;----------------------------------------------------------------------- ; Copyright (C) 2010-2012 ; 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 31DEC11 is TST *********************************************************************** *********************************************************************** 13339. December 1, 2010 Code rollover Eric Changed: PUVD.INC 32768 -> 65536 for MAXCIF, MAXCHA 5000 -> 25000 for XCTBSZ PMAD.INC 32768 -> 65536 for MAXIMG ZPBUFSZ 32768 -> 65536, 6*256 -> 12*256 QRFFT 8192 -> MAXIMG/4 QXFOUR 32768 -> MAXIMG, 8192 -> MAXIMG/4 APLSUB GETRLS.FOR 31DEC10 -> 31DEC11 INSUNIX INSTEP1 31DEC11 SYSPERL install.pl 31DEC11 SYSPERL update.pl 31DEC11 SYSPERL ZXHLP2.PL 31DEC11 SYSPERL LaCook.PL 31DEC11 UPDUNIX UPDCONTROL 31DEC11 SYSUNIX AIPSPATH.CSH 31DEC11 SYSUNIX AIPSPATH.SH 31DEC11 Moved nowhere. 13340. December 1, 2010 RMSD Eric The MACPPC compiler cannot handle a parameter that involves > 1 math operation. Thus PARAMETER (A = B*B/32) does not work whilst PARAMETER (A = (B)*(B/32)) does. Moved to 31DEC10 this date. 13341. December 3, 2010 VLA.GAINS Eric The last Q band data added had a last good date of 20110101 while the last good date for other bands is 21000101. Fixed the Q band date since we have no new/better info. Moved nowhere - goes to all versions automatically. 13342. December 3, 2010 VBGLU Eric Changed: 1. Compatibility tests to say what is incompatible 2. Sort order tests to allow 'Tx' to match 'Ty' 3. HI writing to write up to 4 input file names Moved nowhere - could go to 10. 13343. December 3, 2010 UVFLG Eric Test for EIF > 0 but EIF < BIF and also BCHAN, ECHAN - treat as error. Moved nowhere. 13344. December 3, 2010 TIORD Eric Added test on baseline order. Also help file. Moved nowhere. 13345. December 3, 2010 SNEDT Eric A badly type If statement caused a computation of difference values with EDIFSY (for SY tables) when SHOW ALSO RATE was selected in editing SN tables (and probably in editing UV data). Fixed it to do only SY there in QEDIUTIL. Moved to 31DEC10 this date. 13346. December 6, 2010 IMAGR Eric Made a change to the formula for FACTOR used with the multi-scale options. The old formula could blow up and so would have been hard to use. Changed IMAGR.HLP and QCLEAN. Moved nowhere. 13347. December 6, 2010 LISTR, ANFIND Eric LISTR set 2 READ statuses when OPTYPE not recognized. This messes up ANREFM which clears one only. Changed ANFIND to recognize SAO SMA array. Moved to 31DEC10 this date. 13348. December 7, 2010 SNPLT Leonia Add print (at the plot header) the ploting tone number, if the table is PC. Moved nowhere. 13349. December 7, 2010 VLAPROCS, VLATECR aMY Added a message to VLATECR about the possibilty of the download failing if the VLA project is too recent. Moved to 31DEC10 Dec 8. 13350. December 8, 2010 FITLD Eric FITLD had trouble with a FITS IDI table that expressed a logical variable as a character-string valued variable. Changed GETLG to ignore an open quote mark but then require the value to follow immediately. Moved to 31DEC10 this date. 13351. December 8, 2010 NORMALIZ gain Eric Changed POPSDAT, NEWPARMS to add adverb NORMALIZ. It replaces CPARM(2) in CALIB to control gain normalization. Added it to CLSMO and SNSMO to forbid, allow, or require global gain normalization after smoothing. Previously, these too did a global normalization if the previous table had a non-1.0 normalization only ("allow" in above terms). Changed LISTR to apply the global normalization to lists of gains. Moved nowhere, LISTR change moved to 31DEC10 January 3, 2011. 13352. December 9, 2010 TVFLG Eric Added option LOAD NEXT CH/IF/ST which increments the spectral channel and reloads the TV. If spectral channel exceeds the max, it is reset to BCHAN and the IF is incremented. If the IF exceeds the maximum, it is set to BIF and the polarization is switched possibly with a Stokes flag change. Corrected the help file and HLPTVFLG.HLP, the run-time help file. Moved to 31DEC10 this date. 13353. December 13, 2010 TYSMO Eric All other SY tasks changed to use Tsys rather than Tsys/Tcal so I changed this task too. In so doing, I found a number of issues with the handling of SY clipping - mostly because TY has only one real column and SY has tow columns and a combination and all are related to each other. These include (1) clipping Pdif and Psum before using them to compute Tsys, (2) blanking Pdif, Psum, Tsys if Pdif <= 0 or Psum < Pdif or either Psum of Pdif blanked (e.g. by clipping) (3) blanking Pdif and Psum if Tsys outside range allowed (4) rearrange logic in testing result after MWF smoothing Moved to 31DEC10 December 10. 13354. December 13, 2010 TYSMO Eric Added FLAGVER adverb and subroutine to apply FG tables to TY or SY data. This does the flagging before the clipping and smoothing operations, removing known bad periods before these other operations. Flags that depend on baseline rather than antenna or that depend on spectral channel are ignored. Moved nowhere. 13355. December 15, 2010 CookBook Eric Updated: 4 Added LOAD NEXT to TVFLG section Put in a modern TVFLG figure. E Corrected SY section - Tsys rather than Tsys/Tcal is now used by SNEDT, EDITA, SNPLT, TYSMO. Added remarks about applying the FG table with SNEDT (in 31DEC10) or TYSMO (in 31DEC11). All - changed the figure inclusion macro to specify the size of the figures (for html jpeg usage) and changed chapters 0,1,4,5,6,7,8,9,10,A,D,E to use the macro. Note that the pdf macro package does not specify the size (when one does the correct size is allocated but no picture appears). Moved nowhere. 13356. December 16, 2010 TVFLG Eric Changed option LOAD NEXT CH/IF/ST to LOAD NEXT IF/ST since LOAD NEXT CHAN is already there. Now it increments IF only and when that rolls over swaps the polariztion. Also changed the spacing of the bottom messages and added code to make the image start just above the bottom status lines (plus wedge and wedge label) if it is small enough. Corrected the TVFLG help file and HLPTVFLG.HLP, the run-time help file. Made the spacing changes to SPFLG as well. Moved nowhere. 13357. December 16, 2010 SOUSP Eric Added CALCODE to the adverbs so that identical SOURCE/QUAL entries in the SU table may be differentiated. Moved to 31DEC10 on 28 Dec. 13358. December 16, 2010 REWGT Eric NEW TASK: version of REAMP that scales weights only. Moved nowhere. 13359. December 22, 2010 TVFLG, SPFLG Eric The change made Dec 16 to move the grey-scale down the screen meant that the center of the screen might have no image. The catalog header reads assumed that it would always have one. Added parameter YCENTR to tell where to make the catalog reads and had the TV loading set YCENTR. Moved nowhere - bug not in 31DEC10. 13360. December 22, 2010 TRANS Eric The new TRAROW algorithm to figure out number of planes did not take into account the fact the images have an integer number of rows per disk block or an inetegr number of disk blocks per row. Thus if NX = 401, the actual space used is as if NX = 512. So it thought it could do more than it could and aborted. Moved to 31DEC10 this date. 13361. December 24, 2010 CALIB Eric The subroutine that applied gain normalization to the SN table on an IF/polarization dependent basis messed up by trying to use the same buffer for read and write. Switched to 2 LUNs and buffers with TABSN and it still did not work. Changed to straight TABIBI and TABIO and that does work. Moved GMODIT routine to 31DEC10 this date. 13362. December 27, 2010 SPLIT Eric Copied the PD tables - should not do that. Moved nowhere. 13363. December 27, 2010 UVFLG Eric Add adverbs QUAL and CALCODE which modify SOURCES except when UFLG and SOURCES='-'. They even modify the flags read in from INTEXT files. moved nowhere. 13364. December 27, 2010 EDITA, EDITR, SNEDT Eric Changed NEXT CORRELATOR to NEXT POL/IF since users did not know what I mean by correlator. Includes help files and 7 real time help files. Moved nowhere. 13365. December 27, 2010 RSPEC Eric Added option to write out a signal-to-noise image. Moved nowhere. 13366. December 28, 2010 SOUSP Eric SOUSP did not check for fluxes <= 0 in some IFs. This led to NaNs and so I fixed it. Moved to 31DEC10 this date. 13367. December 29, 2010 DBCON Eric Added messages on appending tables and added SY table to list of those appended. Moved to 31DEC10 this date. 13368. December 29, 2010 SNEDT, EDITA Eric Plot scaling for SY tables was affected by an error in the coding. Moved the minimum plot scaling values to a larger array and set them in the individual data type main routines in QEDIUTIL. This means the e-13 scaling for rates is not affecting Pdif/, etc. Moved to 31DEC10 this date. 13369. December 29, 2010 DBCON Eric Changed it to copy and append SN tables also. Changed it to do all versions of tables in appending with a warning. Thus vers 2 in file 2 is appended to vers 2 in file 1, vers 3 in file 2 is appended to vers 3 in file 1, etc. Moved nowhere - do not move this one. 13370. December 30, 2010 AIPSLetter Eric Put back LET10B in TEX, PS, PDF in $AIPSPUBL. Moved nowhere. 13371. December 30, 2010 PCAL Eric The task kept track of source numbers as SID(source#)=cal# but then used SID in a form that should have been CID(cal#)=source#. Added the alternative list so that internal use and use with the CP table would work. CP table ended up with wrong values for source number. Moved to 31DEC10 this date. 13372. December 30, 2010 RLDIF Eric The corection to the PD table used the same sign when correcting the R and L sides. The hermitian should be used for L polarization. Moved to 31DEC10 this date. 13373. December 31, 2010 DBCON Eric Added calls to make an index table on the fly and dropped the initial suggestion to use INDXR afterwards. Added CALCODE to the tests of how to renumber sources - it was losing some of the distinctions that the EVLA has put in. Moved CALCODE correction to 31DEC10 this date, not NX one. 13374. December 31, 2010 ELINT Eric With large numbers pf sources and large numbers of IFs, the limit of 1000 for history cards could be exceeded perhaps. Changed code to allow 2000 and to test for the limit before adding more history cards. Moved nowhere. 13375. December 31, 2010 RLDLY Eric An array was dimensions MAXANT and was used as if MXBASE which walked on things. Fixed it - why did it seem to work? Moved to 31DEC10 this date. 13376. January 3, 2011 LISTR Eric The GAIN listing applied any gain renormalization prior to testing for magic blank. That made magic blanks not be magic and so messed up the listing. Moved to 31DEC10 this date. 13377. January 4, 2011 SY2TY Eric Created new task to do the Tsys computation from an EVLA SY plus CD table, writing a TY table. These may be useful for diagnostics but will not help in calibrating the EVLA. The raw "visibilities" from the BDF are not even normalized correlation coefficients so the gain calibration is not a scaling by Tsys. Moved nowhere. 13378. January 5, 2011 UVFND Eric Added adverbs QUAL, CALCODE, ANTENNAS, BASELINE to limit which data are tested. Moved nowhere. 13379. January 5, 2011 AIPS Eric Updated Copyleft in AIPS.FOR. Moved nowhere. 13380. January 6, 2011 RLDLY Eric Found an error handling the DOIFS=2 option and probably also the DOIFS=1 method. There was a loop over antennas that should not have been run when spreading the results into all of the IFs except for 1 antenna each time. Moved to 31DEC10 this date. 13381. January 6-7, 2011 LaCookE Eric It did not describe RLDLY correctly - CLCAL is not required after running it. PCAL no longer needs to be run after RLDIF which now correctes the PD table properly. Moved nowhere. 13382. January 7, 2011 FRING Eric 1. FRQTAB needs to do all frequency computation in double precision. With a small increment and large center frequency the resulting list was not accurate enough for FRING. 2. FRING failed to test after QINIT that it got all the memory it wanted. Added tests and messages about SETMAXAP. Moved to patch January 21. 13383. January 7, 2011 CALIB Eric Changed CPARM(2) to signal a desire for a median rather than a mean gain modulus to be used in normalizing the gains. Our new median routine is rather fast so, although this forces a further read of the SN table, it may be preferable to avoid really bad solutions from dragging the mean away from a desirable value. Moved nowhere. 13384. January 18, 2011 TABINI, DBCON Eric Changed TABINI to make NREC an in/out variable. If 0 on write, it will be reset to 100. This defends against bad values in BUFFER(42), the amount by which to expand the file. Changed DBCON to set NREC which should defend it against bad BUFFER(42) values. Moved DBCON NREC value only to patch this date. 13385. January 18, 2011 UVCOP Eric Used only the first 12 of the 16 characters of the source names. EVLA tends to be profligate so this no longer works. Moved to patch this date. 13386. January 18, 2011 CALIB Eric The header for the output file was not previously prepared by UVGET although the data for the file were. This made it possible for a FREQID random parameter to appear in the output file with completely uninitialized values. Changed the file creation to call UVGET and then modify the header as needed to put back compression if the output was to be compressed. Moved to patch 24 January. 13387. January 18, 2011 POSSM Eric POSSM neglected the fact that IFs have different frequencies when applying the SHIFT adverb. Moved the CHNDAT call to the INIT routine and fixed the shift frequency computation. Moved nowhere. 13388. January 18, 2011 RLDIF Eric 1. Add CALIN as a text file containing the Q/U spectrum of the calibration source(s). 2. Improve help file description of the task - it was obsolete. 3. Changed handling of blanked channels to use magic blanks and check for them etc. Should be more reliable. 4. Improved interpolation methods for the channels with no solution. 5. Found that the one-source, continuum method in which the SU, AN, and CL tables are corrected did not work. Required repairs to the addressing. This fix should go to patch. Moved 2,5 to patch 24 January. 13389. January 18, 2011 RLDLY Eric 1. Convert RMS computation to use double precision. 2. Change it to copy the GAINUSE CL table rather than the highest numbered one. 2. Added a phase correction which will line up the phases in some cases (when IFs solved separately or when they have not been given IF-dependent R-L differences by CALIB et al.). Moved nowhere. 13390. January 18, 2011 PCAL Eric 1. Made sure the calibration polarization array is initialized to zero. Moved nowhere. 13391. January 18, 2011 CLCOR Eric Added a correction of the Q and U in the source table when correcting the Antenna table D terms using values from RLDIF. No need to do RLCOR since there is no source table. Moved nowhere. 13392. January 19, 2011 FARS.FOR Leonia Right now FARS allows to put out the CLEAN components or residuals. Now the option of the sum of clean components and the residuals is added. The option of substituting the maximum amplitude (if it is smaller the given value) by the other given value is added if the option of 3 numbers sended to output is used. Moved nowhere. 13393. January 20, 2011 IMERG Eric Added tests to the code to make sure that the input images are powers of 2 in both axes. Cleaned up the code a little in some routines although this is very old, Cornwell-style AIPS. Added warnings to the help file about the power of 2 and the fact that the 2 images must agree in all details of their coordinates. Dropped unused adverbs BLC, TRC but added CHANNEL for the BLC(3) which was used for cubes. Moved nowhere. 13394. January 20, 2011 FARS Leonia Additionally to the FARS correction described at the message 13392, the substituded value equaled BLANK is allowed under control of input parameters. Moved nowhere. 13395. January 20, 2011 OOP Eric Subroutine OUVATT in UVDATA.FOR searches for the catalog number of the file after it is created. For some erroneous reason it searched for ISEQ=0 meaning to get the highest SEQ not the one that we were creating. It then attached the catalog number of the highest SEQ to the object. The next time the file had to be found, the wrong one was found. This caused SPECR to copy the tables to the highest SEQ file rather than the one it had just filled with data unless that one happened to be the highest SEQ. With different code sequences it could cause the new file itself to never be filled (except with 0's) and the old higher SEQ file to be overwritten! Moved to patch 24 January, many OOP tasks must be relinked. 13396. January 20, 2011 PCAL Eric Raise limit of cal source from 20 to 50 Moved nowhere. 13397. January 20, 2011 3C138 X band model Amy Corrected the position for the 3C128 X band model, which was wrong by about 0.25". Moved nowhere. 13398. January 21, 2011 FRING Eric The exhaustive search method suffered a serious error in which one of the parameters giving the number of channels per IF was divided by the number of IFs a second time. This led to correct answers for the first IF but wrong ones for subsequent IFs. Moved to patch this date. 13399. January 24, 2011 FARS Leonia The substitution of RM istead of RE,IM as it was done previously, is carriing out now. The HLP file is modifyied to send the details of the algorithm in the explain part. Moved nowhere. 13400. January 24, 2011 COOKE Eric Corrected the last changed date to match the truth which is now today since RLDIF has changed again. Moved nowhere. 13401. January 24, 2011 SWAPR Eric Ignored the DOACOR adverb - fixed to put that in common for UVGET. Moved to patch this date. 13402. January 24, 2011 TYSMO Eric The Flag version adverb was out of order in the common declaration causing issues for INEXT and FLAGVER. Moved to patch this date. 13403. January 24, 2011 help files Eric BPASS, CPASS, FRMAP, FRPLT, and POSSM inputs files spelled UVRANGE as UVRANG. This disturbs parseltongue although AIPS does not have any problem. Moved nowhere. 13404. January 24, 2011 SPLIT Eric Set the CL version to 0 when DOCAL is false for the HI file writing including the CL to FO conversion. Otherwise it could be using a non existent version. Moved nowhere. 13405. January 24, 2011 CALIB Eric More adverbs set for getting cal data rather than output data - in this case BCHAN and ECHAN. Fix to use all channels for output file. Moved to patch this date. 13406. January 24, 2011 Single/multi source Eric A real multi-source file must have a SOURCE random parameter and one and only one SU table. We have received single-source files with an SU table so the code cannot take an SU table as evidence of a multi-source. Changed MULSDB Added messages FNDPOL Fixed test although of no value GAININ Fixed test - decides SN vs CL on this test DAYFX Changed message - need new CL table and NX Fixed to do things right: ACFIT ANTAB BOXES CLCAL CLIP DBCON FACES FIXBX GETJY LISTR SETFC SHOUV SPLAT SPLIT UVDI1 UVFIX UVHOL UVMLN UVPRM UVPRT DEFLG KRING FRMAP FRPLT LOCIT POSSM UVPLT WIPER Moved nowhere. 13407. January 27, 2011 SYCOP Eric NEW TASK: averages SY table (and CD table) in IFs listed in APARM and puts those average results in IFs listed in BPARM. Moved nowhere. 13408. January 27, 2011 DOBAND Eric BPASET For DOBAND > 1, the logic for testing FREQID and SUBARRAY from the BP file was in error. If the FREQID was 1 in the BP table, but there was no FREQID random parameter, the BP values could never be used (FRQSEL is always -1 then). Added test that both FREQID and FRQSEL > 0 before one cares that they do not match. BPGET For DOBAND > 1, the logic testing FREQID and SUBARRAY demanded that both do not match before excluding a record. The record should be and now will be excluded if either do not match. The FRQSEL part was correct here. DOBAND=1 did the test correctly in BPASET and so had no problem. Moved nowhere. 13409. January 28, 2011 QUOUT Eric NEW TASK to prepare text input for RLDIF giving the polarization angle or Q and U as a function of frequency. It can use either a well calibrated CP or SU table (post RLDIF for that data set) or a spectral image cube containing Q and U. Moved nowhere. 13410. January 30, 2011 QUFIX Eric NEW TASK to find an RLDIF correction by comparing two Q/U data cubes, one well calibrated and one not. Provides an RL correction for the UV data which produced the not well calibrated Q/U cube. Moved nowhere. 13411. January 30, 2011 HUEWEDGE Eric NEW VERB, draws 2 wwedges. One is increasing horizontally and constant vertically, the other is increasing vertically and constant horizontally. The former goes in the hue image plane and the latter in the intensity image plane to generate a true wedge for TVHUEINT displays. Changed: POPSDAT.HLP and NEWPARMS.001 to create the verb, new help file HUEWEDGE and AU6 to implement the verb. Changed IWEDGE to have a PIXRANGE type argument and changed AU5C and TVDEVICE to have thios argument (as 0,0 => ignore). Moved nowhere. 13412. January 31, 2011 SPLIT Eric Added CP and SY as table types not to be copied. Moved nowhere. 13413. January 31, 2011 HUEWEDGE Eric Changed: IAXIS1 Made UNIQUE argument into integer DOZERO which is all it controlled with > 0 => zero full plane, < 0 => zero partial plane, 0 => no zero. AU5B Correct call to IAXIS1 TVFLG Correct call to IAXIS1 SPFLG Correct call to IAXIS1 WIPER Correct call to IAXIS1 TVDEVICE Correct call to IAXIS1 HUEWEDGE help file: add LABEL and GRCHAN adverbs AU6 Fix it to turn images back on as they were found. Added code to label wedges and image correctly and adjusted the default vertical position. Moved nowhere. 13414. February 2, 2011 IMFIT, JMFIT, SAD Eric Corrected all 3 to separate the header Clean beam from the actual Clean beam used in the given image plane. It reads the CG table and finds the beam for the closest frequency to the frequency of the current plane. That is used as initial guess for the widths and in the fitting. Total flux however requires the Clean beam of the header, For IMFIT and JMFIT it now lists the fit in Jy/headerbeam and in Jy/actualbeam. The tasks were correct so long as one understood that the units were Jy per header beam and allowed the fit widths to be any value including < the beam in the header. The main change was to give a better initial gues for small diameter sources from cubes. Using the correct Clean Beam does mean that the deconvolved widths are now sensible rather than failures. The special CfA output now uses the actual Clean Beam and corrects the noise and peak values to Jy/actualbeam. Moved nowhere. 13415. February 2, 2011 FARS Leonia FARS forced erroniously the outputs to zero, if BLANK is expected. The problem has been fixed. Moved nowhere. 13416. February 4, 2011 FARS Leonia FARS set out amplitude to one exit and zero to another one. Change it to send amplitude to one exit and phase to another one. Moved nowhere. 13417. February 6, 2011 FARS Simplify the concept of amplitude/phase additionally to RE/IM. Moved nowhere. 13418. February 7, 2011 helps Eric Changed RLDLY help file to drop the unused SNVER from the help (it was not in the inputs). Also added a couple sentences about the difference from FRING - RLDLY writes a new CL table not an SN table. Changed FITTP help file to include a paragraph of warnings about external packages understanding AIPS tables. Changed UVAVG help file to fix a typo in the Explain section. Added remarks about UBAVG and IMAGR time averaging. Moved nowhere. 13419. February 8, 2011 REWAY Eric Added option to use the incoming INTTIME random parameter as the incoming weight rather than either the actual incoming weight or 1.0. The results appear not to depend much on this option which means that the incoming weights are handled reasonably. Also changed all robust rms computations to use double precision which seems to be required. Moved nowhere. 13420. February 8, 2011 DISKU Eric Changed it to sort always on something, with number of files the default. Changed it to examine all the disk spaces it is about to print. They are then scaled uniformly to kbytes (as before), Mbytes, Gbytes, or Tbytes. Moved nowhere. 13421. February 10, 2011 VLANT Eric The help file stated that the task does not work for EVLA data. While once true, it is no longer the case. Changed to say this. Created VLA.ant.2011 and in the public ftp area also baselines.vlais11 so that old VLANTs will work on old VLA data. These files contain nothing but comments. Moved nowhere. 13422. February 10, 2011 UVFLG, GETANT Eric UVFLG made the assumption that all "null" antennas would have (X,Y,Z) coordinates of (0,1,1) meters. That is not the case if some antenna numbers simply do not occur in the AN file. The flagging on low fringe rate then did use antennas at the center of the Earth. Changed to avoid antennas < 100m from the center. A similar error was made in the shadowing routine. GETANT is supposed to fill parameter arrays for the current subarray. It does not init the variables before it starts so values for missing antennas are not changed. Bad idea - added initialization code. Moved nowhere. 13423. February 16, 2011 Misc Eric SCANLENG help file had wrong contents CLSMO had a HI file format that tried to do a character variable with F10.5. Moved nowhere. 13424. February 17, 2011 APCAL Eric Change the number of input tables allowed from 8192 to 32768. Changed meaning of negative ANTENNAS values - now can have list of omitted antennas. Corrected handling in the omit case of DOFIT - which is not then referenced to ANTENNAS. This occurred numerous places in the code and was wrong previously in the omit only 1 antenna mode. Moved nowhere. 13425. February 17, 2011 PCAL Eric Wrong subscript typo led to model fluxes not being set up correctly from the input adverb giving spectral indices. Moved nowhere. 13426. February 18, 2011 AFARS Leonia The NEW task AFARS that can be run after FARS to make the presentation of the results better. Moved nowhere. 13427. February 18, 2011 Robust rms Eric REWAY made it clear that double precision is a very good idea for robust attempts to determine the rms. Changed: QCLEAN rms in autoboxing, lower case letters, variable name too long FILIT rms in autoboxing, variable name too long, wrong parameter name for INSEQ -> wrong images FINDR rms is several modes FLAGR rms in several modes - 75 % more memory now for more times. EVASN rms in amp and phase difference; corrected useless handling of single-source files EVAUV rms in image rms and UV analysis, lower case GETJY Solution code for average flux SNFLG AMP opcode - find mean gain and rms RSPEC rms of image region, lower case comment characters Moved nowhere. 13428. February 19, 2011 AFARS Leonia A small error at the yesterday .HLP file has been corrected Moved nowhere. 13429. February 21, 2011 Double precision Eric LISTR: changed the matrix and list averaging to do the averaging and rmses in double precision. The LIST averaging did not previously use data weights. RLDIF: changed the LISTR-like section to double precision. Changed the multi-source/spectral section to use double precision and to actually blank the solutions in the solution finding code to match the testing for blanks done in the smoothing and extending codes. Moved nowhere. 13430. February 23, 2011 CALIB Eric Added another thing to keep track of whether we have data by polarization, IF, and antenna in each solint. Count things with no data as CNTSAD rather than CNTBAD and report the 2 separately. Fixed the counting to get all the ones with no data (the weights were 0 and -2 for various reasons). Moved nowhere. 13431. February 23, 2011 TYAPL Eric (1) Added option to avoid changing the data weights for EVLA (SY table) correction. (2) Added code to apply the Gains in the square for data taken before some date and to ignore the Gains after that. SY values were the (Pon+Poff)/Gain when (Pon+Poff)*Gain would have made more sense. We are switching to that on the "some date". (3) We cannot escape assuming that the data weights are equal to the integration time. Moved nowhere. 13432. February 23, 2011 AFARS Leonia The error near cut off the amplitude is fixed. Moved nowhere. 13433. February 24, 2011 AFARS Leonia A small mistake near BPARM. Moved nowhere. 13434. February 24, 2011 TYAPL Eric The units of the SY table have been corrected from a real Psum DIVIDED by a gain to a real Psum TIMES a gain - also Pdif and the recorded gains changed accordingly. TYAPL attempts to handle the previous form and for data taken 2001-02-24 and later will do the new form. Results appear promising for both weights and fringe visibilities close to Jy**(-2) and Jy, resp. Discrepancies may well be due to Tcal issues rather than TYAPL scalings. Moved nowhere. 13435. February 24, 2011 LOCIT Eric Add final plot to inputs comments and allow it to be done even if the ouput files are not being written. Gave it the option to plot all solutions or only the "good" ones. Imppoved the labeling a little. Moved nowhere. 13436. February 25, 2011 OTT tables Eric The definition of OTT tables includes the time as a double precision value in days. It has however been written with single precision on occasion. Changed DOTTV.INC, OTTINI, and TABOTT to trap this case and do the right thing. This caused LOCIT to get wrong answers for the EVLA station coordinates. Moved to Socorro right now, nowhere else. 13437. February 25, 2011 AFARS Leonia The sequence numbers of the two output files have been swaped, if they are not zero in inputs. Fixed. Moved nowhere. 13438. February 25, 2011 USERNO Eric The list in $DOCTXT/USERNO.LIS was rather out of date. The programs that prepared it were lost in the interim. So wrote a new stand-alone program $AIPNOT/USERNO.FOR and used it to update the USERNO.LIS. moved nowhere. 13439. February 28, 2011 RLDIF Eric Changed EXPECT angles to be correct values for 4C48, 3C138, and 3C147. They were half what they should have been. Also fixed bug that kept one from overriding a known source with one's own value for polarization angle. Moved nowhere. 13440. March 3, 2011 RSPEC Leonia The weight (WEIGHT=1/RMS**2) can be sent to the output now additionally to RMS under control of OPCODE. Moved nowhere. 13441. March 7, 2011 FARS Leonia The case of NO FQ table has been added. Moved nowhere. 13442. March 7, 2011 ACCOR.HLP Leonia Corrected the wrong statement for ACCOR application for Australian LBA. Moved nowhere. 13443. March 9, 2011 AFARS Leonia Formats of the output cubes has been changed. The both output cubes has dimension equaled one. The position of the amplitude maximum is sent to the first cube. The amplitude or phase are sent to the second cube under control of OPCODE. Moved nowhere. 13444. March 10, 2011 SNPLT Leonia SNPLT has plotted PHASE using the phase range -180,+180deg, but labeled the vertical axis at the range 0,360deg. As a result, some plots did not appear. The problem is fixed. Moved nowhere. 13445. March 16, 2011 PCCOR Leonia PCCOR has used minimum and maximum of available PC tones. Now, the 2 PC tones may be chosen for each IF under control of the input parameters. Moved nowhere. 13446. March 21, 2011 flags Eric Changed PFLG.INC to allow 30001 flags to apply to a single time, up from 6000. UVCOP raised from 120001 to 200001. Moved nowhere. 13447. March 21, 2011 Stokes in tables Eric The copying of tables has not taken into account any possible Stokes selection. The bad case is when RR is dropped. The tables should then have only one polarization and contain the L values - but they have retained two and R values were then used for LL data! Changed the call sequences and code of the following to select polarization: BLSEL BPSEL CDSEL CLSEL GCSEL IMSEL MCSEL PCSEL PDSEL SNSEL SYSEL TYSEL CLNULL Also fixed handling of BIF/EIF in case the values were not desirable. PCSEL was a mess that required simplification. Then changed the COPTAB subroutine used in calibration tasks to determine BPOL/EPOL and call with the changed call sequences. Also changed AVSPC Added SY and CD tables and made the COPTAB changes in the local COPTAB routine UVCOP Stokes selection not allowed but fixed call seq in local COPTAB routine - SHOULD THERE BE SYFSEL flagging SY table? SPLAT Changed like COPTAB UJOIN To local COPTAB routine, added CD and SY, changed call seq although no STOKES selection here Moved nowhere. 13448. March 21, 2011 TYAPL, SPLAT Eric Fixed weights to use the integration time random parameter if present rather than the input weights (avoids issues with channel averaging). Added 1.178 division in vis scaling for "the 3-level fringe rotation function." Fixed SPLAT which scaled by weight on random parameters that should only have been summed (i.e. INTTIM). Moved nowhere. 13449. March 22, 2011 FARS Leonia FARS has not followed the BLC,TRC calculating weights from the infile. It always read the weight file since the beginning. So FARS has worked correctely only if BLC(1)=0. The error has been fixed. Moved nowhere. 13450. March 23, 2011 SNPLT Eric The correction of March 10 (#13444) was incorrectly conceived. SNPLT failed to handle PC table phases correctly because at one critical point, the data type code is changed to special values just for PC tables. When other operations on SY tables were added, these special values were not updated at every place that they were used. Phase is in fact handle carefully to try to determine if -180 to 180 or 0 to 360 plotting would be better and the better on is chosen. Also increased the number of samples allowed using dynamic memory. Some EVLA SY tables are huge. Moved real correction to patch this date. 13451. March 23, 2011 TRUEP Eric Corrected: 1. Acquisition of source fluxes neglected the fact that the SU table has I,Q,U,V fluxes whilst only I is needed here. 2. Scaling by flux ignored the fact that BIF is not always 1. 3. Clear READ status after reading the data. 4. Fit spectral index to fluxes and use that to get flux by channel for SPECTRAL true. 5. Add APARM(4,5,6) spectral smoothing option. Moved nowhere. 13452. March 24, 2011 SPLAT Eric Changed: 1. Source names were used only as 12 characters (they are 16) in several places. All 16 are now frequently used. 2. The decision on whether to force a SOURCE random parameter was made prematurely. If one omit-this-source name was in SOURCES, it would make the wrong decision and actually create a 2nd random parameter. The data and SU tables would be fine but the NX table would have the wrong source number. 3. The recent change to the xxSEL routines caused a bad call sequence for CPSEL in SPLAT. It does not have polarization dependent columns. The same mistake (3) was made in UVCOP and subroutine COPTAB used by a bunch of FUDGE clones. Moved nowhere. 13453. March 25, 2011 FARS Leonia Using BLC(1), TRC(1) for selection of frequencies was incorrect, if FQ table is absent. The problem is fixed. Moved nowhere. 13454. March 25, 2011 TRUEP Eric It tested for data in one data set and not the other, but, for some reason restricted that test to the continuum case. Thus it attempted solutions when there were no data values from one of the two data sets, leading to NaN's in the answers. Fixed that test to do both line and continuum, with a message for the continuum case and a counter and summary message for the line case. Moved nowhere. 13455. March 25, 2011 RLDLY Eric The phase correction of change 13389 works approximately when DOIFS = 0 (all IFs with independent delays, phases). It did not get correct phases for DOIFS=2 however. Changes: 1. Added messages giving the phases 2. Added tests on whther there was valid data for an antenna when shuffling data around between 2 VLA "IFs" (AC and BD) and the AIPS IFs 1-8 and 9-16 (for example). This shuffling caused many of the answers for phase to disappear. 3. Changed frequency reference to be IFs #1 and # NUMIF/2+1 during computations for the 2 halves of the DOIFS=2 mode. Moved nowhere. 13456. March 28, 2011 UVRFI Leonia UVRFI has reproduced correct phases but amplitudes sometimes still rised under the expected ones at the frequencies of very high RFI. The option SMOOTH allows to improve the amplitude behavior, smoothing the amplitudes at the frequency channels. Moved nowhere. 13457. March 28, 2011 RLDLY Eric Yet another subtlety about breaking the IFs into 2 halves. SWT has to have a first-IF-group copy for use when filling out the returned values. Moved nowhere. 13458. March 28, 2011 BPCOR Eric Corrected the spectral index portion to read the correct frequencies in via CHNDAT and to compute the correction properly as [(Freq chan)/(Freq tilt chan)] ** SpecIndx/2. since BP tables are voltage gains. Cleaned up help file and added comments re BPASS being a better way. Added channel selection for bP normalization with inner 3/4 as default. Moved nowhere. 13459. March 29, 2011 QUFIX Eric Add estimate of uncertainty to printed display. Moved nowhere. 13460. March 29, 2011 UVCOP Eric Added SY table flagging based on the applied flag table. Changed SY, TY, and SN table flagging to be "opt-out" via UVCOPPRM(6) rather than "opt-in." Moved nowhere. 13461. March 29, 2011 SETJY Eric Changed handling of spectral index so for CALC sources one set is found for each source and appears appropriately in the HI file and messages. Previously only the last source appeared in the HI file. Moved nowhere. 13462. March 31, 2011 NOIFS Eric Found an error in setting the UVW scaling parameter. Moved nowhere - patch?? 13463. March 31, 2011 IM2CC Eric NEW TASK: converts an image into a multi-facet set of images with image values as Clean Components in CC tables. Main good use is for models presented as images in JY/PIXEL from casa and other such systems. Moved nowhere. 13464. April 1, 2011 SNPLT Eric The dynamic memory request could be small when there were rather few rows in the relevant table. Added a bunch of memory to avoid this issue. Moved nowhere. 13465. April 1, 2011 CALIB Eric The help file was enhanced with info about reading in source models. Moved nowhere. 13466. April 4, 2011 BDF2AIPS Lorant Added an extra option to select loading of particular calcodes only instead of all data. May speed up operational issues. 13467. April 4, 2011 AFARS Eric Changed the IO around so that the 1-pixel first axis is eliminated entirely. Cleaned up the header info, history. Sped up by eliminating excess max/min testing. Moved nowhere. 13468. April 5, 2011 OFMGET, LWPLA Eric Changed verb OFMGET (subroutine AU5E) to use the COLORS adverb to control whether all colors come from OFMFILE or only one color comes from OFMFILE. In the latter case, the other 2 colors may remain unchanged (COLORS > 0) or be zeroed (COLORS < 0). In test, I found that LWPLA was using a bad formulation of the 'LG' transfer function. I fixed that to be like the 'LG' used elsewhere and added 'L2' and 'N2' as well. Moved nowhere. 13469. April 8, 2011 AFARS Eric Restored the option to control the value used for balnked pixels with -1111 being a code for magic blanks. A separate value is given for RM and phase in BPARM(2) and (3). Corrected the IO which did not account for the omission of the excess axis properly. Changed the magic blank pointer to correctly show that the amplitude image is not blanked. Dropped more unused stuff from the code and added sensible defaults for the output CLASSes. Changed it to read real/imaginary as well as amplitude/phase input cubes. Moved nowhere. 13470. April 8, 2011 FARS Eric Added sensible default output classes. Changed help for this and to correct some typing that got out of alignment. Dropped USERID from code and help. Dropped the APARM(4)=2 output option entirely - AFARS will handle it. Changed handling to FQ table to (1) look in header to see if expected and require one if so, (2) use a correct FREQ axis rather than assume 5 (3) find the freqid coordinate value and the associated pixel by correct computations (no guarantee it has not been modified in some way since FQUBE or MCUBE). Got rid of some excess code. Moved nowhere. 13471. April 8, 2011 DOFARS Eric Wrote RUN file DOFARS.001 with help file to define a procedure that will run RSPEC optionally, TRANS both the Q and U cubes, run FARS, delete the temporary transposed cubes, and, if APARM(4) is not 2, transpose the output cubes, leaving both the zxy and xyz output cubes behind. Moved nowhere. 13472. April 11, 2011 SNPLT Eric Changed allowed range of SysPower gains since they are now much smaller than the old ~512 values. Fixed a comment. Moved nowhere. 13473. April 11, 2011 CookBook Appendix E Eric Added to wording to help with getting ASDM and using obit. Moved nowhere. 13474. April 11, 2011 GETJY Eric Fixed an error that used the previous IF when an IF was blanked in R polarization. The sums in the L polarization were done correctly. Moved nowhere. 13475. April 11, 2011 PCAL Eric The solution for spectral index did not take into account blanked fluxes (in case there are no data). Changed it to skip those with 0.0 flux and fill them in from the spectral index if one can be found. Also changed the fitting to look for any good IF rather than requiring IF 1 to be good. Moved nowhere. 13476. April 12, 2011 DOFARS Eric Dropped APARM(4) not equal 2 test from DOFARS.001 and fixed a comment. Cleaned up the FARS and DOFARS help file a little. Moved nowhere. 13477. April 13, 2011 COPTAB Eric The addition of polarization ranges did not make it into subroutine COPTAB for BP tables - leading to aborts if there were BP tables to be copied. Moved nowhere. 13478. April 13, 2011 TVFLG, SPFLG Eric Counters of points included must now be double precision floats since 2^31 - 1 of integer is no longer enough. Changed GRIDTB, UVIMG, TVFLG, SPFLG. Moved nowhere. 13479. April 14, 2011 OBIT in AIPS Eric Create verbs to run OBIT tasks: Added AUO calls to VERBS subroutine in AIPS, AIPSC, AIPSB. Changed POPSDAT, NEWPARMS to make verbs BDFLIST, BDFAIPS and adverbs ASDMFILE, NCHAN, NIF, CONFIG. Wrote new: AUO Verb subroutine - I see no way around it having to know the OBIT task intimately - a maintenance issue. BDF2AIPS New help file BDFLIST New help file Also new help files for the above adverbs. Moved nowhere. 13480. April 15, 2011 DSORC Eric Contained some strange code in which it did not correct the SU table for source number changes and did do all others. With the change to ALLTAB to refuse to copy NX tables, added special code to copy the NX table and let the source number column fixer solve the changed source numbers. The fix routine has to use a special column name for the SU table since that table uses SOURCE for the character string name rather than the ID number. Moved nowhere. 13481. April 15, 2011 NX table writing Eric RNXUPD wrote the new subarray number in the NX table when dumping records for other subarrays, i.e. when switching from subarray 1 to subarray 2. This messes up CALIB and causes it to omit the last scan of subarray 1 in self-cal. Fixed local copy of RNXUPD in SPLAT also. Moved to patch this date. 13482. April 15, 2011 IMAGR Eric Had an error addressing multiple channels when doing time averaging in a baseline-length dependent way. Led to smears of signal going out radially. Moved to patch this date. 13483. April 17, 2011 OBIT in AIPS Eric Changed AUO to handle the display of the log files better. PRTLIN needed more correct call (NACROS not current string length) sequence, initial page=1 not 0. Put in code to detect OBIT's long time string 'YYYYMMDDTHHMMSS ' and to squeeze it out of the display. Moved nowhere. 13484. April 18, 2011 SOUSP Eric Added DOCONFRM option - to ask if the fluxes in the SU table should be replaced by the fit values. They are listed before the question. Moved nowhere. 13485. April 18, 2011 file opens Eric There are file systems (lustre for one) in which information about disk files may not be entirely current due to the asynchronous way in which the macro information is kept. Changed: ZDAOPN.C If the file size is returned 0, issue message and delay 1 sec, and try again. If needed, repeat one more time. ZEXPN2.C If the expansion shows as <= 0 bytes, do the above. ZMSGOP Change ZDAOPN call to ZDAOPM. ZMSGXP Change ZEXPN2 call to ZEXPNM. ZDAOPM.C New version of ZDAOPN with no MSGWRT calls. ZEXPNM.C New version of ZEXPN2 with no MSGWRT calls. Moved nowhere. 13486. April 18, 2011 TVFLG, SPFLG Eric Got rid of annoyances: in both, change the Stokes flag when selecting the new Stokes to avoid flagging, e.g. RR when displaying LL. In TVFLG, drop the subarray question on ANTENNA-DT when one subarray was loaded out of multiple subarrays. Moved nowhere. 13487. April 20, 2011 CALCODE foibles Eric Data sets come with the same source having 2 CALCODEs. This may be deliberate and useful so it is not forbidden. But it can cause issues: SPLIT Changed CALCODE ' ' for a source to CALCODE='-CAL' SPLAT When not assembling a multi-source file, made above change. DSORC Previous "fix" can cause trouble! One should not update an SU table so that > 1 record has the same source number. Moved nowhere. 13488. April 21, 2011 PRTUV Eric Changed code to make use of NX table. It makes a list of vis record ranges in the data set that match the start vis, SOURCE adverb, and time range. This should make it start up faster and run better. Moved nowhere. 13488. April 27, 2011 SNPLT Eric Added code to read the table and find out how many records are needed for each antenna, setting up a pointer array into the bug data array. Now we survive a table with antenna 89 as the highest (and only) antenna as well as tables with very many records. Moved nowhere. 13489. April 28, 2011 RFLAG Eric New task - examines data in time for rms and in spectral channel for robust mean and rms. It can then plot the histogram distribution of time rms's and of deviations from the spectral fit in sigmas. It can, alternatively, flag on the basis of these. Changed POPSDAT and NEWPARMS to create adverbs DCHANSEL and DOPLOT. Made new help files for these. Changed GINIT to reserve #36 plot type for RFLAG. Moved nowhere. 13490. April 29, 2011 TYAPL Eric When using the integration time with SY tables to set weights, flagged data were unflagged. Bug dates to March 21, 2011. Moved nowhere. 13491. May 1, 2011 RFLAG Eric Corrected it to avoid wiping out a data set when creating plot files. Also add 3 parameters at the end of the adverbs to help EXTLIST. Changed AU2A (PLGET) to know about RFLAG and AU8A (EXTLIST) to list useful info about RFLAG plot files. Moved nowhere. 13492. May 1, 2011 PRTUV Eric The new NX table using code had a minor but serious error. Moved nowhere. 13493. May 2, 2011 SETJY Eric Put in a call to the calc routines with type = current+1 and added code to test if things changed by > 3%. If so, the previous fluxes are printed and the peak difference in %. Changed the numbering of models to be in the sensible time order sequence with Baars >= 4 in both Fortran and help. mention the testing of previous model in the explain. Moved nowhere. 13494. May 2, 2011 IMAGRPRM(18) Eric Changed QCLEAN and the help files for IMAGR and IMAGRPRM to have the code gradually reduce the maximum number of fields imaged at one time to at most 2. It examines the average and rms of the fields imaged at one go and when the rms < average/20, it reduces the maximum allowed by one with a maximum of 4 going to 2 in one step. Moved nowhere. 13495. May 3, 2011 DOOSRO Lorant DOOSRO (ie Do ORSO) is VLARUN for the EVLA with additional fringe fit and syspower table apllication. EVAUV to come. Will also work for simple RSRO, ie 2GHz OSRO in 2011 D-array. 13496. May 3, 2010 IMAGR, POPSDAT Eric Corrections: POPSDAT The DOPLOT adverb was shown as an array of size 0. ParselTongue could not handle that - changed to scalar. QCLEAN The scheme to decrease IMAGRP(18) was way to extreme. A revised scheme that might work has now been tested. VISDFT The address offset for Stokes was not initialized. This made an error only if Stokes I followed usage at Q, U, or V. Moved nowhere. 13497. May 4, 2011 FITLD Eric Changed the local parameter include to contain more of the parameters and put that in SUCHCK rather than declaring MAXSOU as only 1000 while it was 2500 elsewhere. Changed it to 5000 at the same time. Moved nowhere to patch June 24. 13498. May 4, 2011 TIORD Eric Mention baseline-order checks in the help file and move those messages to come out only at the end with a better format. Moved nowhere 13499. May 4, 2011 SPLAT Eric SOLINT was recorded with an F6.1 format in the history. F7.3 is minimal. Moved nowhere. 13500. May 5, 2011 IMAGR Eric Tweaked the algorithm adjusting IMAGRP(18) in QCLEAN. Moved nowhere. 13501. May 6, 2010 Fiddly bits Eric QCLEAN Changed the algorithm for IMAGRP(18) again so that 2 jumps to 4 on a restart so it can come down again SETJY Re-ordered and reworded some of the messages esp when flux has changed. Moved nowhere. 13502. May 9, 2011 PCAL Eric PCAL has never worked for CC models of Q and U which probably should be used for even 3C286 in A array. The pointers that offset the Stokes in the model computation routines were not set except for point models (SMODEL). The base Stokes pointer for the model computations could also have been wrong although that would be very unlikely. Also fixed an error in which it did not get some dynamic memory that it was going to use when DOLINE true in all cases of DOMODEL. Also fixed a place that required a PMODEL for single source neglecting that IN2NAME was specified. Moved nowhere. 13503. May 9, 2011 RLCAL Eric NEW TASK. The right minus left phase of the reference antenna may not be as stable with time as assumed by RLDIF. If there is a good model for the source Q and U, it is possible to solve for this difference as a function of time. The new task does this, averaging over all baselines, channels, and RL and LR* to get the IF-dependent correction within a SOLINT. That correction is applied to the left polarizations of all antennas at that time in a new SN table. Moved nowhere. 13504. May 10, 2011 PRTUV Eric The default handling of the scan list - when an NX table is not read - came before the needed parameter was known. Moved it to after. Moved nowhere. 13505. May 11-12, 2011 FLAGR, FINDR Eric Changed BPARM(10) to mean do robust averaging on the spectra of each baseline/IF/time set rather than simple averaging. Any further averaging over time is done by simple averaging methods. Changed the averaging over spectra to s subroutine call 5 places in FLAGR to implement the option. Channels omitted from the robust average at each baseline/IF/time are not flagged by this task - see RFLAG. Changed POPSDAT, NEWPARMS, AU2A for new adverb DOROBUST. Made new help file for DOROBUST. Moved nowhere. 13506. May 12, 2011 FUDGE Eric the history cards ended up with the task name in the format statement and then pre-pended in the history writing. Dropped the one in the format statements. Moved nowhere. 13507. May 12, 2011 POSSM Eric AGAIN! (1) In plotting tables the selection of Stokes pointers would have treated 'RL' and 'RRLL' as R only. (2) Then it looked at negative error codes (says if all R or all L or both are flagged) from TABBP (et al.) and skipped those records even if some the data were good and wanted. The later code can handle empty (all 0 and/or blank) data groups just fine. Affectd BP, BD, and PD tables. (3) POSSM made decisions about plotting Stokes from autocorrelations without knowing the STOKES the user actually wanted, at least when the user did not set STOKES. Added FNDPOL call ahead of any other use of the STOKES value. Moved nowhere. 13508. May 16, 2011 OBIT in AIPS Eric Added TPUT to all verbs, created new simple access to Imager. Will make fancier ones in time. Changed: POPSDAT Make new verb OBITMAP and adverbs FOV, NTHREAD. WHATSNEW As POPSDAT. AUO Create code for OBITMAP. Add call to AU2 for TPUT with OBIT verbname as an immediate argument. FPTRIM Trims leading blanks and trailing 0's from an Fm.n formated string. ZSHCMD An internal varialble was too short causing memory to be overwritten. OBITMAP Verb help file with full OBIT Imager help file as explain FOV New adverb help file NTHREAD New adverb help file BDFLIST Add TPUT equivalent info BDF2AIPS Add TPUT equivalent info Moved nowhere. 13509. May 22, 2011 DOOSRO Lorant Add a test to defend an array from over-running its bounds. Moved nowhere. 13510. May 23, 2011 DOFARS, AFARS Eric The run file DOFARS.001 had a missing semi-colon which cause a grammar error after the first TRANS. The help file for AFARS had dropped the BPARM adverb that got reinstated later. Added help info although the inputs section was quite adequate. Moved nowhere. 13511. May 23, 2011 OBIT in AIPS Eric Added PRTLEV adverb to OBITMAP. Changed meaning of DOWAIT some. For OBITMAP amd BDF2AIPS, which can take a while to run, 0 < DOWAIT < 2 means to echo any additions to the log file as they occur. This uses tail -f but stops when the ObitTalk process ends. DOWAIT=2 still waits and then echos the log file to the screen. For BDFLIST, all DOWAIT > 0 are set to DOWAIT=2 since it is fast and you may only want to look at one or two pages of the output. Moved nowhere. 13512. May 26, 2011 ZDCHIN Eric Changed the variables used to read NETSP to allow 1024 entries each 192 characters long. The Long lines were killing us. Moved nowhere. 13513. May 27, 2011 TVROAM Eric Changed AU5A to copy the LUTs of channel TVCHAN to all channels to be used in the TVROAM. Changed TVROAM to do IENHNS (black-and-white enhancement) on button A. Fixed help files TVROAM, ROAM (was obsolete), REROAM. Changed OTVROM in TVUTIL to copy the LUTs also (affects FILIT). Moved nowhere. 13514. May 30, 2011 PRTAB Eric Added adverbs ARRAY3, KEYVALUE, KEYSTRNG, DOINVRS to allow selection on a column variable's value, restricting the print to include only or exclude those rows matching the selected value. Moved nowhere. 13515. May 30, 2011 COMB Eric The BPARM(4,5) option to blank on S/N should have used the absolute vlaue of (S/N) since large negative signals are also large wrt the noise. Fixed this in code, HI, and help. Moved nowhere. 13516. May 30, 2011 SNPLT Eric Added option DOSCAN to plot scan boundaries on plots with time as the X axis. Changed POPSDAT and NEWPARMS and wrote a new help file to add the new adverb. Moved nowhere. 13517. May 31, 2011 OBIT Imager Eric Added 3 new verbs to give nearly full access to OBIT's Imager task. OBITIMAG gives multi-scale, calibration, and other IMAGR-like functions. OBITSCAL add self-calibration to this. OBITPEEL piles on peeling operations on top of all the rest. Changed: NVSSVZ.FIT Added to $AIPSTARS - NVSS catalog in a VZ table form for finding outliers POPSDAT Added verb OBITIMAG, OBITSCAL, OBITPEEL NEWPARMS Added verb OBITIMAG, OBITSCAL, OBITPEEL AUO Added code to implement all these verbs New help files for each of these 3 also. Fixed typos and things in BDF2AIPS and OBITMAP help files as well. Some aspects (NVSS catalog in particular) are not working just yet but put these back to get them off my plate for a bit. Moved nowhere. 13518. May 31, 2011 PLOTR Eric The help file contained residue from an older version that limited things to 10 types of data. Moved nowhere. 13519. May 31, 2011 OBIT things Eric Changed AUO to give CatDisk=0 in an attempt to get the NVSS catalog stuff to work. Mention output UV data set in OBIT*.HLP. Moved nowhere. 13520. June 1, 2011 Inbox archaeology Eric Changed: LTESS Explain confused people with references about MEM. Added sentence explaining that LTESS works with VTESS and the references are for the latter. Moved nowhere. 13521. June 3, 2011 Auto Index tables Eric Changed the automatic indexing of output data to use the input indexing if available in most cases. Changed DRNX.INC Add parameters to hold old times and pointers to use them GETNX New subroutine extracted from VPLOT reads in the NX table data. RNXGET Calls GETNX seting up the old info pointers and values for up to 20 subarrays. RNXINI Added one pointer initialization. RNXUPD Added code to use lists of times when present to notice the scan gaps. Also deals with scans not present in the output file at the beginning and later on in the file. Changed all tasks to add a call to RNXGET just ahead of the RNXINI call. The former needs the input file and the latter the output file info, so they have been kept separate. Tasks that were a bit different include DBCON RNXGET call uses output file info and so deduces that there is no input NX info. DBCON changes things enough the input info is not likely to be reliable. VBGLU As DBCON. SPLAT Changed local version of RNXUPD as above. UVRFI The RNXINI call was commented out but the RNXUPD and RNXCLS calls were not. This should have caused curious difficulties. Other tasks changed were AVSPC DIFRL FIXAL FLGIT FLOPM FUDGE NOIFS PHSRF REWAY RLCOR SPLIT SUFIX SWAPR TYAPL UVAVG UVBAS UVCOP UVFIX UVGLU UVLIN UVLSD UVLSF UVMOD UVMTH UVRFI BLCHN CALIB CVEL Moved nowhere. 13522. June 6, 2011 SPLAT Eric The wrong source number ended up in the NX tables when not assembling a multi-source file. Moved nowhere. 13523. June 6, 2011 AVSPC Eric Used the channel count and average of the first IF in building the header even when BIF was set to some higher number. It then did odd things to the FQ table channel and total widths. Fixed to use BIF and to correct the FQ table in a manner consistent with what earlier routines did to it. Changed the help file to emphasize that there is no channel selection in mode SUBS. Moved nowhere. 13524. June 6, 2011 FRPLT Eric 1. Increase allowed number times 2. Do a sensible default for start and stop times 3. One of the plot loops did not set up blanking correctly and so would always join blanked points to good ones. 4. A label had an extra ) character after the polarization. Moved nowhere. 13525. June 7, 2011 OBITSCAL Eric Changed info re SOLINT since 1 minute data with a solint of 1 minute gets solutions every 2 minutes (2 samples separate by exactly 1 minute and so are averaged). Moved nowhere. 13526. June 7, 2011 IMEAN Eric DROUND was declared double precision when it returns integer. This seemingly worked under GNU but not under Intel! Moved to patch June 8. 13527. June 8, 2011 FRPLT Eric Changed: 1. The page-full logic could not handle missing plots and the like. Used the correct POSSM as a model of how to upgrade the code. Previously it would actually over-plot. 2. The Stokes labels had a simple typo error for IQUV. 3. The X axis labeling did not take into account the plotting of multiple polarizations and/or IFs when APARM(9) >= 1. 4. The time binning assumed infinite precision in single-procision floating pooint computations of time. Changed it to use IROUND. 5. The axis labeling was on each horizontal axis when APARM(9) >= 1. This led to overlaps and confusion. Changed it to do every other plot when there are > 2. Changed also to label the axis type with the tick labels. Previously the type appeared on upper plots where no tick numbers appear and the numbers appeared on the lower with no type. 6. Added adverb FACTOR with the same meanings as POSSM. Made the plus signs larger for FACTOR=1. Factor > 100 -> scale by FACTOR-100 and do not connect amplitude points. 7. Add display of antenna numbers and messed a little with the position of interior labeling strings (antenna names, numbers, IF and pol.). Moved nowhere. 13528. June 8, 2011 POSSM Eric Changed FACTOR to match FRPLT above. Also messed a little with the position of interior labeling strings (antenna names, numbers, IF and pol.). Moved nowhere. 13529. June 9, 2011 Y2K Eric Y2KLOAD had a typo causing a bad test in HUGE in residual image 4. Y2KEXEC had one too few YES commands causing the test to quit instead of be very wrong. Moved nowhere. 13530. June 9, 2011 SETPAR Eric Raised the speed limit from 200 to 500. Moved nowhere. 13531. June 10, 2011 PRTAC Eric The list of tasks for Y2KH changed so it oculd not find a 'HUGE' job in the AC file. Moved nowhere. 13532. June 10, 2011 RFLAG Eric 1. Add adverbs BIF, EIF 2. Adjust cumulative plot of time rms to go only to the first 0 box. Moved nowhere. 13533. June 13, 2011 ACLIP Eric NEW TASK: CLIP but for auto-correlations. This means BASELINE, UVRANGE, PHSLIMIT adverbs get dropped and the flagged summaries change a lot. Updated info in CLIP.HLP to mention ACLIP and new flag count limits. Moved nowhere. 13534. June 15, 2011 RLDIF Eric Correct the computation of uncertainties to divide by sqrt of the sum of the weights, yielding one hopes the uncertainty in the answer as opposed to the variance in the data values. Moved nowhere. 13535. June 15, 2011 CALIB Eric Add DOAPPLY adverb to control whether the SN table is applied to a single-source data set with an output file written (> or = zero - so done by default) or not (< 0). Moved nowhere. 13536. June 20, 2011 RFLAG Eric 1. Changed task to use BPARM for all sorts of plot controls including plotting only parts of the X and Y range of things. 2.Changed binning approach - APARM(2) is now the expected time interval and a new time bin is started when a baseline that has already appeared in the current bin is encountered (or 1.1 * APARM(2) has passed). Added minimum expected rms (APARM(5)) used in spectral mode. 3. Changed time rms flagging to flag the full interval by default and the central point only when APARM(10)>=1. Moved nowhere. 13537. June 21, 2011 RFLAG Eric Changed the code to insist that medians be found from an odd number of points - the MEDI and MEDR histograms were assymetric about 0.0 otherwise. Also changed to do the spectral rms only on the central spectrum in each time group. This makes the count of sample much more sensible and saves some excess computing. Added a remark to the help file pointing out that the median estimator forces the histogram to have three large spikes, one at 0.0 and one each at plus or minus 0.675 sigma. These do not affect the flagging under any normal circumstances. Moved nowhere. 13538. June 23, 2011 ZMI2, ZFI2 Eric Changed the error message handling in these bottom level subroutines. C requires formats %lld, %ld, and %d for items longer than long, items longer than int, and int, respectively. On LNX64 at least long is 8 bytes now and the %ld case was not present. The numbers printed with %d looked like the internal variables had overflowed in 32 bits when it was only the printing of them. Moved nowhere. 13539. June 23, 2011 CVEL Eric Programmers got confused in the init subroutine about which variable actually went back to the calling program as the error code. The wrong one was used for many important things like BP file setup, causing the task to plow on in numerous hopeless circumstances. Moved to patch this date. 13540. June 23, 2011 NETSP follies Eric Changed to allow bigger NETSP entries: DISK.INC Raise size of each disk name to 192 (the limit in ZDCHIN) DSKASS Raise size of hollerith temporary variable GNETSP Raise size of line to be read from NETSP Moved nowhere. 13541. June 24, 2011 FITS Table read error Eric The subroutine ATCONV adjusts the list of table keywords for a couple of effects including the presence of ISORTORD. It did not adjust the list of character valued keywords so they no longer corresponded to the list of types, names, pointers, etc. Fixed ATCONV and changed the call sequence in FITLD, UVLOD, IMLOD, CALRD, and TBIN. Wht was this not noticed? Moved to patch this date. 13542. June 24, 2011 BDF2AIPS Eric Changed AUO to make BDFLIST and BDF2AIPS more accepting of logicals in ASDMFILE. If ASDMF(1) begins with either $xxxx/ or xxxx:, the value of the variable xxxx will be translated and the actual ASDMF sent to OBIT will be the concatenation of that translation, a '/', anything left in ASDMF(1) with trailing blanks omitted, and ASDMF(2). Modified help files for ASDMFILE, BDF2AIPS, BDFLIST. Moved nowhere. 13543. June 27 & 29, 2011 DOFARS Eric Changed temporary variable name INDOWA to DOWINAIT. This interfered with INDISK. Moved nowhere. 13544. June 29, 2011 AIPSLetter Eric Mid-year AIPSLetter: $DOCTXT/LET11A.{TEX,PS,PDF}. Moved nowhere. 13545. June 29, 2011 PRTAB Eric Changed to use RPARM and allow up to 7 tests to be performed and, if all are met, print or do not print the row. Moved nowhere. 13546. June 29, 2011 FGPLT Eric The spectral channel setting tested against JLOCIF instead of JLOCF for BCHAN and so used a big channel range when a narrow one was intended. Moved nowhere. 13547. June 29, 2011 UVHOL Eric Changed it to recognize changed formats for the EVLA and do its usual thing. Added ICHANSEL replacing CHANNEL for all the multiple channel data sets. Changed to handle W, QUAL etc based on the array name in the antenna file. Added CHWANT to make a mask of desired channels and DATAVG to average up those channels. Moved nowhere. 13548. June 30, 2011 RFLAG Eric Changed: 1. Setting of Ymin for cummulative plots - with LOG it was sometimes too high (at -5.0). 2. Changed to write a new FG table always, but to replace an existing table at the end if the user requested an OUTFGVER that already existed. 3. Added NOISE as an IF-dependent set of values for APARM(3) to be used when APARM(3) < 0.0. Implemeted by always using NOISE but filling it from APARM(3) when it is > 0. Handled HI properly however. 4. Changed DOPLOT > 0, to return NOISE(i) for i = IFs in current job, leaving the other NOISE values alone. The task gets this by summing the time rmses and their squares in each channel individually. The mean and variance in each channel is found and may be plotted (DOPLOT >= 4). In each IF the median average channel rmses and its variance are added and then scaled by FPARM(9). 0 -> 5 which seems good. 5. Changed APARM to FPARM and BPARM to VPARM to allow up to 30 control and plot parameters, respectively. Added plot scale controls for the spectral cumulative histogram and for the spectrum of the time rmses. 6. Added warning about SMOOTH in the help file. 7. Changed handling of times to find the min and max time within each time bin and use those appropriately in setting the flag time ranges. 8. Added a new variable to hold all the flag indicators for a given set of times and broke the big loop set into two sets. This opens up the possibility of checking across baselines and the like. Added FPARM(10) to flag all baselines if more than FPARM(10) (fraction) of them are to be flagged. 9. Added code to test for filling in channels between flagged channels and for too high a fraction of channels flagged ahead of the FPARM(10) baselines test as well as the code that was already there afterwards. Changed the too high fraction test to count both previously flagged and current about to be flagged. This means if the cutoff is set at 0.4 and half the channels were previously flagged they will all disappear. Moved nowhere. 13549. June 30, 2011 CALIB Eric The DOAPPLY adverb was applied to the writing of the output file but not to the creation of the file and writing of its HIstory file. Added further tests on APPLY to prevent it when it is not wanted. Moved nowhere. 13550. July 1, 2011 SPLAT, UVCOP Eric In order to flag SY, TY, SN tables changed NXTFLG Generally used subroutine needed TABLE argument to decide which flags to take and which to ignore. Channel-dependent and baseline-dependent flags are not applied to tables. DATFLG Added a .FALSE. to the NXTFLG call. SNSEL Did not use the local BIF where it should have. UVCOP Changed NXTFLG - the SOUWAN list is prepared differently by UVCOP and is always sources that are wanted. The usual flag about that said otherwise. This meant that a UVCOP with SOURCES did not apply any flags to UV data or to table data. Changed the xxFSEL routines to include the BPOL, EPOL argument even though it does not matter in UVCOP. Changed xxFSEL to make sure that the flagging and IF/polarization selection still left data which were not flagged before writing a table row. SPLAT Changed COPTAB to apply flagging to TY, SN, and SY tables. Copied the TYFSEL, SYFSEL, SNFSEL routines and their xxFLG subroutines from UVCOP. Note that SPLAT must use the general version of NXTFLG and DATFLG whilst UVCOP uses a special NXTFLG (more flags allowed) and does not use DATFLG. SPLAT Help file: mention tables in FLAGVER help section and the type of averaging done by SOLINT compared to the multiple choices in UVAVG. Moved UVCOP fix to its NXTFLG to patch this date. 13551. July 1, 2011 RFLAG Eric 1. The history writing task had a wrong format and variable for one of the VPARMS and a format one character too long for another. 2. Changed the axis labeling for the plot of rms versus channels. It had residue of other labeling routines that had not been updated. Corrected lack of use and misuse of variables to get consistent labeling. Moved nowhere. 13552. July 4, 2011 BDF2AIPS Eric The script still contained a x.i command to display the inputs. This has started to show a page full, press Enter to continue state, demanding the enter from the terminal. This is confusing in the extreme so I removed the x.i command. Moved nowhere. 13553. July 4, 2011 PCAL Eric Indeterminate solutions are usually caused by user error. However, the data could cause one in marginal cases. Changed the code to go on even in the presence of indeterminate solutions until 10 in a row occur. Also changed the low level messages so that the number of diagonal non-zero terms and the expected number are shown in such cases. Moved nowhere. 13554. July 5, 2011 SNIFS Eric NEW TASK. Sort of like SNPLT but the X axis is IF number for R followed by IF number for L. It plots CL, SN, TY, SY table parameters that depend on IF. It plots one time interval at a time and can do each antenna individually averaged over that time or plot each time sample in the appropriate panel. It can also plot all antennas each averaged individuall of the time interval. Then it goes on to the next time interval. Moved nowhere. 13555. July 6, 2011 OBIT Eric The command x.i for inputs did not work in the OBIT verbs at all. It did work from the script BDF2AIPS but left a page full wait with the prompt in the log file and the user in a quandary. Bill has made the command x.inputs(file=x.taskLog) which he says will work. I have put in (commented out still) in AUO.FOR but added it back to the BDF2AIPS script. Moved nowhere. 13556. July 7, 2011 UPDCVS Eric Improve the $cvscmd cannot execute error message to make it clear what it has tried. Moved nowhere. 13557. July 7, 2011 ANFIND Eric Some idiot somewhere is just making up names for telescopes. There is no way that one can recognize all the possible spellings for even commonly used antennas, esp if case is also a possible variation. But we can add more to the list anyway. Add trap for 0,0,0 for a station location. Moved nowhere. 13558. July 7, 2011 RFLAG Eric Changed: 1. Created new adverb SCUTOFF(64) in POPSDAT, NEWPARMS, and a help file. 2. Added SCUTOFF as an IF-dependent cutoff for deviations from the spectral fit. Changed code to set it with median of mean spectral deviations, scaled with FPARM(10). Put in HI. SCUTOFF is in Jy not sigmas as is FPARM(4) now. 3. Added DOFLAG adverb - if > 0, flagging will be done even if DOPLOT > 0. Moved code that sets the internal cutoffs to the actually upper flag routine (so after the plot side of the code that will compute NOISE and SCUTOFF). 4. Moved FPARM(10,11) to FPARM(11,12). 5. Changed FPARM(5) from a now meaningless lower limit to the spectral rms to a lower rms cutoff (like FPARM(6) is an upper rms cutoff). 6. Added array to sum the deviations from the spectral fit and squares on a per channel basis. Added code to plot it (bit 4 = value 8 in DOPLOT) and set SCUTOFF. 7. Fixed OVERFLOW display to show all over the boxes plotted and to show total count. Cumulative plots show total only. 8. When DOTV true and GRCHAN=0, the channels for line types 3 and 4 are cleared. 9. The channel plots are now labeled with channels counted within each IF. And the IFs are labeled. Moved nowhere. 13559. July 7, 2011 USUBA Eric Raised upper limits on entries and sources by an order of magnitude. Moved nowhere. 13560. July 7, 2011 CALIB Eric Changed HI writing so that the input file gets a little info even when a new file is being written. The date, the name of the new file, and the SN version number are probably enough. Moved nowhere. 13561. July 7, 2011 SNPLT Eric Changed axis labeling for PSYS - it was Tsys/Tcal but it has been TSYS Kelvins for some time. Moved nowhere. 13562. July 11, 2011 USUBA Eric Corrected a place that assumed there were no more than 20 antennas when making a printed disply. Put in a loop with a limit of 20 in any one line. Moved nowhere. 13563. July 11, 2011 RFLAG Eric I7 is not a big enough format for the count of flags! Changed to I12 which includes a blank and a sign plus the largest integer standard 32-bit integers can count. Moved nowhere. 13564. July 11, 2011 OBIT Eric OBIT does not like OUTDISK=0. Changed AUO to set it to the maximum disk number at AIPS startup (not affected by ADDDISK). Moved nowhere. 13565. July 11, 2011 SNIFS Eric Connected line plot for 2nd polarization was one off to the right. Moved nowhere. 13566. July 12, 2011 PBCALC Eric PBCALC had a maximum radius parameter that did not adjust for arrays that are not recognized. Changed it to set RCUT to an enormous number for unrecognized arrays for which beam parameters are provided. This affects a lot of tasks most notably PBCOR, but also SAD, BOXES, IMFIT, SETFC, OOSUB, IMAGR, VTESS, and similar tasks. Moved nowhere. 13567. July 15, 2011 RFLAG Eric Manage time differently a little. Expand the user's time interval by 0.1 percent rather than 10 per cent. Use the time in the test for too much time passed as the current minimum in the time bin rather than the initial time - it should be the same but ?? Changed the time array to (2,*) rather than (3,*) since the init time is not relevant. Expanded the discussion of FPARM(2) in the help file. Moved nowhere. 13568. July 15, 2011 REFLG Eric NEW TASK: compresses FG tables like the ones common from RFLAG, CLIP, and other such tasks. Has option to expand the flagging if too much of a spectrum is flagged at a particular baseline and time. Options to flag all times for a particular channel and baseline if too many are already flagged and to flag all baselines at a particular time and channel are already flagged are also available. Moved nowhere. 13569. July 15, 2011 FGDIF Eric NEW TASK: compares the affects of two FG tables to see if they flag the same data. Moved nowhere. 13570. July 19, 2011 TVFLG, SPFLG Eric These old tasks still used UVINIT/UVDISK to read the data set to determine the times to be used in the grid. This is probably okay if you are gridding the full data set, but does not work well if you are selecting data by time or source and does not reflect times that are already fully flagged. Changed to use UVGET applying all the selection adverbs. In many EVLA cases it may be 60 times faster in the "finding a list of times" section of the code. Moved nowhere. 13571. July 19, 2011 RFLAG Eric 1. Changed help file to allow NBOXES of 0 and show that <20 goes to 200. 2. Added FPARM(12) (and moved FPARM(12) to FPARM(20) for flagging only the center one) to control flagging an antenna when the fraction of baselines to it exceeds FPARM(12). 3. Changed logic (and parameter order) of big FLGITS array so that it is zeroed and passed from above. The lower level routine picks up the previous values for times not yet written to the FG table. Testing on fraction of baselines is now done only on the "central" time(s). Moved nowhere. 13572. July 19, 2011 REFLG Eric 1. Changed all excess tests not to count those cases in which everything was already flagged. 2. Moved CPARM(2-5) to (3-5,7) and added info to the help file. 3. Added test on excess baselines to an antenna as CPARM(6). 4. Added CPARM(2) to flag good channels between groups of flagged channels. 5. Added to help the reminder of how baselines are counted (including autocorrelations and with the maximum antenna number). 6. Changed to use UVGET to read the UV data file for times. Should be faster than UVINIT/UVDISK when getting 1 source which it always does. 7. Changed to delete the FG tables as needed in addition to removing them from the UV data file header. Moved nowhere. 13573. July 19, 2011 FGDIF Eric 1. Add indicator of which FG table had a flag when printing details of the differences. 2. Changed to use UVGET to read the UV data file for times. Should be faster than UVINIT/UVDISK when getting 1 source which it always does. Moved nowhere. 13574. July 20, 2011 DOOSRO Lorant Added a estimate for CHINC in FRING so that it won't fringe on all gazillion channels. Choice is to do either the max of one sample per half a MHz or limit to 32 channels to speed things up. Of course less if less than the max are available. 13575. July 21, 2011 REFLG Eric Added subroutine to sort the output FG table to a scratch table and then write the output over again working through the flags to see if any match in all but IF and allow a range of IFs to be specified rather than just one. It also checks to see if all baselines to an antenna are flagged and combines those flags. Results are disappointing. Moved nowhere. 13576. July 21, 2011 UVCOP Eric The option UVCOPPRM(5) would omit data records from the output if all correlators (all channels, polarizations, IFs) had low weight but did no flagging. It now flags all low weights and passes the record on if any correlator is not flagged. Moved nowhere. 13577. July 21, 2011 RFLAG Eric Added DOOUTPUT and OUTTEXT to write out a text file with channel weights based on the rmses of time rmses and/or frequency deviations. Cleaned up display code a little. Moved nowhere. 13578. July 25, 2011 IMMOD Eric Added option of input text file listing up to 9999 components. Moved nowhere. 13579. July 25, 2011 STFUN Eric Changed it to always write out an "rms" image also. Note that the Structure Function is already a differences squared so the statistics are not simple Gaussian... Moved nowhere. 13580. July 25, 2011 IMAGR Eric Changed QCLEAN: 1. CLSCAL will now write a message whenever rescaling by something other than 1.0. Previously, if the scale was exactly the previous one it siad nothing. 2. CLRMAX default of 5 for the inscribed circle (5 pixels less than the full image). 3. CLGFIT now uses the field for which a beam was computed most recently to provide the beam parameters. Previously it used each field even if ONEBEAM caused that field not to have a beam. As a result it read the header beam value and used it for all planes of the image in scaling, restoring, etc. It did write this bad value in the CG table but who looks? The new value does get into all facets from this point. 4. CLSCRS changed to display all rescaling not = 1.0 Moved to patch this date. 13581. July 26, 2011 CONVL Eric Changed it to: 1. At the start, see if the header Clean beam can be convolved to the desired width. If not and a CG file is to be used, then check to see if any channels in the CG file will work. If at least one will, CONVL goes ahead. It used to die. 2. If, at any channel it cannot find the frequency in the CG file or cannot find a Gaussian which will convolve to the desired width, it now complains and fills the output plane with magic blanks and keeps going. Previously, it convolved with the previous good channel's parameters (usually in a plane filled with blanks) or died. Moved nowhere. 13582. July 27, 2011 EXTLIST, RFLAG Eric Changed AU8A for the revised RFLAG for EXTLIST and AU2A for PLGET of RFLAG plots. Changed RFLAG tp pass back two more extra parameters giving the range of boxes in the plot. Moved nowhere. 13583. July 28, 2011 VPLOT, CLPLT, CAPLT Eric Raised the maximum number of model Clean components to 200000 in all of the above. The limit was 50000 in VPLOT and 25000 in CLPLT and CAPLT. Of course 200000 CCs will be very slow in the model computation.... Moved nowhere. 13584. July 28, 2011 TYAPL, DFLG.INC Eric Rearranged the flagging common to mke 2, one with parameters and one with the data. Changed TYAPL to allow FLAGVER and to have DATFLG and NXTFLG versions with the number of flag commands allowed at one time = 200001 like UVCOP. Recompiled a lot Moved nowhere. 13585. July 29, 2011 install.pl Eric Added code to recognize 64-bit Macs as MACINT. Moved nowhere. 13586. July 29, 2011 VPLOT Eric The routine that picked up the CC models had an error if NMAPS was greater than 64. A wrong variable was used in an IF statement so NCOMPS for field 66 was set to NMAPS and so all other fields were ignored. Updated VPLOT, CAPLT and CLPLT help files for the increased number of CCs allowed. Moved bug fix to patch. 13587. July 29, 2011 CookBook Appendix E Eric Revised Appendix E to suggest a quick early edit, temporary BPASS/CALIB/GETJY, then RFLAG and REFLG, then delete CL etc tables, then TYAPL applying the bug RFLAG flag table, then re-doing FRING, BPASS, CALIB, GETJY, then on to polarization. Corrected some of polarization also since fixed models do not work. Updated LaCook0, I, J the table of contents and indices. Revised the on-line html and linked PDF files. Moved nowhere. 13588. July 29, 2011 IMAGR Eric Added test on ALLOK before initializaing the baseline-length dependent time averaging. Moved that IF group to right after all possible changes of ALLOK to false. Now ALLOK will work with that averaging option. Moved nowhere. 13589. August 1, 2011 CookBook E Eric Pointed out that Hanning smoothing should be applied once and for all with SPLAT since it has gotchas with bandpasses and especially in the auto-flagging tasks like RFLAG. Adjusted COOKBOOK.PDF and web copies. Moved nowhere (31DEC10 gets it automatically). 13590. August 3, 2011 RFLAG Eric Changed the flag writing code and the code calling. It now does the spectral computation and and extension computation only on the time bin or bins (at the end of a source) that will be dropped following the call to the flagging routine. It then writes out the flags only for that time or group of times. This halved the number of flags written in one test without making much difference in what was flagged. Moved nowhere. 13591. August 4, 2011 RFLAG Eric Added option FPARM(13) to clip the data while entering into the histogram section of the code and also while entering into the flagging section of code. The data are flagged by the clip and the pointer to cause an FG entry to be written for the channel/time/BL/IF is set. Moved nowhere. 13592. August 4, 2011 CLCOR Eric Two significant places - opcodes 'EOPS' and 'SUND' - assumed that the antenna coordinate system is left handed. These did NOT get corrected in the fall of 2009 as they should have. Moved to 31DEC10 this date. 13593. August 8, 2011 VPFLG Eric VPFLG was actually changed July 14 but no CHANGE.DOC entry was made. That change added to the meaning of DOIFS so that > 0 meant flag all of an IF if one channel bad and flag all IFS if one bad whan DOIFS > 1. The logic to do this was flawed and has now been corrected. It appears that if DOIFS > 0, the code flagged everything during the past month! Fortunately, the counter of flags was also changed to double precision float. A grammar error in the inputs was fixed. Moved nowhere. 13594. August 8, 2011 REWGT Eric Added INTEXT option to take the OUTTEXT file from RFLAG and apply the weight factors once and for all. Moved nowhere. 13595. August 10, 2011 POSSM, BPLOT Eric Added to POSSM, Stokes = 'V' to plot the difference of the R and L bandpasses. Added to BPLOT, STOKES 'I' to plot the (R+L)/2 bandpass and STOKES 'V' to plot (R-L)/2 bandpass. Added to BPLOT for all table types plotting of real or imaginary parts or the difference of real or imaginary parts. Moved nowhere. 13596. August 12, 2011 RSPEC Eric If the last channel was blanked, the task would die rather than make the plot. Fixed the warning error code to 0 after heading the warning. Moved nowhere. 13597. August 12, 2011 UVHOL Eric Changed it to display real and imaginary by default with an option for amplitude and phase and to display the uncertainties in these means as well - all on HOLG. Moved nowhere. 13598. August 13, 2011 SQASH Eric Rewrote the algorithm The old one did scatter reads which hit the disk these days very badly indeed. Added weighted average (weights found by robust means in SQASH or from a text file), weighted and straight rms to the mean, sum, max, min operations there previously. Moved nowhere. 13599. August 15, 2011 RFARS Eric NEW TASK: It corrects the input cubes to FARS (or DOFARS) for the rotation measure at the maximum as found from the FARS output by AFARS. Corrected typo in SQASH extension file copying. Moved nowhere. 13600. August 17, 2011 FLOPM Eric Changed it to copy all times for the table copying and to copy the CL table if DOCAL false. Previously it copied the tables only for the time range of the correction (plus 15 minutes on each side) and surpressed the CL table (making a null one). Moved nowhere. 13601. August 18, 2011 MC table times Eric Despite indications otherwise, the times in MC tables are start times rather than mid-point times. Added an interval keyword to the MC table definition and corrected the indication about times. Changed MCINI Added DELTAT argument in days, stored as keyword with default 2 minutes. MCTAB Change time meaning pn precursor remarks. MCSEL Added parameter to MCINI call sequences GETMC Added DELTAT as call argument, corrected comments, Changed TRANGE to be DELTAT if > 0, else 2 minutes and change usage so that times in MC table are the start of intervals. TABLEUTIL Made corresponding changes to the OMCINI, OTABMC, and OGETMC pass-through routines. FITLD Add a DELTAT to the calls to MCINI, GETMC. FXPOL Add a DELTAT to the calls to MCINI. INDXR Add a DELTAT to the calls to OMCINI, OGETMC and carry it around in a common. Moved nowhere. 13602. August 19, 2011 Memo 114 Eric Changed AIPS Memo 114 for changes to MC table (DELTAT keyword) and TIME is at start not middle of interval. Moved nowhere. 13603. August 22, 2011 TVFLG, SPFLG Eric They did not clean up after long source names in the CURVALU like display since the blank string used was only 11 characters long. NOw it is 16 which is the upper limit. Moved nowhere. 13604. August 24, 2011 TVFLG Eric Added FLAG BASELINE-DT operation to the Fortran and to the help (explain) and real-time help files. One selects the baseline and time range with the cursor and can do multiple such baseline-dt's before returning to the main menu. Moved nowhere. 13605. August 24, 2011 RFLAG Eric Corrected the labeling of the spectral plots - it was slightly off. Also changed it to put a full channel blank in between each IF. Previously there was no blank space except at the start and end of the full plot. Moved nowhere. 13606. August 25, 2011 TVFLG, RFLAG Eric TVFLG - lost a useful command when there was only 1 spectral channel since I added a command to that column and did not change the column-length correction for single channel. RFLAG - the spectral plot ranges required that the first VPARM of the pair be >= 0.0 and < the 2nd of the pair. Changed the first test to be < Ymax of the plot. Moved nowhere. 13607. August 26, 2011 TVFLG Eric I made things worse yesterday with column 4 losing all its options on about the 7th command. Fixed that and a pointer into the clip function as well. De Sade's law again. Moved nowhere. 13608. August 27, 2011 DOOSRO Lorant Noticed that the channel averaging logic for FRING that I put in last month had twisted logic. Tried to untwist it here. Moved nowhere. 13608. August 30, 2011 Returned adverbs Eric Waiting for adverbs should not require the task returning them to finish if it can return the adverbs sooner. (RQUICK is DOWAIT.LE.0.0.) Changed: AU2 Drop the section that forces DOWAIT TRUE when there are return adverbs INPUT Add local method OUTQAV to return adverbs and resume AIPS when RQUICK is true. FINDR No speed up here - must add call to OUTQAV ahead of the OUT2AV call to handle the 2 possible values (now) of RQUICK. SLFIT Changed it to force QUICK false since there can be no speed win here. SETFC NFIELD is not known until the end so force RQUICK false and do usual resume at end. RLDIF Drop immediate resume on RQUICK. Put quick resume and return 0.0's when spectral line and/or doapply are true and DOCRT <= 0. Force RQUICK false when we are returnig real values in CLCORPARM. SOUSP Drop immediate response to RQUICK and set RQUICK false if confirming a change to the SU table. If the SU table is to change, the terminal is needed to confirm this, so nothing can happen until the task is finishing anyway. Send the adverbs back and resume AIPS if RQUICK is still true after the fitting routine but before the plot routine. GAL Drop immediate response to RQUICK, set RQUICK false on errors, restart AIPS with parameters on RQUICK right after the fitting before plotting, printing, and output image. EVAUV Drop immediate RQICK response, set it false on error, return answers before doing plots if RQUICK. Moved nowhere. 13609. August 30, 2011 RFLAG Eric Changed for above and to correct minor errors: 1. In MAIN routine, force RQUICK false on errors and branch to close calls. If RQUICK, do a call to return adverbs and resume AIPS after the NOISE and SCUTOFF values are known (after plot before flagging). 2. In init routine, resume AIPS on get parameter failure only, ignoring RQUICK initially. 3. Changed test on ignoring of CL or SN to check only that applying to the multi- or single-source nature of the data. 4. Moved the progress messages above the branch back on UVRANGE so that they will appear even if there is a UVRANGE. 5. Corrected typo in UVRANGE test causing them to fail. 6. Dropped a premature change of UVRANGE units internally that could cause failure in the flagging. 7. Moved the computation of end points for the NOISE and SCUTOFF line plots so that it would occur even if the current line does not fit on the plot. Moved nowhere. 13610. August 30, 2011 IMEAN Eric Changed PIXSTD, PIXAVG to solely output adverbs. Changed to get the initial guess for these either from the header keywords or from a robust fit to the image (or a sub-image if the image is too large). Changed to ignore RQUICK at the start, to make it false for error branches, and to return the fit as soon as it is available if RQUICK is true (adverb DOWAIT false). Moved nowhere. 13611. August 30, 2011 IMFIT, JMFIT Eric Wrote new WaWa IO routine TINBEG to start an interactive task ans force DOWAIT to true. A more subtle usage is not needed for IMFIT and JMFIT. Changed IMFIT and JMFIT to use TINBEG and to do printing on the CRT with DOCRT > 0 and line printer with DOCRT < 0 and no printing with DOCRT=0. Moved nowhere. 13612. August 30, 2011 CALIB Eric In working on OOCAL, realized the impossibility of DOCALIB true for single-source files - two SN tables which cannot be combined to write any output file. Added notes in the help file and code and a message in the Fortran forcing DOCALIB to -1.0 for single-source files. Moved nowhere. 13613. August 31, 2011 IMAGR Eric Adjusted the wording in the help file regarding the restoration of multi-scale Clean components. Moved nowhere. 13614. August 31, 2011 IMFIT, ISPEC, etc Eric Changed DOCRT in IMFIT, JMFIT, ISPEC, RSPEC and created this adverb in POPSDAT and NEWPARMS. Initial value 0 => no print. New help file also. Moved nowhere. 13615. September 1, 2011 UVAVG Eric UVAVG computed subarray number with an exagerated sense of the accuracy of floating point numbers and so got the wrong answer a lot. Furthermore, it did not keep track of subarray properly in the averaging routine and so did not average any subarray but the first encountered. The calibration adverbs would not work for all subarrays anyway. Corrected the first two errors and changed the code to loop over subarray. This may result in a data set no longer in time order. Moved nowhere. 13616. September 1, 2011 RFLAG Eric changes: 1. Error in the init routine skipped restarting AIPS - fixed to set RQUICK and branch correctly so that restart happens properly. 2. Clip by FPARM(13) mixed time subscript methods. Wrong time was marked in the to be flagged list. 3. On time RMS clip, set unflagged to be flagged but not previously flagged ones. Also if no samples valid in the channel/IF time group, set only previously unflagged to be flagged (if this matters). 4. On failure of the fit in spectral flagging, set previously unflagged channels to be flagged - they were simply marked as previously flagged which could leave them as unflagged. 5. Drop a debugging parameter. Moved nowhere. 13617. September 1, 2011 REFLG Eric Added a spread of the start and stop times of the flags by 0.1sec. Fixed a status issue in cleaning up the header at the end. Moved nowhere. 13618. September 2, 2011 TVFLG, SPFLG Eric Changed to give times to 0.1 sec in all displays when the interval is < 10 sec. Wrote new subroutine TIDHMS to support this. Moved nowhere. 13619. September 2, 2011 UVCOP, PRTUV Eric Created subroutine FINDVO to make a list of visibility ranges within a data set based on the index table with source, time, subarray, and freqid selection. Dropped FNDVO (the source of FINDVO) from PRTUV and changed it to use FINDVO. Changed UVCOP to use use FINDVO to make life easier and faster when possible. Moved nowhere. 13620. September 5, 2011 VPLOT Eric VPLOT used LUN 30 when doing its flagging option. This collides with usage in the calibration package when applying calibration and/or flagging. Changed the special option LUNs to 70 and 71 which should be safe. Moved nowhere. 13621. September 6, 2011 RLDIF Eric Changed DOCRT to DOPRINT since 0 means none like the other DOPRINT usages but unlike the usual DOCRT usage. Moved nowhere. 13622. September 6, 2011 CookBook Eric Slight upgrade of wordings in Chapter E; updated COOKBOOK.PDF with new version. Moved nowhere. 13623. September 7, 2011 ISPEC Eric Added OPTYPEs for LOG10 and LN in both flux and average. Moved nowhere. 13624. September 7, 2011 IMEAN Eric Corrected typo which made the robust guessing of the mean and rms ahead of the fitting for it failed unless there were more than 2 planes on the 3rd axis (L3 was just 3) Moved nowhere. 13625. September 8, 2011 XMOM Eric Added OPTYPE = 'MAX' to write out images of the value at the maximum in each row and of the coordinate at that pixel. Moved nowhere. 13626. September 8, 2011 IMAGR Eric Changed IMAGR.FOR (and help) so that SUBARRAY=0 will actually work as advertised when DOCAL and friends are used. There are many places that open the UV data set even before it is read to copy to the imaging work file. The request SUBARRAY=0 had to be saved in a parameter passed around and overridden until it could be invoked under loop control. the logic of ALLOKAY also had to be rearranged in the code to allow the looping over subarray. Anyway, it works I think. Moved nowhere. 13627. September 9, 2011 Stars Eric (1) Created new adverb STVERS for STar position version number. Changed POPSDAT and NEWPARMS to create it and STARS, CNTR, GREYS, KNTR, PCNTR, and PROFL to use it. STVERS < 0 => no star plot so the Fortran as well as the help had to be modified. Changed AU2A for DEFAULT value -1. Also new help file. (2) Created missing STINI and TABST service routines. (3) Changed STARPL and STARS to use STINI and TABST. Moved nowhere. 13628. September 10, 2011 JMFIT, SAD Eric Changed IMFIT, JMFIT, SAD to offer the option of writing a stars file attached to the input file. The "stars" are ellipses recorded in a new or pre-existing ST table and may be plotted with TVSTAR and many tasks such as KNTR, GREYS, etc. Moved nowhere. 13629. September 12, 2011 REFLG Eric Added SOURCES and CALCODE to limit the number of sources that have to be read. Changed also so that if there are no records in the scratch area, the number of sources is set to zero. The extra flagging work can help with generic flags too but an error had to be corrected. It should and now does ignore flags that are already antenna dependent rather than baseline dependent. Moved nowhere. 13630. September 13, 2011 FGDIF Eric Found several errors: (1) did not handle IFs=n,0 correctly (is n,max) when reading flags. (2) compted baseline number wrong in one loop over all antennas to a single antenna Changed to adverbs SOURCES, CALCODE, and CPARM(4), the latter as the maximum number of details differences to be printed (with 100 as default). Moved nowhere. 13631. September 13, 2011 IMEAN Eric Changed format in output text file to simply print the double precision counters even if 0 with an F14.0 format rather than do an integer conversion and I9 (good to 1 billion only). Moved nowhere. 13632. September 13, 2011 TVHELIX Eric Added DOPRINT in AU6 to control the printing of the current parameters, the printing of which was added to TVHELX. Also corrected a couple of errors pointed out by Dave Green and added a reference to a published paper on the algorithm. Moved nowhere. 13633. September 14, 2010 FRING Eric The EVLA option (APARM(5)=3) was setting the multi-band delay in the output which is not correct. Only the APARM(5)=1 or 2 should set it. Moved nowhere. 13634. September 14, 2010 IMEAN Eric Changed to compute the histogram even if there is no intention to plot it. This lets the print routine function as intended. Corrected incorrect vaiable name, long variable name, and lower case comment characters earlier. Moved nowhere. 13635. September 19, 2011 RFLAG Eric 1. Added a parameter to remember whether an outout table is time sorted and to mark it as such at the end. It is very unlikely to be of any use. 2. Put test for any valid values in routine that finds the medians of the channel rmses. 3. Rearranged code so that the flags will be in time order. This means that at scan ends the flags will still not encompass the time range of the current list of times. Moved nowhere. 13636. September 19, 2011 PRTAB Eric Added more control over formats for number and rows and row number. Changed the scaling routine to read no more than 5000 table rows. Moved nowhere. 13637. September 19, 2011 Table sorting Eric Changed: ZMIO Allow IO ops up to 128 Mega words ITBSRT ICSORT can do up to 204800 items in a list. Fixed to know this rather than the old 1024 limit. TABSRT Put in an upper limit to the buffer size that is acceptable to ZMIO for use in OTBSRT. Moved nowhere. 13638. September 19, 2011 REFLG Eric The input FG table does not need to be time sorted so dropped that sort. Changed the other use of TABSRT to use a large dynamic memory, potentially large enough to hold the whole FG table in RAM or at least be the max size of the pseudo AP. Corrected a bug handling compression to antenna-based flags which made something of missing antennas. Made sure the output FG table number of rows was controlled by the final write. Old rows were being kept at the end. Changed advice on CPARM(1) to slightly less than the normal interval and dropped a scaling by 1.001 in the code. Changed to issue UVGET ('CLOS'..) always - READ statuses were building up when sources are missing. Moved nowhere. 13639. September 19, 2011 FGDIF Eric Dropped both uses of TABSRT - the input FG table files do not need to be time sorted. Changed the detailed display to show times to 0.1 sec. Changed advice on CPARM(1) to slightly less than the normal interval and dropped a scaling by 1.001 in the code. Moved nowhere. 13640. October 3, 2011 IMEAN Eric Corrected a call sequence issue to MINIT - the full image size was needed not the subimage size. Moved nowhere. 13641. October 3, 2011 TABSRT Eric Dropped buffer and buffer size from call sequence and added KEYSUB(2,2) for subscripts into KEY columns. Added dynamic memory allocation for work buffer. Changed TABSRT call sequence and to give null KEYSUB: BPASET CALADJ CCMERG CLSORT FRQUPD IMSORT SDGET UVGET TABLE(TBLSRT) ANCAL ATMCA BPSMO CLCAL CLSMO FITAB FITLD FQUBE INDXH LISTR MFPRT PCLOD REFLG REWAY TASRT TYAPL TYSMO UVCOP BPASS CALIB CVEL FRING KRING SDVEL APCAL BPLOT IBLED SPFLG TVFLG WETHR Other changes: CALADJ corrected subroutine name in error messages CLSORT new call sequence but no one calls it so forces no other changes FRQUPD Drop WORK and WRKSIZ from call seq IMSORT Drop WORK and WRKSIZ from call seq UVFIX Correct call sequence to FRQUPD, drop WORK buffer. BPASS Correct call sequence to FRQUPD. FITLD Correct call sequence to IMSORT INDXH Correct call sequence to IMSORT, change buffer to size 512 as needed for tables. Made siomilar changes dropping variables and buffers if they are no longer needed. Moved nowhere. 13642. October 3, 2011 UVSRT, TAPPE Eric UVSRT Still had the 1024 limit of initial sort. Performance is improved with more memory and more sorting in the initial pass. Drop local copy of ICSORT. TAPPE Changed to do the sort on the two input files if needed. Then they are read and merged in the sort order with a check to drop duplicates when the sort keys are equal. Moved nowhere. 13643. October 3, 2011 OFMLIST Eric Corrected the formats in OFMLST to manage the higher range of input numbers now available. Moved nowhere. 13644. October 4, 2011 SDIMG Eric Drop local copy of ICSORT and change it to know that 204800 is the maximum number of sorts it can do. Changed image limit from 4096 to MAXIMG. Moved nowhere. 13645. October 4, 2011 TECOR Eric TECOR would die if a source number was mentioned in the CL table and not in the SU table. This can arise from e.g. SPLAT with source selection. Changed it to blank the output. Moved nowhere. 13646. October 4, 2011 TVFLG, SPFLG Eric If the Stokes type is changed on reading the data, such as from RR/LL/RL/LR to I/Q/U/V these tasks would prepare flags as if the data were in the translated form. This would not achieve the desired result. Rather than get fancy, made any translated flagging flag all correlators at all times. Moved nowhere. 13647. October 4, 2011 SETJY Eric Changed the cutoff for display old parameters. Moved nowhere. 13648. October 6, 2011 DOFARS Eric The weights file name had an inadverntant comma. Changed to a period. Moved nowhere. 13649. October 7, 2011 RFLAG Eric Added FPARM(14) option to flag any channel/IF/time with fewer than F(14) baselines on input to the algorithms. This is sort of a "quack" operation. Moved nowhere. 13650. October 10, 2011 TVCOLORS Eric Added values to enter when TV is missing. Moved nowhere. 13651. October 10, 2011 SL2PL Eric Allow PIXRANGE to be 200 times larger than in the data - up from 50 times. Moved nowhere. 13652. October 10, 2011 NOIFS Eric The COPTAB routine was not updated when several xxSEL subroutines had BPOL, EPOL added to them. Moved nowhere. 13653. October 10, 2011 PCAL Eric Changed it to copy with cal and flagging a full IF to a scratch file and then to read that file for the individual channels data. Should be faster... Moved nowhere. 13654. October 11, 2011 TABSRT Eric TABSRT ran into trouble destroying the scratch files it had made since a buffer of 256 words had become only 2 for dynamic memory usage. Made it 256 again. Moved nowhere. 13655. October 12, 2010 PCAL Eric Wrote subroutine CATFIX to force the status of a file to something desired or to e.g. NOT READ but allow no status or WRIT status. PCAL: 1. Changed default SOLINT to 5 minutes 2. Added calls to FIXCAT to defend against READ status when needing to create a PD and a CP file. 3. Added PCLNDX subroutine to examine the NX file to see if there is only one scan with the requested SOLINT. In that case a SOLINT is forced equal to 1/3 of the length of the scan (seemed to be better than 1/2). Moved nowhere. 13656. October 13, 2011 SETJY, SOUSP Eric Changed SETJY to use APARM(4) as the channel to which the fluxes should be referenced in CALC mode with default (1+Nchan)/2.0. The frequencies used will be displayed and put in the HI file. Different ref channels may be used by running SETJY with different BIF, EIF, APARM(4). Changed SOUSP to read the HI file to find the real frequencies to use with the SU data. It takes the last value found for each IF. Moved nowhere. 13657. October 13, 2011 CALIB, BPASS Eric Added calls to CATFIX in CALIB and BPASS just ahead of their creation of output tables to make sure there are no READ flags on the data set. Fixed infinite loop in the history file writing of GAINERR - only done on GCON mode solutions in CALIB. Moved nowhere. 13658. October 16, 2011 DOFARS Eric RSPEC now uses DOPRINT rather than DOCRT. Fixed procedure DOFARS to use the new adverb. Moved nowhere. 13659. October 16, 2011 FARS Eric The initialize routine was very confused as to its error return variable and got it rather wrong in some cases (such as missing weights file). Moved nowhere. 13660. October 16, 2011 POSSM Eric SOLINT affected BP plots but not in any deliberate or good way. Fixed code to set SOLINT=0.0 internally on BP plots. Fixed error messages to say the right subroutine name. Moved nowhere. 13661. October 17, 2011 QUACK Eric Changed the default time adjustment from 5 to 0.5 sec. Moved nowhere. 13662. October 17, 2011 IMAGR Eric Added adverbs to control the TV timeout lengths (first time, later times) as IM2PARM(8 and 9). Implemented in QCLEAN including TELL. Moved nowhere. 13663. October 24, 2011 UVFIT, DVDMIN Eric DVDMIN: cleaned up parameter formats so that lines do not wrap UVFIT: Added INLIST option to read in up to 20 components with initial guesses and whether to hold fixed or solve for each of the parameters. Each line has a spectral channel first and only those lines matching BCHAN are used in any execution. Raised number of allowed vis by a lot. Improved formatting of output. Added default GMAX of 1.0 - the task did not iterate when it was 0 previously. Moved nowhere. 13664. October 24, 2011 ELINT Leonia ELINT has printed out all 3 coefficients of the fitted polynomial at the each plot, but only one (the first) coefficient at the display. It's fixed. Moved nowhere. 13665. October 24, 2011 ZFIO Eric I messed up a delicate balance where ZFI2 was used by message write as well as other things and the message was passed back to higher routines that separated them. It needs a 2-line message so made a ZMSFI2.C for messages and changed ZMSGDK to call it. Moved nowhere. 13666. October 25, 2011 TVFLG, SPFLG Eric Found bug in GRIDTB and GRIDTC which assumed that ICOR0 and NCOR were known for the input data although with the previous use of UVGET (to find times to grid) they now applied to the output. This messed up the actual Stokes displayed - although the task correctly told what it was displaying. You ask for V and get RR. Force correct setting of parameters for the input file through an additional UVPGET call. Moved nowhere. 13667. October 26, 2011 ISPEC Eric Added OPTYPE = 'DERI' with adverb FACTOR to plot the per channel difference of a FLUX plot rather than the FLUX plot itself. Useful for Zeeman plots and the like. FACTOR scales only DERI plots. Corrected all operations so that smoothing is done before the conversion to log, ln, or derivative and the finding of the default plot scale. Conversion to log was done ahead of smoothing previously - probably with curious results. Moved nowhere. 13668. October 26, 2011 CLCOR Eric The code had a "pause" statement in it! Antique math code is really rather bad Fortran. Moved nowhere. 13669. October 27, 2011 CLCOR Eric The subroutine that used the STNHT function declared it to be REAL when in fact it is DOUBLE PRECISION. Changed that a couple of other variables to DOUBLE. Moved nowhere. 13670. October 27, 2011 BPLOT Eric Changed ANTENNAS so that a negative means exclude the list. Changed the plotting which was trying to count the number in a given plot and getting confused by single times being blanked. This does mean that antennas that are all blanked will appear in 'TA' plots unless excluded via ANTENNAS. No plots will be generated that are all blank however. Moved nowhere. 13671. October 28, 2011 TYAPL Eric The routine that went after new SY (and TY) data from the tables had a parameter designed to reduce the work. This caused it to fail to get needed new data when there were gaps in the table for some but not all antennas. These can arise due to flagging which is now applied to the SY and TY tables. This bug was found when it made Nan's in the output for one time - but it incorrectly applied the SY table over a substantial period up to that time. Found a couple old debug messages still in the code and several typos that could have affected the portion of code that undoes a previous correction in order to do a new one. That portion is probably rarely used. Moved nowhere. 13672. October 28, 2011 ISPEC Eric Changed DERI into ADER and FDER which scale as average and flux functions. Moved nowhere. 13673. October 29, 2011 REWAY Eric Added BPARM(1) to set weights to 0 if fewer than BPARM(1) samples occur in the rms computation. Moved nowhere. 13674. October 31, 2011 UVFIT Eric The task used the minimum offset and minimum size to determine prefixes and so used nano-asec and overflowed the output formats. Changed to use maximum. Added FITOUT option to write a text file with a compact but complete display of the fit results. Moved nowhere. 13675. October 31, 2011 Gridded subtraction Eric For Halloween - changed limit on doubling of grid from 4096 to 8192. Gridded models can now be 16K on a side. Subroutine GRDAT sets this and FORCED reflects it in choosing DFT for really large images. Moved nowhere. 13676. November 1, 2011 CALREF failures Eric Users are worried about messages from CALREF about it being unable to relate antenna N to M. These can be caused from the calling tasks asking for re-references that are not needed. In particular, FRING, KRING, CALIB, and CLCAL need to keep a list of reference antennas as a function of subarray. Then in looping over antenna, subarray, polarization they will only ask for antennas that were used as refernces withing the particular subarray. This will save both work and heartache. Subroutines CALADJ and SNAPP used by CLCAL (only) need to have their call sequences changed and to use/set the list by subarray. Moved nowhere. 13677. November 2, 2011 NANS Eric Wrote a NEW TASK to count NaNs in UV data sets. Moved nowhere. 13678. November 3, 2011 XAS Eric Changed init.s and xas.h to have a new user option startIconic with default 0 (start open). Added option to XAS help file. Moved nowhere. 13679. November 3, 2011 BPLOT Eric The help file limits IFs to 1 - 28. Fixed upper limit to 90. Moved nowhere. 13680. November 3, 2011 VPLOT Amy Changed MSAMP from 2000000 to 10000000, because of the size of some EVLA data sets. This will probably will not be sufficient for many experiments. Moved nowhere. 13681. November 3, 2011 FLATN Eric Added code to keep CC files from being copied to the output image. It only copies those of the first input image and so is not useful. Moved nowhere. 13682. November 7, 2011 DBCON Eric Corrected parameter MXANT of 45 to MAXANT and fixed a variable typed in lower case. Moved nowhere 13683. November 7, 2011 PRTAB Eric Corrected the comparison for string to use only the characters given in KEYSTRNG. Add info to help on this. Moved nowhere. 13684. November 7, 2011 RFLAG Eric Corrected copying of old FG table to omit any flag records that were flagged before RFLAG. Previously, it reinstated them. Discovered that UVRANGE caused all data outside the range to disappear - it did not do that at some time in the past but now did. Added code to check for data completely missing and skip them rather than generate flags for them. Moved nowhere. 13685. November 8, 2011 SPLAT, UVCOP Eric UVCOP contained a version of FLGINI for no good reason and it was an out of date version. UVCOP did not die on error from flagging tables, but error was unlikely. SPLAT failed to flag tables since the FG file was not opened. It produced a ton of messages since it did not quit on error. Moved nowhere. 13686. November 9, 2011 QCREATE Eric To create large files quickly but with some dangers as an option other than the slow but safer filling with zeros. DCRE.INC Special include for quick create integer GTPARM Pick up quick create value from GO AU2 Get adverb value QCREATE if possible - else error message suggesting a compress - and pass it on GO. AU2A Default QCREATE is -1 POPSDAT QCREATE added NEWPARMS QCREATE added ZDCHIN Inite QCREAT parameter as -1. ZCREAT Pick up CREAT parameter and pass along to ZCREA2 with all files <= 1 Mbyte false. ZCREA2.C $APLUNIX: changed call sequence. If quick create, do an lseek almost to the requested end of file and write the last buffer+ of bytes to make the file look as long as it is supposed to be. Otherwise do as we always have and fill the whole thing with 0's. GO Help file: add QCREATE QCREATE New help file Moved nowhere. 13687. November 9, 2011 NANS Eric Added testing for INDE (magic blank value). Moved nowhere. 13688. November 9, 2011 NANS Eric Corrected starting vis number - it was one too high causing a read off the end of file sometimes.. Moved nowhere. 13689. November 10, 2011 IMEAN Eric Corrected window setting for the robust fit when DOINVER is true. ABORTs happened before the fix. Also changed it to use the full image in X and Y in the robust fit, honoring only BLC/TRC(3-7) when DOINVER is false. Moved nowhere. 13690. November 10, 2011 ZCREA2 Eric Update $APLGEN/ZCREA2.FOR and $APLDEC/ZCREA2.C for the new call sequence and in the latter for the new function. Moved nowhere. 13691. November 10, 2011 FITTP, FITAB Eric Wrote new ZRLRUV routine which is ZRLR32 but with no checking for magic blanks. It is possible for u, v, w, or visibility data to have the exact magic blank value (3140.89282) accidentally. This then puts NaNs in the FITS file which CASA cannot handle. Changed FITTP, FITAB to use ZRLRUV for UV data but not for images or tables which can have magic blanks. Moved nowhere. 13692. November 10, 2011 SNEDT, EDITA Eric Changed QEDIUTIL - on a narrow TV window with a bunch of points it was possible for FLAG ABOVE and FLAG BELOW to start searches for samples at counters < 1 or > maximum. This led to references to addresses that were unintended and thence to aborts. Put in sensible limitations. Moved nowhere. 13693. November 10, 2011 ELINT Leonia The format E9.3 has been used for the first term of the fitted polynomial. So 3 digits after the decimal points have been printed out. It may not be enough. So the format is changed for E10.4 . Moved nowhere. 13694. November 11, 2011 ZMSFI2 Eric The error messages needed a new-line command at the end. Moved nowhere. 13695. November 11, 2011 AU2 (AIPS) Eric The message in Go about using COMPRESS to get QCREAT was expanded into explicit detail. Moved nowhere. 13696. November 11, 2011 TYAPL Eric The algorithm for finding the needed SY samples in the SY table was not adequate when there is any serious editing. Modified details and added a section of code to make sure the best job possible is done in findng samples to interpolate. Four copies of all changes had to be made (TY/SY and file 1 and 2). Added tests for having to extrapolate - counts all that extrapolate more than 0.03 sec and prints info about those extrapolating by more than 10+ seconds (up to a maximum print count). Moved nowhere. 13697. November 11, 2011 FRING, RLDLY Eric Changed AVERT to use UVGET rather than UvINIT so as to read only the desired data rather than much of the data file. AVERT checks the integration time at the start of these tasks. This simpler version led to a simpler call sequence, so changed FRING and RLDLY to match. In my data set, a 15 minute job became less than 0.5. Moved nowhere. 13698. November 12, 2011 CALIB, FRING, KRING, CLCAL Eric CALREF would report failures to re-reference even though there were no records for the "old" reference antenna and hence none needing re-referencing. Added counter to CALREF to make sure before issuing the error message (and error return). Moved nowhere. 13699. November 14, 2011 PCAL Eric My new copying of an IF's worth of channels to a scratch file followed by copying the needed channel to the actual work file got tripped up by flagging. The task needs to count the actual number of vis records in the IF-scratch file and reuse that count as NVIS for each channel. Moved nowhere. 13700. November 15, 2011 NXTFLG Eric Fixed the format that say too many flags to work. Moved nowhere. 13701. November 15, 2011 TYAPL Eric Added DOACOR = .TRUE. Moved nowhere. 13702. November 16, 2011 FRING, RLDLY Eric Improved the error messages in FRING subroutine FRNALL to be those of FRNSRC to explain when excessive FFTs are requested. Changed both to allow FFTs 2*MAXIMG in size. Change $QPSAP versions of QXFOUR, QCFFT, QRFFT to allow FFTs twice as big as before. Moved nowhere. 13703. November 16, 2011 VLBAUTIL/VLBAFIX Amy VLBAFIX is supposed to separate frequencies if more than one existed in a data file. It did not seperate S/X data properly. Fixed. Moved nowhere. 13704. November 17, 2011 PCAL Eric Raised internal arrays from 20 thousand time averaged cal samples to 50 thousand. The task could not handle having an entire IF flagged. The new line code messed things up and the older portions would still process the "data" with NVIS=0. That was very bad when DOMODEL is true. Now it traps NVIS becoming 0 and handels it fairly well (solutions are 0). Moved nowhere. 13705. November 17, 2011 DOBAND Eric The application of bandpass corrections for autocorrelations used the real part of the BP function only. For LL and RL it needs to use the amplitude of the BP function and for RL and LR it should take into account the phase. That makes autocorrelation bandpass corrections into the same function as cross correlation functions other than the happenstance that the two antennas are the same. Shortened DATBND to make this change. Moved nowhere. 13706. November 21, 2011 SPLIT, SPLAT Eric The code handling CALCODE when looping over sources forgot to handle the special cases ('*', '-CAL') and so these did not work. Corrected. Moved nowhere. 13707. November 21, 2011 RLDLY Eric Added PRTLEV as an adverb to allow for diagnostic displays. Moved nowhere. 13708. November 21, 2011 RLDLY Eric Made local copy of CALCOP and changed it to copy only baselines to REFANT. Makes things smaller. Changed dimensions of one variable from MXANT to MXBASE which caused MACs to fail. This was correct in FRING. Moved nowhere. 13709. November 21, 2011 CLCAL Eric A failed solution in an SN table is shown with 0 for reference antenna and -1 for weight. The smoothing routine which allows getting rid of the blanks checked for <= 0 weights and set them to 1.0. Added code to set the reference antenna also to the desired REFANT. The routine that interpolates SN to CL did not like reference antennas that did not match (as in 0 and REFANT). This led to discontinuities in the "interpolation". Moved nowhere. 13710. November 23, 2011 RFLAG Eric Added option to flag the data and then plot the result. This uses DOPLOT < 0 and applies the flag table just written to generate the plots. The TYAPL code for applying a large FG table was included (DATFLG and NXTFLG with special local include). Moved nowhere. 13711. November 24, 2011 UVFIT Eric Added the start and stop time in days to the FITOUT text. Added more info about FITOUT to the help file. Moved nowhere. 13712. November 25, 2011 maintenance Eric Updated the TAB completion file HLPIT.LIS Corrected the copyright on LSAPROPO.HLP Updated the about files: ADVERB ANALYSIS CALIBRAT EDITING EXT-APPL GENERAL IMAGING INFORMAT INTERACT MODELING PLOT POLARIZA POPS PROCEDUR RUN SPECTRAL TABLE TASK TV UTILITY UV VERB VLA Moved nowhere. 13713. November 25, 2011 CLIP Eric A bit of bad logic kept CLIP from flagging cross-hands because they were bad. It really only got cross-hands when the parallel were bad. Moved nowhere. 13714. November 26, 2011 RFLAG Eric Corrected GINITL It closed the TV on error - attempts to finish the plot thereafter hung out to dry. RFLAG The history routine messed up the values of UVRANGE so the plot routine done following flagging found no data. Changed to use local variables in the history routine. Moved nowhere. 13715. November 27, 2011 TEK, MSG servers Eric Changed XASERVERS and UNIXSERVERS to honor the environment variables AIPS_TEK_HOLD and AIPS_MSG_HOLD. These cause xterms to set -hold to hold the window even if the program inside it fails. This gives lots of time to review any messages before having to remove the window manually with the window manager. The TEK will also come up not in icon form. UBUNTU 11.04 systems seem to hide things that start in icon form and then not run the TEK server properly. Moved nowhere. 13716. November 27, 2011 ACLIP Eric A bit of bad logic kept ACLIP from flagging cross-hands because they were bad. It really only got cross-hands when the parallel were bad. Moved nowhere. 13717. November 29, 2011 CookBook Eric Fixed COOKBOOK.PDF, all html, cook.html web page Changed: 0 Dates, new table of contents 2 Iconified or not, Tek and XAS controls 4 RFLAG, ACLIP, REFLG, FGDIF, TVFLG (BASELINE-DT), PCAL replace PMODEL with DOMODEL and prohibit it, DOPRINT in RLDIF, SU table corrected by RLDIF, DOAPPLY=-1 in 2nd run of rLDIF, SNIFS 5 CALIB's DOAPPLY & NORMALIZE & self-cal approach, RFLAG, OBIT imaging verbs 6 STVERS, IMFIT et al make STARS file, RSPEC output options, SNIFS, HUEWEDGE 7 FARS, AFARS, RFARS, DOFARS; IMEAN initial guess, STVERS in IMFIT/SAD, UVFIT improvements 8 RFLAG, REFLG, IMEAN initial guess, XMOM 'MAX' op, ISPEC derivative, RSPEC output options 9 SNIFS, flag TY tables in UVCOP/SPLAT, UVFIT limit 2.5 million now and other options. 12 QCREATE 13 several of the ABOUT files E PCAL does models only if 0, QUOUT/QUFIX, RLCAL. I, J update Moved nowhere. 13718. November 30, 2011 FRING Eric The change to AVERT (use UVGET to speed data access) meant that SUBARR=0 no longer was acceptable. Changed AVERT to set SUBARR=1 temporarily when it is 0 so that AVERT can work and ten set it back to 0. Dropped ISUBA from call sequence. Corrected the call sequence in FRING and RLDLY. Moved nowhere. *********************************************************************** *********************************************************************** Changes while 31DEC11 is NEW *********************************************************************** *********************************************************************** 13720. December 12, 2011 help files Eric UVFIT: correct limits on visibilities and parameters, fix some bad grammar REWAY: correct minor typos and things easily mis-read SNIFS needed details for SNIFS rather than SNPLT in the summary sections. Moved to 31DEC11 this date. 13722. December 14, 2011 APCAL Eric APCAL did not check if the TY antenna record had SUBARRAY <= 0 which is supposed to mean "any". The FQ tests were also in error on a similar point. Moved to 31DEC11 this date. 13723. December 14, 2011 FARS Eric Changed HI writing to explain all 9 of the APARM parameters. Changed handling of some as well - there were cases that would escape the code. Moved to 31DEC11 this date. 13724. December 14, 2011 REWAY Eric Added BADDISK Moved to 31DEC11 this date. 13725. December 16, 2011 CLIP Eric The history file using a generalized CALHIS recorded only the last source considered. It was prepared to do otherwise but did not. Took advantage of the preparation to prepare the source names and numbers for the history routine. Moved to 31DEC11 this date. 13726. December 19, 2011 SPLIT, SPLAT Eric Changed code to supress messages on UVGET CLOSe calls but to do the close calls even when no data were found. The FG table was opened and left open although other files were not open - a close is needed but it is best to avoid the file not open in FTAB messages. Also moved the RNXCLS to make sure that an NX table is closed even on error. Blocked continuing on the next source. Moved to 31DEC11 this date. 13727. December 20, 2011 FRING, KRING, UVFLG Olaf Wucknitz/Eric Using LOFAR data, some limits have been found that are incorrect. Changed the history writing in FRING and KRING to handle ANTWT is a general fashion. Changed the baseline limits in UVFLG to MXBASE (= 90*91/2.) rather than 500. Moved to 31DEC11 this date.. 13728. December 20, 2011 SLICE Eric Added X, Y, Z physical coordinates to the output text file. Allow multiple outputs to go to the same text file and added a header line at the start. Moved to 31DEC11 this date. 13729. December 20, 2011 SQASH Eric Changed to allow it to sum even axis 2, producing a 1-row image if summing all axes >= 2 or an image with 1 point on the 2nd axis otherwise. A trivial TRANS may be needed on the latter to make it more visible. Moved to 31DEC11 this date. 13732. December 27, 2011 SLICE Eric Made a bad array reference - the subscript got set to 0 by the coordinate call - so the text file outputs had zero for slice value. MOved to 31DEC11 this date. 13734. December 30, 2011 AFARS Leonia/Eric A small problem has existed near BLC, TRC. It has been fixed. The fitting of parabola to determine position of maximum (with beter accuracy) is not carried out now if maximum is too close to the edges of data. Changed coordinate computation to use header parameters rather than make assumptions about the coordinates. Moved to 31DEC11 on Jan 2. 13738. January 2, 2011 UNIXSERVERS Eric Removed debug message inadvertently left in Moved to 31DEC11 this date. 13739. January 2, 2011 IM2CC Eric Changed it to center each piece of the input image in an output image that is a power of 2 pixels on a side. Moved to 31DEC11 this date. 13740. January 3, 2011 FARS Eric FARS had the meaning of APARM(4) backwards for the HI file. It also did wierd things (nothing???) when APARM(4) > 1.5 - a holdover from when FARS had an AFARS kind of output. Moved to 31DEC11 this date. 13741. January 3, 2012 Imaging Eric Changed GRDAT, GRDSET, SETGRD to insist on powers of 2, rounding up as needed. The first allowed images up to and including 8192 to be doublesd in gridded subtraction, while the second limited them to 4096 and the third to 2048. Fixed the latter two to allow 8192 doubling. Changed SETDO3 to understand DO3D false imaging - it forbad multiple facets in 3D false. Also have it set FLDSZ which needs defaulting sooner than later. OOSUB was failing badly. Moved to patch 16 Jan. 13743. January 4, 2012 FARS Leonia FARS has had several mistakes near conversion RE,IM to AMP/PHASE especially when RESIDUAL sent to the output Moved to patch 16 Jan. 13744. January 4, 2012 Imaging Eric Added progress messages to VISDFT, ALGMEM, ALGSTB. Moved to patch 16 Jan. 13745. January 5, 2012 FARS Leonia Add shifting residual phase to the initial lambda^2 set. Moved to patch 16 Jan. 13746. January 5, 2012 FARS Leonia Misprint at the previous installation Moved to patch 16 Jan. 13748. January 6, 2012 FITAB Eric Did not write the TNULL keyword for compressed uv data. It is -32767 rather than the obvious -32768 so it is important. Moved to patch 16 Jan. 13749. January 6, 2012 SQASH Eric A WRITE statement for a history line had too many arguments for its format. Dropped the unwanted argument. Moved to patch 16 Jan. 13750. January 7, 2012 UVFIT Eric Add option to have a comment string following the parameters in the INLIST file. This comment appears at the end of the line in the FITOUT file. Moved to patch 16 Jan. 13752. January 9, 2012 AFARS Eric Found several errors in the code: (1) If the quadratic maximum fit failed or was too close to the edge, the BLC(1) correction was not added. (2) If the fit was once too close to the edge, it no longer did the quadratic fit at all (IRET never reset to 0). (3) If the input row was all zero or blank, the task used a pixel value that was not properly tested. Now it will use the blanking value. Moved to patch 16 Jan. 13753. January 9, 2012 UVFIT Eric Found lots of places still dimensioned 4 rather than XMXCMP (20) so the commputation of lots of things should have gone into the weeds and did when there were > 4 model components. Adde a couple of spaces to the output file format and changed how time is handled so the actual minimum and maximum times used are what is reported. Moved to patch 16 Jan. 13754. January 16, 2012 CLCOR Eric The GMRT uses a backwards phase convention to the rest of us. Changed a few places - esp the place that moves antenna and source positions - to reverse sign for GMRT. Moved to patch this date. 13755. January 16, 2012 CLCOR Eric Tediously removed all of the horribly non-standard illegal Fortran from the code. Corrected bug put in previously today. Moved to patch this date. 13777. March 1, 2012 VLANT Eric The code reversed the sign of the Y correction for EVLA for the delay computations. It also neglected to shift the hour angles from local to Greenwich which it must do for the EVLA. I guess these errors compensated in tests in 2009. Moved to patch this date. 13780. March 6, 2012 SCMAP Eric Changes to Fortran: 5. Previously, uncompressed input files were fed directly into UVMDIV for model division. This meant that flag tables were not applied. Compressed data were copied first to make them uncompressed but also to apply flags. Changed two places to do this for both compressed and uncompressed data. Moved fix 5 to patch March 7. 13781. March 7, 2012 SCIMG Eric Made the same changes to SCIMG as to SCMAP. Moved fix 5 to patch this date. 13865. June 15, 2012 Mountain Lion Eric The system command ps in Mountain Lion now issues an error message if there is an LD_LIBRARY_PATH set (which is required in binary installations). Changed $SYSUNIX files STARTPMON, TVALT, UNIXSERVERS, and XASERVERS to send the error output to /dev/null and to test for succesful completion rather than for error output. Moved to NEW/patch 19-Nov.