran and help: add adverb DO3COL EDITR Fortran and help: add adverb DO3COL SNEDT Fortran and help: add adverb DO3COL TVDEVICE Add TVDLN3 to draw lines in grey levels in 3 colors in channels CHAN to CHAN+2. Added OPTYPE='TV3C' to turn on 3 consecutive channels in RGB. QEDIUTIL Added keyword DO3COLOR which is allowed to be true only when CROWDED is also true. Added ED3CIN to turn on channels 14-16 is RGB and initialize a set of colors for all the IFs and polarizations. Added code to trun off these channels at the end. Added ED3LIN to call TVDLIN or TVDLN3 depending on DO3COL. Added lots of places to set COLOR to point at the current polarization/IF being plotted. In EDPLT1, changed to call ED3LIN for the edit area plot and to init these grey-scale planes to zero. In EDPLTP, added calls to ED3LIN when plotting the edit data. Corrected NEXT POL/IF functions to include all polarizations and all IFs in the cycle as appropriate. Moved nowhere. 14623. July 13, 2015 gfortran Eric Sites trying to compile their own versions of AIPS have had some difficulties. Changed: OPTIMIZE.LIS APCLN to be done at level 1 (commented out for main distributed version) PAPC.INC Parameter PKPWRD added, defined as 2, commented version 200000000 DAPC.INC APCORE is size PKPWRD rather than 2. QINIT.FOR Add code for a fixed size AP when PKPWRD>1000 with error message if request larger than allowed. Dynamic allocation remains otherwise. ZTXOP2 gfortran makes the file unwritable when the end-of-file is read but offers an APPEND option in the OPEN. g77 does not allow this option so changed code on append to read to EOF counting records. Then close, re-open, and read just the right number of records leaving it just before the EOF. Moved nowhere. 14624. July 13, 2015 CALIB Eric Fixed one RPOL/LPOL label to also do XPOL/YPOL when appropriate. Moved nowhere. 14625. July 14, 2015 gfortran Eric Upgraded the options a bit in $SYSLNX64 versions of FDEFAULT.SH and LDOPTS.SH. Moved nowhere. 14626. July 14, 2015 gfortran Eric gfortran does not save local variable values from one invocation of a subroutine to the next. This may be a wide-spread issue. Fixed so far including warning fix ups: AVER work routine did no SAVEs and needed many BASRM buffer way too small for possible numbers of channels, may not work for compressed data (INCX ignored) BLAVG SAVE was done but some variables missed, buffer too small too BSMAP SAVE was done, some missed UVMOD SAVE needed for number IFs, etc. Moved nowhere. 14627. July 16, 2015 Side band Eric TABFQ actually forces the sideband code to +1 and, if it encounters a negative sideband code on input, forces the channel increment to negative. Some places still multiply the increment by the sideband code, so fix these in ACTFRQ DBCON FXVLA HAFIX PCLOD SETJY TBAVG UVFIX UVLIN UVLSD UVLSF Also changed: AVSPC Averaged channels in the SUBS option from N to 1 by -1 for "lower sideband" including all ATCA data. This makes the header and FQ table be incorrect. CVEL Shift direction based on sign of frequency increment rather than band code. PCVEL Shift direction based on sign of frequency increment rather than band code. POSSM FFT direction based on sign of frequency increment rather than band code in auto- and cross-correlation plots. Moved nowhere. 14628. July 17, 2015 gfortran Eric More SAVES needed: DESCM DESCDO needed 2 initialized at 1st call, lots of code upgrades to avoid GO TOs, unused statement numbers. DTSIM GETGST needs to remember values from previous call (DTPM.INC needed TAB characters removed.) FILLM VLMDAT needed some SAVEs, unused arguments, excess statement labels. FILLR VLRDAT needed some SAVEs, unused arguments, excess statement labels. HAFIX correct typo from yesterday (not Fortran error) IMMOD Drop 2 unused, add SAVE for component positions upgrade coding style dropping unused and unnecessary statement labels MWFLT A large number of variables needed to be SAVEd since they are computed once per plane only. Moderate coding upgrades too to reduce warnings and GO TOs. NINER Required some variables to be SAVEd at the start of each plane. Required serious rearrangement of code to make it readable. KRSH required two corrections - the code stepped around the matrix too much and so did only 4 of the 8 cases and then it took max(1,result) as the result. That would be okay I guess if the data were >> 1 butthey are not. Corrected help file too. NNLSQ The working routine required a few SAVEs, code needed IF-THEN-ELSE rather than GO TOs with lots of labels. SHADW DIDDLE required a few SAVEs SPLAT AVGTIM required a few SAVEs SPMOD SPMODL required 2 more SAVEs TBAVG TBAVER required 1 additional SAVE UVFIX Correct typo in TRUEF computation from yesterday - why did most compilers accept this bad code??? Moved nowhere. 14629. July 20, 2015 gfortran Eric UVAVG SAVE variables TCHK, LSTSUB. Drop unused formats and unused call sequence variables in the bottom UVAVDO routine UVDI1 SAVE variables TCHK, LSTFQ UVFIL SAVE variables LUN, FIND in the sample code. Leave Unused varaibles in call to FIDDLE for user use. UVFIX SAVE variables OBSPOS, POLAR, IORBIT, EPOCH. UVLIN SAVE variables REPORT, RESCON, DOSHIF; drop unused call variables in UVLINF UVLSD SAVE variable DOSHIF. Drop unused formats, labels, variables in call sequence of UVLSDF UVLSF SAVE variable DOSHIF. Drop unused format, labels, variables in call sequence of UVLSFF UVRFI SAVE variables TCHK, LSTSUB. Drop unused format, labels, variables in call sequence of CMPEXP and UVAVDO VBCAL SAVE variable CALWT VBMRG SAVE variables XCOUNT, TLAST, MCOR. Remove unused labels and modernize code. VPFLG drop an unused label WTMOD SAVE variables WTMAX, WTMIN. Remove unused labels. Moved nowhere. 14630. July 21, 2015 WIPER, UVPLT Eric WIPER had typos causing it to have 0 as the antenna numbers and so to treat everything as an auto-correlation. WIPER and UVPLT both made the mistake of setting the imaginry part of cross-hand auto-correlation visibilities to zero. WIPER also had a bug that caused it not to find data for code 17 (projected spacing along a PA). Moved nowhere. 14631. July 21, 2015 gfortran Eric BSROT Removed excess call sequence arguments. MODIM SAVE variables CPHI, SPHI, drop unused formats SDMOD SAVE variables RRA, RDEC. Drop unused arguments in call sequence SPCOR SAVE VARIABLES DOREF, DX, DY, DT TAFFY SAVE variable LROW plus comment, clean up code some AIPS SAVE variables LUNCRT, IFIND Moved nowhere. 14632. July 21, 2015 PRTMSG, PRTHI Eric Changed AU7 and PRTMSG to read through the HI and MS files to count the lines to be printed and ask for confirmation only if the number exceeds some limit (reduced a bit) after application of the selection adverbs. Moved nowhere. 14633. July 22, 2015 EDITA, EDITR Eric Fixed bugs in QEDIUTIL that caused DO3COLOR false plots with CROWDED true to fail to plot anything when the polarization and/or IF became 0 after the first time. Also fixed bugs that kept EDITR from doing 3-color. Moved nowhere. 14634. July 22, 2015 WIPER Eric Add option to NOT flag cross-hand polarizations when flagging parallel hand. Moved nowhere. 14635. July 22-24, 2015 gfortran Eric MBDLY SAVE variables TIMPR, PTITLE 2 places. Drop unused labels, formats, call seq arguments. DAIP SAVE variables LUNCRT, IFIND Drop unused labels, formats, call seq arguments. IMVIM SAVE variable NPOINT; modernize code Moved nowhere. 14636. July 24, 2015 UVHOL Eric Add options to also average over moving antennas and to do scalar amplitude averaging (over time and antennas, spectral averaging remains vector). Moved nowhere. 14637. July 27, 2015 EDIT class Eric Flag commands in QEDIUTIL are assigned times which attempt to be half way between samples. This did not happen at the first and last sample so fixed it to use the time just after/before the first/last sample to see how much to extend outwards. Also corrected all 4 class types to turn off the grey-scale memory (DO3COL true) on EXIT as well as ABORT. Moved nowhere. 14638. July 27, 2015 TACOP Eric Added FLAGVER adverb to flag SN, SY, and TY tables if requested. EDITA does not edit the work table and so this option (or UVCOP with FLAGVER) is needed. Moved nowhere. 14639. July 27, 2015 gfortran Eric Added SAVE of IXC, IYC to TVFLG, SPFLG, FTFLG. Moved nowhere 14640. July 27, 2015 VPFLG Eric Added OPCODE option 'CROS' to flag RL and LR (or XY and YX) whenever one or both of RR and LL (or XX and YY) are flagged. It still does the V polarization extreme flagging otherwise. Moved nowhere. 14641. July 28, 2015 APCAL Eric In the weather routine and one other place the "I don't care" value for subarray was 0. It must be <= 0 since TABWX returns -1 when there is no subarray column. Include PGCV was full of TAB characters. Moved nowhere. 14642. July 29, 2015 VLBARUN Amy Improved automatic scan selection, by testing all antennas are present in selected scan and if not going to next calibrator in calibrator list. If no scan is present with all the antennas on a calibrator, then script stops and asks the user to input timerange. In plotting: 1) added UVPLT plots after all the calibration; 2) fixed bug for final phase (VPLOT) plots to plot only baselines to refant; 3) fixed bug when creating plots for html file to output images for all the targets rather than just the first one. Moved nowhere. 14643. July 30, 2015 RLDLY Eric Added option to use multiple scans and to write the results in an SN table rather than directly into a CL table. This will allow for time variable R-L delay difference. The task chooses the mode based on the number of calibration scans chosen by TIMERANGE, CALSOUR, et al. Moved nowhere. 14644. July 31, 2015 TYCOP Eric Renamed SYCOP to be like TYAPL, TYSMO and added TY tables to the operations. Added messages saying what it did (already in history, but not previously in message file). VLB data will be able to use the new option for RFI-blasted TY values. Moved nowhere. 14645. July 31, 2015 gfortran Eric CONFI SAVE LABEL, LTYPE; drop numerous unused formats, some statement labels ELINT SAVE LABEL, LTYPE PBEAM SAVE LABEL, LTYPE SNFIT SAVE CSAVE, LTYPE; drop unused formats SNPLT SAVE CSAVE, LTYPE; drop unused variables VPLOT SAVE LTYPE; type mismatches in call sequences, DATA strings too long, unused labels. WETHR SAVE LTYPE; shorten long strings, unused labels and formats BPASS SAVE in BPSHFT variables NFREQ, OLDSOU; cll sequence mismatch, unused formats Moved nowhere. 14646. August 3, 2015 ANTAB Eric Increased the maximum NX table size to 512K records. Moved nowhere. 14647. August 3, 2015 gfortran Eric Changed CVEL, PCVEL, and SDVEL to SAVE variable OLDSOU. Added SAVEs also in DEFLG and KRING. Also changed type mismatches in CVEL and PCVEL, unused variable in SDVEL, unused label and formats in DEFLG, and KRING required adjusting array types across call sequences. Moved nowhere. 14648. August 4, 2015 gfortran Eric Discovered that large variables (e.g. 1.5 Mbytes) are very expensive to implement as automatic variables even when there is no re-use of values between calls to the subroutine. UVPRT required 55 seconds of CPU with one such buffer in DATGET and, when that buffer was SAVEd, only 0.1 seconds! Added SAVE statements for buffers dimensioned n times MAXCIF in $APLNOT BPASET BPREFM BPSEL CPSEL DATGET DGETSD DTBAND FINDUV PDSEL POLSET TABBD TABBP TABCP TABPD Moved nowhere. 14649. August 4, 2015 gfortran Eric More large buffers SAVED: $APLNOT BPSHFT DATBND SMOSP WTSHFT GRIIDTB UVDOUT UVTIME $APLSUB ISLAND $QOOP QCLEAN, APUTIL, PBUTIL, QEDIUTIL, QUVUTIL (and fixed some call sequence issues) $QSUB UVWAIT $QNOT MAKMAP, VISDFT Other: $APLOOP TVDEVICE - MAXIMG now larger than alternative for buffer size, fix some DUMMY variables to DUMMY(1), fix real error in TVBPUT declarations. Moved nowhere. 14650. August 5, 2015 documentation Eric Updated the documentation files: 1. TAB completion (new and deleted things) 2. APROPOS help file (deleted things) 3. ABOUT FILES: ADVERB ANALYSIS CALIBRAT CATALOG EDITING IMAGING PLOT POPS PROCEDUR RUN SPECTRAL TABLE TASK TV UTILITY UV VERB VLA VLBI Moved nowhere. 14651. August 5, 2015 gfortran Eric Added required SAVE variables in $QPSAP routines QXFOUR and QRFFT and advisable SAVE for large variable in QCFFT. Moved nowhere. 14652. August 6, 2015 Index table Eric Changed the index table creation include file DRNX.INC to raise the number of allowed subarrays from 20 to 50. Moved nowhere. 14653. August 6, 2015 gfortran found Eric Fixing things that may be bugs or confuse code: IMAGR The call to IMACPY to make a 2nd data copy for filtering had an old call sequence and an incorrect dummy variable. Unused labels too. FITAB Had 1 call with too many arguments that could have messed up a pointer. Unused labels too. APCLN Numerous unused labels, Modernize code some. CALIB Unused label, rank mismatch in call. COMB Unused labels, a little modernize of code. SUBIM Make types match in call UVMAP Unused labels, modernize a bit FITTP Bad declaration caused error counts to fail. Modernize code, drop unused labels. Subroutines: RLOCAT Function not set on error return TKGMPL A declaration (1) causes confusion. FRMT Did not return a length when number was zero LSTHDR Numerous unused labels TINBEG Unused variable WXINI CENTIGRADE too long a string XTRFIT Unused labels, formats, ... ARRAY Internal mismatches, unused labels CLEAN Lots of internal mismatches CXIMAGE Internal mismatches, fixed by making DUMMY(1) EDIT Internal mismatches, fixed by making DUMMY(1) - character part of GET/PUT needs to be an array (*)*(*) as in the other OOP ones so far IMAGE As above OBJMNGR Mismatches corrected. POSITION Internal mismatches. PRINTER Internal mismatches, unused labels, etc. TABLE Internal mismatches, GET/PUT declaration. UVDATA Internal mismatches, GET/PUT declaration. UVDESC Internal mismatches, GET/PUT declaration. ZCLOSE Unused labels ZMSGDK Unused variables YSCROL Unused variables CLAB2 DATA string error CONDRW Unused labels ITICS Unused formats DFMACH Modernized code, fixed so always returns something IMINIT DATA string too long JPLEPH Init call needed arrays of 0's OTBSRT Unused labels PHINI DATA string too long TSINI DATA string too long, unused format Moved nowhere. 14654. August 7, 2015 gfortran found $APLPGM Eric gfortran apparently does not put the start of a named COMMON on a double precision address unless you force it to!!! SETSP Character in call sequence length error, couple unused FORMATs AHIST COMMON length seemed to change between routines - fixed to include only what was needed. CPYRT COMMON length error fixed by omitting unused OLD8, numerous unused labels. DIFUV Had upgraded the error formatting in the message generation but had not changed the FORMATs to match. DSORC Call seq type mismatch HISEQ Call seq type mismatch MCUBE Dropped section of unused/obsolete code, corrected missing argument in one call bad type in another, numerous unused labels MODIM Call sequence had argument not needed. MODSP Call sequence had argument not needed, unused FORMATs NANS Call seq mismatch OTFIN Call seq mismatch, unused labels PBCOR Call seq mismatch, unused labels PRTPL call seq mismatches from antique declarations and in 2 cases with inadequate buffers. Unused labels too. PRTTP Call seq mismatch, lots unused labels QMSPL Call seq mismatch, unused labels REMAG Call seq mismatch, unused labels, common alignment RESEQ Call seq mismatch SAD Call seq mismatch, DMACH possible uninitialized, unused labels SCLIM Excess args in call SDMOD Call seq mismatch SLFIT Call seq mismatches, clean up code a bit SPIXR IO buffer was way too small and integer rather than float! How did this work? SQASH Call seq type STACK Unused formats STFND Numerous unused labels TAFFY call arg mismatch UVDIF Numerous unused labels WTSUM Unused labels, call seq mismatches XMOM Unused labels, call seq mismatches (buffer wrong type and very small) XSMTH Numerous unused labels XSUM Numerous unused labels, call mismatch Moved nowhere. 14655. August 7-10, 2015 gfortran found $YPGM Eric BLANK Fixed numerous unused labels, but one of them truned out to be an error made in 1991! The FLUW opcode was changed to flagging all pixels < DPARM(3) and the code to implement it correctly was unreachabble. Cleaned up that subroutine extensively. BLSUM Unused labels, minor mismatch GREYS Unused labels, call arguments needed correct dimensions either in the CALL or in the subroutine HUINT One unused label, character string 'MA' passed in to subroutine expecting integer from 0 to 2! Got away with it apparently. IMEAN Unused labels, mismatch, issue with EXTERNAL. KNTR Unused label, correct dimensions in call sequence LWPLA Unused labels, string lengths PANEL Unused labels, clean up GO TOs PCNTR Unused labels PROFL Unused labels RMFIT Unused labels, call mismatches, EXTERNAL issue, usage after the mismatches were solved SERCH Allignement of COMMONs not done properly by gfortran so force it TAPLT Unused labels, rearrange code to reduce GO TOs. TVHLD Unused labels, mismatch TVPL Unused labels, minor mismatches TVRGB Couple unused labels, mismatched and unused call argument TVSAD Unused labels, mismatches, bad call sequence in getting rms to see if fit ok. XBASL Numerous unused labels XGAUS Unused labels, EXTERNAL issue, mismatches XPLOT Unused labels ZEMAN Unused labels, a variety of mismatches. Moved nowhere 14656. August 10-11, 2015 gfortran $YPGNOT Eric ANBPL Labels, mismatch of one call argument many places APCAL the call for the spillover correction reversed the frequency and elevation arguments - the answer had to wrong. Cleaned up typing of many statement labels. BPLOT Unused labels, mismatch. CAPLT Unused labels, mismatches CLPLT Unused labels, mismatches DFTPL Unused labels FRMAP Unused labels, mismatch FTFLG Unused labels, mismatches GAL Mismatches - there were 3 different ideas (100, 200, 500) of the maximum ring number in the different routines. IBLED Numerous unused labels, one unused and mismatched argument IRING Mismatch, minor typing PFPL1 Unused label, clean up PFPL2 Unused label, clean up PFPL3 Unused labels, clean up PLCUB Unused labels, clean up PLOTR Unused label, mismatch POSSM Numerous unused labels RFLAG Unused labels, mismatches SNIFS Unused format SPFLG Fixed a variety of call sequence mismatches, unused labels TARPL One label, 2 call mismatches TVCPS Integer passed in to routine expecting logical. TVFLG Call mismatches UVHGM Unused labels, bad declaration for external UVPLT Unused labels, bad string DATAs, bad declaration modernize code some Moved nowhere. 14657. August 11, 2015 gfortran $QPGOOP Eric BLING Mismatch, unused format BSGRD Mismatch, unused code CXCLN One unused label EDITA Two unused labels EDITR Two unused labels SCIMG Unused labels, mismatches SCMAP Unused labels, mismatch SDGRD Unused labels SNEDT Unused labels WFCLN Unused labels, rearrange code a little Moved nowhere. 14658. August 7,12, 2015 "gfortran" corrections Eric Boo-boos in my code changes: IMAGE The change caused the scratch file number to be zero in IMGSCR and hence to die in IMAGR. UVDATA The changes caused OCLONE to make the output work file in IMAGR be compressed - a disaster to the later routines. Moved nowhere. 14659. August 13, 2015 OMFIT Eric OMFIT had a vast number of FORMATs that were not used and a variety of labels that were also unused. It creates dynamic memory variables and then calls subroutines with them. Fixed the call sequence types via making EQUIVALENCEd variables of the correct type to the variables referenced for dynamic memory. There was one variable that is used heavily as a real array in one routine and then COMPLEX in many below that. So one warning has been left. Also found a significant place where a SAVE is required. Moved nowhere. 14660. August 14, 2015 PCAL Eric When doing models (at least), the task failed to notice when a channel was fully flagged and died later. Changed so that the initial data copying would notice this problem. I suspect that the route through the task for solving for a model noticed the absence of data later and handled it. Also fixed unused labels and, with global equivalences, some mismatches. Should go to patch. 14661. August 14+17, 2015 gfortran $QYPGNOT Eric Subroutine $QNOT/APIO was found to require SAVE for a couple variables and to have unused labels. APGS Unused labels, COMMON misaligned, modernize some APVC Unused labels, COMMON misaligned, modernize some CONPL Unused FORMATs EVAUV Unused labels, mismatches SDCLN COMMON misaligned, unused labels, modernize extensively STEER Unused, COMMON misalignment, modernize Moved nowhere. 14662. August 17, 2015 gfortran $QPGNOT Eric BLCAL Fixed call type matching BLCHN unused variable CONVL COMMON misaligned, unused labels CPASS Unused labels. A mismatch turned up an array used as an array that was declared a scalar and never initialized called IPIVOT. i have no idea what to do with this other than to declare it a large enoough array and let the compiler zero it. FFT Numerous unused labels FRING A few unused labels, one minor mismatch IMERG Unused labels LPCAL Unused labels, type mismatches fixed with equivalences RLCAL Unused labels, bad DATA SDIMG Unused labels UVCON Unused labels, some modernizing, mismatch Moved nowhere. 14663. August 17, 2015 more gfortran Eric gfortran 4.4 finds unused arguments in subroutines TVHELX did not use BUFFER AU6 call to TVHELX TABCCM NUMV arg unused - change to test on it some APDEVICE Unused labels, unused call argument in APDOPN fixed to appear in an error message, logic clean up and typos fixed Moved nowhere. 14664. August 19, 2015 more gfortran Eric Unused arguments in local subroutines in KNTR, RMFIT, TVSAD, XGAUS, ZEMAN, IBLED, RFLAG, VPLOT DLINTR call sequence unused arguments: AU6A GRLUTS HIENH TVHELX TVMOVI TVDEVICE GRBOXS GRPOLY GRSLIC IENHNS OFMCON OFMGAM OFMMOD TVALUE TVFIDL TVFLAM TVMENU TVPSUD TVROAM TVSCRL TVZOME BLANK BLSUM HUINT RMFIT TVHLD TVHUI TVRGB TVSAD XGAUS ZEMAN FTFLG IBLED SPFLG TVFLG WIPER CONDRW call sequence unused arguments PCNTR GREYS CNTR CCNTR MODPLT call sequence unused arguments and add an argument VPLOT CAPLT CLPLT GTIC call sequence unused arguments PCNTR GREYS KNTR Moved nowhere. 14665. August 20, 2015 gfortran found Eric Unused arguments in call sequences can be confusing or misleading and may indicate an error when the argument should have been used. We retain some such cases so that very similar routines may have very similar call sequences (e.g. some OOP basic routines, table I/O routines) or that arguments required by some devices (TVs mostly) are retained even if the latest versions (e.g. XAS) do not require them. Changed: BTBSRT unused in call seq TABSRT unused in BTBSRT call seq CSLGET unused in call seq CGASET unused in CSLGET call seq DATCAL unused in CGASET call seq FLGSTK unused in call seq FLAGUP unused in FLGSTK call seq FRTAB unused in call seq R3DTAB unused in call seq, modernize RWTAB unused in call seq, modernize REFCNT unused in call seq REFFAZ unused in call seq UVUTIL unused in REFCNT, REFFAZ call seq, AVGVIS call seq REIMIO unused in call seq SNAMBG unused in call seq TABAPP unused in error in call seq, clean up TABPH unused in call seq TABTS unused in call seq TPSHFT unused in call seq BPSHFT unused in TPSHFT call seq UVFLG unused in FLGSTK call seq, also unused arg and labels VPLOT unused in FLGSTK call seq CALRD unused in R3DTAB, RWTAB call seq, unused labels IMLOD unused in R3DTAB, RWTAB call seq, unused labels, modernize UVLOD unused in R3DTAB, RWTAB call seq, unused labels MAPBM unused in AVGVIS call seq PLROW unused in REIMIO call seq, modernize ATMCA unused in SNAMBG call seq BPASS unused in TPSHFT call seq FITLD unused in FRTAB R3DTAB RWTAB TABPH TABTS call seq Moved nowhere. 14666. August 20, 2015 gfortran more Eric More unused call arguments in subroutines ALGSTB unused in call seq GRDSUB unused in ALGSTB call seq CALRES unused in call seq GRDCRM unused in call seq CCSMEM unused in GRDCRM call seq CCSGRD unused in GRDCRM call seq UVTBUN error since unused in call seq (DISKO/CNOSCO not used output wrote over input regardless) UVTBUW error since unused in call seq (DISKO/CNOSCO not used output wrote over input regardless) PBUTIL unused in private call sequence UTESS unused in CALRES call seq VTESS unused in CALRES call seq Moved nowhere. 14667. August 21, 2015 flagging default Lorant The default flagver of zero in TYAPL does not do flagging. I meant to apply flagging. Set to maxvers in PIPEAIPS/DOOSRO Moved nowhere. 14668. August 21, 2015 gfortran tasks Eric ATLOD used a floating variable in a DO loop that should have been the integer equivalenced variable (so increments were left 0.0). Subroutine illbase had unused arguments and a real logic error. BSFIX COMMON alignment, unused format, unused args BSMOD COMMON alignment, unused labels/formats, unused and misleading arguments M3TAR COMMON alignment: TMPBUF too small for one equivalence also MK3IN COMMON alignment: TMPBUF too small for one equivalence also, unused FORMATs, unused args PHSRF COMMON alignment, unused args, unused FORMATs UVGLU COMMON alignment UVHIM COMMON alignmnet, string DATA value VLBIN COMMON alignment, unused args, unused FORMATs REUSE 1 unused args in each 2 routines DIFUV 8 unused args in 1 routine SAD 1 unused arg in 1 routine SDLSF 5 unused args in 1 routine, unused format SPIXR 1 unused arg in 1 routine STFND 2 unused args in 1 routine UVPOL 7 unused args in 1 routine XSMTH 1 unused arg in 1 routine XSUM 1 unused arg in 1 routine ACFIT 1 unused in 1, 3 unused in second, error code needed in 3rd ACLIP 2 unused in each of 2 routines, unused labels/FORMATs AFARS 1 unused arg in 1 routine. unused labels BASRM 1 unused arg => fixed task to handle compressed data CALWR 2 unused args to 1 routine, unused labels CANPL 1 unused arg to 1 routine, unused labels CLCAL Fixed unused labels, left 2 unused args to ILAENV since it is called 100 places and is a lAPACK function. CLIP 2 unused args to 1 routine, 1 to a second, unused labels CORER 4 unused args from 1, 1 arg from a second CUBIT 1 unused arg to 1 routine, unused labels DBAPP 1 unused arg to 1 routine, unused labels DECOR 7 unused args in 1 routine, unused labels DQUAL 6 unused args in 1 routine, unused labels FACES 3 unused args in 1 routine FARAD 1 unused arg in one function FARS 1 unused arg in one function, some unused labels FITTP 2 unused args to 2 functions => remove count of value zero written for UV data (no longer possible) Moved nowhere. 14669. August 22, 2015 SETJY Eric One routine had 3 different references to format 1000. Merged two of them and made a separate format for the other. Raised MAXSOU to 5000. Moved top patch this date. 14670. August 21, 2015 gfortran tasks Eric FIT2A Numerous unused labels FIXRL One unused arg in one routine FQUBE 2 unused args in one routine FXTIM 6 unused args in one routine, unused formats GLENS Ellipticity not used in 2 models (one not actually used in this task), labels GSCAT 2 unused args in 3 routines, 1 unused arg in 3 routines, many unused labels HA2TI 1 unused arg in 1 routine IMFIT 1 unused arg in each of 3 routines, many labels, modernize some IMLIN 1 unused arg in 2 routines, labels, modernize MATHS 1 unused arg in 1 routine MEDI 1 unused arg in 2 routines, numerous unused labels MOMNT 2 unused args in 1 routine NOIFS 1 unused arg in 1 routine PGEOM 1 unused arg in 1 routine, modernize some PHNEG 7 unused args in 1 routine QUFIX 1 unused arg (error code) in 1 routine REGRD 1 unused arg in each of 2 routines Moved nowhere. 14671. August 23, 2015 gfortran warnings Eric RLDIF i unused arg in each of 3 routines RM One arg unused in 2 levels of routine, but answer possibly was used. Fixed to block blanking on correlation coefficient whn max RM not specified. RMSD 1 unused arg in 1 routine, minor clean up SBCOR 5 unused args to 1 routine SETFC 3 unused args to 1 routine SKYVE 1 unused arg in 5 routines SLICE 1 unused arg in 2 routines SNFLG i unused arg in 1 routine SNREF 1 unused in 2 routines, one needed new precursor remarks STFUN minor clean up STRAN 1 unused arg in 1 routine TAPPE 1 unused arg in 1 routine, error in DATA of AG tables TARS 1 unused arg in each of 3 rroutines TBIN 1 unused arg in 1 routine TBOUT 1 unused arg in 1 routine, modernize a little TYAPL 1 unused variable TYSMO 1 unused in each of 3 routines: 2 of them should have been using the smoothing time for gains and were using the time for Tsys, the 3rd caused the error code not to be set UV2TB 4 unused args to 1 routine UVCOP 1 unused variable UVCRS 1 error arg unset but tested UVDGP 8 unused args to 1 routine UVMLN 3 args to 1 routine, 1 arg to another UVNOU 5 args to 1 routine UVWAX 4 args to 1 routine XTRAN 1 arg to 1 routine Moved nowhere. 14672. August 24, 2015 gfortran $APGOOP Eric BLAPP Fixed 1 unused arg in 1 routine, there are about 86 unused args in a variety of routines because many call sequences include all SN data arguments declared with a local include. Lazy but effective (and annoying). BSCOR 1 unused arg in 1 routine CCEDT 1 unused arg in 1 routine CCRES 1 unused arg in 1 routine CL2HF Bad HI string lengths, 4 unused args in 2 routines, 1 unused in 2 routines, 2 unused in 1 routine FINDR 1 unused arg in 1 routine FXPOL 1 unused arg in 1 routine, truncated DATA strings HF2SV 3 unused arg in 1 routine HFPRT 1 unused arg in 2 routines RFI 1 unused arg in 1 routine TBDIF 1 unused arg in 1 routine VLABP 1 unused arg in 1 routine Moved nowhere. 14673. August 24, 2015 more gfortran Eric $AIPNOT REBYTE 1 unused arg in 1 routine $QYPGNOT EVAUV 1 unused arg in 1 routine $QPGNOT CALIB 1 unused arg in 1 routine CPASS 1 unused arg in 1 routine CVEL 1 unused arg in 1 routine DEFLG 1 unused arg in 1 routine FRING 2 unused args in 1 routine, 2nd routine could use an arg that was "unused" IMERG 1 unused arg in 1 routine KRING 3 unused args in 1 routine, leave messy math one LPCAL 3 unused args in 1 routine, leave messy math one PCAL 1 unused arg in 3 routines, leave STOPX PCVEL 1 unused arg in 1 routine RLCAL 1 unused arg in 1 routine $QPGOOP BLING 1 unused in 2 routines, 2 unused in a 3rd CXCLN 1 unused arg in 1 routine OOSUB 2 unused arg in 1 routine which also needed true precursor comments SCIMG 1 unused arg in 1 routine called several times SCMAP 1 unused arg in 1 routine called several times OMFIT 33 routines had unused arguments - the attempt to make the various model routines have the same arguments was dropped. Moved nowhere. 14674. August 24, 2015 gfortan Y programs Eric IMEAN 1 unused arg in 1 routine LWPLA 2 unused args to 1 routine PANEL 1 unused arg in each of 2 routines SL2PL 1 unused arg in 1 routine XBASL 1 unused arg in 3 routines XPLOT 1 unused arg in 1 routine ANBPL 1 unused arg in 3 routines, 2 unused in 1 CONFI 1 unused arg in 1 routine FGPLT 2 unused in each of 2 routines FRMAP 3 unused args to 1 routine GAL 1 unused arg in 1 routine POSSM 1 unused arg in 1 routine called a lot SNFIT 1 unused in 4 routines SNIFS 1 unused in 2 routines, 2 unused in a 3rd SNPLT 1 unused in 3 routines SOUSP 1 unused in 1 routine Moved nowhere. 14675. August 25, 2015 RFLAG Eric The dynamic memory allocation for DOSCALE in 31DEC15 was enough only for one source. If more than one were included an abort of some sort would occur. It is supposed to be for 10 sources and the task enforces that. Moved nowhere. 14676. August 25, 2015 TYSMO Eric The corrected call sequence to SMOTY did not make it to the calling routine so an infinite loop was generated. Moved nowhere. 14677. August 25, 2015 gfortran Eric UVUTIL Messed up the dropping of unused variables in FRSOLN Moved nowhere. 14678. August 26, 2015 SPECR Eric A call sequence was not updated when doing the conversion from BASELINE to ANTENNA1 etc random parameters. It then aborted. Moved nowhere. 14679. August 28-30, 2015 gfortran $APGNOT Eric Nuisance mismatches of type and rank fixed in: FXVLA MSORT OTFBS OTFUV RLDIF UVFRE UVIMG VBGLU AFILE ACLIP CANDY CANPL CLIP FETCH FILLM FILLR FITLD FLOPM GLENS IMFIT M3TAR MANDL MK3IN MWFLT NINER NNLSQ PATGN SDTUV SHADW SLICE SPLAT SWPOL UNCAL UVCOP UVFIL UVMLN VLBIN BPWAY BSCAN CLCAL DBAPP DBCON DFQID DQUAL FIXRL FXTIM MATCH RGBMP RLCOR SPCAL TACOP TAPPE USUBA VLAMP VPFLG Actual errors: ACFIT Call not corrected fro unused args which were dropped BSMOD One final call missing argument HOLGR Integer passed back and used as float in last call to PHAZIP LISTR TIME array too small in some calls M3TAR Double-word boundaries were not guaranteed, so did RCOPY to equivalenced variables before conversion to local doubles. RLDIF Dynamic memory needed EQUIVALENCE to real*8 UVDEC Correct 1 recently changed call sequence, nuisances IMLOD Call sequence to history routine missing an argument one place (old format type) SNSMO Declarations changed to make it clear that the time arguments are actually arrays. TYSMO Declarations changed to make it clear that the time arguments are actually arrays. VBCAL Was missing the RESULT argument for uncompressed data in the call to CALIB and in the final call to CALIB. Moved nowhere. 14680. August 31, 2015 more GFORTRAN Eric Warning messages solved by equivalences and the like CCEDT FLAGR Actual errors: LISTR Number mismatch fix earlier led to format errors on times in 2 places WARP Make EXTERNAL function call consistent, clean up code a little CONFI One call was not corrected when the unused argument was removed. CL2HF The routine that initialized HF record values declared one double as a single and the set it to double blank. FINDR OTABXX really needed a new call seq rather than write really different things in the "same' column in the 2 table types. Moved nowhere. 14681. August 31, 2015 still more Eric ANCAL Bad array declaration, erroneous error message WRITE statement SHOUV TIME has only 4 values, task confused about time ranges which were not used COHER Many unused FORMATs and labels, one poor declaration led to size warning BSMOD MAIN called "doit" routine with useless arguments from my dropping unused args OFLAG History routine did not have any args SNREF Another failure to drop an arg in a call UVFND Unused labels, excess argument probably from a dropped as unused one JMFIT Many unused labels, DMACH could fail to set return value BLOAT Unused labels, excess string in DATA statement HLPA Excess string in DATA statement HLPB and HLPA have a 1 character string defined as '\\' since some Fortrans require the escaping of this character. gfortran does not seem to like this. Moved nowhere. 14682. September 3, 2015 SAD, TVSAD Eric The creation of the ST table caused the new MF table to disappear from the header. Changed to use the correct header version in the STINI call. Moved to patch 7 Sept. 14683. September 9, 2015 XG2PL Eric NEW TASK: plots the Gaussian fit at a pixel along with the Zeman fit at that pixel. The data, the model, the model components, and/or the residuals may be plotted for the I data and XGAUS model plus the V data and the ZEMAN model. Also changed: ZEINI Pulled from ZEMAN and given an extra argument DOGAUS which indicates how the derivatives are computed. TABZE Pulled from ZEMAN. ZEMAN Gave up the above and adjusted to the new ZEINI call sequence. Found bad test in slopes computation for one-sided. Moved nowhere. 14684. September 9, 2015 SPLIT Eric OUTNAME for single-source is INNAME - added sentence to help file to say so. Moved nowhere. 14685. September 9, 2015 UVGET Eric UVGET set fatal error codes when the catalog headerread worked but with warnings. Reduce its response to warning codes. Moved nowhere. 14686. September 14, 2015 PCAL Eric The SPECTRAL false mode was broken - it did not handle the end-of-file as a normal condition but as an error. Moved correction only to 31DEC14 this date. 14687. September 14, 2015 XG2PL Eric (1) Fixed status handling to make cataloged plot files. (2) Added ability to print solution values at up to 100 pixels (3) Print will include all channels so BLC1 is always 1 and XBAR=-CATR(KRCRP) so drop any providing of XGAUS BLC/TRC. (4) Change APARM(9,10) to APARM(7,8) to specify averaging area and add APARM(9) to select inclusion of pixels with no model (default is to omit them). (5) Fix scaling of V-only plot. (6) Correct character count for characters to left to include the max from each of the 2 plots. (7) Fix bad error code variable. (8) Fix selection of lines to plot in V (used I codes). (9) Put polarization label inside plot, not on Y axis. (10) Fix inputs comments on BPARM and add help on CPARM(4) the x-axis type. Moved nowhere. 14688. September 16, 2015 RM2PL Eric NEW TASK: like XG2PL but to plot models from RMFIT rather than those from XGAUS/ZEMAN. Changes: RMFIT Extract RMINI and TABRM subroutines. RMINI Extracted from RMFIT, add DOSPIX value as keyword. TABRM Extracted from RMFIT. XG2PL Help - correct averaging comments, mention printing of model values at each pixel in the average XG2PL Find auto-scale values only from plotted range of channels. Correct scaling initialization used to count characters at the left for using only part of the vertical range. Moved nowhere. 14689. September 16, 2015 UVPRT, PRTUV Eric If a non-zero weight is scaled to "0" it is shown by these two tasks as "-1" or "1" depending on sign. Added a note to each of the help files to clarify this. Moved nowhere. 14690. September 16, 2015 AIPS Memo 118 Eric Revised AIPS Memo 118 to mention XG2PL and RM2PL in some detail including example plots. Include TEX, PDF and a gzipped PS file but not the horribly large (126 Mbytes) uncompressed PostScript file. Moved everywhere since $AIPSPUBL is for all versions. 14691. September 20, 2015 OMFIT Eric Dynamic memory seems to fail - tried fixing double precision addresses. Moved nowhere. 14692. September 20, 2015 RM2PL, XG2PL Eric Fixed default XYRATIO to use current size of screen, not absolute sizes of full TV screen. Moved nowhere. 14693. September 25, 2015 XGAUS Eric Changed to (1) Make the XG table for the full input image - output residual image will be full size. (2) Use the user's BLC/TRC as a subimage from the start. (3) Dropped BCHAN and ECHAN and added ORDER to control the solving of no baseline, DC baseline, or linear baseline. The initial guess for baseline is always zero. Changed handling of baseline parameter(s) many mnay places. (4) In TV display of images, added SET and RESET WINDOW to allow the user to look at sub-images in detail. (5) Corrected handling out OUTSEQ. The same value was required previously for all parameter images even when OUTSEQ = 0. (6) Changed help about XG table full size and SET/RESET WINDOW. Also HLPGAUS.HLP about the SET and RESET WINDOW options. Moved nowhere. 14694. September 25, 2015 ZEMAN Eric Changed to (1) Make the ZE table for the full input image - output residual and corrected image will be full size. (2) Use the user's BLC/TRC as a subimage from the start - the previous version did the full ZE-table image always on restarts. (3) In TV display of images, added SET and RESET WINDOW to allow the user to look at sub-images in detail. (6) Changed help about ZE table full size and SET/RESET WINDOW. Also HLPZEMAN.HLP about the SET and RESET WINDOW options. Moved nowhere. 14695. September 25, 2015 RMFIT Eric Changed to (1) Make the RM table for the full input image - output residual images will be full size. (2) Use the user's BLC/TRC as a subimage from the start - the previous version did the full ZE-table image always on restarts. (3) In TV display of images, added SET and RESET WINDOW to allow the user to look at sub-images in detail. (4) For writing out images, corrected the setting of the flags so that parameters will be written, corrected the setting out the OUTSEQ so that all output images do not have to have the same sequence number (if OUTSEQ 0), and corrected the computation of the uncertainties in Q0 and U0. (5) Changed help about RM table full size and SET/RESET WINDOW. Also HLPRMFIT.HLP about the SET and RESET WINDOW options. Moved nowhere. 14696. September 25, 2015 PATGN Eric Added OPCODE 'INVB' to output 1 / primary beam Moved nowhere 14697. September 25, 2015 MATHS Eric Added OPCODE 'DIVP' to output a polynomial in 1 over the input image values. Moved nowhere. 14698. September 25, 2015 RMSD Eric Put in code to check for no valid pixels in a window and return blanks instead of NaNs. Curtailed excessive message, printing once plus a summary of the number of times it recurred in the least-squares. Robust works better than histogram. Moved nowhere. 14699. September 26, 2015 MFPRT Eric 1. Changed to scale in micro-Jy when appropriate 2. FLUX cutoff changed to apply to beam-correct total flux 3. Units test changed to be case indepepndent in determining Jy and Jy/beam. 4. Removed vast number of unused FORMATs and statement labels 5. Upgraded comments in help file. Moved nowhere. 14700. September 26, 2015 OMFIT Eric Changed Fortran to avoid writing 3 records into one string (made it 3 writes and ZTXIO calls) and to make NOISE(1) <= 0 be +1.0. Previously it set it -1.0 and then scaled CHI-squared with the actual post-fit rms which makes good and bad fits look equal. Changed help to indicate that NOISE(1) is that part of the noise not reflected in the weights. If the data weights are calibrated in 1/Jy^2 then NOISE(1) should be 1.0. If they are 1.0, then NOISE(1) should be the real (average) rms noise in Jy. All of this really only to find the CHI-squared values. Moved nowhere. 14701. September 28, 2015 TYSMO Eric Changed maximum number of times to 110000 (> 30 hours at 1 sec). Added messages if one exceeds that. Moved nowhere. 14702. September 28, 2015 BPASET Eric DOBAND init routine does not work optimized with gfortran. Changed OPTIMIZE.LIST for LNX64 since the routine is not called all that much. Moved nowhere. 14703. September 28, 2015 UVMOD Eric Corrected number parameters adding 1. FQCENT was always zero previously. UVPRT - dropped unused formats, labels Moved nowhere. 14704. September 30, 2015 Units Eric MFPRT Needed units=' ' for primary-beam and bandwidth correction factors Also changed to handle the units in a case-insensitive fashion: MFINI FACSET COMB MCUBE PRTIM FQUBE IM2CC IMFIT JMFIT CONVL FFT IM2UV BLSUM IMEAN ISPEC RSPEC Moved nowhere. 14705. October 1, 2015 BP cal Eric The weight scaling for bandpass was not applied to the cross-hand polarizations in DATBND. Moved nowhere. 14706. October 8, 2015 AIPS Memo 118 Eric Revised the memo to discuss the 31DEC15 version of XGAUS, ZEMAN, RMFIT, XG2PL, RM2PL, and the modeling tasks. New text is in red, some figures were updated. Moved everywhere in $AIPSPUBL. 14707. October 8, 2015 TYAPL Eric Help portion of TYAPL.HLP did not mention INEXT. Moved nowhere. 14708. October 8, 2015 SNPLT Eric Had not learned that dispersive delay is now in SN tables. Units of DDLY are sec/m/m also. Moved nowhere. 14709. October 9, 2015 DELZN, CLCOR Eric In DELZN, added a SAVE, dropped unused labels and formats, cleaned up typing. Added DISP opcode to solve for dispersion as another 1/sin(elevation) parameter allowing it to plot and write a text file for CLCOR. Clarified what is plotted in the help file. Added DISP OPCODE to CLCOR to read the output from DELZN for dispersion and apply it to the CL table. Everything was tested with a real DELZN data set modified by a special version of UVMOD which makes clock error delays plus 1/sin(elev) delays and dispersions. Moved nowhere. 14710. October 12, 2015 install.pl Eric Watching a large number of VLB students install AIPS produced suggestions to attempt to reduce difficulties: (1) Setting $AIPS_ROOT, tell user not to include blanks and then check that there were none. (2) Data area names: tell not to use ones including NRAO or OARN because they have special meaning and make other suggestions. (3) Tell user that blanks are not allowed in data area paths. (4) If no data areas are given, make the suggested area be the full list. (It gets created and given SPACE kater.) (5) Save services information to repeat at the end if needed. Put repeated info about defining services at end. (6) At end, add lines meant to draw attention to the text and add /etc/sysctl.conf details when arch is MACINT. Also mention .bashrc, .tchsrc, and .chhrc files. (7) Changed wording in laptop question and made Y the default. Moved nowhere. 14711. October 12, 2015 UVCOP Eric UVCOP tried to avoid work and so did not adjust the FQ offsets when copying 1 FQ table from BIF=1. If the offset of IF 1 was not zero, this was bad. Changed to adjust always even if by zero. 14712. October 12, 2015 FRING, RLDLY Eric Added warning in help files. APARM(5) or DOIFS > 0 will not work if the frequency increment between IFs in a group of IFs has the opposite sign from the frequency increment between channels within the IFs of that group. Moved nowhere. 14713. October 12, 2015 SNIFS Eric Did not do the page full test properly and so died trying to open an already open TV. Moved nowhere. 14714. October 13, 2015 FLOPM Eric Added OPTYPE 'IFON' to flop only IFs leaving channels alone. Moved nowhere. 14715. October 26, 2015 VBGLU Eric The double precision offset into the dynamic memory was not updated when the dynamic memory was reassigned. This led to an abort in the table copying. Moved nowhere (31DEC15 bug). 14716. October 29, 2015 DBCON Eric Changed: 1. Added option to use the previous NX (Index) tables to define the scans as best as possible. Added subroutine DBNXGT to set up the tables a la RNXGET. 2. Made unrecognized sort codes emerge as '*' 3. Fixed frequency ID setup which mixed a counter of number of FQ tables in the 2 data sets with a pointer to FQID number which needs to be a separate and much larger array. Moved nowhere. 14717. October 29, 2015 Mac El Capitan Eric/Josh Malone Changed install.pl to mention a new script that should allow users of Mac OS X 10.11 ("El Capitan") to run AIPS. The script was saved in $SYSMACIN as fix_aips_elcap.sh. Moved nowhere. 14718. October 29, 2015 FRING Eric Put in a check for frequency increments of opposite sign between channels in a group of IFs and the IF center frequencies. The task does not get good solutions in such cases and so should quit and recommend FLOPM. Moved nowhere. 14719. November 2, 2015 RLDLY Eric Added APARM for control parameters: APARM(1) SNRMIN, 0 _ 5 which was the previous fixed value APARM(2) > 0 do not write CL table when have 1 cal scan APARM(3) Max delay error in ns allowed to be included in averaging, 0 -> 1000000. The task now always writes an SN table. IFs with failed solutions now get delay and phase of 0 not magic blanks to avoid deleting data. The display of the answers written to the SN and CL tables now includes not just the formal error but also the rms over the reference antennas. Moved nowhere. 14720. November 9, 2015 PUTHEAD, GETHEAD Eric The full range of available PTYPEn was not shown in the explain files. Moved nowhere. 14721. November 9, 2015 FITTP Eric FITTP did not handle random parameters after ANTENNA1/2 in the list properly. If CORR-ID was after ANTENNA2 then the CORR-ID parameter was labeled ANTENNA1 bit contained the corr-id values (usually all 1). Moved nowhere. 14722. November 10, 2015 OOP Eric UVPGET returns pointers to the last occurrence of the various strings - e.g. if ANTENNA1 occurs twice in the list of random parameters, UVPGET returns a pointer to the second one which turns out to be what one wanted. In UVDESC.FOR subroutine UVDFND returned a pointer to the first occurrence. Corrected it to do the same as UVPGET. The ANTENNA1 bug in FITTP thus affected BLING, IMAGR, OMFIT, DTSUM, and RFI only. Moved nowhere. 14723. November 10, 2015 multi-panel plots Eric The plot buffer needs to be SAVEd to make sure that it is preserved between calls to plot the panels of a multi-panel plot. They have seemed to work without this, but there is no harm in it and the SAVE should be necessary. Tasks changed: BLSUM ANBPL CONFI DELZN ELINT GCPLT PBEAM POSSM SNFIT SNIFS SNPLT VPLOT WETHR SAVEd also CHOUT in BLSUM. The rest had put CHOUT in an included COMMON as a rather global SAVE. Moved nowhere. 14724. November 10, 2015 IMSTAT Eric The limits in the help file for BLC did not include BLC(1)=-1 for a circular aperture. Fixed that and a typo. Moved nowhere. 14725. November 11, 2015 ALVAR Eric NEW TASK (old ALVAR renamed ALVPR) - plots Allan Variance of complex or phase with true mean or robust mean computation. Averages over channels and IFs and can plot baseline by baseline or the global average of the Allan Variances over all selected baselines. Plots can be linear in X or Y or log in X or Y. Moved nowhere. 14726. November 11, 2015 general help Eric Updated tab completion file HLPIT.LIS and the ABOUT files ANALYSIS MODELING ONED PLOT POLARIZA SPECTRAL TASK TV-APPL UV Moved nowhere. 14727. November 12, 2015 SNSMO Eric Changed info in inputs section about NORMALIZ. Moved nowhere. 14728. November 13, 2015 install.pl Eric Move the $SYSMACIN/fix_aips_elcap.sh information to just before the attempt to RUN FILAIP (where the fix will be needed). Add a pause (hit return when ready) line to allow time to do that operation in another window. Moved nowhere. 14729. November 13, 2015 ALVAR Eric Fix up file status so that it can make plot files. Moved nowhere. 14730. November 16, 2015 ATLOD Eric The LUN for the tape drive was not corrected when ZTPOPN was changed to use lun = 129 - INTAPE and so the ZTPOPN failed. $AIPNOT/RDFITS also had this problem. Moved nowhere. 14731. November 17, 2015 ATLOD Eric Fixed the task to get INTAPE to the TAPIO initialization routine as well as to the file skipping routine. Fixed buffer mismatch a bunch of places to reduce the number of warning messages considerably. The INTAPE thing required moving the variable to another include. How did it work when the TAPIO call always assumed tape device 1? Moved nowhere. 14732. November 18, 2015 ANTAB Eric Removed the prohibition against multiple subarray data, but left warning messages in the Fortran and the help files. It is tempting to write subarray 0 for all subarrays, but I am afraid that APCAL might not support that fully. Moved nowhere. 14733. November 18, 2015 DOVLAMP Eric Edited the help file to reformat long lines and to replace "CLCOR" with "CLCAL" which is what is actually used. Moved nowhere. 14734. November 19, 2015 EVAUV Eric Added SMODEL adverb and functionality to allow comparing a point model to the data instead of the image CC model. Much of the code already supported this, but DOPTMD was always false so it was not used. The one thing lost when using SMODEL is the analysis of the true RMS of the images. Moved nowhere. 14735. November 20, 2015 VLBARUN, VLBAUTIL Amy Incorporate new amp cal scheme described in VLBA Scientific Memo #37 (Craig Walker), which changes to VLBARUN and VLBAUTIL. Changes to VLBAUTIL: 1) new proc VLBACCOR - runs ACCOR, SNSMO and CLCAL 2) new proc VLBABPSS - runs BPASS, ACSCL, SNSM0 and CLCAL 3) new proc VLBAAMP - runs APCAL and CLCAL 4) added message to VLBACALA saying you shouldn't use this. Changes to VLBARUN: 1) remove VLBACALA 2) replace VLBACALA and change order of calibration to: - VLBACCOR -- digital sampling corrections - VLBAMPCL (or VLBAPCOR) -- remove instrumental delay - VLBABPSS -- do bandpass and correct auto-correlations - VLBAAMP -- a priori amplitude calibration 3) not related to amplitude calibration, improved several things about how the images were produced (e.g. auto-boxing was not set up properly). 4) not related to amplitude calibration, fixed adding images/ uvplots to .html file. 5) not related to amplitude calibration, added the ability to correct data with a DELZN produced file. Help files: 1) added warning to VLBACALA help file that it should not be used. 2) added new procs to VLBAUTIL help file Moved nowhere. 14736. November 23, 2015 VLBARUN, VLBAUTIL Amy While editing to cookbook I realized that I created VLBABPSS and VLBAAMP with not the most logical seperation of tasks (except to maybe for Craig since he would prefer APCAL by itself). I took ACSCL (and SNSMO and CLCAL) out of VLBABPSS and put them in VLBAAMP. This required a small change to VLBARUN and the help files for VLBABPSS and VLBAAMP. Also edited VLBAMPCL.HLP and VLBAPCOR.HLP so that they give the correct recomendation of the order to run the VLBAUTIL procedures. Moved nowhere. 14737. November 23, 2015 UVMOD Eric Did not address visibility data correctly when BIF > 0 in the model computation routine. Frequently this would mean that no model was applied to the output data. Moved nowhere. 14738. November 23-24, 2015 IMFIT Eric Found a bunch of variables that needed SAVEing in IMFMOD and 2 in FCN itself. Modernized the code style in IMFMOD, IMFERR, and elsewhere. Moved nowhere. 14739. November 24, 2015 CLCOR Eric EOPS is for all data sets with a proper CT table, not just the VLBA correlator. This now includes DiFX correlators and, recently, even the EVLA (when data come into AIPS vis BDF2AIPS = OBIT's BDFIn). changed comments and messages in Fortran and help files. Moved nowhere. 14740. November 25, 2015 TYSMO Eric When there was only 1 polarization, the setting of the pointers to the second one was -1. When the IF number was added to this to do IFs 2-n, the number was no longer negative and the lower level routines attempted to do things with non-existant values. The SY table then ended up trashed. Changed the -1's to -1000's. Improved labeling of Psys and corrected some minor issues in the help file. Made GETCDS check for number polarizations and not complain about the non=existant 2nd polarization. Moved nowhere. 14741. November 26, 2015 PCAL Eric LUN 28 over used - changed all 27 and 28 to other values. LUN 28 already open was reported. Moved nowhere. *********************************************************************** *********************************************************************** Changes while 31DEC15 is NEW *********************************************************************** *********************************************************************** 14743. November 30, 2015 install.pl Eric Corrected grammar in my recent if $ARCH statements. Moved to 31DEC15 this date. 14744. December 1, 2015 FDEFAULT Eric Changed the options for gfortran to those I would recommend. The -Wall makes a message for every integer = float statement for no good reason, so dropped it among others. Moved to 31DEC15 this date. 14745. December 1, 2015 install.pl Eric Add warning message to changed section re laptop. Moved warning to 31DEC15 Dec 6. 14746. December 6-7, 2015 fix_aips_elcap Eric The Xwindows libraries had an extra close parentheses in their names for no good reason. The nice loop that Josh wrote gave a file name of *.dylib for one of the links so changed to make it do the 3 files by the names they are known to have. Dec 7: Josh wrote a proper test on Xwindows rather than linking the files bu force. His loop looks okay but gives odd results on El Cap so left my explicit file linking. Moved to 31DEC15 this date. 14747. December 8, 2015 EDIT class Eric Changed QEDIUTIL at NEXT ANTENNA. It failed when there was only one antenna other than the edited one and that antenna was the next antenna. Now it will start upping the second antenna number to one more than the edit one. Moved to 31DEC15 this date. 14748. December 11, 2015 CLVLB Eric It treated the fact that one CL record was below the horizon as a fatal error. Changed to count the below horizon indicators and report that number plus the total at the end. Moved to 31DEC15 this date. 14749. December 15, 2015 install.pl Eric Add copies of fix_aips_alcap.sh to $SYSLOCAL and $AIPS_ROOT. Moved to 31DEC15 this date. 14750. December 16, 2015 UVFLG Eric The text-file reading got confused about times which it set to minus the day number. Then it tested if both were zero. Added test for both = minus the day number and changed the complaint to allow time1=time2 which should not occur. INTEXT reading did not re-initialize all adverb values between commands leading to later commands inheriting parameters from the previous ones. Simplified the re-initialization to be all inclusive. Moved to 31DEC15 this date. 14752. December 17, 2015 ABOUT lists Eric Updated HLPIT.LIS and the about files for AP, CALIBRAT, PROCEDUR, SPECTRAL, TASK, UTILITY, and VLBI. Moved parts from 31dEC16. 14753. December 18, 2015 PCAL Eric Found out that data points with some but not all polarizations flagged would persist in the data selection and cause problems including the infamous 'POLARIZATION MODEL NOT PROVIDED' message which caused the task to stop. Moved to 31DEC15 this date. 14754. December 18, 2015 PCAL Eric Changed to do linear polarization model only when the data are linear and CPARM(5) > 0. That model does not seem to work and circular (APPR) does when the source model has zero for polarized flux. Moved to 31DEC15 this date. 14755. December 18, 2015 UVHOL Eric A choice of text LUN=89 was unfortunate - changed to 91. Fixed it to do I only with linear polarizations in data set. Moved to 31DEC15 this date. 14756. December 21, 2015 CookBook Eric/Amy Updated chapters: 3 : PRTMSG (OUTPRINT, line count check), re-word tape section 4: TVFLG LOAD LAST PIECE, SOUSP fix SNs, EDITA DO3COLR. de-emphasize tape, drop UVRANGE table 5: antique wording, OBOXFILE temp fille, DO3COL in EDITR/EDITA, reduce priminence of mag tape 6 : Memo 119 (TVSAD), Memo 118 revised, TVHLD, TVMOVI (DOALL), UVIMG, XG2PL, RM2PL, XGAUS (baseline setting) Downplay magnetic tapes at the start 7 Memo 119, TVHLD, MCUBE forgiving, STACK, Memo 118 revised, XG2PL, RM2PL 8. MCUBE forgiving, TVMOVIE DOALL, Memo 118 revised, UVFRE 9. ACSCL, DOVLAMP, DIGICOR in FITLD, dispersion in MBDLY and FRING, RLDLY, TYCOP, DELZN, EDITx CROWDED with color Overhauled for VLBARUN/VLBAUTIL new cal method and to modernize 13. Update C Overhauled for VLBARUN/VLBAUTIL new cal method and to modernize E. FRING dispersion, RLDLY SN tables, fix antique wording places, FRING divide in > 2 parts, TVFLG groups, SOUSP SN tables, UVRABGE -> break IFs into separate files, SYCOP renamed TYCOP Moved nowhere. 14757. December 22, 2015 VPFLG Eric Added calls to make NX (index) table on the fly. Moved to 31DEC15 this date. 14758. December 22, 2015 IRING Eric The geometry was questioned by a user and was reviewed in detail. The radii were increased erroneously by a division by cos(inclination) but the rest of the geometry was found to be correct. Changed the task to have OUTTEXT='NONE' suppress both the text and message file outputs. Also changed it to honor the Y axis range in CPARM(9,10) when CPARM(8)>0. Previously it honored it only when the values did not exceed the stated range. Moved to 31DEC15 this date. 14759. December 28, 2015 PBEAM Eric Changed the task the handle data in any order. Previously it demanded that the data lie in columns on the sky, although there was some confusion with a row-ordered plan also. Changed the plots to be contour plots and to write out the data. model. and residual images. Previous plots were row traces. Corrected a minor issue in EVAUV's contouring. Moved to 31DEC15 this date. 14761. December 29, 2015 PBEAM Eric Found a variable that was not initialized which caused the task to fail to read the data the second time in which the normalization was supposed to have been done. Added APARM(9) to control the width of the convolving function which was too narrow before. Re-named the axes and made the X axis right handed. Made less drastic contour intervals for the residual image and made the data sample markets slightly bigger. Moved to 31DEC15 this date. 14762. December 29, 2015 HOLOG Eric Correct the amplitude image by the pixel-dependent sum of weights making the image be an interpolated one rather than a convolved one. Also normalize that image and make unsampled pixels be magic blanks. Moved to 31DEC15 this date. 14763. December 29, 1015 holography Eric Further changes: UVHOL Was limited to documenting only 14 averaged antennas. Changed to write more than one line (like ref ants) HOLOG Changed to subtract the phase at the pixel of peak amplitude from the phase image, thereby "normalizing" phases as well as amplitudes. PBEAM Changed parsing of text file to free format and added parsing of the Averaged antennas and Averaged RefAnts input lines. Changed display of antennas and refants to support the full (possible) lists. Moved to 31DEC15 this date. 14768. January 7, 2016 GC tables Eric Changed PGCV.INC back to allow up to 200 values (per polarization, per IF, per antenna) for the Y and gain. Some correlators write out 200 values (almost all magic blanks) even though they have no use for them. Moved to 31DEC15 (patch) this date. 14770. January 12, 2016 PBEAM Eric Corrected a subscript error that made the addition of RR and LL to make Stokes I get scaled incorrectly. Also fixed a format issue telling the user the re-scaling factor. Moved fixes only to 31DEC15. 14771. January 13, 2016 RLDLY Eric The output SN table was written to "CLOVER" which could be set to some non-zero value if a new CL table was also written. Fixed to always write a new SN table and improved the history writing as well. Moved to 31DEC15 this date. 14772. January 15, 2016 FITLD Eric Two subroutines GNINI and TABGN were not compiled when PGCV was changed. They include an include that includes PGCV and so have to be compiled. Then FITLD needs to be relinked. Moved to 31DEC15 this date. 14774. January 19, 2016 EDITA, EDITR Eric Changed QEDIUTIL to make the all-source status false to begin with. Corrected the expansion of times in flags to avoid going a long way in one direction due to a gap in the selected data in the work space. Moved to 31DEC15 (patch) this date. 14781. January 28, 2016 DBCON Eric DBCON reinstated flagged table rows. Changed to drop them completely. Number of output records had to be fixed in output table header. Moved to 31DEC15 (patch) this date. 14789. February 9, 2016 DTSUM Eric Typo meant that it did not interpret the new UV format correctly and assigned all baselines to the same autocorrelation. Moved to 31DEC15 (patch) this date. 14814. April 15, 2016 PCAL Eric The baseline pickup for one of two data copies had the handling of the new antenna format wrong! IA2 was not set so I do not see how it could work. Moved to 31DEC15 this date. 14816. April 16, 2016 PCAL Eric Changed the handling in RPCALC and LPCALC of times for which the I model has value 0. It reports it but keeps going so long as the weight is also zero. This was caused by the ionospheric Faraday rotation being blanked for reasons I cannot discover - probably relating to flagging being applied to the CL table. Moved to 31DEC15 April 29 without DPARM. 14820. April 29, 2016 PCAL Eric Found what was wrong - the Faraday rotation values were lost entirely when an IF was copied to a scratch file before pulling out individual channels. Changed it to: (1) Not delete the ION file when killing other scratch files except between IFs. Zero scratch pointers when deleting the scratch file. (2) Create a new ION scratch file when needed before the copy of the individual IF to its scratch file and write the ION as that copy is done (3) Do the ION file with the data read only when not doing spectral solutions as opposed to always doing it. Values are not available if the IF was moved to a scratch. Subroutine $QNOT/GRDSUB has to defend itself when pointers to SCRGRD and SCRWRK point at improper values (outside NSCR or disk or cno <= 0). Moved to 31DEC15 this date. 14826. May 10, 2016 IMAGR Eric Changed QCLEAN - it had some wrong subscripts when testing islands for making automatic boxes. If IMSIZE(2) < IMSIZE(1) then islands would be found only for X values less than IMSIZE(2) - i.e.the left half, quarter, etc of the image. Moved to 31DEC15 this date. 14829. May 19, 2016 Edit Class Eric The routines used by SNEDT and EDITA to read in the table data allowed source number column values <= 0. But the flagging routines with "ONE SOURCE" would not write a flag record if no source number > 0 was found. Fixed EDGT SN, TY, SY in QEDIUTIL to force the source number to be 1 or greater. Moved to 31DEC15 this date. 14842. June 16, 2016 UVFIX Eric UVFIX goes to a lot of work to find the actual frequency of each data sample. However, the definition of U,V,W is wavelengths at the header frequency not wavelengths at the real frequency. Changed scaling of U, V, W accordingly. Normally, AIPS keeps the offset of the first IF zero so this would not matter normally. But CASA/ALMA has no such scruples. Moved to 31DEC15 this date. 14787. February 5, 2016 install.pl Eric Changed error messages around the make of XAS to point out that fixing Makefile and doing a make in $YSERV/XAS is all that is needed. Also added $CCOMOPT in the rewrite of the Makefile before the make. Moved to 31DEC15 June 23. 14847. June 23, 2016 install.pl Eric I wanted to show the user $CDTST but perl insists on translating that so changed the execution of fix_aips_elcap.sh from $SYSMACIN to $AIPS_ROOT which will translate properly. Moved to 31DEC15 this date. 14848. June 24, 2016 ALGMEM Eric If the last pass through this routine started with a model number greater than one but only included a single range of frequency channels, then the array of scaling numbers to step through the frequencies was not set properly. Moved to 31DEC15 this date. 14891. August 25, 2016 UVLOD, FITLD Eric The conversion of BASELINE to SUBARRAY was inaccurate and failed in some cases. Adding roundup of 0.0001 was not enough, changed to 0.05. Moved to 31DEC15 this date. 14893. August 26, 2016 SPLIT Eric When averaging IFs as well as channels, SPLIT got the wrong output frequency. UVGET had been called so the header frequency was adjusted for BIF; the computation did not take this into account. Use OLDFRQ instead. Moved to 31DEC15 this date. 14896. August 30, 2016 UVLOD, FITLD Eric Subroutine RUVTAB did not handle compressed data properly when DOKEEP was false. Added a check that went over the buffer properly. Moved to 31DEC15 31 Aug. 14910. October 6, 2010 SLICE Eric The operation to drill down through a cube with a single pixel in X and Y and the slice running along Z had a problem if the X and/or Y pixel was not exactly an integer. Bad slices resulted. Corrected the code to interpolate in X and Y correctly during this operation. Moved to 31DEC15 this date.