;----------------------------------------------------------------------- ; Copyright (C) 2006-2008 ; 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 31DEC07 is TST *********************************************************************** *********************************************************************** 12219. December 3, 2006 31DEXC06 -> 31DEC07 Eric Changed GETRLS, INSTEP1, ZXHLP2.PL, UPDCONTROL (comment), AIPSPATH.SH, AIPSPATH.CSH, install.pl, update.pl to make the TST version 31DEC07. moved nowhere. 12220. December 3, 2006 Pseudo AP and dynamic memory Eric Changed all pseudo-AP code to use dynamic memory. If we keep the pointers wrt APCORE(base) small enough that 32-bit integers will work, then only base and any direct subscripts of APCORE need be LONGINTs. This should make the problem tractable. Changed: DAPC.INC Include DAPM.INC DAPM.INC Define SPAD and new pointer values PSAPNW and PSAPOF (the LONGINT base) ZDCHIN Include DAPM.INC and set PSAPNW and PSAPOF to 0. Also drop registration trash, fetch $SITE for sysnam. Just typing mods: QCFT1D QCFT2D QCFT3D QCFT1I QCFT2I QCFT3I QCTLUT QGRDCC QGRID QGSP QLSORT QMCALC QMENT QOSORT QPERMA QRLSE QUVINT QWAIT QWD QWR QXFOR2 QXFOUR Also QCLNSZ use PSAPNW now QCVSMS drop numeric IF, fix pointers APUTIL Direct APCORE usage now with pointer QCLEAN Direct APCORE usages changed to use pointer. QUVUTIL drop equivalence with APCORE, make dimension MAXCIF QEDIUTIL rewrote to use dynamic - non-ap - memory. Changed tick algorithm to use same units in all plots of the same type.and extended the plot range. Various pointers into APCORE Q1FIN Q1GRD QBAKSU QBOXSU QCFFT QCGAUS QCLNSU QCRVMU QCSQTR QCVCMU QCVCON QCVEXP QCVJAD QCVMAG QCVMMA QCVMOV QCVMUL QCVSDI QCVSMA QDFT QDIRAD QFINGR QGADIV QGASUB QGET2 QGET QGRD1 QGRD2 QGRD3 QGRD4 QGRD5 QGRD6 QGRD7 QGRD8 QGRDFI QGRDMI QGRIDA QHIST QIHIST QINTG QINTP QLVGT QMAKMS QMAXMI QMAXV QMINV QMTRAN QMULCL QPHSRO QPOLAR QPTADC QPTDIV QPTFAZ QPTSUB QPUT2 QPUT QRECT1 QRECT QRFFT QRFT QSEARC QSORT QSPDIV QSPSUB QSVE QSVESQ QTAPER QUVIN QVABS QVADD QVCLIP QVCLR QVCOS QVDIV QVEXP QVFILL QVFIX QVFLT QVIDIV QVINDE QVLN QVMA QVMOV QVMUL QVNEG QVRVRS QVSADD QVSIN QVSMAF QVSMSA QVSMUL QVSQ QVSQRT QVSUB QVSWAP QVTRAN QVTSMU QWTCNT QWTGET QWTGRD QWTSUM QXXPTS Moved nowhere 12221. December 3, 2006 KAPWRD Eric The AP size is now PSAPNW and only after QINIT. Changed QCLEAN SR CLGRID to set desired size, call QINIT, then do normal ap size arithmetic using PSAPNW. SR CLRGRD done similarly. Add error tests after QINIT calls. SR CLNSRT change GRDFIT call seq. Add message supression, re-init AP if needed at appropriate size. QUVUTIL SR OUVIMG find NEED earlier, call QINIT, test if grid fits silently. If not, QRLSE and new QINIT and new GRDFIT. These QINIT calls are new. Code limits the request to 1/4 or 1/2 Gbyte. SR ODFT determine a need and then call QINIT. QARRAY Fixed up QINIT call in ARRFFT. GRDFIT Add new return argument, the size needed to avoid a sort. Use PSAPNW. UVWAIT Moved QINIT call and forced needed to be available, avoiding disk usage. APROLL Will not need but save size AP for QINIT call. AEGRID Move QINIT call and force memory to needed. ALGSTB Compute what id needed and call QINIT with that, then do existing computation. ALGSUB As ALGSTB but simpler. CONV Estimate min needed and put it in QINIT calls. CORR Estimate min needed and put it in QINIT calls. Fix error messages - this is not CONV. GR3DUV Estimate needed memory and QINIT for it. If that fails try smaller and quit if 2nd fails. GRDCOR Estimate tiny needed memory and put in QINIT call GRDMEM Fixed PSAPNW and the error messges about AP too small. GRDSUB Estimate sizes needed and use in QINIT call. GRIDUV Like GR3DUV. IMGMEM Just added comment around null QINIT call. MAKMAP Compute what is needed for QINIT for DSKFFT, drop QINIT and QRLSE for GRDCOR which does those itself. OLDMAP As MAKMAP SDGRID Estimate desired space and call QINIT, if that fails try a smaller space SRTIN Add comment only re UVSORT setting QINIT size UVSORT Add estimate of size needed to do all in core and try a QINIT. If that fails try smaller twice. Do a QRLSE afterwards. UVGRID Estimate needed and try a QINIT, If fails, try a smaller one. Drop the QRLSE and nearby QINIT leaving the initial QINIT in place. UVGRTB As UVGRID. UVMDIV Compute needed AP memory and call QINIT/QRLSE to get it. UVTBGD Use the NEEDED parameter to call QINIT. UVTBUN Use the NEEDED parameter to call QINIT. UVTBUW As UVTBUN. UVUNIF Compute needed size for QINIT call VISDFT Compute needed size and set with added QINIT call. Use later too. VSCDFT As VISDFT but had to reorder - fixed a test on a variable that was not set as well. VSFDFT As VSCDFT. Simply change KAPWRD+KAP2WD to PSAPNW: PASS1 CCSGRD CONV1 CONV2 CONV3 CONV4 PASS2 Added code to manage small output images. It was writing row 2 to N/2 then row 1 and row N/2+1. This is okay if a row occupies a whole disk sector but fails on short rows. In this case, the image is kept in core and dumped at the end. Moved nowhere 12222. December 3, 2006 Q Tasks Eric Tasks: CXCLN Change to LONGINT in references to APCORE; changed KAPWRD ref to PSAPNW and fixed QINIT call to ask for what is required. IMAGR Dropped memory-saving equivalence w APCORE. WFCLN Dropped memory-saving equivalence w APCORE. RSTOR Changed to use PSAPNW and to specify what is desirable to QINIT. APCLN Changed to use PSAPNW and to specify what is desirable to QINIT. Corrected errors where MAXPCH did not test the AP size and where it could then change to new values w/o control. APGS Added size quess for QINIT call, dropped unused variables. APVC As APGS. CONPL Just a comment at QINIT. MX Add size setting and QINIT call in MXIN, moving all APSIZ references to after that. Changed KAPWRD refs to PSAPNW. SDCLN As APCLN. STEER As APGS. UTESS Put size estimate in FILES subroutine, sets rest of QINIT calls. Fix error branch to return error code in GETIN on test of NOISE. UVMAP Put size estimate in INI subroutine to set. Remove unused variables. VTESS As UTESS. DEFLG Replace KAPWRD w PSAPNW and change APCORE subscripts to LONGINT offset by PSAPOF rather than 1. BPASS Request a small AP for shifting VLBI BPs twice CONVL Estimate needs and put in the QINIT calls. CVEL Request a small AP for shifting VLBI BPs FFT Request a good size AP and settle for less before the QINIT call. FRING Move QINIT call from FRNFIT to FRNSRC and FRNALL. have these 2 compute NEED and call QINIT rather than the complex set of ZMEMRY calls there before. IM2UV Request a good size AP and settle for less before the QINIT call. IMERG As FFT. SDVEL As CVEL. SDIMG Replace KAPWRD with PSAPNW, change APCORE subscripts to LONGINT offset by PSAPOF, add QINIT/QRLSE in INI routine to set a desirable size for the AP. UVCON Changed local QXXPTS to longint refs to APCORE, changed local VISDFT like VISDFT. Moved nowhere. 12223. December 3, 2006 pre-existing errors Eric 1. Created QPUTIN, QPUTI2, QGETIN, QGETI2 routines to do QPUTs, QGETs with increments. Changed the TV loading function in APCLN and SDCLN to use QPUTIN. Otherwise it did TXINC=1 with any TYINC. Changed APGS, APVC, STEER, UTESS, and VTESS to do the same. They avoid the INC problem but did the function way into memory that it should not have accessed. 2. Changed UVMAP to do the same TINC in both X and Y; it did different ones previously. 3. Changed file size computation for CONVL. It needs the larger of NX*NY and 2*NY*(NX/2+1). Moved nowhere. 12224. December 5, 2006 IMAGR Eric Changed formats in PBUTIL - it was using I2 for field number. This stopped the frequency-dependent model subtraction if there were more than 99 fields. Moved to NEW this date. 12225. December 8, 2006 3C138_L.MODEL Amy/Eric Put back Amy's latest calibrator source model. Moves to NEW etc automatically. 12226. December 13, 2006 VLATECR Amy Fixed the bug in DAYNUM subroutine that VLATECR uses, same ask the fix to VLBATECR makde on Nov. 27, 2006. Moved to 31DEC06 this date. 12227. December 13, 2006 install.pl Eric Added rsync of the $HIST area which was overlooked. Moved correction to 31DEC06 12228. December 15, 2006 Flagged CCs Eric Changed GRDCRM to ignore flagged CC lines rather than die on them and changed FACSET to ignore them rather than use them. Moved correction to 31DEC06 12229. December 15, 2006 QPTDIV Eric Model division for point-source CC files was broken by an error in the overhaul of Q routines. Wrong values of U and V were used. Moved nowhere. 12230. December 16, 2006 SPLIT Eric Changed default of CHINC to NCHAVG. Moved to 31DEC06 this date. 12231. December 18, 2006 CLCAL Eric The phases of IFs 2-N were being changed for smoothing done on delays of those IFs as Delta(phase) = -2pi * (F(ifn)-F(if1)) * Delta (delay of ifn). This code was commented out as well as a non-trivial amount of supporting code. Moved to 31DEC06 this date. 12232. December 19, 2006 FILLM Eric FILLM was changed to parse the IF control bits word to see if the data have been scaled by the nominal sensitivity or not. All data for format numbers 24 and less are assumed to have been scaled and all recent data should be scaled. However, new data from the E/VLA starting in 2007 will not be scaled and so will need scaling by FILLM. FILLM checks the bit and makes sure that it is the same for all IFs and antennas and scales any data that have not been scaled. Moved to 31DEC06 to support 2007 VLA data. 12233. December 21, 2006 FLAGR Eric Fix it to flag singletons when it is supposed to do so. Moved to 31DEC06 this date. 12234. December 21, 2006 AVSPC Eric Correct alternate reference pixel location for frequency averaging. Moved to 31DEC06 this date. 12235. December 21, 2006 UVGLU Eric The test on correctness of frequencies was (F1-F2) / (DeltaF) > 0.1 This has to be absolute value > 0.1. Moved to 31DEC06 this date. 12236. December 27, 2006 IMAGR Eric QCLEAN had an error in which a pointer was used in arithmetic without taking into account its new (large + or - offset) value. Changed that computation top use correct values. The error affected some operating systems (RedHat 9) when filtering out weak isolated components. moved nowhere. 12237. December 31, 2006 FLAGR Eric Changed: 1. Summing buffers were only large enough for 2 polarizations if the numbers of antennas and IFs were large. Doubled them. 2. VIS variables were sized for 1 polarization, quadrupled them. 3. Doubled summing buffers yet again to allow for double buffered averaging in RFI, HIST, ANTE, TIME, GAIN modes. Changed those modes to use two summing buffers. This allows the second buffer to be checked when it is done to see if all correlators have only 1 sample. If so, it is added to the first one, making that buffer one sample longer than normal. This beats having to flag or not check all singletons. It also allow more times in 4. Changed RFI/HIST to sum RMS rather than RMS^2 to check for questionable correlators. Had it set RMS=0 for singletons so that flagging on excess amplitudes and weights could go on as normal and then set flags for singletons based on BPARM(9). Otherwise the RMS appeared large (= amp) and so caused questionables/bads without checking BPARM(9). 5. Changed ANTE to sum RMS rather than RMS^2 when checking for questionable correlators. Improved formats for reporting errors. Made some adjustments to the help file as well. Moved nowhere. 12238. December 31, 2006 TV Servers Eric Changed UNIXSERVERS and XASERVERS in a way that will probably produce screams but let's see. Pat found that putting "LD_LIBRARY_PATH=$LD_LIBRARY_PATH $LOAD/TEKSRV.EXE" as the command seems to avoid the need to reference LOGIN.SH in the users' .bashrc et al files. Now if the " marks don't kill us. Moved the SOL versions of these which have no quotes to SUL in 31DEC07 and 31DEC06. Moved rest nowhere. 12239. January 1, 2007 AIPS Eric Update copyright statement Moved nowhere. 12240. January 2, 2007 INDXR Eric Changed it to allow for NX tables in single-source files. CALIB will take advantage of this in self-cal. the help file already claimed that the task worked for single-source files! Added remarks to it anyway. Moved nowhere. 12241. January 2, 2007 CALIB Eric Discovered an error in which the NX table times were extended by nearly 5 seconds in each direction to compute the integration start and stop times. This does not matter when doing scan averages, but does when a true SOLINT is set. An extension of 0.1 sec if probably okay. Moved to 31DEC06 patch this date. 12242. January 3, 2007 XASERVERS, UNIXSERVERS Eric Changed them back yet again. The combination of setting an environment variable and then doing the command within ""'s may well work. But the setting of the variable depends on the nature of the user's shell and so is broken about half the time. Even the use of compound commands in double ""'s wasn't that bad. Leave the -hold and the environment setting as suggestions in comments. Moved to AOC and CV this date by forced MNJ. 12243. January 4, 2007 TAPLT Eric Fixed task to recognize single-precision floating magic blanks. Changed scaling to include only bins with data in them; the others are plotted at the bottom anyway so 0 is not a real value. Fixed Y string to omit the overall scaling which applies to X only. Moved nowhere. 12244. January 4, 2007 VPLOT Leonia Added OPCODE='IFRA' to plot out the ratio of the two IFs BIF/EIF. Moved nowhere. 12245. January 10, 2007 UPDCONTROL Eric Amazingly, this proc still used the old sort syntax. We finally hit a machine that refused it visibly. Since SUN4 is no longer a current OS, simply commented out the old syntax and replace the lines with the new. I think the operation that it was to do is not significant. Moved to 31DEC06 patch this date. 12246. January 17, 2007 OOSUB Eric New task to do UVSUB with the frequency-dependent primary beam and spectral index options of IMAGR. Changed QUVUTIL routines to read in the extended SMODEL source parameters if they have been set. Moved nowhere. 12247. January 17, 2007 servers again Eric Changed XASERVERS and UNIXSERVERS to test the user's shell and do the correct setenv type command. This still may hit trouble with sites that cannot handle " "'ed command sets in an xterm. Moved nowhere. 12248. January 17, 2007 MNJ Eric changed installation procs MAKE.MNJ and MAKE.BMNJ to add some things to the .cvsignore files, namely MASTER and COAOARN to the $TST/$ARCH/SYSTEM/.cvsignore and LIBRDBG to all systems areas. Changed the output of UPDCVS to go to the log file and added explanatory messages. Added onager to UPDCONFIG. Moved nowhere. 12249. January 17, 2007 LOCIT Eric Added options to use the small frequency difference between IFs rather than than the large absolute frequency in cases of really bad antenna locations. Also added the option to fit a phase slope. Moved nowhere. 12250. January 19, 2007 UVFLG Leonia Craig Walker payed attention on the difference using different supression factor (APARM(6)). The bug has been found: the same name TST, TEN used for the time interval inspite of the fact that this interval is adjusted when compared with the low fringe rate interval. As a result the initial time interval of the scan was corrupted at the cycle of baselines and IFs. The information about the found flagged time interval for each IF has been added at the end of running of UVFLG. Moved nowhere. 12251. January 22, 2007 install.pl Eric Changed rsync -ruv to $rsync which is equal rsync -auvz --timeout=120 The z will do some compression, the a = rlptgoD for a full archive representation. Tested this on primate. Moved nowhere. 12252. January 22, 2007 $APGNOT/UVDI1 Leonia UVDI1 subtracts UV data(having averaged up to one time) from the other UV data. This task serves to the same purpose as UVAVG with OPCODE = 'SUBT', but alows sutract one averaged source data from another source data. Moved nowhere. 12253. January 22, 2007 LWPLA Eric Changed it to defer the open until the first PL file is open and the first record read successfully. This saves creating empty files in case of error. Moved nowhere. 12254. January 22, 2007 LABINI Eric LABINI refused to change the reference pixel location or axis type for FELO axes when the user asked for sub-image relative coordinates. (FELO = regularly gridded in freq but expressed in velocity units.) This is too fussy in general for this sort of usage. Removed the refusal. moved nowhere. 12255. January 23, 2007 New verbs Eric Changed ASSGN, KWICK, and EQUIV to handle a type 12 temporary numeric array. In KWICK have PRINT, MAX, and MIN treat it as a 1-dim array. Changed the code for MAX/MIN to allow a single argument, e.g. MAX(ANTEN), although this can be confused by e.g. print 99, max(anten) which will say 99 only. Note that print max(anten), 99 will do what is wanted. Changed POPSGN and NEWPARMS.001 to define verbs VLA, EVLA, VLBA, and HSA which will read the AN file and put the selected antenna numbers on the stack as a temporary array. Thus ANTENNAS = VLA will make an array of the non-EVLA antenna numbers. Adjusted the MAX and MIN help files. Changed AU3B to perform these functions. Moved nowhere. 12256. January 23, 2007 name collision Eric An unknown run and help file called VLBA has been renamed VLBAIT. It purports to reduce VLBA data. Moved nowhere. 12257. January 24, 2007 UPDRSYNC Eric Added z and --timeout=120 to the rsync options used, as in the install.pl. Moved nowhere. 12258. January 24, 2007 FILLM Eric Fixed the coordinates of the master pad from meters to ns which is what they are internally. Now it will recognize the master pad. Moved nowhere. 12259. January 26, 2007 VLA, EVLA, etc Eric I missed a routine MASSGN that needed to change to handle ANTENNAS = EVLA, 3, 4 and missed the fact that ANTENNAS = EVLA, VLA PRINT EVLA VLA do not work. POPS defends itself from verbs so 2 temp arrays cannot be created in one command. Updated the 4 help files. Moved nowhere. 12260. January 30, 2007 IMAGR Eric Changed: IMAGR A message should have said multi-res nut multi-field. TVUTIL The inscribed circle should have been about the numeric center not the reference pixel. The latter is way off the field even when do3d=-1. QCLEAN CLNUV2 forced DO3D true which is way too late since only 1 beam and all the images have already been made. QCLEAN CLBHIS changed beam histogram to stay within a radius = half of the larger axis size. This should keep the histogram from being wierd because of really bad corners. QCLEAN CLNTV rearranged the menu to puit the 3 options that stop things at the start away from the things usually used repetitively. Avoids accidental aborts et al. Moved nowhere. 12261. January 30, 2007 ANBPL Eric Suppress messages from, and further attempts at, scaling when data are not going to be plotted. Moved nowhere. 12262. January 31, 2007 IMAGR Eric There was a path (ALLOKAY=2) through the code that led it to use an AP size before one was allocated. Changed QCLNSZ to call QINIT rather than return 0. Moved nowhere. 12263. February 1, 2007 OOSUB Eric Changed PBUTIL, OOSUB, OOSUB.HLP to use BPARM instead of IMAGRPRM and to implement a new function. This function allows the user to omit from the OOSUB all CCs that are inside the highest frequency beam, outside the highest frequency beam, inside the lowest frequency beam, or outside the lowest frequency beam. This will allow the subtraction of partial models to enable better imaging of the inner portions of the data and will probably have other uses when we get clever. Moved nowhere 12264. February 8, 2007 RMSD Leonia Increase the upper limit of the strip size from 8912 to 16384 Moved nowhere 12265. February 9, 2007 LPSEC, UVFIX Leonia The leap second subroutine LPSEC is updated until the last correction at Jan 2006, and the task UVFIX (only one using LPSEC) is comlnked. Moved nowhere, moved to patch 2/27? 12266. February 9, 2007 MAKE.MNJ Eric Changed the two procs that prepare for MNJs to make a few more .cvsignore files. There needs to be one (with a *) in $TST/$ARCH/SYSTEM/COAOARN since that area has a CVS directory. Similarly, there needs to be a .cvsignore containing PAPC.INC in the $INC/NOTST/$ARCH area. Changed them to specify one thing per line. Older installations will have to add these by hand since they are created during installation. moved nowhere. 12267. February 13, 2007 IMAGR Eric Added more handling of pathological beams in CLBHIS in QCLEAN.FOR. The inscribed circle was adequate for the particular case, but lets handle more bad cases if needed. Moved nowhere. 12268. February 20, 2007 CLCOR Leonia OPCODE='SUND' has not worked probably since the time when OPCODE='EOP' had been installed. The GO TO 999 was missed Moved nowhere, moved to patch 2/27. 12269. February 21, 2007 Gaussian models Eric Corrected QGASUB and QGADIV which had one subscript to the visibility data not converted to the dynamic memory offset. So Gaussian CCs should have been ignored more or less in imaging and modeling. ANY MULTI-RESOLUTION IMAGING with 31DEC07 HAS BEEN WRONG. Moved nowhere. 12270. February 23, 2007 KNTR Eric Changed it to allow ZINC < 0 to mean do the image planes in the reverse order. Fixed help file for this too. Moved nowhere. 12271. February 26, 2007 VLBAPCOR.HLP, FITDISK.HLP Amy Added INVER to VLBAPCOR, per Craig's request. Removed references to DOALL from DITDISK help file. Moved nowhere. 12272. February 26, 2007 VBGLU Amy Fixed merging of the STATE count column of in the PC table. It used to not expand to the correct number of values in the table cell because it had twice the number of values as all the other 2D columns. Moved nowhere, moved to patch 2/27. 12273. February 26, 2007 Misc Eric IMAGR.HLP fix typo re ANTENNAS UPDCVS add C is error re local version Moved nowhere. 12274. February 27, 2007 CVEL Eric The call to CALINI used the same dummy argument to several output variables. This caused the number of polarizations to be set strangely and thus to produce warning messages galore. The final call to DSHIFT had the wrong number of arguments causing the task to abort (with the GNU compiler anyway) at the close down stage (before extension file copies). Moved nowhere, moved to patch this date. 12275. March 1, 2007 SETFC Eric It was discovered that the formula used to determine the maximum facet radius was in error by sqrt(3), or, alternatively, that the max phase error was 3 times larger than stated. Corrected the formula and changed the default to 60 degrees, giving the same output as previously. Changed the help file to give correct formulae, to change the pahse error default, and to give a reference in the explain section. A radius of <= 0 confused the facet size setting since that would be less than the max allowed facet size. Fixed the confusion for <= 0. Moved nowhere. 12276. March 6, 2007 VLANT Eric New task: reads (downloading if needed) the VLA antenna location text files and sums up any baseline corrections. Then modifies the new CL table (copied from an old one) for these corrections and up0dates the ANtenna file. Put VLA.ant.year for year = 1992-2006 in $AIPSTARS. Fixed copyright statement and minor typography in CLCOR. Moved nowhere. 12277. March 7, 2007 DELZN Leonia The new OPTYPE='rate' has been added. This optype uses column 'rate' at the input SN table instead of 'mdel'. This new option may be usefull at the case of small bandwidth when accuracy of the delay measurement can be low. Moved nowhere. 12278. March 8, 2007 UVFIX Leonia Originally differential aberration has been added always. Now I put it under control of UVFIXPRM(18). Moved nowhere. 12278. March 8, 2007 SMODEL adverb Eric Changed VISDFT and VSFDFT to apply the SMODEL parameters including the W*Z term in the phase. That term was always in the low level Q routines but Z was made 0 except in VSCDFT. The 3D imaging with Clean Components depends on the W term being used (where 2D explicitly does not use it) so the Q routines are already constructed carefully. Probably should go to patch, but this has been there forever. Moved nowhere. 12279. March 12, 2007 IMAGR Eric Changed QCLEAN to implement two new TV options - to turn off a facet now and to turn it back on. Changed HLPCLEAN and IMAGR help files accordingly. Fixed some FORMATs in QCLEAN to I5 for field number. Changed TSKHLP to accept XXXXXXXXX as limiting the match to those characters prior to the . Changed HLPCLEAN.HLP to use this for the new options and the old SELECT FIELD and DISPLAY ONLY. The latter did not work since the spellings had changed. Moved nowhere. 12280. March 13, 2007 FILLM Eric Changed the subroutine that checks for scaling by the nominal sensitivity. It will now examine only the antennas in the current subarray and it will report any that differ from the setting for the first antenna in the array. Added code to detect Z mode in which the first spectral channel is not the on-line computed average of the central channels ("channel 0") but the actual zero-lag channel. The number of channels and reference pixel output in this case are one higher and the zero lag channel is retained. This might allow re-apodization of the data. Moved nowhere. 12281. March 13, 2007 VBGLU; VBGLU.HLP Amy VBGLU glued datasets with multiple FQids without complaining but the resulting uvfile would be a single FQid and the data would be glued in a nonsensical way. VBGLU now refuses to glue multi FQid data. Note of this is made in the help file. Moved nowhere. 12282. March 13, 2007 SETFC Eric changed: (1) Formats to be clearer and handle VLBI better (2) Add message when user radius overrides the size (3) SETCEL error did not place return values for AIPS, leaving it to get garbage instead. (4) Changed BCHAN to 1 + 0.4*Nchan; ECHAN=BCHAN. VLBI data frequently have channel 1 flagged. (5) When the user's radius limits the area, the CELLSIZE was set and then only the IMSIZE rounded to a power 2. Changed to adjust the CELLSIZE after the rounding some although not as drastically as when the two are set by maximum phase error. (6) Changed the round up levels substantially: if J > 1.05N recommend 2N where N is a power of 2 and J is the needed size w/o considering powers of 2. Previously it did not round up until rather larger values of J. Moved nowhere. 12283. March 14, 2007 SNSMO Amy There was an error introduced last year in the call sequence to the subroutine SMORAT. This caused odd smoothing of the rates for the first IF, including flagging solutions. Fixed. Moved nowhere, went to patch April 26. 12284. March 14, 2007 VBGLU Eric Corrected minor Fortran gotchas in 12281. Added code to copy with reformat of IF number the FG tables from each file. The resulting FG table can be several times larger. Fixed file closing for CT table tests. The LUNs were needed for the FG function after the CT tables were tested. Moved nowhere. 12285. March 16, 2007 BLCAL Eric Code has never worked on multi-source data. It required that all source IDs have a non-zero flux in the SU table, not just those included by the adverbs. Moved the divide by zero test to the inner loop where all source selection will have occurred. Moved nowhere. 12286. March 17, 2007 VLA, EVLA, et al. Eric Changed help files for VLA, EVLA, VLBA, and HSA to use adverb SUBARRAY rather than INVERS. They are equivalent in this usage but SUBARRAY is what the user should be thinking. Changed AU3B to use adverb SUBARRAY and to set <= 0 to 1. Changed adverb help files for ANTENNAS and BASELINE to mention these verbs. Moved nowhere. 12287. March 20, 2007 CUBIT Eric/Judith Irwin New task to fit a full HI cube. CUBIT was originally written and maintained by Judith Irwin at University of Toronto. She has given me permission to modernize the code and release it with the regular AIPS. I have retained much of her code but changed the memory handling to be dynamic so any size/shape cube can be fit. The old-style AIPS code Judith inherited and used has been updated to reduce the prevalance of GO TOs and otherwise make for more modern looking AIPS code. I have also added the needed adverbs to POPSDAT and NEWPARMS and made new help files for FPARM, RPARM, VPARM, DCODE, and VCODE. ERF.FOR was modernized too to be used in printing summaries of the results. Moved nowhere. 12288. March 20, 2007 VLBAUTIL (VLBAEOPS, VLBATECR) Amy Added two more possible download command to VLBAEOPS (curl and ncftpget), the same as VLBATECR. Added a '-r' to the wget commands in VLBAEOPS and VLBATECR, the other commands automatically overwrite. Moved nowhere. 12289. March 21, 2007 UVFIX Leonia UVFIX has not calculated REFCHN for uncompressed data. It has provided mistake at the frequency calculation when source shift was involved. Moved nowhere, went to patch April 26. 12290. March 21, 2007 CUBIT Eric Correct two minor Fortran errors and add Judith Irwin's reference to both the Fortran and the help file. Moved nowhere. 12291. March 21, 2007 CLCOR Eric The source position correction updates the SU table first so that the source positions taken as the "input" positions were really the "output" positions. This makes a second-order error in the correction. the SOUMOD routine had a 3rd order error in that it used the output apparent declination to scale the DX term when it should be the input apparent declination. Moved nowhere, went to patch April 26. 12292. March 21, 2007 SMOOTH Eric This adverb was left out of 17 tasks that do calibration. They are CLIPM, FLGIT, SPCAL, UVFND, UVMLN, FINDR, FLAGR, MAPBM, RFI, BLCAL, DEFLG, LPCAL, PCAL, OMFIT, SCIMG, SCMAP, FRMAP. Moved nowhere. 12293. March 22, 2007 VLANT Eric The task worked but only when both polarizations were present. Otherwise it did odd things - fixed it to match the polarization handling in CLCOR. Moved nowhere. 12294. March 23, 2007 VLBAEOPS, VLBATECR Amy Remove "-r" from wget lines. Caused file to be put in non- existant directory. Don't know why. Moved nowhere. 12295. March 23, 2007 VLANT Eric Removed the -r from wget command - when use -O file-name the -r is not needed and cariies some other meanings. Added a subroutine to get the start time from the NX table so that the observation will be set in time more exactly. Changed the handling of the stokes parameter to a view of such things that is more modern than what is hidden away in CLCOR. There is only 1 set of cal columns with 1 polarization no matter what that polarization is (R or L). For 1 day it got the single L case wrong, if that ever occurs with the VLA. Position corrections (large) have been entered for EVLA antennas in the data file which will be fetched automatically when VLANT is run. Moved nowhere. 12296. March 28, 2007 UVAVG.HLP Leonia A paragraph about multisource file has been added. Moved nowhere. 12297. March 29, 2007 UVDI1 Leonia UVDI1 has been modified to make it working with both single and multi source file for the first UV data input. The help file is updated. Moved nowhere. 12298. March 29, 2007 Misc Eric ACFIT: The offset term was printed with a format F9.5. Changed it to have several more digits. CUBIT: Incorrect subscripts in the explain file Moved nowhere. 12299. March 30, 2007 FLATN Eric Added OPTYPE to allow images of noise or weight to be computed instead of real data. Discovered error - the NOISE values were ignored in mosaicing so all pointings had same NOISE effectively. Renamed DOWEIGHT since a value of 1 is catastrophic in mosaicing (1.0e-4 is better) - will use WEIGHTIT instead which defaults to 0. Changed help to emphasize - warn about - this parameter. Changed STESS.HLP to warn about the geometry needs and to say that FLATN can now do STESS's job. Moved nowhere. 12300. March 30, 2007 PRTUV Leonia Add option of print out RE, IM instead of AMP, PHASE under control of DPARM(7). Moved nowhere. 12301. April 2, 2007 RMSD.FOR Leonia RMSD should ignore the blanked pixels but the logic error has leaded to substituting them by zeros. Fixed. Moved nowhere. 12302. April 2, 2007 RLDIF Eric The averaging of each scan was done carefully in real and imaginary parts. However, the averaging between scans was done on the phase values directly. This is not a good idea with noisy data or phases close to 180. Changed it to do real and imaginary correctly. Moved nowhere. 12303. April 2, 2007 xTESS Eric Changed VTESS, UTESS, LTESS, STESS to test whether the incoming images for a mosaic are actually on the same geometry as these tasks require them to be. Changed STESS to correct the call sequence to PBCALC that omitted the array name and so was all wrong (and probably the cause of the aborts). Moved nowhere. 12304. April 2, 2007 VLARUN Lorant Changed adverbs to have some more descriptive meaning in VLARUN.001 and VLARUN.HLP Latter was rewritten by Amy. Moved nowhere. 12305. April 3, 2007 TABED Eric Users never notice OUTDISK so blank name parameters other than OUTDISK sent the output to the highest catalog number file on OUTDISK not where expected. Changed the defaults to individually match the input, reducing the likelihood that a false match will occur. Moved nowhere. 12306. April 5, 2007 FILLM Eric Changed it to have 10 frequency bands, adding S and A, with contiguous frequencies so that no possible observing frequency would be assigned band = '?'. Changed the frequency divisions to be those of the official EVLA list. Moved nowhere. 12307. April 5, 2007 PBCALC Eric The message about using VLA for array XXXX should only be issued if it is defaulting. Moved how that message is issued to a more appropriate place. Moved nowhere. 12308. April 5, 2007 FITLD Eric Change to support IDI bandpass tables: TABHDR so look for BANDPASS TABLE and call it a BT (was just BANDPASS and BA which was no where supported) BTINI to use # polarizations in managing the number of columns for file creation and warning messages FITLD to drop warning in BT2BP and to set BPIF so that there would be data. Moved nowhere. 12309. April 6, 2007 FITLD Eric Changed TABHDR to allow either BANDPASS or 'BANDPASS TABLE' for a BT table. Fixed BTTAB which was off by one in the 1st polarization in copying the BP and very far off (wrong pointer on write, correct pointer on read but both off by 1 BP size). Moved nowhere. 12310. April 9, 2007 FLATN Eric Upgraded it to fill in NOISE from the image header parameter ACTNOISE when it is present. Explained this in the help file. Moved nowhere. 12311. April 9, 2007 FILLM Eric Change OPACTY in $APLSUB to recognize the new band codes. Change INDXR to use the same band edges as FILLM so that it calls OPACTY with the new band codes. Change WETHR to use the new band codes for some of its operations. Moved nowhere. 12312. April 11, 2007 MULTI Eric Fixed the call sequence to JPRECS - it did not do the general relativity correction which we normally do and it used a large rather than tiny interpolation parameter. Moved nowhere. 12313. April 11, 2007 FILLM Eric Changed help and fortran to use the back-end Tsys for EVLA antennas no matter what the band and the setting of the CPARM(2) bit might be. There is no front-end Tsys at this time for EVLA antennas. Changed it to print only so many bad exponent messages and then print a summary at the end. Moved nowhere. 12314. April 12, 2007 EDITR, EDITA, SNEDT Eric Changed QEDIUTIL to test the error return from reading user-provided numbers more carefully. The IERR = -n would fall through to a place that treated all non zwero errors as fatal. Added error messages and resetting of IERR to 0 when the user enters numbers badly (or out of range). Moved nowhere. 12315. April 13, 2007 Debris Eric Problems discovered and repaired while considering extending the calibration system to more tasks, incl non T* sorts: CLINV Help file referred to input and output "image" when it works only on UV data sets. DBCON Help file had incomplete sentence which failed to tell reader what happens to CL > 1 from file 2. DIFUV Help file was silent about tables. Changed to state that no tables are applied and only those of the first data set are copied to the output. DFQID Did not renumber FQIDs in tables. Stole routine from DSORC with appropriate revisions. DQUAL Did not renumber sources in tables. Fixed this for all but the SU which was/is written separately. FXTIM Fix the TIME and TIME RANGE in all tables too. UVLSD Do not copy cal tables (CL, SN, BP, BL) since a division produces stuff that should not be calibrated. IBLED Fixed comment that wrongly referenced NXTFLG. CLPLT Removed unused references to DFLG.INC ANBPL Removed unused references to DFLG.INC VPLOT Removed 1 unused reference to DFLG.INC, others are used. Moved nowhere. 12316. April 13, 2007 CALWR, FITTP, FITAB Eric These should not write the BLANK keyword when writing floating point FITS data. Changed it to comment about NaNs. Moved nowhere. 12317. April 20, 2007 AVSPC.HLP Leonia The AVSPC.HLP has not explained clearly what should happen when ICHANSEL describes averaging explicitly for some IFs but skips other IFs. The paragraph describing this situation is added. The start channel of the 75% of the band has been erroniously determined as NCHAN/8. It should be NCHAN/8 + 1. It has been changed. Moved nowhere. 12318. April 20, 2007 VLBA routines Lorant Found a bug in VLBAFQS when S/X or 90/50 mode and frequency-ID is not the row number. Corrected part in VLBAARCH, but Amy has VLBAUTIL checked out so VLBAFQS and VLBAFIX will be done later. Moved nowhere. 12319. April 22, 2007 UVFLG Leonia The bug is found at the flagging of time interval near zero fringe rate (APARM(6) > 0). UVFLG has considered the fringe rate (as U) zeros repeat with period 24 hours. Actually this period is 12 hours. This mistake leads to the possible skip of some flagging due to fringe rate close to zero (APARM(6)>0). Fixed. Moved nowhere. 12320. April 23, 2007 VLBATECR, EOPS, FIX and FQS Amy Changed VLBATECR and VLBAEOPS so when wget is used it will overwrite files, rather than rename the input file. Also simplified logic a bit. Fixed VLBAFQS and VLBAFIX as described by Lorant in change #12318. Moved nowhere. 12321. April 26, 2007 BLCAL, UVFND Eric The cal system has an integration time parameter DXTIME which is the integration time in DAYS. If it is not set, then the random parameter is consulted or 1 sec taken. It controls the gain correction made for averaging when there is a delay and rate. A 1 day integration makes for a large correction! Two tasks, BLCAL and UVFND, set this parameter to 1.0. CALIB set it to TINTG which was an unused variable carried in a local common. Changed all to set it 0.0. Changed BLCAL to clean up code a bit and to use dynamic memory rather than a fixed and rather modest set of buffers. Moved DXTIME correction to patch this date. 12322. April 26, 2007 UVCON Leonia UVCON used the wrong sign at the telescope longtitude when elevation of the telescope is calculated. As the result wrong time interval was skipped when limit of elevation is used. Moved nowhere. 12323. April 27, 2007 UVPLT Eric When UVPLT does hour angle, azimuth, and elevation without a REFANT it averages the relevant angle as measured at each telescope. For VLBI, these can be quite different and a vector average is required. Corrected code to do that. Moved nowhere. 12324. April 27, 2007 Source models Eric Changed BLCAL to be like CALIB in handling source models which might have IF-dependent fluxes in a source table. Previously, BLCAL scaled all IFs as if they had the source flux of IF 1. Found minor fixes to put in CALIB - wrong CC ver saved and IF-dependent code entered when using an image model which is not useful. Lots of other tasks should have this done. Moved nowhere. 12325. April 27, 2007 PCNTR, KNTR Eric Changed error message in POLDRW so that coordinate errors are separate from plot errors with the former suggesting that FACTOR is too large. KNTR was very short on error reporting. Moved nowhere. 12326. April 28, 2007 Model fluxes Eric More changes to handle different fluxes at different IFs when using a model of a calibrator: PCAL is okay - does one IF at a time usually LPCAL Corrected to pass through IFNO rather than using 1 always in FACSET. Is usually 1 IF at a time. FRING Added code to do 1 If at a time when appropriate. KRING Added code to do 1 If at a time when appropriate. BPASS Added code to do 1 If at a time when appropriate. CPASS Added code to do 1 If at a time when appropriate. CALIB Got rid of useless comments and variable BLCAL Got rid of useless comments Moved nowhere. 12327. April 30, 2007 UVCON Leonia I have found some bugs near creation of AN table specifically when input configuration is VLA or VLBA. Fixed. Moved nowhere. 12328. May 1, 2007 Model CC version Eric Corrected BLCAL, BPASS, CPASS, FRING, KRING, and LPCAL. All of them defaulted the CC file version to 1 rather than the highest. Other tasks like CALIB, UVSUB, etc got it right. Moved nowhere. 12329. May 1, 2007 LISTR, RLDIF Eric LISTR and RLDIF did not average spectral channels when doing MATX displays - they just used BCHAN. LISTR also set EIF=BIF in doing MATX. Changed LISTR to loop over IF at the outside (main routine) rather than somewhere inside. Fixed a bad computation of line length in the DEC-10 format GAIN display. Made data buffer the correct size (3 * 4 * MAXCIF). Moved nowhere. 12330. May 3, 2007 Missing antenna # Eric If an antenna number is missing from an AN file, some bad things happen. Fixed: UVFLG The translation of station name to number used min match so a 0-length string matched all. Station names should be an exact match. DTSUM It greatly confused antenna number and count. Rewrote handling of antenna number so that true antenna number if recognized. Reads all data once through to get which antennas actually appear in the data somewhere. Those are then printed. It can optionally print the total summary, optionally followed by the scan-by-scan summary. Added controls of this printing and of the number of digits used. Changed scaling to avoid overflows to be a simply right shift of the counts (with rounding). help file updated for options. Moved nowhere. 12331. May 4, 2007 Man pages Eric The man pages for aips got lost. The Mac systems did not allow us to have both aips.l and AIPS.l in the same directory, so I deleted AIPS.l. Unfortunately AIPS.l had all the text and aips.l was just a link to AIPS.L. Fortunately, we have an antique version of aips still on disk and I was able to recover to file, now stored as aips.l. man AIPS will not work but man aips does. Moved to patch this date. 12332. May 4, 2007 VLANT Eric Added a section to try "curl" should wget fail. curl is usually on Macs while wget is not. Moved nowhere. 12333. May 4, 2007 VLBAPCOR Amy Changed CLCAL interpolation to SELF. Moved nowhere. 12334. May 9, 2007 VLAPROCS, VLACALIB Amy The new calibrator model downloading that I added to VLACALIB last year did not work for 1950 data. Fixed. Moved nowhere. 12335. May 10, 2007 AMBPL Leonia AMBPL has plotted wrong elevations when elevation is plotted. It happenned because the subroutine GETSOU is never called. The condition prevented it. The condition is wrong but I am not sure how to modify it. So I removed the condition at all until Eric corrects me. MERLIN is treated by default at subroutine GETANT as VLBI (LH with origin at the Earth center) And Merlin actually follows such a coordinate system. Moved nowhere. 12336. May 14, 2007 CLCOR Leonia Some VLBI data correlated not at Socorro have the source RA's at the range of -180, +180 degrees while Socorro correlator creates SU table with RA ar the range 0,360 degrees. CLCOR appliing correction to the source position ('antp', antc') reading negative RA (RAEPO for example) recalculates it to the positive RAAPP (add 360) and comparing it with the negative SU table RAAPP reports the correction of the order 360 deg! I fixed it adjusting RA output of the SU table to the range 0-360 always. Moved nowhere. 12337. May 21, 2007 GRIPE Eric Changed ZGMAIL.FOR to refer to daip@nrao.edu not egreisen. Changed ZGMAI2.C to drop an invalid mail address, leaving only daip@nrao.edu. Moved nowhere. 12338. May 22, 2007 ANBPL Eric The initial value for SCANUM of -1 matched the value of the NX table record when there is no NX table. So the header was never used to get source parameters and elevation failed. Adjusted the initial value of the "scan number" and removed the commenting out of the IF (new scan) around the SOUINI calls. Moved nowhere. 12339. May 22, 2007 TASKWT Eric AIPS calls TASKWT while it is waiting for a task to restart it. Changed this task to be tolerant of operating system glitches that tell it that the TD file does not exist. It must or this would never get called so the OS must lie 5 times in a row before we quit in future. This will not prevent other file open failures in bad operating systems but reduces one vulnerability. Moved nowhere. 12340. May 23, 2007 TV servers Eric Changed XASERVERS and UNIXSERVERS to have a case on $ARCH as well as on $SHELL for setting the library path. Macs require DYLD_LIBRARY_PATH rather than LD_LIBRARY_PATH which is all we had there before. Moved nowhere. 12341. May 23, 2007 CUBIT Eric Judith Irwin has provided more information for getting started using this fitting task. Moved nowhere. 12342. May 24, 2007 FILLM Eric There was an extra call to MRECO (read routine) when an EOF occurred that caused it to skip the first data record in the next file. This led to real problems when the next file was incompatible with the previous, but also caused it to lose a record when the two were compatible. Moved this fix (not other FILLM changes) to patch this date. 12343. May 24, 2007 CTINI Eric For some reason ICHAR rather than ITRIM was called to determine the length of a string. Moved nowhere. 12344. May 24, 2007 LISTR Eric The outer loop over IF did not exit on a user's quit command but attempted to do the next IF. Since the files were not closed properly on a quit, this then generated error messages. Fixed it to pass back the quit and honor it. Moved nowhere. 12345. May 25, 2007 Cal system Eric In order to allow flagging and some calibration on data sets that are not in time order: DFLG.INC Add variable for start time of interval (FLGTST) and logical TIMORD true if in time order. PFLG.INC Raised MAXFLG to 6000. UVGET Set TIMORD, test for time order ignoring DOPOL and DOBAND=1 and remove DOFLAG test. (Flagging non time ordered will fail if too many flags for the buffer are found.) POLSET Changed handling of no pol info to be clearer. Pol type is normally null rather than blank when it has never been set so added messages. DATPOL Changed it to allow time to move in any direction and to recompute if it differs by 0.1 sec. Before it had to move forward and any change led to a recomputation. DATGET Changed the test that forces time order to ignore DOPOL and DOBAND=1. TIMORD should be true for DOCAL etc., but matters if DOFLAG assumes it. DOFLAG was not tested previously. Added includes PFLG and DFLG. DATFLG When not TIMORD, test if time is in flag's interval. NXTFLG Bail out of read only if TIMORD. Set and maintain FLGTST, the start time. Add to message when buffer overflows. SDGET Set TIMORD, test only DOCAL to force time order. DGETSD It used to force time order in all cases. Now it forces time order if TIMORD and DOCAL or DOFLAG. Moved nowhere. 12346. May 25, 2007 Tasks and cal system Eric Changed UVCOP Made same changes to local versions of DATFLG and NXTFLG. Changed ..IN routine to set TIMORD. NOTE: The task did not check on time order previously and so could have failed badly. Changed number flags to 60000.o AVSPC Similarly, changed local versions of DATFLG and NXTFLG and SETFLG. It had the same bug as UVCOP. CVEL Did not allow flagging for single-source files. Changed local versions of flagging routines and INI to do non time ordered flagging. SDVEL Changed local versions of flagging routines and INI to do non time ordered flagging. FLAGVER Updated the help file with info about time order and limits and UVCOP. Moved nowhere. 12347. May 25, 2007 UVPLT Eric Changed the display when elevation et al are plotted to show the REFANT or lack thereof in a line at the bottom. UVPLT replaced the normal antennas, IFs, channels display with REFANT display previously and turned of DO3COL. This is not right since one might display e.g. amplitude versus elevation rather than elevation versus time. Moved nowhere. 12348. May 30, 2007 TASKWT Eric I forgot to reset the message suppression in this subroutine so AIPS would begin to hide errors after the first task was run. Moved nowhere. 12349. May 30, 2007 SNPLT Eric The Hour angle computation had a mysterious -PI/2 causing all hour angles plotted to be too low by 6 hours. Moved to patch June 14, nowhere else. 12350. May 31, 2007 QUACK Eric The EVLA currently needs to QUACK differently for different IFs and polarizations. So added STOKES, BIF, EIF to it. Moved nowhere. 12351. May 31, 2007 UVPLT, WIPER Eric Fixed WIPER to average hour angles, elevations, and azimuths correctly using sines and cosines like UVPLT. Then fixed both to do the same with paralactic angles. Moved nowhere. 12352. June 6, 2007 UVAVG.HLP Leonia Description of OPTYPE='SUBT' has not been clear. Tryed to improve. Moved nowhere. 12353. June 8, 2007 install.pl Eric The script asked for the desired group id and then never used it. Added system ("chgrp $GROUP file") type commands everywhere a chmod was done. moved nowhere. 12354. June 8, 2007 FILLM Eric FILLM required that the stop time exceed the start time in TIMERANG or it used -99 to +999 days. The defaulting should be by halves in TIMERANG according to the (TIMERANG) help file so changed the code. Added warning message when non-zero TIMERANG is overriden because the user spcified it wrong. Moved nowhere. 12355. June 8, 2007 FILAIP Eric The max number of ME files allowed was limited to 15 not 35. We now use extended hex so raised the limit. Moved nowhere. 12356. June 8, 2007 ALLDEST Eric One question used an I1 format for disk number - I3 is probably better. Changed several formats in AU3A. Moved nowhere. 12357. June 10, 2007 SPLIT, SPLAT Eric Errors occurred when TIMERANG limited the times and SOURCES included more sources than were in the TIMERANGe. Changed: INDXIN Ignored TEND and so selected many NX records that were infact not useful. Also used the NX time interval when it should have used a half interval. SPLIT If UVGET ('INIT' returned an error code, DOCAL et al were left in the temporary (false) state for fetching the new source. SPLAT Fixed up the logic to match SPLIT in setting and resetting cal parameters and in keeping track of whether the UV data set is open. Added code to deal with missing data in subarrays which has been in SPLIT for a long time now. Moved SPLIT fix simply to patch, nowhere else. 12358. June 10, 2007 Linear polarizations Eric There were various assumptions about linear polarizations in various places. The cal system seemed to be willing to accept tham some places with the assumption that DOPOL would translate the linears to circular and that DOPOL would always be set. FNDPOL basically did not allow you to have linear polarization however. This all seems odd to me. FNDPOL Re-wrote an incompetent mess. It should select polarizations more competently now although why DGINIT and thence FNDPOL have so many limitations on the allowd Stokes axis values I do not know. DGINIT Re-arranged to allow X and Y type polarizations. DGHEAD Modified to support X and Y types. SDGET Fixed nasty error where the line LRECIN=LREC was lost accidentally. PRTSD Changed to display the X type polarizations and other longitude and latitude axes. SDIMG Allow XX and YY polarizations in Fortran and help. SDGRD Fortran okay - changed help. Moved nowhere. 12359. June 11, 2007 Cal adverbs added Eric Change tasks to allow calibration and flagging where not previously expected: COPTAB New subroutine to do copying of tables for such tasks including channel, IF, source, et al selection. CALHIS New subroutine to fill HI file with calibration adverbs. PCINI Limit warning messages allowing for # tones = 0. UVLSF Moved from APLPGM to APGNOT and added cal adverbs. Continuum output always uncompressed; line output matches input. Continuum output no longer the weight of the input CHANNEL but rather the sum of the weights of the channels used to find the fit. UVLSD As UVLSF. FUDGE Moved from APLPGM to APGNOT and added cal adverbs. Moved nowhere. 12360. June 12, 2007 FILLM Eric Fixed the TELL code defaults for TIMERANG to match those used on the original inputs. Moved nowhere. 12361. June 13, 2007 SAD Eric Extended the DPARM(9) option to use a noise image. If >= 2, the image will now be required to match the the first image and will be read as islands are searched. In this case the CPARM cutoffs plus ICAT and some of the DPARMs will be taken as being in units of S/N rather than image units. Corrected error that caused coordinates to be based on the 2nd image rather than the first in the output display. Moved nowhere. 12362. June 14, 2007 VLARUN Lorant Amy noticed a feature when using calibrator models with more than eight characters (ie non-observe names, eg fast-switching or sched used). Fixed to allow for 16 characters in the name. Moved nowhere. 12363. June 14, 2007 VLAPROCS, VLACALIB Amy As above. The real issue was trying to compare calibrator names that were longer than 8 characters with the standard flux calibrators. Anyway, fixes so up to 16 characters are allowed. Moved nowhere. 12364. June 14, 2007 SNPLT Eric The inner plotting routine changed the phas code for PC tables to a special value and then did not test that special value when checking for phase wraps. So lots of points potentially did not get plotted. Moved to patch this date. 12365. June 16, 2007 VLA station names Eric There has been a protest over the use of a blank in the VLA station names so I have changed it to an _. Let us hope that this does not mess up labeling in AIPS where _ is changed to an unshrinkable blank on occasion in string handling for display. Changed FILLM to do the _ and changed PRTAN, VLANT, LOCIT, FRPLT, and POSSM to support all possible VLA station naming conventions. FRPLT especially but also POSSM did not handle the previous conventions entirely correctly. FILLM did not test the possibilities of names properly a couple of other places. Moved to patch (except VLANT) this date. 12366. June 19, 2007 Stokes values Eric We need to allow a few more Stokes values in calibration tasks, e.g. RRLL, XXYY, RLLR, XYYX where the first 2 are synonyms for HALF but the other two were not allowed. Furthermore, formal I should be allowed to be requested. So Stokes values F, FQU, FQUV, and FV have been added although the I versions of the last 3 are already formal I (both RR and LL or XX and YY must be present). DGINIT Implement the new set of 21 PMODEs including formal I and careful interpretation of the words HALF, CROS, and FULL. DGHEAD Adjust the interpretation of PMODE adding the new RLLR and XYYX. FNDPOL Add the new codes. STOKES Upgraded the help file. Moved nowhere. 12367. June 19, 2007 UVPLT Eric Remove all restrictions on STOKES - anything that FNDPOL and DGINIT (the calibration software system) will take will be plotted. Moved nowhere. 12368. June 19, 2007 WIPER Eric Similarly, removed all restrictions on STOKES. The flagging follows sensible rules as to which correlators get flagged when various displayed correlators are marked for flagging. Added code to keep track of which baselines are in which plotted points. Up to 2 will be plotted during interactive flagging operations with a "+" sign if there are > 2 baselines within the point. (Isn't dynamic memory convenient?) Changed the axis labeling so that sub-windows will also be labeled. Changed the format handling in the display of coordinate values during flagging. Moved nowhere. 12369. June 20, 2007 TVFLG, SPFLG Eric These 2 tasks had a strange set of rules for handling FLAGVER including permanently changing the input file if it was a single source file with no pre=existing FG table and FLAGVER was 0. Changed them to copy the input FG table (if any used) to a new table and then append the new flags. Moved nowhere. 12370. June 20, 2007 UVMTH Eric This old and unknown task has been upgraded to average the first input data set over all times after application (at the users' discretion) of all calibration/flagging options. It then subtracts, adds, multiplies, or divides the average (on a correlator by correlator, baseline by baseline basis) to the second data set. That data set remains untouched by any calibration and flagging. This should allow the removal of additive correlator bias (if it can be measured) which has been seen on one VLA bandwidth (25 MHz) and also at the GMRT. Dynamic memory is used for the averaging buffers in UVMTH. A typo in FUDGE was also corrected. Moved nowhere. 12371. June 20, 2007 VLBAUTIL, VLBAFIX Amy Changed VLBAFIX so it sorts after it splits the data into different frequencies. The former case (where it sorted first) caused problems when multipass data was loaded into the same file. Moved nowhere. 12372. June 21, 2007 TVFLG, SPFLG Eric When clipping other data with the pattern of 1 clip command, an absence of data in a particular channel or IF led to the FC file remaining open and a serious error on the next attempt to open it. Rearranged the code to avoid this problem. TABINI did not respond to an error from ISTAB suitably and tried to create a table on LUN IN USE and open it a second time. Changed TABINI to honor an I/O error in ISTAB and quit right there (it would have on READ but did not on WRIT). Moved nowhere. 12373. June 22, 2007 POSSM, FRPLT Eric There were typos in the display string handling due to the VLA station name change. The string length was too short to handle non VLA names as well. POSSM and FRPLT did their own interpreting of STOKES as well as FNDPOL which led to conflicts and failure. Moved the display fix to patch this date. 12374. June 22, 2007 FILLM Eric Changed the output axis order to that used most of the time in AIPS. The natural order to a telescope is complex, frequency, then either Stokes or IF. FILLM output a variety of orders depending on the observing mode. Changed it to reorder the data just before writing it out and to have headers in the aips tsandard order complex, Stokes, frequency, IF. Found a bug in the Z-mode data handling. The main consequence of the error was to mess up the weights of channel 1. I don't think this mode has ever been used except for a test run to debug FILLM. Moved nowhere. 12375. June 26, 2007 FILLM again Eric Changed: 1. DOALL adverb was removed. It was confusing and overlooked and all adverbs have values which produce the same result. 2. Found that the code which tried to correct for Q-band plus Pie Town requiring a new VLA "center" to bring fringe rates into a manageable range, failed. The frequency is not known at the point is was used. So changed code to look for antennas found or not and try the center shift if none found. This also covers the fact that the center shift came after the first observations at Q with Pie Town. 3. Found that there really was not much code checking for moved antennas. Now the antenna coordinates are read from all possible concatenation files and checked before doing a concatenation. Also the current antenna coordinates from the latst written AN file are kept to compare to the current coordinates. Any change over 0.01 ns forces a new file. I had to add a new BAIL variable to manage to get this information high eneough to get the desired actions taken. I bet I use that variable some more later.... 4. Corrected the help file to show all recognized bands and their new definitions. Moved nowhere. 12376. June 26, 2007 VLANT, VLBAUTIL Eric Discovered that an option advertised in man curl is not present in modern versions of curl (see curl --help). This caused it to fail when wget had already failed. Also removed --connect-time 120 from $RUNSYS VLAPROCS and VLBAUTIL. Moved nowhere. 12377. June 27, 2007 FILLM Eric 1. Put a new bit in CPARM(2) = 2048 to request that the data be written as correlation coefficients rather than being scaled by the nominal sensitivities. This means no scaling in the new data form from the ModComp-less system (IDCAF) or a division in data from the ModComps. Note that this division is the same as the old MCSOLR subroutine which is now replaced by the MCSCAL that takes DIVIDE as an argument. 2. Put in a test for a change in program code when all codes are acceptable. BAIL gets set and a message issued. In general, FILLM will continue to write to existing files if possible despite this. Moved nowhere. 12378. June 27, 2007 UVLOD, FITLD Eric The variables BSC and BZE were not initialized. If the incoming UV data FITS file did not have a BSCALE keyword, then the output data were all zero. Added defaults. Moved nowhere. 12379. June 28, 2007 PRTUV Eric Discovered the the display of real and imaginary did not take into account the widths of the formats being used and so printed lines wider than the window. The column labels did not align at all. So, fixed the typing which had wandered from easy legibility, the testing of widths and setting of format types, the actual formats (im did not change when re did), and redid the column labeling for both Amp/phas and Re/Im. Discovered that the printing of random parameters has been broken for some time. It ignored the first 5 random parameters and then would only display something at all if there were 1, 2, or 3 floating-point random parameters after that. Changed it to display all random parameters but time and baseline, which are also properly formatted in the display anyway. Adjusted the formatting so that it can respond to the width of the screen. Moved nowhere. 12380. June 30, 2007 AIPSLetter Eric Put back LET07A.TEX and .PS. MOved nowhere. 12381. July 1, 2007 DBCON Eric The Stokes conversion which can be done to make an RR/LL file agree with a I,Q,U,V file had a serious error in handling compressed data, writing out bad things. It also did a weighted average of RR and LL, which is not done elsewhere in AIPS. Corrected the addressing of compressed data, chhanged the I computation, and changed the I and V weights to be the ones one should use in the unweighted case. Moved nowhere. 12382. July 3, 2007 Adding cal adverbs Eric Corrected a format in HI routines and removed redundant task message already done by HISCOP in UVMTH, UVLF, UVLSD, FUDGE. Added adverb DOACOR to all these tasks. UVMTH had an error in which it assumed LREC and NRPARM were the same in both data sets. This caused it to walk through the second data set at incorrect intervals and so to fail to recognize baselines. Moved nowhere. 12383. July 3, 2007 UVAVG Eric Added calibration and flagging and removed the SUBT option (it worked for normal inputs, but had several gotchas for other relatively normal possible input data sets). Made the averaging buffer dynamic as well. Moved nowhere. 12384. July 5, 2007 NX tables Eric A lot of these UV copying tasks lose the NX table. Created DRNX.INC Include with tables for re-indexing a file RNXINI Initialize re-indexing including making NX table RNXUPD Update tables, writing NX as needed RNXCLS Update NX with any remnants in tables, close NX. The above retain the error code if they start not to work as well as an internal flag to tell them not to do anything more. RNXUPD has a useful bit of code to determine the proper gap which would indicate a scan break. COPTAB Drop copying of NX table entirely. Trick from UVCOP added - expand time range by 15 minutes each way to make sure all relevant records are copied in tables. FUDGE Add calls to RNX routines UVMTH Add calls to RNX routines with another UVPGET needed to point to input data at RNX time. Do not copy NX table. UVAVG Calls to RNX routines, RNXRET in common so that the error if any may be reported further at the end. UVLSF Add calls to RNX routines. UVLSD Add calls to RNX routines. FLGIT Add calls to RNX routines. UVCOP Needs its own COPTAB since it renumbers subarrays and carries the time range in different variables. Add calls to RNX routines and drop any copying of NX table. Moved nowhere. 12385. July 5, 2007 FILLM Eric Commented out two sections of code that did nothing useful, but appeared to care whether the time was the time at the center or the end of the integration. Changed the time MCIATI to be the mid-point time by subtracting 1/2 of the integration time. Added message about that to FILLM and FILLM.HLP and UVFIX.HLP. Corrected error welcoming user to on-line FILLM when reading an old local disk file. Moved nowhere. 12386. July 6, 2007 POPS numbers Eric Changed: PRTMSG Automatically deleted all messages for POPS numbers greater than 15. Changed the 15 to 35, the not so new limit. AU1A Added PRNUMBER, NUMTELL, and BATQUE to the list of adverbs that have host-dependent limits. Also GPOS and FPOS. AU2 As AU1A. AIPSC As AU1A, drop CCBOX which was erroneous. changed it to ignore the fact that it is running AP-using tasks. I kept the code since it is so delicious, but 2001 AP tasks must now be run to invoke it. Edited help files: ZZBATCH, BATCHJOB, BATCH, BATQUE, JOBNUM, BATLIST, BATEDIT, SUBMIT, JOBLIST, and UNQUE to bring them up to date. Moved nowhere. 12387. July 5-6, 2007 FILAIP Eric It made the TD data file for only 15 POPS numbers - should be 35 now. A test on max POPS number also was fixed. FILAI2 needed these changes and a MIN test also. Moved nowhere. 12388. July 6, 2007 CCRES Eric The defaulting for the 2nd name parameters was not present - added 32 lines of code (OOP is wordy) to make sure that all blanks and 0s are replaced with the equivalent 1st name parameter. Moved nowhere. 12389. July 8, 2007 install.pl Eric The print area used by batch jobs was not created by install.pl. Added it and commented out all of the stuff having to do with the pseudo-AP size. Dynamic memory means never having to say you're sorry. Tar ball size is a lot bigger than 45 Mbytes - like 89.6 at the moment. Moved nowhere. 12390. July 9, 2007 WIPER Eric Fixed IAXIS1 to guard against changing the time in hours axis to wrong units (milli, ...) in LABINI. Moved nowhere. 12391. July 9, 2007 FILLM Eric The help file was upgraded to tell the truth about CPARM(2)=1. That ONLY masks the Tsys fluctuating error condition, not any of the others. To ignore others, e.g. source change, subflrctor position, off source, modules not locked, filters misset, total power out of range, LO not locked, flagged by operator, one must set CPARM(3) somewhat blindly. Moved nowhere. 12392. July 10, 2007 UVFLG Leonia/Eric UVFLG has never worked with elevation flaging. The relevant sections were written very unclear and too much complicated. So the problem is fixed removing several subroutine and writing the new one. Moved nowhere. 12393. July 11, 2007 CXCLN Eric Fixed: QCVMMA It did not adjust an AP address by PSAPOF and so saved a longint in an int and left an address that would shortly be biased a second time and so be illegal. Moved nowhere. 12394. July 11, 2007 FUDGE Eric Added STOKES adverb - this is now a pretty powerful UVCOP task. Moved nowhere. 12395. July 10, 2007 Break CLIPM Eric Added STOKES adverb. Messed with cross-bar logic a lot. Dropped display when nothing in a correlator is flagged. Moved nowhere. 12396. July 12, 2007 CLIPM Eric Changed CLIPM to handle single-source files too. Added PHSLIMIT to clip on phases. Changed History to use CALHIS. Moved nowhere 12397. July 12, 2007 NXTFLG Eric Did not return an error code when could not handle number of flags => infinite number messages. Changed to cause death of task when too many flags. Also fixed UVCOP and AVSPC. Moved nowhere. 12398. July 13, 2007 DIFRL, FUDGE Eric Changed DIFRL to use calibration and flagging and data selection adverbs. While doing so, spotted a problem with the history copy call in FUDGE and also changed the first scratch buffer to be a small (256) one. Moved nowhere. 12399. July 16, 2007 UVFLG Leonia UVFLG has not worked if subarray=0 at the case of elevation flag. Put subarray=1 indepedent on the subarray input, if elevation flags. This is a temporal treatment while UVFLG serves only one subarray for elevation flag. Moved nowhere. 12400. July 30, 2007 Output FG version Eric The changes to TVFLG to make it more like all the other flagging tasks have caused a variety of complaints. To add to the din, therefore, I have created a new adverb OUTFGVER, changing POPSDAT.HLP and NEWPARMS.001 to create it. The adverb will specify the output FG version number. If it is <= 0 or > FGmax, then a new file version FGmax+1 will be created and the FLAGVER (if any) used on input will be copied to it along with any new flags. If it is a pre-existing table, then FLAGVER is not copied even if FLAGVER does not equal OUTFGVER. Changed rather simply help and fortran of: CLIPM DEFLG FLAGR SPFLG TVFLG For OOP EDIT class a new parameter COP_FGVER is needed to say when to copy an input version to an output version, changing EDFCAP in QEDIUTIL and changing EDITR and EDITA who now also need to know the current maximum FG number. Moved nowhere, more to come. 12401. July 31, 2007 More flagging Eric Changed UVMLN, WETHR, and WIPER to support OUTFGVER in the same manner as CLIPM et al put back yesterday. Moved nowhere. 12402. July 31, 2007 IBLED Eric Changed: 1) Added OUTFGVER and dropped the application of the flags to single-source data files. It will now always write a FG table. 2) Added option PREV BASELINE to code and to HLPIBLED.HLP. 3) Corrected the interactive editing function which never could have worked. There was great confusion in units between "image" and TV pixels. Moved nowhere. 12403. July 31, 2007 SCIMG, SCMAP Eric Changed QEDIUTIL to mark the copy FG version as done so as to avoid copying a second time and repaired an error that put the new out version as a "copy" version rather than an "in" version. Also change the in version to the out version. Changed TVUTIL to be forgiving about the XWINLOAD parameter - the graphics plane to show an inscribed circle. Changed QCLEAN to make sure FINISH is true after the components have been restored. Changed SCMAP and SCIMG to support OUTFGVER and to have a rearranged self-cal menu so I don't kill it so easily. SCIMG needed added DIM setting to avoid a death on the EDIT option. Moved nowhere. 12404. August 1, 2007 AIPS Eric Fixed test in AU2 on size of TG file which now has 4 records per task and it was testing as if 3. Fixed number of adverbs for POSSM in AU2A (verb PLGET). Moved nowhere. 12405. August 1, 2007 CLIP Eric Moved CLIPM to CLIP, replacing the old limited task that everyone uses with the one I wish they would use. Changed CLIPM help file to point users at CLIP. Moved nowhere. 12406. August 2, 2007 CPASS Eric Fixed: 1. Format confusion - two different messages in BPSIN used the same format number. The uncommon one would have aborted. 2. BP file access for an initial BP guess did not close the BP file on error and so left things to die much later. Moved nowhere. 12407. August 3, 2007 FLGIT Eric Added adverb OUTFGVER and code to write an FG table instead of writing out the edited and calibrated vis data. If OUTFGVER<0, the edited/calibrated vis data are still written. Moved nowhere. 12408. August 3, 2007 VLBAPIPE.001 Lorant Changed CLIPM in VLBAPIPE to CLIP (see #12405 on Aug 1st). Moved nowhere 12409. August 7, 2007 PRTAB Eric Added option DOFLAG to request display of flagged as well as good rows. It is useful in debugging if nothing else. Moved nowhere. 12410. August 7, 2007 UVFLG Eric Changed FLAGVER to OUTFGVER and made task always use an FG table. There will be no more flagging of the data directly in place so droped EDROP and BDROP and APARM(1-3), moving APARM(4-7). Changed the elevation code to loop over SUBARRAY, to test FQ correctly, and to support the ANTENNAS adverb. Changed the handling of the history file to report some essential parameters like OUTFGVER, OPCODE, elevation range if used, PC table FACTOR if used, and number flag table lines. Kept the elevation and PC codes from generating some false flag info for the HI file. Worked over the help file extensively, esp for the removal of the in-place flagging. FUDGE will allow in-place application of flagging. Moved nowhere. 12411. August 8, 2007 MAX, MIN Eric Changed KWICK to fix up STACK(SP) so that the equals operator is not confused after doing a MAX or MIN on an array. This has helped understand the parser limits on MAX(array) -- a usage with only one argument. That works with PRINT sometimes, but any other adverb in the line confuses things so that i = max(blc) actually takes the max oi and blc and leaves a result on the stack. Moved nowhere. 12412. August 8, 2007 STALIN Eric Changed help file - the verb is not friendly with defaults, unlike what was assumed in the help. The user needs to give reasonable HISTART and HIEND values. Moved nowhere. 12413. August 13, 2007 REBYTE Eric New service program - translates a full data area from one byte order to the other. It works from Solaris and MacPPC to/from Intel Linux and MacINT taking the format of every file into account. Fortunately most files are tables which are self defining. Copy the binary data to the new computer and then RUN REBYTE which will copy to a new data area for safety. A new Z routine ZBFLC2 was written to flip bytes and words but only if the 4 or 8 byte word is not all character. Also wrote a help file. Moved nowhere. 12414. August 16, 2007 IBLED Eric Changed task and run-time help (HLPIBLED.HLP) to simply swicth IFs if there are ony 2 or switch polarizations if there are only 2 when the select options are chosen. Moved nowhere. 12415. August 16, 2007 LWPLA Eric Found minor error which would arise when skipping 3-color grey scale records because there was no record to init such data. It used the wrong number of pixels (factor 3). Moved nowhere. 12416. August 16, 2007 REBYTE Eric We should be very nervous about floating point numbers matching 4 prinatble characters. So, changed swap of UV data and most image data to not check for character value. Changed magic-blanked image data to check only for magic blank with new C routine ZBFLM2.C. Changed flips of SL and PL to walk through the files using the known format. XChanged flip of table header to use known format rather than ZBFLC2. Changed flip of table data to separate reals and ints from characters thereby avoiding another use of ZBFLC2. We are left with PL file headers (used by PLGET and EXTLIST), TPUT/TGET, and SAVE/GET files for which ZBFLC2 is used. Updated REBYTE.HLP with this warning. Moved nowhere. 12417. August 17, 2007 VLARUN Lorant Changed VLARUN to also accept an underscore in the VLA antenna pad locations when determining the array configuration after an earlier fix by Eric on June 16. Moved nowhere. 12418. August 17, 2007 EDITA, EDITR Eric Changed QEDIUTIL to make NEXT ANTENNA and NEXT BASELINE not just recycle existing antenna numbers but drop the present one and find a new higher one. This will allow one to step through all the data pretty well. Change also the HLPEDI??.HLP files where ?? = CL, SN, TY, TS, UV to reflect this change. Moved nowhere. 12419. August 20, 2007 OOSUB, UVSUB Eric The CHANNEL adverb in these two tasks did not work as advertised and caused OOSUB, which tried some places to implement it, to crash. Changed the advertising for the adverb - it means subtract/divide this channel but copy all channels. Cleaned up the handling of the parameters in UVSUB (one gotcha if the IF axis precedes the FREQ axis) and put similar handling in OOSUB. Fixed OOSUB to handle compressed data (DOUVCOMP required in UVIN) and to copy tables at the exnd (except AN and FQ which are already copied). Moved nowhere. 12420. August 21, 2007 EDIT class Eric Changed QEDIUTIL to handle ENTER IF under crowded allowed by rotating through IF 0, BIF, EIF when there are 2 IFs (EIF=BIF+1). Affects EDITA, EDITR, SNEDT, SCMAP, SCIMG. Moved nowhere. 12421. August 21, 2007 GETJY Eric Fixed it to record the source name, fitted flux and error in the history file. Moved nowhere. 12422. August 21, 2007 IBLED Eric Added option FLAG ALL TIME to Fortran, help, and HLPIBLED run-time documentation. Moved nowhere. 12423. August 22, 2007 TVCPS Eric The read-from-disk option did not multiple the LUTDAT by the bit that said whether a color was on or not. This led to a balck and white picture. Changed the code to zero LUTDAT for colors/planes that are not on. Moved nowhere. 12424. August 22, 2007 OOSRT Eric Did not copy the input file's tables. Moved nowhere. 12425. August 22, 2007 Misc Eric MX Added comments to the help file pointing out the lack of FG and SN table application and other advantages of IMAGR. UVDEC Made local version of FGSEL that includes an increment in frequency to adjust channel numbers of flags. Moved nowhere. 12426. August 23, 2007 GREYS Eric With some LTYPE values, the task refused to plot the Clean Beam. Used another location common group for plotting the beam that would not be changed for plotting the main axes. Moved nowhere. 12427. August 28, 2007 UVHOL Eric Corrected it to handle polarization states other than FULL. It assumed that all 4 Stokes are always present and Rick has RR only data. Moved nowhere. 12428. August 28, 2007 MATCH Eric The header information for AN, SU, and FQ tables came from the master file rather than the input. This messes up the date in AN files and other significant parameters. Played a simple game to fix this. Moved nowhere. 12429. August 29, 2007 HAFIX Leonia This task is a modification of UVFIX designed to diagnose a possible time shift error at the original UV data. The input UVdata are the output of the task TI2HA, which converted original IAT time into hour angle. HAFIX reads a row of the input data, stores the original U,V,W, converts the given hour angle back to IAT time, recalculates U,V,W for this time and for the shifted time and finally adds the difference of the two sets of U,V,W to the original U,V,W. Moved nowhere. 12430. August 29, 2007 LISTR Eric The gain routines changed the frequency from Hz to GHz so with a loop on the outside over IF the 2nd IF had a very poor knowledge of frequency indeed. Fixed it to use an unchanging idea of the frequency in Hz at the start of each time through each routine. Also fixed the error return from the print routines which enabled them to respond to a quit command from the user but kept them from looping over IF. Moved nowhere. 12431. August 29, 2007 CL2HF, UBAVG, SCIMG, SCMAP Eric For some reason many of the averaging subroutines in UVUTIL did not do weighted averages. Changed SCNAVG, SCNHAY (used by CL2HF), UVBAVG (used by UBAVG), UVTAVG, and UVAMPS (used by SCMAP and SCIMG) to do weighted averaging. The last required a change in an input histogram array from (4,*) to (5,*) since counts as well as weight sums are needed. So changed SCIMG and SCMAP themselves to set and use the new dimensionality. Note that all of the averaging routines output the weight as the sum of weights even though that is not correct for unweighted averages. It is now correct. Moved nowhere. 12432. September 3, 2007 HAFIX Leonia Two errors are found at the recent installation of HAFIX: 1. TWOPI was forgotten at convertion of days into radians. 2. UT1-UTC and IAT-UTC were erroniously egnored at UVW calculations. Moved nowhere. 12433. September 4, 2007 CORER Eric Rewrote task to use the full set of calibration adverbs and to offer the choice of printing or not and the choice of writing a flag table or writing out the flagged data. In the last case, the cal adverbs are not applied to the data. Moved task from $APLPGM to $APGNOT. Moved nowhere. 12434. September 4, 2007 SPFLG Eric There was an oversite that allowed the input FG table to be copied to the output even if the output was not a new file. Copying onto oneself is an error causing the task to die. Fixed to only copy to new files (as the help says). Moved nowhere. 12435. September 4, 2007 IMAGR Eric Added BOXFILE and OBOXFILE to the HI file. Moved nowhere. 12436. September 4, 2007 BLAVG Amy In weighted average the visabilities were being multiplied by the weight twice, so when the weights were calibrated and became very small the average vis became nearly zero. Fixed. Moved to patch Sept 5. 12437. September 4-5 gfortran found Eric TAB characters: PUVD.INC PCQV.INC POBV.INC FITLD MX CTREFM WXSEL PLAXES PLCLOS PLDRW PLINIT PLLINE PLPGO PLPNT PLRANG BPCOEF BPFSEL BPINI CHEBYD CHEBY CLSEL COPTAB CQINI CQMAKE CQSEL CVLDOP DATCAL DATGET DATPOL DGNRAO DSMEAR DTBAND DTCALS DTINDX DTMATC DTPARM FGSEL FXSEG GCINI GCSEL GETANT GNINI IKEY KEYIN MCSEL NXTFLG OBINI OBREFM PCSEL SELINI SN2CL SUSEL TABCQ TFILTR TYSEL VLBPOL QGRDFI QGRIDA Other: INTERP Tabs, variable in call seq unused. No routine calls this so dropped unused and fixed up code. ITBSRT Tabs, variable in call seq unused. Also unused labels. Changed call seq and fixed up code. TABSRT Changed for call seq to ITBSRT. SNSEL JSUB argument was not honored - added use. TABGN Changed call seq for unused variable. Tabs. TABGP Changed call seq for unused GPNUMV. Tabs. TABLEUTIL For changed call seq of TABGP. TABOB Changed to use the NUMV variable (several not 1), Tabs. QCLEAN One tab, several unused labels and formats, one unused call seq arg Tasks: FITAB Variables in call seq not used. Unused labels galore. TPMON Unused variable in call, unused labels. MX Unused variable in call, labels. FITTP Unused variables in call, unused formats FITLD Many unused variables in calls, many unused labels and formats. Call seq to TABGN Moved nowhere. 12438. September 6, 2007 UVDI1 Eric Corrected: 1. Main routine did not stop on error but did something the second time around. 2. Removed many blank lines and altered non-standard typing 3. Saved CNOs in common arguments so as to avoid using FCNO. The correct ones were referenced only by resetting NCFILE which would have made error recovery fail. 4. INDISK was referenced for the second file, making the averaging of the second file fail if IN2DISK not INDISK. Since this error was not tested, the task then copied the first input file to the output file subtracting 0.0. 5. Added 2nd input file to history. Moved nowhere. 12439. September 6, 2007 Unused call args Eric Changed: 1. STLTOU ( scrtch ) and PREAD calls it 2. TKGGPL ( inpts ) and AU9B calls it 3. TVSLAC ( iblk ) and AU6E calls it 4. FITBM ( ierr ) and BMSHP, QCLEAN, QUVUTIL, MX, APCLN, SDCLN which call it 5. Remove ACINI2, ACINIT, fix commented out call in ZDCHIN and erroneous reference in GAINI 6. SNSMOO ( soukol, frqkol ) and UVUTIL (SLFSMO) calls it 7. LPCLOS ( scrtch ) and PRINTER, OFMLST, PRTMSG, AU1A, AU7 PRTIM, PRTTP, HFPRT, DTSUM, COHER, TBDIF, OTFIN, SAD, PRTAB, PRTAN, MFPRT, UVFND, UVPRT, IMFIT, PRTSD, SHOUV, TXPL, RLDIF, UVHOL, CORER, JMFIT, LISTR, PRTUV, PRTCC, PRTAC, UVDIF, BLSUM, ANBPL, ISPEC that call it 8. TXTMAT ( scrtch ) and thence VERMAT ( names ) and PRTAC, AU2A, AUA, AU1A, HELPS, AU2, AIPSC that call them. Moved nowhere. 12440. September 12, 2007 servers Eric In UNIXSERVERS and XASERVERS, the bash setting of library path for the tek and message servers should have included the word export. For Macs in XASERVERS for the message server the library path variable had a spelling (case) error as well. Moved nowhere. 12441. September 12, 2007 EDITR Eric NEXT ANTENNA had 2 lines left in that should have been deleted. This caused the current antenna number to be wrong and would cause flagged antennas to be wrong and the next NEXT ANTENNA to be quite strange. Fixed QEDIUTIL. Moved nowhere. 12442. September 17, 2007 Tabs and messes Eric Changed for tabs and unused labels only: RESEQ ANCAL ANTAB BLAVG BPCOR DBCON DFCOR FXVLA HAFIX LISTR OBEDT PCLOD SETJY SHOUV SPLIT UVFIT UVFIX UVGLU VLOG DTSUM BLCAL CONVL FRING VTESS IRING LOCIT PBEAM SNPLT SPFLG Changed for tabs, unused labels, unused call seq arguments: DTCHK DTSIM FITLD SNSMO SPCAL SPLAT UJOIN USUBA UVDEC UVDI1 VBGLU BPASS KRING APCAL ELINT Changed more seriously: STUFFR Added statement about no FG, SN etc applied to help. UVGET Added message about which FG version every 20 init calls. AVSPC Fixed unused labels, unused call arg, added message about flagging version or not CLCOR All of the above, plus the code from calc is a horribly nonstandard mess. Cleaned up large number of unused variables and a little of the typing and other messes. M3TAR Tabs, truncated strings, unused formats, converted a big data statement of pointers (many unused at times) to a parameter include. UVCOP Tabs, unused labels/formats, unused argument, added message to say which FG table or none. VTEST Tabs, unused labels, DO loops with floating-point loop counters no longer acceptable (replaced with hand-rolled loops) CPASS tabs, unused labels, call seq arguments unused (in math routines), replace obsolete ASSIGN n TO NEXT; GO TO NEXT, (n1, n2, n3, n4) usage in one match routine - This is terrible spaghetti code, no wonder it does not work. UVSIM Replace one floating point DO loop and some unused labels. CVEL Tabs, unused labels, added message about FG version. SDVEL Tabs, unused labels, unused args in call seq, added message about FG version. FRPLT Tabs, unused labels, character strings of wrong length as temporaries for standard strings Moved nowhere. 12443. September 18, 2007 IMAGR Eric Improved the format in WINDF to allow for field numbers > 999. Also changed GR3DUV, GRDMEM, GRIDUV, IMGMEM, MAKMAP, UVGRID, UVGRTB, UVTBGD, FITBM, SCIMG, and UTESS to make the Field number be more readable. Moved nowhere. 12444, September 18, 2007 USUBA Eric Yesterday's update had an error test backwards causing it to die at start up. Moved nowhere. 12445. September 21, 2007 FILLM Eric Changed response to nominal sensitivites of 0.0. Previously, 0.333 was used for weights but 1.0 was used for data scaling. That should have been sqrt (0.333) to be consistent and so leads to weights wrong by factors of 2 for baselines having one such antenna and 4 for baselines with both antennas affected. Also corrected weights output when writing correlation coefficient data - the nominal sensitivity should not be included in such weights although the bandwidth and so forth should be used. The ModComps would not have allowed nominal sensitivities of 0.0, so this is a recent problem. moved nowhere. 12446. September 21, 2007 TAFLG Eric The task wants to record the reason for the flag but that area is limited to 8 such records. Added test to avoid going over this limit - tell the user but say unimportant. Moved nowhere. 12447. September 24, 2007 OGEOM Eric The functions in INTERPLATE called internal function IMINST with a plane number on the output image which was taken to be the same as a plane number in the input image. This is not necessarily so. Fixed it to add the desired plane to the BLC plane of the input image. Moved nowhere. 12448. September 24, 2007 TAPPE Eric NEW TASK: append two tables elliminating duplicate rows as well as possible. Contains local version of TABSRT which supports subscripts into column arrays. This should go to all TABSRT usages with a change in table header format plus added options in TASRT, TAMRG??? Moved nowhere.. 12449. September 25, 2007 TI2HA, STUFFR Eric Changed TI2HA to allow the full range of calibration adverbs including allowing multi-source files as input. Only one source may be selected however. Changed STUFFR.HLP to allow DOCALIB and FLAGVER, both of which act as on/off switches. The highest numbered SN and FG tables will be applied when selected. Changed STUFFER.001 in anticipation of UVSRT or UBAVG acquiring these adverbs someday, deselecting both at those stages. Moved nowhere. 12450. September 26, 2007 gfortran Eric Added gfortran to FDEFAULT.SH and LDOPTS.SH for LINUX. Moved nowhere. 12451. September 27, 2007 MAXCIF Eric This parameter is the max allowed product of number polarizations * number spectral channels * number IFs. The max row in a table is based on it using XBPRSZ. Checked all uses of MAXCHA and MAXCIF and changed: $APLNOT/UVGET : not incl pol in test $APLNOT/SDGET : npol ignored $APLNOT/BPINI : buffer size comment wrong $AIPSUB/TBDATA : XBPRSZ not MAXCIF $APGOOP/INDXR : comment wrong, code ok $APGOOP/FLAGR : all VIS overkill by 4 $APGNOT/DFQID : XBPRSZ not 3*MAXCIF $APGNOT/FXTIM : XBPRSZ not 3*MAXCIF $APGNOT/FITTP : XBPRSZ not MXLREC $APGNOT/FITAB : XBPRSZ not MXLREC $APGNOT/DQUAL : XBPRSZ not 3*MAXCIF $APGNOT/FLGIT : WASBAD overkill by x4 $APGNOT/DBCON : XBPRSZ not 3*MAXCIF $APGNOT/TBIN : XBPRSZ not MAXCIF $APGNOT/TBOUT : XBPRSZ not MAXCIF $APGNOT/CALWR : XBPRSZ not MXLREC $APGNOT/LISTR : overkill x4 $APGNOT/UV2TB : buffer = MAXCIF +3 not *3 $YPGM/TAPLT : XBPRSZ not MAXCIF $YPGNOT/FRPLT : comments wrong $YPGNOT/TVFLG : unused usage $YPGNOT/IBLED : unused ref $APLPGM/PRTAB : XBPRSZ not MAXCHA $QPGNOT/SDIMG : MAXCHA should be MAXCIF $APGNOT/RLDIF : overkill x4, misuse $YPGNOT/WIPER : overkill x4, misuse $YPGNOT/UVPLT : excess x4, misuse Moved nowhere. 12452. October 1, 2007 JMFIT Eric In trying various solutions around the baseline fits, it did not return G (the answer) to its actual value. If the final image was printed, then G did get returned, otherwise not. Added an extra call to FXDVD to make things right all the time. Moved nowhere. 12453. October 1, 2007 FUDGE Eric Fixed some wording anomolies in the help file and added KEYCOP to the Fortran. Modified the typing some too. Moved nowhere. 12454. October 1, 2007 tables Eric Every TABxx subroutine should test every column pointer and number to make sure they are okay. This is really excessive but it is good practise. Changed the TABWX routine to do this since VLB tables lack one expected column (subarray) and have no values in 2 others - or at least used to. Moved nowhere. 12455. October 1, 2007 AVSPC Eric Added full calibration and normal flagging. Local version of CALCOP leaves out all FG and BP tables. Moved nowhere. 12456. October 3, 2007 CHKNAME Eric Changed AU8 and CHKNAME help file. Previously it would only look for 1000 name matches. Now it will look for 1000000. Moved nowhere. 12457. October 3, 2007 FILLM Eric Corrected the bandwidth put in the SU table. It was taking the total bandwidth and dividing by the actual number of channels out. It needs much more complicated computation to get the channel width - copied the increment computation from the header creation location. Fixed the history file; it used I2 for NFILES which might not be enough. Moved nowhere. 12458. October 3, 2007 CPASS Eric Found one FORMAT lacking enough format codes for the number of items. Fixed many FORMATS with no format codes to be simple MSGTXT = 'string' statements. Moved nowhere. 12459. October 4, 2007 CPASS Eric My eyes are getting bad (?) - an I1 was typed as 11 in a WRITE statement. This caused an abort under Intel anyway. Fixed long lines in inputs section of CPASS.HLP also. Moved nowhere. 12460. October 5, 2007 slices Eric I found that slice plots default to the pixrange of the slice itself but then use the image max/min for scaling. This caused no end of trouble when the max and min are large but the range of the slice is seriously less. Changed all scaling of slice plots to use the slice max/min exclusively and to change the "pixel" range by a factor 10. Changed the image catalog header to use the pixel range in physical units in the IRRAN, IRRAN+1 locations as a way of making the information available within the current program and any other that access the plot later. This set of changes was made in subroutines AU6E, AU9A, SLOCIN, TKGGPL, TKGMPL, TKRSPL, TKSLPL, TVSLIN, TVSLPL, and TKSLIN. Tasks SLFIT, SL2PL, XBASL, XGAUS, XPLOT, and ISPEC also were changed. The Xxxxx tasks were found to make the plot over the full size of the TV except that the left out the boundary box and some of the data when they were not visible. Fixed it to use the current size of the TV which can be changed between plots. Moved nowhere. 12461. October 9, 2007 UVCON.HLP Leonia Vincent Fish pointed out that default for CMODEL is not described correctely. Fixed. Moved nowhere. 12462. October 9, 2007 UVHOL Eric Added the options to average data to all the reference antennas and/or to average in time within a pointing. Moved nowhere. 12463. October 10, 2007 UVAVG Eric Changed code top force DOACOR true when MERG and AUTO modes are specified. Changed help to emphasize the new DOACOR adverb. Moved nowhere. 12464. October 10, 2007 UVHOL Eric Changed default on BDROP to none (0 means 0, same as 1). Moved nowhere. 12465. October 11, 2007 FILLM Eric The antenna has moved test was too tight for Pie Town - made it twice as tolerant for Pie Town only. Moved nowhere. 12466. October 11, 2007 UVHOL Eric The [hase inversion was not applied correctly in the new code arrangement which allows for averaging. Fixed it I hope. MOved nowhere. 12467. October 15, 2007 SPLIT, SPLAT Eric The computation of the IF frequency of averaged channels was incorrect because it used FINC(BIF) and FINC(BIF) was already scaled by the degree of averaging. This affects all arrangements of IFs, not just the mix of USB and LSB for which the error was initially entered into the code. SPLIT was dufferent in that it scaled the FINC(IF) first and then used it with the scaled FINC(BIF). For IFs with the same FINC, no error was produced (unlike SPLAT), but a worse error was made when the FINCs had different values. Moved to patch this date. 12468. October 15, 2007 FITAB Eric FITAB "cheated" by making AIPS HISTORY info in the main header that made it easy for AIPS' UVLOD to reread the data. Too easy if the data have been out to obit and back. Dropped all this history from FITAB and made all loaders smarter. DBHD.INC Added CROTA, EPOCH, OBSRA, OBSDEC, BUNIT, OBJECT SETDEF Zero CROTA GETINT New routine to do GETNUM only for integers GETNUM Checked for s.p. INDE before d.p.INDEFINT and so would always fail to find the d.p. GETSSY New version of GETSYM that allows a single integer follwed by a letter as a keyword name. GTWCRD Changed call seq to allow for coordinate version codes and to handle 'nKEYWnn' type keywords EXTREQ Changed for GTWCRD call sequence change TABHDR Parse for correct table coordinate keywords (iKEYWnnV) plus BUNIT, EPOCH, EQUINOX, OBSRA, OBSDEC, OBJECT. Use new GTWCRD call seq. RUVTAB Changed call sequence to include the piece and vis number values to allow a call to RUVFIL. Use scaling and offsets in case these data not from FITAB. RUVFIL New routine - fixes UV header from table keywords and replaces dummy file or expands initial UV file. Also fixes units for scaling columns. UVLOD Changed call sequence to RUVTAB. Added array USED to avoid resetting parameters from the parser if they are already set and the new keyword is an HISTORY AIPS keyword. When UVTABL, do not build a new header or expand UV file in UVFHDR on 2nd piece - info not known properly there. FITAB Dropped all of the main header HISTORY AIPS stuff that had standard FITS keywords etc. IMLOD Add USED to avoid resetting parameters from the parser if they are already set and the new keyword is an HISTORY AIPS keyword. FITLD Add all IMLOD and UVLOD changes. Also fix GTWCRD call sequence. Added ERROR adverb as in IMLOD and UVLOD. Moved nowhere. 12469. October 16, 2007 UVHOL Eric Changed UVHOL to: 1. Drop BCOUNT and UVRANGE which were not meaningful 2. Change comment symbol in HOLG to #! rather than !! 3. Label U and V columns properly (l and m etc) 4. HOLG: add comments about antennas averaged and about avergae number of samples per printed sample 5. Scaling now uses the full ref antenna and time averaging options to get proper scaling foer weight. Changed HOLGR to accept either ! or # in column 1 as a comment. Moved nowhere. 12470. October 17, 2007 UVMTH, UBAVG Eric Added info on the default OPCODE in UVMTH and the meaning of "field of view" in UBAVG help files. Moved nowhere. 12471. October 18, 2007 BL tables Eric BLGET treated FRQSEL=-1 as a meaningful value rather than "don't care" and so failed to get BL values and so flagged RR and LL when BLVER=1. Why FRQSEL was -1 is a good question but the routine was in error anyway. COPTAB forgot about BL tables. Added BL table copy when DOBL false. BLSEL new routine - no one ever cared before. No selection on time or source since they are not really meaningful usually. Selection on IF critical and allowed also selection on antennas and correction of FQID. Tasks AVSPC, UJOIN, UVCOP, and SPLAT have their own versions of COPTAB to which I also added BL tables. 12472. October 18, 2007 NX table generation Eric The NX table generation could be fooled by subtle time differences in tasks like UVAVG. Changed it to ignore and time differences < 1 sec which seems to be enough. Moved nowhere. 12473. October 22-23, 2007 errors Eric UVCOP and UJOIN should have said CATOLD rather than CATUV in the call to BLSEL. Probably okay in the former, not the latter. Moved nowhere. 12474. October 23, 2007 SPLIT Eric SPLIT still did not get things right. There was a very long-standing error in the header reference pixel computation when averaging channels but still making a multi-channel output file. SPLIT leaves the ref value alone for frequency and so the offset frequencies in the FQ table should not be changed - a more recent error. SPLAT does things differently, puting the ref pixel to 1.0 and changing the ref values. Moved to patch this date. 12475. October 23, 2007 FLOPM Eric New FUDGE-clone task to reverse the sign of the phase and the spectral channel order without changing anything else. This fixes a VLA on-line system error where side-bands were confused. Moved nowhere. 12476. October 24-25, 2007 UVAVG Eric Changed the code to set time intervals for TB sorted data from the first sample of the interval. BT remains in the integer*YINC from the start of the day intervals. Improved the help file to make this clear and to drop size limitation discussions since dynamic memory is now available. Moved nowhere. 12477. October 25, 2007 UVLOD, FITLD Eric RUVFIL did not handle the case of an old FITAB header well. It did not have the file name and did not do all the comparisons properly. Fixed it to force the output compression to be that of the input table, not DOUVC, and to add weight and scale parameters if needed and so forth. Changed UVLOD to have compression only on DOUVC > 0 (used to be >= 0, which differred with FITLD) and fixed help files to say that FITAB uvtables have the compression of the input table. Fixed error message in RUVTAB. Moved nowhere. 12478. October 25, 2007 UVDIF Eric Fixed it to allow one file to be compressed and the other not and still do a proper comparison. Moved nowhere. 12479. October 26, 2007 UVAVG.HLP Eric Changed the help file to emphasize the importance of well chosen YINC. In particular, setting YINC to the interval between samples will now, with TB data, cause deep troubles for the MERG operation. Moved nowhere. 12480. October 26, 2007 FILLM Eric Added a header keyword to specify whether the data are correlation coefficients (CORRCOEF 1) or not (CORCOEF -1). Found an error where the tolerance for appending data was 0 for the first file when comparing antenna locations. Fixed the tolerance to use the date. Also fixed the conversion of AN file locations to FILLM units - that should have been fine even though it said (1.E-9 * VELITE) but (1.D-9 * VELITE) produces the correct results. Moved nowhere. 12481. October 29-30, 2007 Web address Eric The web address has changed from www.aoc.nrao.edu/aips to www.aips.nrao.edu. Changed the files ZXHLP2, ZXHLP2.PL, CDSETUP, CookBook chapters 0,1,3,C,Z, and helps for FXVLB, VLOG, Y2K, and SETJY. Redid all web pages with references to the old location, redid the CookBook html and PDF versions, and redid the cgi scripts on the new cgi location. Moved nowhere. 12482. October 30, 2007 DCONV Eric DCONV neglected to convert OPCODE from hollerith in the inputs to the local variable. So the second image was tested in all circumstances and used if it accidentally passed inspection. Moved nowhere. 12483. October 30, 2007 LISTR Eric The matrix averaging routines, when they were switched to properly weighted averages, retained an RMS test on COUNT being larger than 2. That would be fine if COUNT were a simple count, but it is now the sum of real weights and so can be scaled any old which way. Changed several compares from 2 to 0.0. Moved nowhere. 12484. October 31, 2007 FITTP, FITAB Eric Changed VKEY.INC so that EQUINOX will be written rather than the deprecated EPOCH. Moved nowhere. 12485. October 31, 2007 UVLSF Eric Changed fitting code (uses XBASL-like code now) to allow up to 4th order orthogonal polynomials in the fit and added the option to add the continuum of channel CHANNEL back in to all spectral channels equally. Moved nowhere. 12486. November 2, 2007 UVLSF, UVLSD Eric The header of the continuum output went back to the original input header - a really wrong thing to do these days. It led to an empty FREQSEL parameter for example. Moved nowhere. 12487. November 2, 2007 VPLOT Eric Changed default SYMBOL. Added axis types for hour angle, elevation, azimuth, and parallactic angle. Added FACTOR to control size of SYMBOL and to allow connect the dots when not plotting a model or 2 of u,v,w. Changed subroutines POLCHK and POLVIS to allow Q/I, U/I, V/I. Moved nowhere. 12488. November 2, 2007 SNPLT, LISTR Eric Add words re TY columns for VLA: "TSYS" is nominal sensitivity and "TANT" is a front or back-end system temperature to these help files. Moved nowhere. 12489. November 5, 2007 SPLIT, SPLAT Eric I do not believe this - both tasks had the number of words/correlators to copy wrong (3 times too large). This can cross array boundaries messing up needed data and can certainly get wrong results for the scaling of compressed data. Moved to patch this date. 12490. November 5, 2007 random numbers Eric Changed RANDIN to make seconds of time the major initiator rather than hours and increased all of the multipliers by quite a bit. It is still not a great random number generator, but it is better this way. Moved nowhere. 12491. November 6, 2007 SPLIT, SPLAT Eric I further don't believe how I messed yesterday's correction to copy too few for compressed data. Moved to patch this date. 12492. November 7, 2007 TYSMO Eric New task to clip and smooth TY tables. A task to remove and/or apply the TY (nominal sensitivity) to UV data is still to come. Moved nowhere. 12493. November 7, 2007 FLOPM Eric Added OPTYPE='VLAE' to do the VLA special fix, with any other OPTYPE flopping the channel order and fixing the header and FQ table. Changed to copy all times, fixing the TIMERANG requested. Moved nowhere. 12494. November 8, 2007 UVFIX Eric the array for holding source parameters was limited without test to 200. Raised that to 2000. Moved nowhere. 12495. November 8, 2007 VPLOT Eric Changed the time averaging to be (1) on normal not integer-step intervals and (2) to allow all axis types. Moved nowhere. 12496. November 12, 2007 FLATN Eric The code wrote the noise squared and the weight squared out rather than the correct units. Moved nowhere. 12497. November 12, 2007 UVLOD Eric Changed UVLOD, IMLOD, and FITLD to ignore the EXTEND card that used to be required. Little use was made of it except for IMLOD covering for an empty image if there standard tables following. Fixed UVLOD, FITLD, IMLOD to do the blocking of AIPS HISTORY correctly for required cards. Fixed UVLOD and FITLD to not be unhappy with headers it sees that are mixed - some axis labels with bad values when UVTABLe is true. Moved nowhere 12498. November 13, 2007 TPHEAD, PRTTP Eric FWRITE Changed simply to give up on AIPS UV-tables and say header info later in file (so TPHEAD can't give it). This applies to old FITAB output too due to adding the USED tests to prevent HISTORY AIPS from overriding keyword values FPARSE modernized code style, added USED to call seq and use it to prevent an AIPS HISTORY card from overriding a regular keyword card. GEYKEY moved from $APLNOT to $APLSUB, cleaned up style SDTCRD moved from $APLNOT to $APLSUB, cleaned up style TABF3D moved from $APLNOT to $APLSUB. PRTTP Should not use FPARSE on table headers; added parts of TABHDR to parse the UV-tableheader. Also use some incoming parameters parsed from the HISTORY AIPS main header. Moved nowhere. 12499. November 13, 2007 VPLOT Eric The real and imaginary parts got reversed when SOLINT>0. Fixed it and Rick is happy. Moved nowhere. 12500. November 14, 2007 3C286 X-band model Amy Put back a revised model - for some reason the CCs were offset 1 pixel in X from the center. Moved nowhere. 12501. November 14, 2007 PRTTP Eric Found a few more header parameters to display correctly for uv-table data. Moved nowhere. 12502. November 26, 2007 CONFI Leonia Having got the request of the LWA people, I have added the constraint of the outer configuration border at the shape of ellipse with the given ratio of the ellipse axes. Moved nowhere 12503. November 26, 2007 UVLSF Eric Changed the response to data for which there are not enough selected samples to fit the spectrum. Instead of dying, it just omits that spectrum. Changed the handling of the flagging on residuals - the continuum was added back (when selected) before the test on residuals was performed. Rearranged things to test the true residuals and then add back the continuum. The new code flagged all samples following the first bad one. Both the new code and the old code did not reset the counter for number of samples in the rms. Therefore, the flagging based on rms in the fit region never worked - but should now. Moved nowhere. 12504. November 26, 2007 UVMOD Eric The explain file used numbered cases which implied that those numbers were the CTYPE values (but did not say so). Changed the case numbering to start with 0 to match the CTYPE values associated with each function. Moved nowhere. 12505. November 27, 2007 COPIXEL Eric Added the ERROR adverb to COPIXEL so errors (COORDI not give pixel inside image) will not stop things but allow control to be passed. AU7B has the code. Moved nowhere. 12506. November 28, 2007 polarization cal Eric A bug was discovered that caused polarization calibration to be incorrect when the calibration routines are initialized for a second time. This is common - LISTR loops over IF with an init each time, SPLIT loops over source, UVPLT reads the data for scaling and checking before rereading for plotting, etc. The result is grossly wrong data because the polarization cal tables do not get reinitialized. Fortunately, the bug applies only to non-approximate modes and then only when the parallactic angle does not change which would be rare except for equatorial telescopes which have no parallactic angle. Thus WSRT data are almost the only data affected by this bug. Changed DATPOL to init the "old" parallactic angle to an impossible value whenever the source changes. Moved nowhere. 12507. November 28, 2007 SPLIT, SPLAT Eric When averaging all channels but not mixing IFs, these tasks bizarely allowed the frequency to stay the same. This perhaps avoided some multiplies but at the cost of incorrect sizing in the uv plane. Changed it to allow the ref frequency to change as it should. Moved nowhere. 12508. November 30, 2007 TYAPL Eric NEW TASK: It can remove one TY table (Tsys col) from a data set and then apply another TY table to it (e.g. a smoothed one). Each of the 2 operations are optional although it will refuse to remove a TY from a data set marked as CORRCOEF=1 (true) and will not apply a TY unless the data set was, or has been converted to, CORRCOEF=1. Moved nowhere. 12509. November 30, 2007 AIPSPATH Eric Changed both. The AIPS libraries for binary distributions need to be pre-pended to the user's LD_LIBRARY_PATH not post-pended. If the user has an incompatible version of the compiler in his path, a post-pend will be ignored. Moved nowhere. 12510. December 2, 2007 lists Eric Checked and updated help files LSAPROPO (apropo list), HLPLST (tab completions), and ABOUT files ADVERB, AP, CALIBRAT, EDITING, GENERAL, IMAGING, INFORMAT, INTERACT, MODELING, OBSOLETE, OOP, POPS, PROCEDUR, TABLE, TASK. UTILITY, UV, VERB, VLA, VLBI. Movved nowhere. *********************************************************************** *********************************************************************** Changes while 31DEC07 is NEW *********************************************************************** *********************************************************************** 12513. December 5, 2007 RECAT Eric Added code to the stand-alone program RECAT to detect that it is being run as a task and to forbid that. Moved from TST this date. 12516. December 10, 2007 install.pl Eric MACINT was missing from the list of extra architectures for which diretories must be created. It was also missing from the MAC case issue when making a link file named AIPS as well as the one named aips. Moved fixes from TST this date. 12519. December 25, 2007 VLBAEOPS Eric Cleaned up typing of help files. Lines were too long causing them to be truncated for the user displays, etc. Moved from 31DEC08 this date. 12520. December 28, 2007 AIPSLetter Eric Released LET07B.{TEX,PS,PDF} for 31DEC07 final revision. Moved nowhere ($AIPSPUBL is in all versions). 12521. December 28, 2007 CookBook Eric Changed $SYSUNIX/AIPSWC to include more areas and count PDF to skip. Changed chapters: 1 Updated line and task counts, shipping info, mailing address, aips++ info, etc. 3 Comments re message server, Web and street addresses, REBYTE mention. 4 FILLM (corr coef), TYSMO/TYAPL, TAPPE, OUTFGVER, CLIPM -> CLIP, WIPER, VLANT, FITAB changed, VLACALIB does models 5 Numerous wording changes - esp CLIP usage different, EDITR has FLAGVER and OUTFGVER, IMAGR has two new menu options, new OOSUB mentioned. Weights from FILLM are good not poor as stated. WIPER more general now with baselines known. 6 Numerous wording changes - esp TVLOD FUNCTYPEs, CORER calibration added, COPIXEL error return, IMEAN doinver. 7 IMCENTER, IMEAN, reword some 8 Wording changes, CUBIT, TAPPE, UVLSF 9 DBCON does NOT fix antenna numbering (fixed several errors on this point). Numerous wording/typo fixes. UVCOP now does NX tables. VLBA verb in MkIII section. TYSMO added. LISTR corrected. OUTFGVER with UVFLG. Drop SLIME. Add WEIGHTIT in CALIB self-cal section. 10 Wording improvements, time order for flagging, OUTFGVER 11 Updated wording, FITAB 12 Tab completion, COPIXEL, FUDGE, move data to new computer section incl REBYTE 13 Updated every list to be sure A OUTFGVER in UVFLG, add major section on VLARUN., use model in VLACALIB B Tape -> archive disk files, FILLM parms now default, lots more rewrites incl SETFC. C CLIP, OUTFGVER, UVCOP (NX table), numerous minor wording changes F Dropped 9-track tape section, FITAB added some and 31DEC07 change in it, fixed size of CL table Moved nowhere ($AIPSPUBL is in all versions). 12522. December 29, 2007 CookBook Eric Changed chapters: LaCook*.tex in $DOCTXT, COOK*.PS in $AIPSPUBL. Copies in ftp site. 0 Updated D Add SNFLG, correct INDXR remark, emphasize models, VLANT, correct editing of SN tables (EDITA does not, CLCOR does CL), other re-wordings. I Updated J Updated Recipe chapter rearranged and reorganized as needed. $AIPSPUBL/COOKBOOK.PDF replaced as well as the full html version. Moved nowhere ($AIPSPUBL is in all versions). 12532. January 9, 2008 VLANT Eric Moved the - I hope - completed VLA station corrections file for 2007 to $AIPSTARS and saw to a dummy 2008 file being created in the appropriate mysterious places. Moved nowhere, but 31dEC07 gets it automatically. 12533. January 9, 2008 REBYTE Eric The table translation made a hash of tables with long rows - IM tables from VLBI and potentially BP tables for significant numbers (~ 120 or so) of channels times IFs. Moved to 31DEC07 this date. 12542. January 18, 2008 FITLD Eric Corrected errors I put in the call sequence to TABWR so that weather tables can be properly interpreted. Included use of the number of parameters to handle weather tables missing WVR and ION values. The default reference pixel is 0.0 not the center of the image. Changed only the IMLOD portion of FITLD. Added defenses against strings with null bytes in tables and elsewhere. Moved above fixes from 31DEC08. 12570. March 4, 2008 model division Eric The correct weights after model division are multiplied by amplitude squared, not amplitude. The gridded modeling was corrected years ago, but the DFT was not until today. Fixed QPTDIV, QGADIV, and QSPDIV. Moved to patch March 5. 12572. March 5, 2008 FILLM Eric FILLM has been scaling and weighting cross-hand data incorrectly in those cases in which the antenna order in the on-line system is reversed from the antenna order implied by the numbers painted on the sides of the telescope. The reversal took place before the scaling and weighting and it must take place after since we cannot change the antenna numbers themselves. Note that this affects ModComp-era data when converted to correlation coefficients and post-Modcomp data when converted to visibilities. It affects weights of Modcomp-era data when not converted to correlation coefficients. The affects on solar data cross-hands are more extensive since they are converted to correlation coefficients. Moved to 31DEC07 patch this date. 12610. April 29, 2008 DFT modeling Eric VISDFT found the model type from the first facet, first component and then ignored any model types not matching without comment. Changed it to check all facets and to use Gaussian or Sphere for point models if there are also Gaussian or Sphere models. It will complain if both Gaussians and Spheres occur, but points can be either with 0 width. It takes more space and computing, but this is better than ignoring most of the flux! Moved to 31DEC07 this date. 12632. June 19, 2008 FILLM Eric The setting of the CORRCOEF keyword was all messed up. It should only care about the setting of the ISSUN and RICKS options rather than whether a correction needs to be made. Moved to patch this date. 12649. July 8, 2009 FILLM Eric/George M As far back as our records go (12/1990), FILLM has confused which IF goes with which in applying on-line flags for modes 2BC, 2CD, 4, PA, and PB. This means some data were flagged that should not have been and some were left unflagged that should have been flagged. Note that the new OF table with PRTOF and OFLAG would have pointed out this error had we looked at appropriate data sets. Moved to 31DEC07 patch this date. 12690. August 13, 2008 GPOS, FPOS Eric System adverbs need to have standard usages - both FPPOS and GPOS are used as absolute pixels, relative pixels, and arc seconds in different places. Removed both from the system adverbs list. Moved to 31DEC07 this date. 12700. September 5, 2008 FACSET Eric FACSET confused the subscript order in the alternative names and so used the wrong radius for names for the cal sources. Moved to patch Sept 10. 12531. January 7, 2008 XAS, START_AIPS Wes/Eric The new Mac operating system "leopard" has a clever scheme in which $DISPLAY is set to a named pipe of for "/tmp/gobbldegook" which confused aips significantly. Made small changes to XAS.SHR (init.c) and START_AIPS and START_TVSERVERS. Moved to new 26 Sept, 2008. 12748. November 18, 2008 FILLM Eric The subroutine that is used when CPARM(4) is not 0 or 25.0 had an error in computing sqrt (u^2 + v^2) due to a misplaced close parenthesis. abs(u) + v^2 was examined instead. This matters since apparently the on-line system has been setting the shadow-flag bit wrongly and that bit is used when CPARM(4)= 0 or 25.0. Clarified this in the help file. Moved nowhere.