----------------------------------------------------------------------- ; Copyright (C) 1995-1997 ; 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 the master version of AIPS kept in Charlottesville. 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 to 15JAN96 *********************************************************************** Changes while 15JAN96 was TST: *********************************************************************** *********************************************************************** 8670. July 3, 1995 $SYSUNIX/LINK Pat Fixed a subtle but nasty bug in LINK. If the LINK environment variable set in the $SYSLOCAL/LDOPTS.SH file contained a qualifier that started with -n (e.g. -native on SunOS), the bourne shell command "LINK=`echo $LINK | sed -e ...`" to change it to something else will not work as expected. "echo" will see the "-n" and apparentyly ignore the rest of the string, thus resetting the variable to empty. In the LINK script, this was happening to the LDOPTS variable when it tried to "squeeze out extra blanks" via a sed command like the above. Removed the offending command (we can live with a few extra spaces). Moved from 15JUL95 this date. 8671. July 4, 1995 ZACTV8, FILEBOX Eric A couple of fixes I could not do during the freeze: AU5C Fix the default naming for the 1st field ZACTV8 Be self-defensive on the returned path name Moved to 15JUL95 this date. 8672. July 10, 1995 VBGLU Chris Increased MAXBAS, which dimensions some bookeeping arrays, to 1,000,000. The previous value of 30,000 was far too puny for for some VLBA experiments. This is a band-aid solution: there must be some way of getting around the need for these arrays. Moved nowhere. 8673. July 11, 1994 Help files Eric Fixed the one-liner for WTSUM and re-ran SHOPH. This stated that changes were needed to 8 ABOUT files: ANALYSIS, ASTROMET, IMAGING, SINGLEDI, SPECTRAL, TABLE, TASK, and VLBI. Updated these help files and changed COOK13 and COOK0 and RECIPE.TEX to adjust for it all. Moved to 15JUL95 as needed. 8674. July 12, 1995 FILEBOX Eric Redisigned FILEBOX in AU5C to write a complete backup file in the user's area with a related (but scratch) name and then zap and re-write the output file. The backup is kept if there is any error and its name reported to the user. The $FITS area is used only if the user does not include a logical area in the input file name. Moved to 15JUL95 this date. 8675. July 12, 1995 $APGNOT/CLCOR.FOR Leonia I have discovered a confusion with left/right hand coordinate system in OPCODE=ANTP. Left handed coordinate system is used in VLA. Right handed coordinate system is used in VLBA. But the task required opposite coordinate system for both VLA and VLBA to provide the correct correction. I think the task worked succesfully with VLA only because users used default value of CLCORPRM(4)=0. It provided right hand coordinate system (in accordance with HELP) which is wrong for VLA but fits with the formulas in the task. Users have not used probably the OPCODE=ANTP for VLBI, that is why the wrong definition of the coordinate systems had not detected early in this case. I have changed the codes to have an agreements with HELP and make the task provide a correct correction with default value of CLCORPRM(4)=0 for VLBA and VLA. The relevant change of HELP has been done as well. Moved nowhere 8676. July 12, 1995 $APGNOT/UVCRS.FOR Leonia The task has not worked if a start day is more than 1. I have fixed it Moved nowhere 8677. July 12, 1995 $YPGNOT/VPLOT.FOR Leonia The task plots different baselines for ANTENNAS=0, BASELINE=list and BASELINE=0, ANTENNAS=list. Inspite of exhaustive comment in HELP complains occurs somtimes. I have added a couple of warnings. Moved nowhere 8678. July 12, 1995 ZACTV9 in APLUNIX Chris Stopped ZACTV9 changing the process name of the debugger to that of the task. Revision 3.0 and later of the SPARCworks debuggers use the process name to search for icon files and to obtain license-manager tokens and will not run if the process name is messed with. Moved to 15JUL95 on this date. 8679. July 12, 1995 SPLIT Chris Added a trap to detect when SOUNDX has stripped all of the requested sources from the SOURCES adverb because they don't match the frequency ID selection. The program will now terminate when this happens instead of interpreting the now-empty list as a request for all sources. Moved nowhere. 8680. July 12, 1995 GNU Eric Created a GNUGPL.HLP from the short text in the help and the full GPL text in the Explain sections. Changed AIPS.FOR to put out some lines witb info. Moved to 15JUL95 this date. 8681. July 13, 1995 PP and PROG Pat (separate issues). Modified PP so that you can debug the preprocessor itself by saying PP DEBUG module-to-process. It will use a PP.EXE in current directory if it's there, otherwise the PP.EXE in $SYSLOCAL. Modified PROG so that it uses PRINTENV, not printenv. Moved to 15JUL95 this date. 8682. July 13, 1995 PP Eric Removed "new version" message - it's not that new anymore. Corrected variable name that caused it to include the gnu copyleft stuff at first level includes. Moved to 15JUL95 this date. 8683. July 13, 1995 SCMAP Eric Corrected saving of WINdows - it used the old 50 boxes dimension rather than the parameterized one (now 500). Later routines were correct and OOP defended itself. Moved to 15JUL95 this date. 8684. July 13, 1995 PLAYR Eric Corrected minor errors in help file. Typed up a new HLPPLAYR.HLP file since the old one was lost in a putbck screwup. Corrected the code to allow for HLPPLAYR and to see that graphics planes are off when empty. Moved to 15JUL95 this date. 8685. July 14, 1995 OTFUV Eric Changed OTFUV to add time smoothing - an averaging interval and an output increment - to the task. This should save disk space without comromising the imaging. It has suprisingly large effects on SDGRD's output, but after baseline removal an averaging of 2 or 3 has only a little effect. Moved to 15JUL95 this date. 8686. July 14, 1995 $APGNOT/ACCOR Leonia The task has put unit in the SN table in the case of data absence for a given time interval. It provided some errors in result of application of output SN table near gaps of the data. I have changed the unit on blank. Moved nowhere 8687. July 18, 1995 EGETNAME Eric Fixed AU8 to do EGETNAME again. It got lost in the code clean-up accompanying the new GET4NAME... Moved to 15JUL95 this date. 8688. July 18, 1995 IMAGR, SCMAP Eric Fixed both tasks to force the IMSIZE to be a power of two. Changed SCMAP to set NITER to 300 if the user sets it <= 0. Moved to 15JUL95 this date. 8689. July 18, 1995 AIPS Eric Tried centering the GNU message to see if that's more readable. Moved to 15JUL95 this date. 8690. July 18, 1995 $APGNOT/PCLOD.FOR $HLPFIL/PCLOD.HLP Athol Upgrade PCLOD, which reads VLBA pulse-cal information from an ASCII text file, to correctly handle data with multiple freq. ID's. In addition, now reads the NX table to assign source ID's to the SN table entries and is more robust in handling mixed state count/pulse cal. modes. Moved to 15JUL95 this date. 8691. July 20, 1995 $YSERV/XAS.SHR (Makefile*) Pat Made some changes to the Makefile that avoid the need to edit the file on most systems. Ultrix (4.3) needs its own; its version of make is more primitive. Otherwise it works without editing on all our other architectures. It now includes the -R option for Solaris 2.x systems to burn in the library locations. Moved to 15JUL95 this date. 8692. July 20, 1995 ZMSSRV Jeff/Eric Corrected the length argument to gethostbyaddr. Giving it too much was okay with everybody until Solaris 2.4. Moved to 15JUL95 this date. 8693. July 20, 1995 HISTORY, SCMAP ERic Added more adverbs to history from SCMAP and corrected HISTORY which had too small a buffer for possible adverbs to go to history. Raised it from 50 to 500 - CLBOX requires 200. Moved to 15JUL95 this date. 8694. July 21, 1995 FITLD Phil We have known for a few weeks now that AIPS has problems in some areas, especially fringe-fitting, if the reference frequency is not in IF 1. FITLD would always sort the incoming VLBA data into increasing freq. order but would not adjust the reference IF from that labelled at the correlator. Now it performs the adjustemnt so that IF 1 always contains the ref. freq. Also increased the size of a couple of buffers to deal with the large number of channels being routinely generated by the VLBA & ATNF. Moved nowhere. 8695. July 21, 1995 $QPGNOT/LPCAL.FOR & LPCAL.HLP Kari Leppanen via Athol New polarization calibration task, contributed by Kari Leppanen. Algorithm is described in his thesis, also Leppanen, Zensus and Diamond, 1995, ApJ (in press). Moved to 15JUL95 this date. 8696. July 21, 1995 $APGOOP/CCEDT.FOR CCEDT.HLP K.Leppanen/ Athol New features in CCEDT which are required by new polarization calibration task LPCAL. These include the ability to write multiple CC output tables and an option to automatically select CC boxes satisfying a flux constraint. Moved to 15JUL95 this date 8697. July 21, 1995 $APGNOT/FITLD.FOR Phil When FITLD encountered a file with a large number of FQIDs the output FQ table did not correctly describe the data due to an indexing bug in labelling the rows of the FQ table as it was being written. Fixed, also increased the maximum number of FQIDs to 100 from 10. Moved nowhere yet. 8698. July 24, 1995 $APGNOT/PCCOR.FOR Leonia The new task PCCOR corrects instrumental phase and delay offsets using pulse-calibration information in the PC table. Pulse-calibration tones are injected into the VLBA receivers and, in conjunction with the cable calibration system, can be used to measure and track instrumental variations in delay and phase between separate baseband converters. The measured phases of the tones allow the connection of the phase across all the IFs. The phase difference between two tones in the same IF allows the estimation of the instrumental single-band delay for that IF. The measurement of single band instrumental delay is however subject to ambiguity due to the 2pi*N ambiguity in measured phase. To resolve this ambiguity, the task uses observed visibility data taken on a calibrator. The ambiguity factor is determined by the condition of maximum coherence. The ambiguity factor is determined in this manner for each polarization, IF and antenna relative to the reference antenna. The adverbs, TIMERANG, CALSOUR and FREQID select both UV and PC table calibrator data. Currently PCCOR assumed only two tones are measured for each IF. Each FREQID and SUBARRAY must be processed separately. Moved nowhere. 8699. July 25, 1995 $AIPPGM/RECAT.FOR Pat Modified it to give very obvious warning messages if a Catalog Block file is either zero length or generates some other error when it tries to read it. Also warn the user that files may need to be deleted by hand if one or more CB files generate problems on the read call. Moved nowhere yet, may move to 15JUL95 jury duty permitting... 8700. July 27, 1995 $HLPFIL/CLCOR.HLP Leonia The confusion with left/right hand coordinate system which I discovered in OPCODE=ANTP (see measage 8675) is a difference in the definition I was taught in Russia. I am sorry about the additional confusion. Right handed coordinate system is used in VLA. Left handed coordinate system is used in VLBA. I have edited HELP a litle bit to have an agreement with the correct definition. 8701. July 28, 1995 New adverb: DELCORR Phil Added a new adverb to control the generation of the CQ table in FITLD. The default is to create the CQ table. Once generated, and when delay corrections are applied to VLBA data at a later point, the 2nd order corrections described in Athol's AIPS Memo 90 will be applied to the data. Moved to 15JUL95 today. 8702. July 28, 1995 FITLD Phil Added the DELCORR adverb and ensured that CQMAKE is called if necessary. Moved this and all previous changes to 15JUL95 today. 8703. July 28, 1995 FITLD Phil Enabled FITLD to select data by frequency and/or bandwidth. Moved nowhere. 8704. July 30, 1995 Update HLP files Athol Update some help files in preparation for the new release: FXVLB.HLP: Add EXPLAIN section ANTAB.HLP: Document DPFU=RCP,LCP mode more clearly PCLOD.HLP: Change contact person for PCLOD text files. Moved to 15JUL95 today. 8705. 30 July 1995 VLBA delay decorrelation Athol Perform delay decorrelation corrections for VLBA data using information in the CQ table. These include spectral averaging and segmentation loss corrections, which are further described in AIPS Memo. 90. This method replaces the previous use of the SPEC_AVG keyword in the AN table so that these corrections are implemented in a more robust form that is less sensitive to changes in the frequency structure of the uv-data during calibration. Several options which are possible in the VLBA correlator, but not yet implemented, have been anticipated in the new format. For further information, consult the AIPS Memo. mentioned above or type EXPLAIN FXVLB. Changes include: $INC/DSEL.INC - Add new arrays to contain the information required for the segmentation corrections. $APLNOT/SELINI.FOR - Reset new DSEL.INC variables. $APLNOT/GAININ.FOR - New call format to DSMEAR $APLNOT/DSMEAR.FOR - Re-written to use CQ information $APLNOT/FXSEG.FOR - New routine to calculate segmentation losses. $APLNOT/DATGET.FOR - pass subarray number to DATCAL $APLNOT/DATCAL.FOR - add segmentation corrections, also warning if VLBA data with no CQ information but delay solutions are being applied. Move to 15JUL95 on Aug 1, 1995. 8706. 30 July 1995 $APGNOT/SPLIT.FOR Athol Disable copy of the GC and CQ tables. Also add a history record noting whether the CQ table in present and hence VLBA delay decorrelation corrections are activated. Move to 15JUL95 on Aug, 1, 1995. 8707. July 30, 1995 Convolving functions Eric Add circular functions and a type 6 = BESSJ1(x)/x * exp. Changes: CONVFN Add type 6 and circular storage GRDFLT Add type 6 and circular; clean up code. BESSJ1 Steal from UVMOD. Change help files to mention type 6 (only): CONPL GRIDR HORUS MX UVMAP CXPOLN WFCLN IMAGR SCMAP MAPIT_MX Changed code to know type 6 name: CONPL GRIDR HORUS MX UVMAP Changed GRIDR help to recommend SDGRD and fixed XPARM variable name in PGEOM help to APARM. Moved nowhere. 8708. July 30, 1995 SDGRD Eric Changed QGRD6 and QGRD7 to do circular as well as separable convolving functions. Changed SDGRD to allow for circular functions, Bessel functions, and functions specified in arc seconds rather than cells. Moved to 15JUL95 as SDGRD with revised subroutines appended. 8709. July 31, 1995 OTFUV Eric Change output header to put the SD 12m beam width in the Clean beam major and minor axis. Moved to 15JUL95 this date. 8710. July 31, 1995 LWPLA Eric The grey-scales have been put into characters ! through ~ - 94 levels. Unfortunately, this leaves a finite chance that a % sign will be the 1st character in a line. PostScript says that % is a comment, so PostScript-like things which have not done a full POstScript interpretation of the file may think that these are comments (or even the %'s inside lines). The TeX tool psfig is one of these not-too-intelligent processors. Changed LWPLA for now to use characters & through ~, 89 levels instead. We should probably use 255 levels via HEX, but the present change is enough for 15JUL95 to be safe. Moved to 15JUL95 this date. 8711. August 1, 1995 LWPLA Eric Changed from 89-level character representation of grey scales to 255-level hexadecimal representation. This doubles the size of the file (or the grey part anyway), but gives a smoother looking output. Moved nowhere. 8712. August 1, 1995 $APGNOT/UVCOP.FOR $APLNOT/CQSEL.FOR Athol Modify UVCOP to renumber the IF's and FQ_ID's in the CQ table if UVCOP applies frequency selection. New routine $APLNOT/CQSEL.FOR added for this purpose. Moved to 15JUL95 this date. 8713. August 1, 1995 $APGNOT/VBGLU.FOR Athol Disable copy of CQ table. Moved to 15JUL95 this date. 8714. August 2, 1995 $APGNOT/FITLD Phil When loading multiple files FITLD depends on the reference frequency keyword when deciding whether files should be concatanated. In one case we had 2 freqids and most files had a ref. freq corresponding to the first freqid, but the final file had the second as a ref. freq. In this case FITLD created a 2nd file. Fixed it to check all extant freqids when deciding if concatanation is possible. Moved nowhere. 8715. August 2, 1995 $YPGNOT/APCAL.FOR Athol Found a typographical error in an error message regarding missing gain curve information. The antenna number was reported incorrectly. Request pending to move to NEW. 8716. August 3, 1995 FITLD Phil Somehow some TAB characters crept into the code changes I made yesterday. The Sun compiler didn't mind but the IBMs caught them. Fixed. Moved nowhere. 8717. August 1 and 3, 1995 TVRGB and friends Eric Added the adverb RGBGAMMA to POPSDAT.HLP and NEWPARMS.001, wrote a help file for it, and changed AU5E, OFMGAM, TVDEVICE, and GAMMASET.HLP to allow RGBGAMMA as an (optional) output adverb from GAMMASET. Changed TVRGB and its help to provide the option of writing out a color PostScript output file (printed or saved) using all 24 bits. Moved nowhere. 8718. August 3, 1995 $APLNOT/CQMAKE.FOR Athol Fixed typographical error in buffer usage. Request pending to move to NEW. 8719. August 3, 1995 FITLD Phil Minor change to now overwrite CQ table when necessary. Will work now that Athol has fixed CQMAKE (see #8718). Request pending to move to NEW. 8720. August 3, 1995 $APGNOT/SPLIT.FOR Athol Update to #8706, disable CQ table copy. Request pending to move to NEW. 8721. August 7, 1995 Buffer size Eric Raised buffer sizes (soma as small as 512) in $APGNOT for CANDY DCONV FETCH GLENS IMFLT IMLIN IMMOD MANDL MATHS MODVF MOMFT MOMNT MWFLT NINER NNLSQ PADIM PATGN POLCO RGBMP RM SHADW SKYVE SMOTH SUMSQ WARP Changed grossly long strings in LWPLA. Moved nowhere. 8722. August 8, 1995 Excess copylefts Eric Removed excess Copyleft statements from $AIPPGM: AIPS AIPSB AIPSC $APGNOT: BLWUP CLCAL FXPOL STRAN UVCOP $QPGNOT: CVEL $QYPGNOT: UTESS $YPGNOT: FRMAP SNPLT TVFLG VBPLT VPLOT Also removed the "All rights reserved" and "Copyright 1988" statements found in REQBAS, CVEL, FRMAP, VPLOT. These are not legal and never were. Moved nowhere. 8723. August 8, 1995 KWICK ERic Changed the end-of-loop test which was subject to floating-point inaccuracy problems. Moved nowhere. 8724. August 9, 1995 $SYSAIPS/AREAS.DAT Jeff Added $INCRL as $INC/NOTST/READLINE for GNU readline library include files. I'm also regenerating the AREAS.{SH,CSH} scripts to instantiate the change. (I'm pretending I speak C++ there...) Moved nowhere. 8725. August 10, 1995 $INCRL, $INCGNU, AREAS.* Jeff Moved $INCRL to $INCGNU (as $INC/NOTST/GNU). Updated $SYSAIPS/AREAS.DAT, $SYSUNIX/AREAS.{SH,CSH} to reflect change in logical. Moved nowhere. 8726. August 11, 1995 $AIPSUB/RLHIST.FOR $APLGEN/ZGRLHI.FOR Jeff RLHIST is the top-level routine for initializing (recalling) and saving cross-session command-line history for the GNU readline library's history-recall functionality. $APLGEN/ZGRLHI is simply a stub routine that returns IERR indicating that no history was saved or recalled. Neither of these routines actually access any readline library functionality. Moved nowhere. 8727. August 11, 1995 $AIPPGM/AIPS.FOR Jeff Added RLHIST calls to open/close 'RL' (GNU readline command- line history) files across sessions. This functionality is stubbed until I put libreadline.a in place ($LIBR/GNU) on all arch's and check in the $APLUNIX/ZGRLHI.C routine. (Other preparatory changes are needed as well, such as to LIBR.DAT files.) Moved nowhere. 8728. August 13, 1995 OOP Eric Changed UVDESC to declare REFFPIX, the input uv data set's reference pixel. Changed UVDATA (OUVOPN) to set it and UVUTIL (UVREFQ) to use it. Relinked all of OOP. Moved nowhere. 8729. August 14, 1995 $APGNOT/ANTAB.FOR Athol Code that was designed to deal with experiments going over year-end failed when the uv-data reference date was later than the first day number found in the calibration file. Corrected this error. Task would fail reporting an invalid time stamp. Moved nowhere (should go to the patch area). 8730. August 14, 1995 $APGNOT/TABED.FOR Athol The table buffer sizes were inadequate to deal with the larger table records encountered with current data. This is particularly true of VLBA BP tables. Increased the buffer sizes from 2048 to 8192. Task would fail requesting a larger table buffer. Moved nowhere. 8731. August 14, 1995 $APGNOT/SNCOR.FOR Athol Option XFER, which transfers solutions from one IF to another with frequency scaling, had been injured in a previous fix. Antenna number (KANT) was no longer defined in the correction routine XFER. Mis-addressing then added 1 radian of phase to all solutions. The task ran to completion without any outward signs of failure. Moved nowhere. 8732. August 14, 1995 TVMENU Eric Changed the call sequence to TVMENU, adding a parameter specifying how far down the page to begin the menu. Changed TVDEVICE (TVDMEN) similarly to pass the parameter through. Changed TVRGB, QCLEAN, SCMAP, PLAYR to the new call seq. Moved nowhere. 8733. August 14, 1995 TABLE object Chris Added a public subroutine (TABEXI) to test for the existence of a table. Moved nowhere. 8734. August 15, 1994 $APGNOT/FITLD Phil The recent changes concerning frequency selection etc have gone to great lengths to ensure that the first FREQID value written to the output file is always 1. However there was one case that slipped by. When the user specified source and time-range selection the output FREQID was not trapped. Ensured that it now is and that if FREQID 1 has not been written that renumbering occurs. Moved nowhere. 8735. August 15, 1995 $APGNOT/VBGLU Phil VBGLU was failing on compressed data in that the data in file #1 were being scaled on output with the scale appropriate for the input file rather than the output file. This caused offsets to appear between IFs, although the data within the IFs still contained the correct relative scales. Moved nowhere. 8736. August 15, 1995 OOP Eric Changed INPUT.FOR, adding a routine called AV2INT which is AV2INP but for interactive tasks. Actually created a private subroutine INSTRT and made AV2INP and AV2INT call it and then just handle the AIPS restart parameters. Also added RESAIP to resume AIPS in the middle of a task that used AV2INT. Changed PLAYR to call the new routine. Moved nowhere. 8737. August 16, 1995 XAS Eric XAS had a bug that caused it to skip displaying the TV screen when zooming with no grey planes turned on. Fixed it by trapping cur_chan = 0 when the channel was > NGREY and doing the display. Moved nowhere. 8738. August 16, 1995 ZPRMPT.C ZGRLHI.{FOR,C} LIBR.DAT Jeff GNU readline versions of ZPRMPT.C and ZGRLHI.C put in $APLUNIX. Non-readline (old) versions of ZPRMPT.C put into $APL* for ARCH's that are as yet not using GNU-RL (ALLN, CRI, CVEX, DEC, IBM3090), ditto for stub versions of ZGRLHI.FOR (same file as is found in $APLGEN). LIBR.DAT files for ARCH's that now support GNU-RL (SUN4, SOL, IBM, ALPHA, HP, LINUX, SGI) modified to include links to $LIBR/GNU/libreadline.a and -ltermcap for all $AIPPGM tasks. Moved nowhere. 8739. August 17, 1995 $QPGNOT/FRING.FOR Athol In copying the FQ table from the input u-v file to the calibrated scratch file, FRING did not correct for possible changes in the reference frequency made by UVGET:DGHEAD. As there is no IF selection in FRING this only affects datasets where the freq. offset of IF.1 in the FQ table is non-zero, or there are significant source-based freq. offsets (and CALSOUR set explicitly to select a single source). FITLD only produced files where IF.1 had a non-zero offset for a short period of time and such files should be rare. This error affected the scaling of the fringe rates, which were then not properly removed when applied. Also added unrelated code to print the coarse search spacing in delay and fringe rate for print levels greater than 1. This is useful in determining how the AP memory is being subdivided. Moved nowhere. 8740. August 17, 1995 $AIPSUB/AU8A.FOR Athol EXTL failed on VPLOT files (plot type 21). Several IF statements checked only for VBPLT rather than VBPLT and VPLOT. EXTL would fail with an exception or print garbage, depending on the operating system. Moved nowhere. 8741. August 17, 1995 $APL{DEC,ULTRX}/{ZPRMPT.C,ZGRLHI.FOR} Jeff Moved non-GNU-RL version of ZPRMPT.C and stub of ZGRLHI.FOR from $APLDEC to $APLULTRX so that $ARCH=ALPHA would pick up the readline functionality. This corrected a screw-up on my part. Moved nowhere. 8742. August 17, 1995 $SYS{SUN,SOL}/LDOPTS.SH Pat/Jeff Fixed for USESHARED use of GNU readline; it has to be handled (linked) differently if AIPS is built to use shared lib's. Doing this in LDOPTS.SH, vice in LIBR.DAT.SHARED, makes more sense. Moved nowhere. 8743. August 17, 1995 TVFLG Eric Changed a few comments in TVFLG and SPFLG related to flag tables and single-source files. Changed the typing for longer lines in the help area (72 character lines are ok). Moved nowhere. 8744. August 18, 1995 $APLUNIX/ZPRMPT.C Jeff Changed so that if input to AIPS is not coming from an interactive tty (as in from a shell-script "here document") then readline() will not be called and AIPS will fall back to the old input method of fgets(). Moved nowhere. 8745. August 18, 1995 $APGNOT/AVSPC Phil AVSPC ran into a subtle problem if there were > 1 FREQIDs and a FG table was applied. With > 1 FREQIDs CHNDAT would object when reading data if the fqid had not been set, it would then jump out without closing the FQ table. The FQ table had been opened with LUN 30 and the routine that opened the FG table also used LUN 30, so we had a clash. Unfortunately MSGSUP was set such that all useful messages were not printed. Fixed. Moved nowhere. 8746. August 21, 1995 $SYSUNIX/LINK LIBR.DAT.SHARED LDOPTS.SH Jeff/Pat Cleaned up linking of GNU readline library when using shared libraries under SUN4 and SOL. It was acting a bit senile when AIPS was compiled DEBUG with shared libraries. Moved nowhere. 8747. August 21, 1995 $APLUNIX/ZGRLHI.C Jeff Unbinds TAB key for libreadline to prevent the confusing and useless TAB-expansion of filenames in the current directory under AIPS. This key (and any other key-sequence in fact) can be overridden/changed in ~/.inputrc. Moved nowhere. 8748. August 21, 1995 $APGNOT/FITLD.FOR Phil A minor indexing error caused the FITLD frequency selection to fail under some circumstances. Fixed. 8749. August 21, 1995 $HLPFIL/MOVE.HLP Athol Correct line which remained in the file from the SUBIM template, incorrectly describing the task function. Moved nowhere. 8750. August 21, 1995 $APGNOT/UVCOP.FOR Athol Memory overwrite errors arose in UVCOP if the number of sources exceeded 500, the dimension of the array used to track all sources copied. The task failed silently, but the CL table consolidation was corrupted in antenna selection. The symptoms were operating system specific. Increased the maximum number of sources allowed to 16384 and added a warning if more are found. Will move to patch area. 8751. August 22, 1995 MCUBE Gustaaf/Eric When the program created a SEQ.NUM axis it failed to insert the number of points on this axis in the input image header (1) and so looped 1 to 0 which is not at all now that we have Fortran 77 compilers. Either headers used to have 1s in the unused dimensions (a distinct possibility) or this option was last used under a Fortran 66 compiler. Changed the message level about blanking so that it also goes to the message file. Moved nowhere. 8752. August 22, 1995 $SYSUNIX/AREAS{C,}SH.SED Pat/Jeff Now allows (read: handles correctly) logicals with lower-case characters in the path; lower-case is usually required for a directory to be usable for $MANPATH-accessed man-pages (man is not a forgiving animal). Moved nowhere. 8753. August 22, 1995 $SYSAIPS/AREAS.DAT Jeff Added $SYSMAN logical ($SYSUNIX/manl). Re-generated $SYSUNIX/AREAS.{C,}SH. Moved nowhere. 8754. August 22, 1995 $APGNOT/FITLD Phil When performing the adjustment of the ref. freq. to IF 1 I had overlooked the case of mixed lower/upper sidebands. In such cases the ref. freq would have been in IF2 not IF1 and files that were concatanated to the first file loaded would have had all their sources labelled with freqids 2 not 1. Fixed. Moved nowhere. 8755. August 22, 1995 $APGOOP/SPECR Chris/Phil Chris repaired SPECR which was dying on Suns with a FP exception since I was trying to overwrite a DATA'd integer. Moved nowhere. 8756. August 24, 1995 $QPGNOT/CVEL.FOR Athol The number of words read by GTPARM was incorrectly set at 192 rather than 170. On Solaris GTPARM then overwrote some crucial file handling variables and CVEL failed on start-up, reporting message file problems. Fixed this; also doubled the BP buffer size to 8192 to cater for large current BP tables. A patch may be warranted. 8757. August 27, 1995 ZTREAD Eric Changed output buffer from CHARACTER*80 to CHARACTER*(*). The former does not allow for variation, which there may be. The latter should ALWAYS be done. Moved nowhere. 8758. August 28, 1995 $SYSUNIX/F2PS.C Pat As the old K&R C compilers (like /bin/cc on SunOS 4) do not permit aggregate initialisation of automatic variables, F2PS could not be built on a system lacking an ANSI C compiler (acc, gcc on SunOS 4). Moving the page coordinate information up and making the variables static makes the program palatable to both flavours of C compiler. Moved nowhere. 8759. August 28, 1995 Real-time helps Eric Changed TSKHLP to take a string input rather than the column and row number. This means menus no longer have to have a fixed format and the HLP*.HLP files will not need so much editing when one changes a menu. The new HLP*.HLP format is a C-------- line separating each section, with the string naming the section left justified in the next line. The comparison is case sensitive. Changed TVMENU to pass the string to TSKHLP and to pass out not the row and column numbers but only the subscript in the CHOICS list that was selected. Calling routines should do their case statement on the strings, not col and row numbers anyway. Also changed TVMENU call seq to take a list of logicals to say whether the menu is turned off after the selection or not. Changed TVDMEN in TVDEVICE to match and changed TVFLG, SPFLG, IBLED, TVHUI for the change in TSKHLP, TVRGB for the change in TVMENU, QCLEAN, PLAYR, SCMAP for the change in TVDMEN. Changed the real-time help files HLPCLEAN, HLPIBLED, HLPPLAYR, HLPSCMAP, HLPSPFLG, HLPTVFLG, HLPTVHUI, and HLPTVRGB to the new format. Changed TVRGB's handling of options and changed TVHUI to use TVMENU with better handling of options. Moved nowhere. 8760. August 28, 1995 SNEDT Eric New task to do interactive editing - a la IBLED - of SN and CL tables. This is mostly the EDIT class utility library. New: SNEDT New task - a calling routine mostly SNEDT Help file for new task QEDIUTIL $QOOP: edit utilities for this form of editing HLPEDISN Real-time help file for EDITSN in QEDIUTIL. Changed: EDIT $APLOOP: add TBEDIT as known internal, fix error msg TVTPOS $YSUB: a <= was supposed to < so infinite wait reads became quickies TVDEVICE $APLOOP: add capabilities of (1) return visible corner pixel numbers, (2) interact with TV cursor in usual way with calling routine to handle results, (3) read and change (write) zoom parameters. Correct precursor comments and channel handling in TVDCHR. Moved nowhere. 8761. August 28, 1995 SNPLT Eric Changed help and Fortran to have FREQID = -1 mean all, before the help lied and and FREQID got made > 0. Moved nowhere. 8762. August 28, 1995 MULTI Chris Fixed a subtle bug that caused the sequence number to be be written as zero in the header for the output file. This caused some confusion using the OOP UVDATA interface and would, I assume, have been potentially dangerous if a catalogue containing an output file from MULTI needed to be rebuilt. Moved nowhere. 8763. August 29, 1995 $SYSUNIX/START_AIPS Jeff Added a "norl" startup option that will disble GNU readline extensions if someone wants to fall back to the old behavior for some reason. (Also makes a good troubleshooting aid.) Moved to $AIPS_ROOT in CV this date. 8764. August 30, 1995 OMFIT Ketan Added new task akin to UVFIT that has the following advantages: Simultaneously model-fits and self-calibrates UV-data Uses Singular-Value-Decomposition to better determine parameters Employs Levenberg-Marquardt Non-linear convergence methods Will be able to solve for multiple self-calibration models [one for each model component] comments to kdesai@nrao.edu 8765. August 30, 1995 OMFIT.HLP Ketan HELP file for OMFIT, described above, This program uses an INFILE, a sample INFILE is appended to OMFIT.HLP comments to kdesai@nrao.edu 8766. August 30, 1995 FITLD Phil The updated correlator code is beginning to be tested. One of the new features is that the FFT twiddle tables are much better and the level of FFT artifacts in the total-power spectra are reduced to the level of invisibility. We added an FFT_TWID keyword to the MC table, FITLD now checks for this and if it has a value of 1 will no longer perform the artifact corrections. Also ensured that the subarray frequency written into the AN table is corrected for roundoff error. Moved nowhere. 8767. August 31, 1995 ZDCHIN Eric Correctd the display of the GNU copyleft. It was writing a 132-char line to terminals - 70 would have done and resulted in less double spacing. Moved nowhere. 8768. August 31, 1995 SNEDT Eric It forgot to get the parameter DODELAY, so did not display delay at all. Moved nowhere. 8769. August 31, 1995 UVSORT Eric It zapped the SC files it made, but did not lower NSCR, the count of SC files. So, if it is called a lot, it can blow the bank of 128 SC files. Moved nowhere. 8770. August 31, 1995 CookBook Eric Changed the COOK82.MAC and COOK82G.MAC macro sets to write out index files and added \indx, \Indx, \tndx, \Tndx, \iodx, \Iodx, \todx, and \Todx macroes to index a word or phrase. The ones with t or T put the word in tt font, the capitals put the page number in bold type (for primary references), the odx ones do not put the word in the actual document, the others do. Also made up some macros for tex setting the Index and a Fortran program COOKNX.FOR to read the sorted index list and a procedure COOKIT to run the sort, COOKNX, and tex set of COOKNX.TEX. Changed chapters 0, 1, 2 and 10 so far. Changed 1 to mention the index and 2 to correct for new XAS parameters. Moved nowhere. 8771. September 1, 1995 OPTIMIZE.LIS, FDEFAULT.SH Pat Knocked the optimization for OMFIT down to OPT1 on the IBM's. The optimiser dies a horrible death otherwise. Also increased the maxmem parameter for IBMs from 32k to 64k. Moved nowhere. 8772. September 1, 1995 DFIL.INC Eric Changed DFIL.INC to put the dimensions of the arrays into parameters. Then changed $APLSUB/ MAPCR MAPSNC SCREAT $APLNOT/GRIDTB IMCREA INTMIO SDGET SETGDS UVGET $APLOOP/ IMAGE UVDATA (128 -> parameter) to check these limits and die if they are exceeded. Corrected an error handling the scratch count in CCMERG which caused scratch files not to be deleted. MOved nowhere. 8773. September 4, 1995 $APLNOT/SNSMO.FOR Athol Subroutine SNSMO (called by CLCAL) did not check for preceding blanked values in extrapolating the phase to remove 2pi ambiguities. Moved nowhere. 8774. September 5, 1995 $APGNOT/UVCOP Phil Due to a logical array being declared to be of size 1024 UVCOP was restricted to copying files whose product of IFs and channels did not exceed that. Changed it to a large value that should not be exceeded by any existing correlator. Moved nowhere. 8775. September 5, 1995 $APGNOT/UVFIT.FOR Athol UVFIT had convergence problems related to the initial guess for the inverse Hessian diagonal terms as passed to the fitting routine DVDMIN. These were estimated as 0.001 times the initial values. Although DVDMIN appears fairly insensitive to this input parameter, it did not converge or explore a sufficient region of parameter space in some cases, particularly when fitting VLBI components. Changed UVFIT to estimate the approximate order of the Hessian terms from the gradient. This rough approximation gives broader convergence for both VLA and VLBI test data. Moved nowhere. 8776. September 6, 1995 ZGRLHI Jeff Reworked the "hierarchy" for ZGRLHI.C (readline version) and ZGRLHI.FOR (stub version). There's now a copy of the readline-capable version in the $APL directory for each OS that readline is supported under, and a copy of the stub in $APLGEN and $APLULTRX (to override the $APLDEC, a.k.a. Alpha, version). Added handling of readline-disabling (via starting AIPS with the "norl" option) in ZGRLHI.C. Moved nowhere. 8777. September 6, 1995 ZPRMPT.C Jeff As with ZGRLHI, I re-did how the hierarchy works for ZPRMPT.C. The non-readline version now resides in $APLUNIX, and the readline-capable versions, for OS's that libreadline is supported under, now reside in $APL$ARCH. ($APL3090 and $APLULTRX have to have additional copies of the non-readline version from $APLUNIX due to the layout of the hierarchy--I forgot about the 3090, deleted the non-readline version of ZPRMPT.C, and then had to restore it again.) The readline-capable version of ZPRMPT.C now sets signal-handlers for SIGTTIN and SIGTTOU just before the readline() call, and resets these signals back to SIG_DFL immediately after readline() returns; this appears to eliminate some, ahem, interesting behavior under SunOS and AIX (especially AIX, where a backgrounded AIPS session would lock the terminal hard once it wanted input again). The signal-handler does a kill(getpid(),SIGTSTP), effectively redirecting the other signals into one that appears to behave better in this situation. (Who says that programming is not an empirical science?) libreadline's attempt at this sort of signal-handling does not appear to be very smart, and behaves very inconsistently across different OS's. There are now two versions of the readline-capable ZPRMPT.C scattered among the readline-supported OS's $APL$ARCH directories; one re-prep's the terminal (fixing the echo) within the signal-handler immediately after the kill() call, and one doesn't. Again, this is due to varying behavior among OS's--behavior that does not appear to follow any real pattern (BSD, SYSV, etc.)--it's a real scatter-shot. HP/UX appears really hosed on all of this, so I have temporarily removed readline support for it until I can get it figured out. Disabling ^Z under AIPS would eliminate all of this nonsense, but we'd really need to add some sort of shell-out functionality to make up for its loss. Also added handling of readline-disabling (through the "norl" startup option). Moved nowhere. 8778. September 6, 1995 $SYSUNIX/AIPS.L $SYSMAN/{AIPS,aips}.l Jeff Moved man-page from $SYSUNIX/AIPS.L to $SYSMAN/AIPS.l. Added sections on basic GNU readline functionality and the new "norl" startup option. Did other minor nroff cleanups. Created aips.l file that simply sources the AIPS.l page. Users with $SYSUNIX (not $SYSMAN!) in their $MANPATH will pick up any stuff that we put in $SYSMAN (which is $SYSUNIX/manl). Moved nowhere. 8779. September 6, 1995 $APLHP/{ZPRMPT,ZGRLHI}.C Jeff Restored readline support for HP/UX--but without attempts at signal-handling for SIGT{TIN,TOU,STP}. If someone backgrounds AIPS under HP/UX with GNU readline enabled, it could behave strangely. (This is in the "functions but still needs some work" category right now.) Moved nowhere. 8780. September 6, 1995 $AIPSPUBL/COOK9.TEX & COOK9.PS Athol Put back an updated VLBI chapter, which should be complete up to 15JUL95. Moved nowhere. 8781. September 6, 1995 $APGNOT/UVCOP Phil Following Eric's suggestion I have changed the size of the logical array to use MAXCIF rather than UVBFSS. Moved nowhere. 8782. September 6, 1995 $YPGNOT/VPLOT Leonia VPLOT has been modified. 1. Now VPLOT can be used for editing data to recognise the points which have a big amplitude deviation. The amplitude of each individual point in the average interval is compared with the mean value. The rms is estimated for the solution interval sumulteneously. If the deviation of the amplitude from the mean is larger than the specified multiple of the rms then that point is marked in a new flag table. Plots are not generated in the bad points recognition mode. It is rather comfortable to flag the bad points and look the result by the same task. 2. Optionally the vertical lines that specified the minimum and maximum of the amplitudes found in the average interval are ploted together with the mean amplitudes. These lines help to identify the average intervals with bad points. 3. Run time has been decreased approximately in 2 times in any case and even more in the case of TIMERANG is less than the time interval of the data. 4. Scalar average has been added to the vector average which has already been present. The information about the type of averaging has been added to the header of the plots. Moved nowhere. 8782. September 12, 1995 CookBook Eric Finished putting back the CookBook. Changed Chapter Y to have it mention DATs and Exabytes and to be written better. Created COOKI.TEX and a template COOKI.TEMP. The latter is used by $AIPNOT/COOKNX.FOR to write the former. A bit of henad editing is needed to getr everything nice afterwards, unfortunately. Moved nowhere. 8783. September 12, 1995 ZSTRTA, ZACTV8 Eric Several character strings in these were limited to 48 characters even though they could carry the fully qualified path name. Changed them to 256. Moved nowhere. 8784. September 13, 1995 $HLPFIL/READLINE.HLP Jeff Checked in initial revision of the 'readline' help-file. I'll be polishing this up and adding information to it as I find free time. (Hahaha! ) Moved nowhere. 8785. September 13, 1995 LWPLA, TVCPS, TVRGB Eric Added the COPIES adverb to these. It applies only when the POstScript is going directly to the printer and is implemented with the "copypage" PS command. Moved nowhere. 8786. September 14, 1995 Missing help files Eric Created help files that were missing. They are adverbs BOXFILE CHANNEL CMETHOD CMODEL COPIES DONEWTAB DOTWO ERROR NOISE NUMTELL OUT2CLAS OUT2DISK OUT2NAME OUT2SEQ PBSIZE REFDATE Procedures: BREAK FEW MANY QUIT STOP TELFLM Moved nowhere. 8787. September 14, 1995 EHEX, REHEX Eric Changed AU8, NEWPARMS.001, POPSDAT.HLP and created help files EHEX and REHEX to make these procs also available as verbs. Moved nowhere. 8788. September 15, 1995 $APGNOT/FILLM.FOR Gustaaf FILLM's behavior with data the frequencies of which lie outside the official boundaries of a certain band is erratic. The Q-band receivers appear to give useful data outside the original range from 38 -- 52 GHz; this range has now been extended from 33 -- 56 GHz. Moved nowhere. 8789. September 16, 1995 $APGNOT/PCCOR.FOR Leonia Testing PCCOR for the case of ungomogenious IFs I have discovered some errors which affect this case. The errors have been fixed. Moved nowhere. 8790. September 19, 1995 $APGNOT/MK3IN.FOR Athol MK3IN was vulnerable to an array boundary error in one of the visibility buffers for very large record sizes. This error discovered when loading data with 28 IF's and 220 spectral channels per IF. Increased this buffer as well as an FQ table buffer and added a check on the buffer size to prevent this happening in the future. Moved nowhere. 8791. September 19, 1995 $YPGNOT/APCAL.FOR Athol Initialize the MB delay field to zero in the SN table. Was set to FBLANK but this upset the geodetic software. Moved nowhere. 8792. September 19, 1995 UTESS Eric Changed number of parameters to GTPARM. It was 198, but the common only had 195 words. The extra 3 replaced the POPS number and the user number and caused a little havoc. Solaris seems always to put the message common after the inputs common. Moved nowhere. 8793. September 19, 1995 CHNDAT Eric When the FQ table has one entry, FQMATC is its wisdom sets the desired FQ number to -1 (=> all ?). Then CHNDAT sees that -1 is requested, there is only one, sets the request to 1 and then dies if the FQ in the file is not 1. Changed the latter beghavior to select the 1 FQ in the file no matter what its value, so long as the requested FQ is <= 0. Errors are returned if the request is > 0 and not present in the file or the request is <= 0 and there is more than 1 in the file. Moved nowhere - link a whole lot. 8794. September 19, 1995 $APLNOT/CHEBY.FOR Athol New routine to calculate a Chebyshev polynomial sequence using the standard recurrence relation. Needed by the forthcoming polynomial BP software. Moved nowhere. 8795. September 19, 1995 $APLNOT/BPCOEF.FOR Athol New routine to compute a complex polynomial bandpass. Phrased such that it can be called from any point where BP conversion may be required (eg. DATBND, POSSM; this code centralised in one routine for BP fitting and application. Moved nowhere. 8796. September 19, 1995 FIXWT Chris New task that makes a copy of a data set with weights derived from estimates of the standard deviation of amplitudes in the data. The resulting weights should be a better approximation of the inverse variance weights than those given by most input sources. Moved nowhere. 8797. September 20, 1995 OTFUV Eric Changed it to allow for more simultaneous scan numbers - it was limited to a range to 4096. Now it is limited to a range to 32768. Added adverbs BCHAN and ECHAN in case people are so foolish as to not want all their channels. Added test for INFILE = IN2FILE which cannot be right. Moved nowhere. 8798. September 20, 1995 PATGN Eric Added operation to make single-dish beam patterns - like PBCOR. Moved nowhere. 8799. September 20, 1995 AIPS.HLP Eric Updated to include the latest from the man page - the readline stuff, tpok, tvok, changed remote, etc. Moved nowhere. 8800. September 22, 1995 CookBook Eric Chapter 2: Add the readline and other aips man changes Chapter 3: Hex -> extended hex and mention EHEX, REHEX Chapter 5: Mention FIXWT Chapter 12: Hex -> extended hex and mention EHEX, REHEX Chapter 0: WWW more prominant, chapter 2 Chapter I: Index always changes with anything Moved to ftp/www area. 8801. September 22, 1995 CLCAL Eric The call to FQMATC passed in the user adverb XFREQ, a real*4. But FQMATC expected a REAL*8 so the user value of -1 led to a bad multiply and an abort on the DEC machines. Moved nowhere. 8802. September 22, 1995 VPLOT Leonia The min/max lines which were installed for amplitude early (#8782 September 6, 1995) have been implemented now for phase, real and image. The plotting of real and image was available only for SOLINT=0. Now these options function in the average mode also. Moved nowhere. 8803. September 23, 1995 ZGRLTC,ZGRLHI,ZPRMPT,RLHIST Jeff For all 7 OS's that support GNU readline in $APL$ARCH: Removed TAB-key unbinding in ZGRLHI.C. Added TAB-key expansion functionality via new file ZGRLTC.C. Changed ZPRMPT.C to check prompt character and to disable the TAB key when prompting for user number (prompt character of '?'). Added hook in $AIPSUB/RLHIST.FOR to initialize and activate TAB-key expansions immediately after initialization of GNU readline and history recovery. Added stub routine $APLGEN/ZGRLTC.FOR for OS's without readline support. Returns IERR=1 to indicate that TAB-key expansion is disabled. Moved nowhere. 8804. September 24, 1995 VPLOT Leonia The COMLNK of VPLOT (#8800 September 22, 1995) failed on the IBMs. The reason was in a careless line I added recentely and in another one which existed since the time when VBPLT was modernized to VPLOT. It is strange there was no a complain on the second careless line during several modernisation of VPLOT. I have fixed the both lines. Moved nowhere. 8805. September 25, 1995 ZGRLTC.C Jeff Made TAB-completion interface smarter about quoted arguments (it no longers diddles the case on them), disabled all hooks to filename-completion for arguments (to prevent user confusion). Also cleaned up code and made it generally more efficient, fixed bug with finding the file containing the completion list. Moved nowhere. 8806. September 27, 1995 Polynomial banbpasses Athol Modifications to allow the application of polynomial bandpasses as generated by the forthcoming task CPASS. This is achieved by added a keyword BP_TYPE to the BP table. If this string is blank or not present then the bandpass is assumed standard, else the keyword identifies the type of polynomial coefficients stored in the BP table. BPASS generates standard bandpass solutions and CPASS generates polynomial bandpass solutions. There is no impact on the DOBAND or BPVER adverbs from the user perspective; the bandpass type is determined during calibration and polynomial bandpasses are expanded as needed. For data correlated at the VLBA correlator, frequency shifts arising from antenna-based fringe-rotation are included before the bandpass is computed. Specific modifications: $INC/DCVL.INC - added LTYPBP*8 identifying the BP_TYPE keyword. $APLNOT/BPINI.FOR - add new keyword BP_TYPE; change calling sequence accordingly. $APLNOT/BPASET.FOR - modify BPINI calls. $APLNOT/BPREFM.FOR - modify BPINI calls. $APLNOT/DATBND.FOR - expand polynomial bandpasses on the fly; incorporate VLBA specific fringe rotation frequency shifts directly in expansion. $APLNOT/BPFSEL.FOR - renumber only IF's and not frequency channels when a polynomial BP is subject to frequency selection. Modify IBCHAN to reflect that the polynomial BP is shifted. $APLOOP/TABLEUTIL.FOR - change calling sequence to OBPINI to reflect new BP_TYPE keyword. $YPGNOT/POSSM.FOR - modify subroutine POSSBP to deal with polynomial bandpasses when plotting BP tables. $APGNOT/BPCOR.FOR - disable smoothing of polynomial bandpasses in BPCOR for now. $APGNOT/UNCAL.FOR - add blank BP_TYPE keyword to BPINI calls. $QPGNOT/BPASS.FOR - modify BPINI calls; do not allow user to append standard BP solutions to an exisitng table containing polynomial bandpasses. $QPGNOT/CVEL.FOR - modify BPINI calls in subroutine CVBPST to mirror BPASET. Modify CVBAND to match new DATBND. Correct pre-existing buffer offset error in VLBA FFT shift (found by PJD); also fixed minor shift error for VLBA XP bandpasses. $APGNOT/FITLD.FOR - modify subroutine BA2BP to write blank BP_TYPE keyword. Note: BA tables not yet written by correlator. Substantial re-linking. Moved nowhere. 8807. October $APGNOT/PCCOR.FOR Leonia The task has not coppied source ID from a PC table row to the SN table row leaving it equal zero. It's fixed. Moved nowhere. 8808. October $YPGNOT/VPLOT.FOR Leonia During last modernisation I put a mistake in the block of frequency/IF averaging. As a result VPLOT has not plot the average amplitude for some range of bchan, echan. It's fixed. Moved nowhere. 8809. October 4, 1995 $QYPGNOT/UTESS.FOR Dave Caught UTESS up on some changes which had been put into VTESS but not UTESS over the years. Specifically, item 7690 (adding code for when NOISE=0 to avoid arithmetic exception), 7722 (clearing read status of the beam), and 8048 (writing the history file to the output maps). Moved nowhere. 8810. October 5, 1995 $APGNOT/LTESS.FOR Dave Similar change as yesterday for UTESS; the read status of the first map was not being cleared. Changed 'READ' to 'REST' in the second call to CATIO. Moved nowhere. 8811. October 6, 1995 QCLEAN Eric Changed the TELL section to put the beam parameters into each Clean field. Putting them in the parent Clean object did not cause them to end up on disk. Moved nowhere. 8812. October 7, 1995 $APLNOT/BPCOEF.FOR Athol Force BPCOEF to require FBLANK termination of polynomial coefficients in BP tables. No user impact. Moved nowhere. 8813. Oct 7, 1995 $QPGNOT/CPASS.FOR Athol New task to determine polynomial bandpass response functions. Writes BP tables which are applied with adverbs DOBAND and BPVER in the standard manner. Moved nowhere. 8814. October 10, 1995 JMFIT ERic Changed the convergence EPS from 1E-10 to 1E-14. Images with goood signal to noise and peaks of 100's of microJy iterated too fast and never solved for widths etc. Moved nowhere. 8815. October 10, 1995 $APGNOT/TBOUT.FOR & TBIN.FOR Athol Increased some internal buffers to cope with larger tables, such as the BP tables produced for spectral line VLBA datasets. Moved nowhere. 8816. October 12, 1995 $APGNOT/GSTAR.FOR Dave IncorrectNPARMS was causing problems. Changed to 12, things seem to work now. Moved nowhere. 8817. October 13, 1995 MX Eric MXSEL did not correctly determine the max frequency when averaging spectral channels and so excluded data for no good reason at all. UVLIMT messages did appear when this happened. Moved nowhere. 8818. October 13, 1995 FFT Eric Happily created cubes and then FFT'd the 1st plane only. Changed to loop over planes. Moved nowhere. 8819. October 13, 1995 $APLPGM/SQASH.FOR Gustaaf SQASH would determine sum or average over one or more planes of a cube. I added the two options to write out the maximum or the minimum along the specified axis/axes. I also changed the default: DPARM(1)=0 to specify average, and DPARM(1)=1 to specify sum; this is more in line with the SUMIM convention. Added plenty warning messages to alert the user. Other DPARM(1) values lead to min and max. Moved nowhere. 8820. October 14, 1995 $APLNOT/BPCOEF.FOR Athol Make more resistant to bad inputs. No user impact. Moved nowhere. 8821. October 14, 1995 $QPGNOT/CPASS.FOR Athol Optimize the code by computing the polynomial terms less frequently. Moved nowhere. 8822. October 16, 1995 UVPRT Eric Once again I forgot that UVGET changes the uv structure. A multi-source in is not a multi-source out if only 1 source is included. So MULTI true but ILOCSU = -1. Fixed source number test. Moved nowhere. 8823. October 16, 1995 CVEL Eric Corrected it to use the buffers that were actually there. Two of them had BUFSZ = 4096 * 2 but were 8192 words long. There are some really large buffers declared, but then a lot of small ones are used. With 4096 words, a 512 channel, 2 IF data set has trouble. So the doubling I did just postpones the problem. Moved to siamang today. 8824. October 17, 1995 UVPRT Eric Changed it to do a better job with the source info. The choice of MULTI should be made only after UVGET ('INIT' call. The 1 source will then be in the header if there is only 1. Moved nowhere. 8825. October 17, 1995 ALGSTB Athol/Eric Corrected an addressing bug. In the case with tons of uv samples on a single row of the grid, the buffer could hold more samples than the ap and too many would get loaded to the ap. Changed ILENBU to MIN (MAXREC and LENBU). Moved nowhere. 8826. October 17, 1995 IMAGR Eric Did not read circular windows from the window text file correctly. It turned them into rectangular from the BLC. Moved to siamang today. 8827. October 17, 1995 DADEVS.SH, DADEVS.PL, DATOUCH Pat Modified the scripts to allow for site-specific DADEVS.LIST and .dadevs files as follows: if $NET0/DADEVS.LIST.$SITE exists, it is used in place of $NET0/DADEVS.LIST. Likewise if a user has a file ~/.dadevs.$SITE it will be used in place of ~/.dadevs. This is dependent on the "SITE" variable as used in HOSTS.LIST; here in CV we have two sites: NRAOCV for big-endian systems and VCOARN for the little-endians (Linux, Alpha). Thus I will now be able to have separate DADEVS files for these two incompatible "sites". Done at the request of Debbie Haarsma at MIT. Moved nowhere yet. 8828. October 17, 1995 $APGNOT/VBGLU.FOR Athol VBGLU concatenates samples which are aligned in time by less than a tenth of an integration period. The integration time was stored in the wrong units. Also modified subroutines GETPTR and DAMATC; some records were not matched correctly. Moved nowhere. 8829. October 17, 1995 $APGOOP/POLSN.FOR K.Leppanen/Athol Added Kari Leppanen's modifications to POLSN to the system. These include: i) Compute R-L delay and phase average for each solution interval; ii) ensure amplitude of each solution is normalized. Moved nowhere. 8830. October 17, 1995 imaging Eric Found a bug applying to gridded subtraction when field 1 is empty. Changed GRDAT to set CELLSG on any field, CCSGRD to not init TFLUXG, and GRDSUB to init TFLUXG. Moved nowhere. 8831. October 18, 1995 GRIPE Eric Changed the gripe verb to e-mail the gripe to aipsmail (and me at least for now). Changed AUC and CHGRIP to create and write a text file and call ZGMAIL. Created ZGMAIL ($APLGEN stub and $APLUNIX) and ZGMAI2.C; modified GRIPE.HLP; updated GRIPR. Moved nowhere. 8832. October 19, 1995 $YPGNOT/VPLOT.FOR Leonia The English of the message printed after the task starts up was not good enough. It has been improved following Athol advise. Moved nowhere. 8833. October 20, 1995 SYSTYP in DDC.INC Eric Changed meaning of SYSTYP which used to be only VMS vs UNIX into a full $ARCH sort of value. Changed ZDCHIC to take this as an input only argument - we will set it for keeps in ZDCHI2. Changed: ZDCHIC $APLGEN (stub), $APLUNIX, $APCRI - also put in Z_systyp[5] C common ZDCHIN $APLGEN (typos), $APLCRI, $APL3090 - update to modern form ZDCHI2 Changed 'UNIX' to something else closer to $ARCH $APLCRI -> CRI $APL3090 -> 3090 $APLHP -> HP $APLLINUX -> LINX $APLSOL -> SOL $APLSGI -> SGI $APLALLN -> ALLN $APLCVEX -> CVEX $APLDEC -> DEC $APLNRAO1 -> CVEX $APLIBM -> IBM $APLSUN -> SUN $APLVLAC1 -> CVEX $APLVAX -> VAX ZDCHI2 Removed $APLBELL created $APLMASC, $APLULTRX ZTPOPD Removed extraneous $APLDEC, changed $APLGEN to check all Digital SYSTYP's (VMS, DEC, ULTR) and change LRECL to words Moved nowhere. 8834. October 20, 1995 GRIPE mail Eric Changed AUC to leave out all To: Subject: etc lines. Changed $APLUNIX/ZGMAIL to pass an extra argument to ZGMAI2 saying how to insert the subject (1 insert in file for ALPHA and SOL, 0 others use 's " "' option), change ZGMAI2 to use mail rather than rmail, to use pipes instead of system calls, and to use unlink to delete the file rather than a system call. Moved nowhere. 8835. October 20, 1995 VLBIN Eric I doubt that this old chstnut still works. Nontheless, it was assuming that all SYSTYP = 'UNIX' had to be translated from Vax floating to IEEE and then used IEEE and also called ZI16IL for some reason for Unix machines. SYSTYP no longer hjas this value so I changed all of it to examine the actual floating format codes and to do the changes correctly (I hope). Moved nowhere. 8836. October 20, 1995 START_AIPS in SYSUNIX Chris Added the CDE desktop's terminal emulator (dtterm) to the list of TERM values recognized as belonging to X11 terminal emulators. Moved nowhere. 8837. October 21, 1995 $SYSPERL/check_version.pl Jeff Added Perl version 5.000 to list of "tested" versions; they're running 5.000 under Solaris (at least) here at the AOC, so I guess you could say that things are quite well tested under that version by now. Moved nowhere. 8838. October 21, 1995 PCNTR Eric Corrected the drawing of polarization lines and Clean beams. They both tried to plot simple angles on the page. One cannot do that in the presence of rotation or non-square pixels. All such plots have been wrong. Changed PLTBEM and PCNTR. Moved to patch area tomorrow. 8839. October 21, 1995 $SYSUNIX/DADEVS.PL Jeff Made slight (functionally insignificant) modifications to Pat's enhancements for site-specific data-disk files; "dot" concatenation of strings is somewhat inefficient in Perl, is not needed in this case for the @DADEVS array assignments, and makes the code look kinda' funky. Moved nowhere. 8840. October 23. 1995 TPMON Pat Added a somewhat crude but effective host-based authentication to TPMON. The (new) file $NET0/TPHOSTS will now contain a list of hosts from which TPMON will accept requests. ALL others will be rejected. The format of this file is fairly simple: one host per line, blank and comment (#) lines ignored, and you can have a SINGLE wildcard as the first character of a hostname, e.g. *.aoc.nrao.edu accepts all AOC machines. The comparison of hostname with entries in this file is performed regardless of upper/lower case. Also added a help file for TPMON, mainly for AIPS managers. Moved nowhere yet, will move to patch area shortly. 8841. October 24, 1995 CookBook Eric Rewrote Chapter 10 and 11 into 1 chapter called 11. EXITING FROM, and SOLVING PROBLEMS IN, AIPS. The description of Gripes was changed to include the e-mail on GRIPE and GRDROP. The problem solving was altered to include modern machines and conditions . Changed references to 10 in Cpaters 1, 3, and, Z. Fixed chapter number references to other chapters in 7. Started on Chapter 10 SINGLE-DISH DATA IN AIPS enoughh to do preliminary Table of Contents and Index. Putback chapters 1, 3, 7, 11, and Z, but am still holding 10, 0, and I. Moved to WWW, nowhere else. 8842. October 24, 1995 GRDROP Eric When a grip is typed in, it is e-mailed to aipsmail. If the user then decides he/she has made an error, he/she can GRDROP it, but we still won't know that. It would be best if he/she entered another gripe saying ignore previous I goofed or whatever. Changed AUC to e-mail to aipsmail the text of any gripe being dropped with GRDROP. That at least will tell us that the user appears to have changed his/her mind. Moved nowhere. 8843. October 24, 1995 SCMAP Eric Changed SCMAP, IMAGR, and UVUTIL to check for a max u baseline in the header before copying the data. Otherwise, the header ended up with the max u of the last subarray (or whatever) appended to the file. This led to failures in weighting, gridding, and gridded subtractions. Changed message levels in CATKEY - they were anoying when they should be suppressed. Moved nowhere. 8844. October 25, 1995 KNTR Eric Dashed contours were being drawn with a ddudu (d-> pen down, u-> pen up) pattern. With LPEN=3 in LWPLA, the PostScript pen width more or less filled in the pen up parts of the plot. Changed the pattern to dduu or dduuu or dudu for bigger images. Usrs who do too many frames of large picture must use LPEN=2 or even 1 if they hope to see the dashed contours. Moved nowhere. 8845. October 26, 1995 IMAGR Eric Fixed: (1) Boo-boo made 2 days ago in which I passed in a constant to get an output value having to do with max baseline. (2) $QPSAP/UVSRT - MAPCLR changes the number of files to zero as it ends for no good reason. This means that each call must have a newly initialized count. We cannot put that count in a DATA and assume it is good. Sigh...lots of SC files left behind. (3) $QOOP/QUVUTIL - the order of the code was such that taper was not applied to NAtural weighted images. Corrected the order. (4) $QOOP/QCLEAN - add field clean flux to clean(field) object and total flux to overall clean object (5) Read in the field and total clean fluxes and number of components for the field and add them to the history. (6) Upgrade SCMAP history to be more like IMAGR Moved nowhere. 8846. October 27, 1995 ZVTPO3/ZVTPGC (TPMON) Pat Modified these routines to check all aliases for the connecting host as well as the official name. Also, for any entry beginning with a digit in TPHOSTS, assume it's an IP address and compare the connecting IP address with it (not the hostname or aliases). Finally, allow crude wildcarding by IP addresses, e.g. 192.33.115.* (only a single trailing wildcard allowed, but this also covers, e.g. 192.33.* if desired). Modified ZTPOPR so that the error messages "PERMISSION DENIED" and "FILE ALREADY EXISTS" are issued in appropriate circumstances. Moved nowhere yet, perhaps should update patch area... 8847. October 27, 1995 IMAGR, SCMAP Eric The new history routines would fail if TFLUX had not been added to the Clean object by Cleaning. In IMAGR, at least, NITER=0 is common and that provoked the failure. Corrected it to handle the missing TFLUX as a signal to omit the Cleaning HIstory rather than a signal to fail. MOved nowhere. 8848. October 27, 1995 Alignment Eric REAL*4 and REAL*8 are not interchangable. For one, the 1st 32 bits usually do not have the same format and for two the use of a REAL*4 where a REAL*8 is expected causes the following 32 bits to be overwritten and a possible alignment error (on Dec which cares to point it out - most machines have some trouble with misaligned values). SOURNU declared FREQO and RESTFQ to be real*4 and than called TABSOU which (very sensibly) assumes thes to be REAL*8. Unfortunately, this will require an INSTEP4 to repair since ACTFRQ calls SOURNU, FQMATC calls ACTFRQ, and UVGET and lots of others call FQMATC. Moved nowhere. 8849. October 27, 1995 SNPLT Eric The error above cannot be regarded as really accidental, but this next one is deliberate sabotage: INP = 0 TEXT(INP:INP+3) = ' ' // TYPE // ' ' causes 1 byte ahead of TEXT to be overwritten. This made illegal plot files on Dec machines and overwrote God knows what on other machines. This gem appeared 3 times in SNPLT. Character strings and most everything else begin at 1. Moved nowhere. 8850. October 27, 1995 SDGRD Eric Forgot to add round convolving functions to QGRD5 and QGRD8. This forced the addition of the increment parameter to the call to QGRIDA which forced additions to UVMAP and GRIDR. Corrected a message actually in SDGRD and the call to QGRIDA. Moved nowhere. 8841. October 30, 1995 $APGNOT/SETJY.FOR Athol SETJY did not correctly incorporate source-based frequency offsets found in the SU table when calculating the total frequency used in the velocity calculations. Fixed this. Impact is limited to those cases where the SU table has a non-zero frequency offset which is significant with respect to the rest frequency. Moved nowhere. 8842. October 31, 1995 $APLUNIX/ZVTPO3, ZVTPGC (TPMON) Pat The loop over possible aliases in the logic to match hostnames was faulty; the relevant pointer was not being incremented. This caused infinite loops and puzzled workstation owners. I'm still not sure I've nailed the problem but at least it outright rejected me instead of sitting there dumbly... more to follow no doubt. Moved to my login area at the AOC for testing... 8843. October 31, 1995 GRCLEAR Eric Fixed the inputs section of the help file to allow a value of 0 (all). Added a couple of comments. Moved nowhere. 8844. November 1, 1995 OTFUV Eric Added a routine to make an antenna table to attach to the output uv data set. This should keep some programs happier and will provide the antenna coordinates should azimuth or elevation need to be computed. Moved nowhere. 8845. November 1, 1995 INDXR Eric Changed it to be forgiving on sort order (T* not TB required) and to handle a missing source table/parameter happily. Added the ability to write CS tables for single-dish and use the scan random parameter if present. Corrected a bug: if more than one subarray was present, then the CL table written was wrong. It would make entries for only one subarray at any one time and then wipe out the full antennas used list rather than only the part of that list for the present subarray. Moved nowhere. 8846. November 1, 1995 $APLUNIX/ZVTPO3, ZVTPGC (TPMON) Pat The logic that was deciding whether a host was authorized got broke when I added the loop over host aliases ("auth = 0" was set inside this loop, not outside). Fixed, and also fixed the printout TPMON gives (to stderr) when access is denied to print the official hostname and also the IP address. Moved to the AOC (aguila only) this date. 8847. November 1, 1995 ALTDEF Eric Changed AU7 so that the correct result would be found when ALTDEF is done on a header having velocity in the main part of the header. It referred to KDARV+I-1 but the alternate ref value is not an axis and should never have the +I-1 part. This never worked (except if velo was axis 1). Moved nowhere. 8848. November 2, 1995 Misc Eric Changed LSTHDR and KWIKHD to display the clean beam parameters even for uv data. This location is for the single-dish beamwidth in sd data. Change ZGMAI2.C to mail the gripes to additional addresses incl aipgripe@zia.aoc.nrao.edu. Moved nowhere. 8849. November 2, 1995 TPMON/ZVTPO3/ZVTPGC Pat Fixed bugs in the string comparison that made it impossible to use if there was a case mismatch in the alias and the contents of the TPHOSTS file. Also added extensive debugging info bracketed by appropriate #if directives. Moved nowhere. 8850. November 2, 1995 $SYSUNIX/DADEVS.PL Jeff Added small sanity check to ensure that calling scripts/programs pass a PID (or something "unique") as the first argument. DADEVS.PL now dies (with an explanation message) if not. Moved nowhere. 8851. November 4, 1995 $APLNOT/BPCOEF.FOR Athol Removed stray TAB character from source code. No user impact. Moved nowhere. 8852. November 7, 1995 $APLNOT/GETSOU.FOR Leonia/Athol Precess single source coordinates to return correct apparent coordinates. Was previously returning mean coordinates of epoch as apparent coordinates in the single source case. Moved nowhere. 8853. November 7, 1995 SDMOD Eric new task. Modifies single-dish uv data replacing data with up tp 4 Gaussians plus noise or adding Gaussians to data. Moved nowhere. 8854. November 7, 1995 $APGNOT/PCCOR.FOR Athol Contained dormant local routine TBTIME which was never called. Removed this dead code: no user impact. Moved nowhere. 8855. November 8, 1995 $HLPFIL/HFPRT.HLP Ed F Changed a few words for clarity in HFPRT help file. No changes made to code in $APGOOP/HFPRT.FOR. Moved nowhere. 8856. November 8, 1995 $APGOOP/HF2SV.FOR Ed F Fixed time round off to prevent seconds = 60. Changed experiment number of 0. Moved nowhere. 8857. November 8, 1995 $APGOOP/CL2HF.FOR Ed F Fixed several bugs: correction of phase rate; added table sort; fixed uv-scaling, fixed phase difference. Moved nowhere. 8858. November 8, 1995 $APGOOP/MBDLY.FOR Ed F Improved a few output formats. Moved nowhere. 8859. November 8, 1995 $HLPFIL/IMFIT.HLP Ed F Claified zero level use in help file. Moved nowhere. 8860. November 8, 1995 $APGNOT/FITLD.FOR Phil Made a series of minor changes. Increased some buffer sizes to parameterized values. Ensured that the FFT artifact correction message is only written to the HI file when it is applied. Tidied up the printing of the frequencies found when FITLD runs. Moved nowhere. 8861. November 8, 1995 DBCON Eric Changed to set DOPOS to 0 when single-dish data and when there is no position difference. Otherwise it was talking about shifts even though (one hopes) nothing much shifted. Moved nowhere. 8862. November 9, 1995 DISMOUNT.HLP Eric Corrected typo, improved wording. Moved nowhere. 8863. November 9, 1995 UVPLT Eric Corrected code that depended on UVGET defaulting because it might be invoked before UVGET is called (if the plot scales are set by the user). Changed it to recognize single-dish data and call things more sensible names and units. Corrected bug handling plots including more than one spectral-line channel. It had a bad error in figuring out the default scaling since an array of (3,*) was called an array of (2,*). This affects interferometer data too. Moved nowhere. 8864. November 10, 1995 $QPGNOT/CPASS.FOR Athol Corrected time mis-labelling in the BP table which could occur for SOLINT=0. Moved nowhere. 8865. November 13, 1995 $APLNOT/UVTIME.FOR Athol Add new subroutine UVTIME, which determines the start and stop times in a multi-source or single-source uv-data file which is sorted in TB order. This is a copy of local routine TBTIME which was in use in five tasks in almost identical form (including an upcoming change to UVFLG). The local TBTIME routine was removed from the following tasks and the call to TBTIME replaced by a call to UVTIME: $APGNOT/VBGLU.FOR $APGNOT/ACCOR.FOR $YPGNOT/FRMAP.FOR $YPGNOT/POSSM.FOR - pending. Moved nowhere. 8866. November 14, 1995 $APGOOP/MBDLY.FOR Pat Removed EXTERNAL references to unused BLAS routines ISAMAX and SMIN; they're not used in the code and generated a link error on the Alpha. Moved nowhere. 8867. November 14, 1995 $YPGNOT/FRMAP.FOR Leonia FRMAP was wtitten especially for VLBI data to use fringe rate for mapping spot sources distributed over large scale. Originally I thought this program was not useful for VLA data because the fringe rate's angular resolution is not high for VLA. But having reduced the methanol masers observed in the A configuration, I discovered the spots distributed on the scale of tens of arcseconds. In this case FRMAP can be effective to determine the windows for clean mapping. So I've updated FRMAP for VLA data, taking into account the RH system of coordinates (instead of LH for VLBI) and direction of X axis in the local meridian (instead of greenwich for VLBI). Moved nowhere. 8868. November 14, 1995 POSSM Eric Correct output format remark and get rest frequency from header if it is not in the source file to improve text file output. Have autocorrelation data treated as reals not amplitude (which people will think are > 0). Have it read an antenna file if there is one - now have them with single-dish. And have it recognize single-dish data and plot real only and use K rather than Jy. Correctd output text file to have a column of frequencies and to attempt to have the correct frequency and velocity for each of the (possible) multiple IFs. The velocities were wrong before and the frequencies were not given Moved nowhere. 8869. November 14, 1995 UVWAIT Eric Some multipliers were not computed when doing natural weighting, causing the actual taper applied to be zero. Moved these multipliers out of the IF areas. Moved nowhere. 8870. November 14, 1995 $APGNOT/PCCOR.FOR Leonia Bill Cotton has demonstrated PCCOR does not work with REFANT=8. I have discovered that PCCOR does not work with any REFANT>5. The reason was a wrong definition of the antennas sequences in a baseline if REFANT>5. I've fixed it. Moved nowhere. 8871. November 14, 1995 $APGNOT/UVFLG.FOR & UVFLG.HLP Leonia/Athol Added facility to flag data below a specified elevation threshold. Moved nowhere. 8872. November 15, 1995 RECAT Eric Added a line of code to replace all file sequence numbers <= 0 with 1000 + CNO. This should repair the seq 0 files which appear occasionally. Moved nowhere. 8873. November 15, 1995 OTFUV Eric Fixed it to work on byte swapped computers (e.g. PCs and DEC Alphas). It detects that the first control words have the wrong value and then tests to see if they get correct values when swapped. If so, all data are swapped from canonical (IEEE, big endian) to local. Moved to the patch area. 8874. November 19, 1995 FITLD Phil Increased the buffer size of the vis. array in the random groups FITS routine UVFDAT. It was too small for large AT line data sets. Also fixed a minor Fortran error that Pat Murphy discovered. Moved nowhere. 8875. November 20, 1995 FILLM Gustaaf FILLM integrates visibilities according to user supplied parameters. To this end, for each time stamp, the "end of current integration" is read from tape. Before 1988, when the online system was completely rewritten, a different value was used than after 1988. FILLM until now only read the latter value, and gave inconsistent results for data from before 1988. This became evident when the online integration time was not an integral multiple of 10 seconds. FILLM now checks the version number of the online software, and reads the appropriate "end of current integration time" value accordingly. 8876. November 21, 1995 IBLED Eric Continuing to work on single-dish in AIPS: IBLED's reaction to single-dish data (and other kinds) was poor. Changed: (1) Changed to require T? not TB sort order. (2) Changed to dummy info if no antenna file. (3) Changed to average Reals not amplitudes for autocorrelations and to compute rms for decorrelations. (4) Changed to allow decorrelation display. (5) Changed to avoid re-loading data when things have not changed. There were a few tests, added several more. (6) Dropped most of the turn-on cursor calls, changing them to read the cursor (if CURSON else turn it on). Dropped all turning off of the cursor. This should speed up user interaction, since the cursor on moves the cursor away from wherever the user has pointed it. (7) Select frame did not work and now does - I do not know what I did to change this. (8) Fixed header of work file - the RA and Dec axes got lost if there was not an IF axis to begin with. (9) Updated the help file, making the help and explain lines longer and trying to clarify the integration time parameter. 8877. November 21, 1995 $SYSUNIX/ZLPCL2 Pat At a user's request, modified this script so it permits printer output to go to a PS previewer or other script in $SYSLOCAL. Moved nowhere. 8878. November 21, 1995 FRING Chris Fixed the AP size check to estimate AP usage correctly when averaging frequencies within IFs (APARM(4) > 0). Moved nowhere. 8879. November 21, 1995 PRTUV Eric Tested ILOCT > 0, should be >= 0. Moved to NRL, nowhere else. 8880. November 21, 1995 SAD Eric Corrected the handling of position angles. The deconvolved results were not going through the coordinate handling routines and hence were not referred to North but instead to the right in the image pixel plane. Corrected an erroneous use of ABS functions elsewhere and changed the error computation for position angle to match the theory developed by Condon. Corrected the position angle error computation (forgot a sqrt) on November 25. Moved nowhere. 8881. November 22, 1995 $APGNOT/MK3IN.FOR Athol Modified MK3IN to allow AC lag functions starting at lag 1 rather than lag zero. The starting lag number depends on the assignment of lag modules at the correlator. Previously AC lag data usually included lag 0 but recent data have included instances of AC lag data starting at lag 1. Modified subroutine ATLAG to deal with this case, and skipped the check for blanked data on AC lag zero in ATREC. This is set to one later in AC normalization anyway. Moved nowhere. 8882. November 27, 1995 POLSN.FOR Bill Cotton Fixed problem in format which made it blow up instead of give error message. Moved nowhere. 8883. November 27, 1995 CROSSPOL.001 Bill Cotton Modified to run the cross polarized FRING multiple times, once per "other" antenna to get better statistics on the crosspolarized delays and to average out some of the variations in instrumental polarization. Moved nowhere. 8884. November 27, 1995 IMFIT, JMFIT Eric Corrected both tasks to use the coordinate routines to determine component widths and position angles when quoting the sizes in arc seconds. The position angles were wrt to something but not necessarily North. Added code to do estimates of noise based on the true rms. Unfortunately, the internal estimates of uncertainties are usually larger than these theory estimates. The tasks give both estimates anyway. Added PRTLEV to JMFIT to cut off the excess printing of fit information. Moved nowhere. 8885. November 27, 1995 Table sorting Athol The calibration system was having difficulties in sorting large BP tables due to inadequate array declarations in several routines: $APLNOT/ITBSRT.FOR - previously assumed a maximum table record size of I*4(2048). Increased this to 16384 and added a check to prevent array overruns. $APLNOT/TABSRT.FOR - buffers BUFFER and TABUFF are adjustable arrays but were declared fixed length. Generated errors when compiling with array bounds checking. Fixed. $APLNOT/OTBSRT.FOR - as for TABSRT. Symptom: SPLIT would fail with a segmentation violation in applying a 1024-channel, dual-polzn, single IF bandpass. Moved nowhere. 8886. November 28, 1995 UVCOP Leonia The selection criteria have been added when coppiing PC table (antennas, IFs, FREQID, sources, time). To do it: PC table has been excluded in COPYHI from the list copied by ALLTAB. A new subroutine PCSEL has been added in COPTAB. Moved nowhere. 8887. November 28, 1995 Large BP tables Athol Implement changes to allow larger BP table records than presently permitted. In particular, accomodate ATCA 2048-channel, dual polarization bandpasses. Double the BP table I/O buffer to 16384 in the following routines and tasks: $APLNOT/BPINI.FOR $APLNOT/BPREFM.FOR $APGNOT/BPCOR.FOR $APGNOT/UNCAL.FOR $APGNOT/FITLD.FOR - pending $QPGNOT/BPASS.FOR - change several format statements to accomodate 4-digit channel numbers. Moved nowhere. 8888. November 28, 1995 SL2PL Eric Changed the setting of the max and min so that one can get the default for min or max and set the other (by setting the one reasonably and the other way off). Moved nowhere. 8889. November 28, 1995 INDXR Eric Changed it to use CPARM(5) to cover up for a missing AN file for single dish data only. CPARM(5) = max antenna number. Moved nowhere. 8890. November 28, 1995 CSCOR Eric Changed the sort test to 'T?' rather than 'TB'. Added CPARM to give the antenna longitude and latitude when the AN file is missing. Moved nowhere. 8891. November 29, 1995 OTFUV Eric The integration time is not the best way to get the number of samples - the data length gives that better since it is not rounded up to the next integer * 512. Moved to the patch area this date. 8892. November 29, 1995 SDCAL.HLP Eric Corrected it to refer to what it does, not SPLIT. Moved nowhere. 8893. November 30, 1995 QCLEAN Eric Fixed it - this time I hope - to store the TELL Clean beam parameters on the output image headers in IMAGR, WFCLN, ... Moved nowhere. 8894. December 1, 1995 ACTNOISE Eric Changed SAD, IMFIT, JMFIT to use the value of header keyword ACTNOISE (if present and > 0) instead of fitting the histogram of the full image to determine the rms. Changed POPSDAT.HLP, NEWPARMS.001, and AU7A to implement a verb ACTNOISE to put ACTNOISE into the image header. Created an ACTNOISE help file. Changed IMEAN to use ACTNOISE as an estimate if PIXSTD = 0 in order to avoid 2 passes. Also changed IMEAN to use histogram cells of 1/8 of PIXSTD (or ACTNOISE) in the first pass and 1/60 of the source-included rms if a second pass is required. Moved nowhere. 8895. December 5, 1995 SAD Eric Corrected computation of plane in input MF file and in window for subtracting components on input and at the end. Moved nowhere. 8896. December 5, 1995 MFPRT Eric Changed it to allow the user to control the sort order and to use reasonable units for coordinates depending on the cell size. Moved nowhere. 8897. December 6, 1995 TEKSRV Pat This failed to start under OSF/1 (Digital Unix) on Alpha AXP ONLY. Other architectures and systems seem unaffected by this. The default protection on Dec Alpha windows (xterm, dxterm, dtterm, whatever) seems to inhibit the ioctl() call that set the terminal characteristics to raw. This is nominally to prevent the often long character strings sent to the "terminal" (tek4010 code) from being broken up with unwanted newlines or 's. Taking out this whole section of code seems to work, though very complex plots may have problems. Moved to the patch area today. 8898. December 6, 1995 RSTOR Eric Changed the test on NX and NY to admit that 8192 is a power of 2. Also doubled the buffer sizes. Moved to rhesus today. 8899. December 6, 1995 UVWAIT Eric Removed a message that is meaningless when uniform weighting is not done. Moved nowhere. 8900. December 6, 1995 REBOX Eric Corrected AU5C, REBOX verb. It referred to the wrong variable rather than the one returned by GRBOXS for the number of boxes. It then branched to the wrong place for closing down and produced all sorts of messages appropriate for FILEBOX. Moved to solaris and ibm today. 8901. December 6, 1995 KNTR Eric Changed AXSTRN to do leading zeros properly in sexagesimal seconds format. Changed KNTR to display the third axis coordinate correctly no matter what third axis it is. Changed it to offer the absoulote value of the coordinate or relative to the reference pixel. Moved nowhere. 8902. December 6, 1995 FILLM Gustaaf FILLM creates a CL (and TY) table entry under three conditions: 1) at the beginning of a scan, 2) at the end of a scan, and 3) when CPARM(8) minutes have elapsed since the previous entry. In the particular case where a scan consists only of one integration, two identical entries were written due to 1) and 2). Logic was added to FILLM to check before writing and 'end-of-scan' entry whether a 'beginning of scan' entry already was written at the same time. 8903. December 6, 1995 $QPGNOT/CPASS.FOR Athol Place tighter restrictions on the default parameters for the least squares fit, to provide a more robust user interface. Correct a value/reference error in calling subroutine BPCOEF; this error surfaced recently by triggering an adressing exception on certain architectures. It did not introduce BP errors. Moved nowhere. 8904. December 6, 1995 $HLPFIL/CPASS.HLP Athol Add a detailed EXPLAIN section. Also made changes relating to the new default input values (#8903). Moved nowhere. 8905. December 7, 1995 $APLNOT/DATBND.FOR Athol Correct simple logic error affecting non-VLBA data when a polynomial BP is applied. This case can to light with some ATCA data. No other bandpass application is affected. Re-link all BP routines and tasks. Moved nowhere. 8906. December 8, 1995 BPASET Eric Changed it to test BPCNO and BPVOL to see if the scratch file already exists and, if so, to expand it if needed and skip creating yet another one. This was blowing iterative tasks out of the water after 120 channels or so. Moved nowhere. 8907. December 8, 1995 QCLEAN Eric The TV option was loading the first channel of the data cube, not the current channel plane. Moved nowhere. 8908. December 8, 1995 GRDFLT Eric Fixed it to issues its messages only once. Moved nowhere. 8909. December 8, 1995 $YPGNOT/POSSM.FOR Athol Fix call to BPCOEF in subroutine POSSBP. Only half the bandpass was plotted in certain limited cases. This affected only BP plotting, not BP application. Moved nowhere. 8910. December 11, 1995 $APGNOT/FITTP.FOR Athol Increased the buffer size for BP tables from 8192 to 16384+128, to accomodate 2048-channel ATCA bandpasses. Moved nowhere. 8911. December 11, 1995 $YPGNOT/FRMAP.FOR Leonia The previous version of the task had direction of RA axis to right. It confused the people because other task produce maps with direction of RA axis to left. So I have changed the direction to bring to conformity FRMAP with other tasks. Moved nowhere. 8912. December 11, 1995 CookBook Eric Changed chapters 4, 5, 6, and 8 for cross-refernces to Chapters 7 and the new 10. Changed Chapter 13 for new lists up to date. Changed the index and table of contents for 7, 10, 13. Wrote a new chapter on single dish. Added \boxt and the indexing to COOK82.MAC. Moved to the www site. 8913. December 11, 1995 ABOUT files Eric Ran the SHOPH proc and updated what it said to update: AP CALIBRAT CATALOG GENERAL IMAGING INFORMAT INTERACT MODELING OOP PLOT SINGLEDI SPECTRAL TAPE TASK UTILITY UV VERB Moved nowhere. 8914. December 13, 1995 $APGNOT/FITLD.FOR Phil 2 minor bug fixes and 1 major enhancement. Fixes: The correlator can write duplicate entries in the incoming frequency table (only if the correlator scripts contain such entries). FITLD had a built in assumption that the highest numbered unique FQID was the one to use (without me really thinking about it) but in fact the first found should be used. Fixed so that this is now the case. Luckily the project that fell into this hole was the first one with this feature. The second bug concerned antennas switching in and out of the array, one project lost one and gained two more in the space of a single scan. FITLD couldn't handle this, it now can. The major enhancement is the implementation of Leonia's correction for the correlator's saturation effect in self-spectra. The result of this change is that all data spat out by FITLD from the VLBA correlator now have all known digital corrections applied to them. Moved nowhere. 8915. December 14, 1995 CONVL Eric Added OPCODE DCON to deconvolve a Gaussian from an image. Cleaned up the wording in the help file which lied about the use of BMAJ. Moved nowhere. 8916. December 14, 1995 IM2UV Eric Added options to shift the phase center, control the peak amplitude, set the central weight, and taper the weights. Moved nowhere. 8917. December 15, 1995 CLIP Eric Cleaned up the code a little and changed the detailed flag info format to show the times in intelligible form. Moved nowhere. 8918. December 15, 1995 *APT.INC Eric Changed VAPT.INC and its declarer DAPT.INC to list all AP tasks, but then to say there are none. This lifts the ban on AP tasks in batch queue 1. Moved nowhere. 8919. December 16, 1995 POPSDAT Eric TAU0 adverb was spelled with a zero in APCAL and NEWPARMS but with a letter O in POPSDAT.HLP. Fixed POPSDAT. Moved nowhere. 8920. December 16, 1995 COMPRESS Eric Wrote a garbage collector for POPS. Changed: K2TEXT new - write the K array to a text file OERROR to delete file on error and restore normal input PREAD to read file, delete on error/EOF and restore normal input STORES to call K2TEXT, set up special input mode, RESTORE 0 AIPS to suppress messages in this mode. AIPSB to suppress messages in this mode. AIPSC to suppress messages in this mode. COMPRESS updated help file with the good news (RUN NEWPARMS is now obsolete - this is better). Moved nowhere. 8921. December 18, 1995 $QPGNOT/CPASS.FOR Athol Fix catalog reference to dimension of frequency axis in subroutine BASOLV. Moved nowhere. *********************************************************************** Changes while 15JAN96 was NEW *********************************************************************** *********************************************************************** 8922. December 21, 1995 ZDIE Eric Created a new Z routine to delete the /tmp/. files at the end of a task. They get deleted other ways too, but occasionally one has bad luck and the pid matches some long running pid on the system. AIPSPROTO.H to reference zdie ZABORS.C UNIX, BERK, SUN to call ZDIE ZDIE.FOR GEN to build task name and call ZDIE2 ZDIE2.FOR GEN stub ZDIE2.C UNIX to do the file deletion. DIETSK to call zdie Moved from 15OCT96 this date. 8923. December 21, 1995 ZWHOMI Eric ZWHOMI was depending on the SETPAR value for the number of TV devices. We can't depend on that being right with the ease with which they are added these days. Changed ZWHOMI to translate the AIPSTVS symbol used in assigning REMOTE TK numbers in the first place. Moved from 15OCT96 this date. 8924. December 21, 1995 Help and CookBook Eric I forgot ZZANALYSIS and COMPRESS changes ZZPOPS and ZZPSEUDOVERB. Also the CookBook Chapter 13 changes. While we are at it I reworded the 1-liner of COMPRESS to say more. Corrected typographical (spacing) errors in chapters 2 and 4 of the CookBook and an error in a reference to extended hex in Chapter 3. Moved from 15OCT96 this date. 8925. December 21, 1995 SPLIT Eric Fixed two errors: (1) The loop over sources in SPLTUV did not alter SELQUA. Thus, if there were > 1 qualifier for a source name, the cal routines would think they were writing a multi-source file. Fixed by setting SELQUA to the current qualifier. (2) The routine that makes a list of desired source numbers had an error test for the list going from > 0 to 0 due to a check on the index table. The wrong variable was used for the error code so the task continued rather than dying as was intended. Moved from 15OCT96 Dec 22. 8926. December 27, 1995 Guest TVs Eric There is no reason why a "remote" computer (not in the TVHOSTS list) cannot run XAS and receive TV signals from the server computer. Even if the 2 are reversed in byte order the TV, TEK and MSG servers all work. The only complication is in the IC and ID files - guests on a server get a guest IC and ID which are not the same as on their machione. If multiple computers address a screen being used by a guest TV then some confusion about images and whioch channel is on etc. will occur. Changed: AIPSEXEC To set the remote TK at 36 higher than before, i.e. to AIPSTVS + 37 through 71. TVDEVS.SH To allow for unknown TVHOST = TVDISP to pass through. ZWHOMI To gets AIPSTVS (# known TV hosts) and recognize the guest usage. It then determines which guest account (if any - it looks for available ID files) is available, creates a /tmp/GUESTVn lock file, does any needed assignments of logicals for TV and TK. ZDIE To call ZGUEST to kill the guest TV if needed. ZGUEST New: to create / delete the /tmp/GUESTVn files. Moved from 15OCT96 this date. 8927. December 27, 1995 ZACTV9 Eric Called the function "creat" which does not care whether the file being created already exists or not. Changed to call "open" with the addition of an excl flag to require that the file not exist. Moved from 15OCT96 this date. 8928. January 3, 1996 CookBook Eric Putback a revised Chapter 12 (advanced POPS etc) including new capabilities for remote use, COMPRESS, etc. Changed Cross references in Chapters 3 and 7 to chapter 12. Added AHISTM HISEQ MF2ST, SKYVE to Chapter 7. Redid the table of contents COOK0 and the index COOKI to match all this. Moved to 15JAN96 since there is but one area. 8929. January 3-4, 1996 ZSSS*2.C Eric Changed the name of 4 external variables to be different from those used by ZVTP*2.C. Moved the external variables declaration inside the code as "external" in ZSSSX2 and ZSSSC2. It is improper to declare the variables externally more than once, although it took SGI's latest linker to complain about them rather than doing the right thing. Moved to NEW this date. 8930. January 4, 1996 $HLPFIL/APCAL.HLP Athol Revise EXPLAIN section for APCAL. Add information about source flux densities and the format of the external weather file. Moved from 15OCT96 this date. 8931. January 11, 1996 $SYSUNIX/READLINE.SHR, etc. Pat Created three new AIPSSHAR format archives to hold the three directories needed to build the GNU ReadLine library: READLINE.SHR, RLDOCS.SHR, and RLEXAMP.SHR. The last two should be expanded into docs/ and examples/ subdirectories of wherever the READLINE.SHR file is expaneded; INSTEP1 needs to take care of this for source-only builds. Moved from TST this date. 8932. January 11-13, 1996 Misc from Alpha Eric The alpha compiler is good at warnings. With the usual 0 initialization the following are not serious: RANDIN Changed Common /RANDOM/ to /RANDMM/ to avoid conflict with library function name RANDUM /RANDOM/ -> /RANDMM/ QMSPL /RANDOM/ -> /RANDMM/ CANPL /RANDOM/ -> /RANDMM/ STFND KOFF was used without intialization - variable dropped. ACFIT USERID set to IUSER an uninitialized local variable DCONV NMAX declared before and after the PARAMETER statement. LTESS Used left over adverbs not present in task to set Common parameters which were not used MK3IN In AT10XX, IBASE referenced in debug section without initialization. GUESSED that an equivalence was needed to TMPBUF(2) as in other routines. SELSD Removed left over error test, no longer meaningful. SNSMO Record offset in RFFXFG routine not initialized - 0 seems what was desired. STESS Used left over adverbs not present in task to set Common parameters which were not used UFVFIL Alpha complained wrongly about 2 out variables not initialized; added inits to fool it and to be defensive UVIMG Removed line using F which was not DATAed. DOSMTH set correctly a few lines later. IM2UV Changed a double comma to a single comma in a FORMAT. FRMAP OLDSRC = SRCDUN where SRCDUN never set and OLDSRC never tested or used; dropped both variables. TVFLG Droped a test against TEMP (=0) since always true. VBPLT Dropped a bunch of unused code that gave warnings a counted junk. VPLOT Dropped a bunch of unused code that gave warnings a counted junk. ITICS Dropped an unused error message section. **** But the following may have caused an error: GSCAT OUTIND not initialized - meant buffers addressed from 0 IMFIT A bad number of good samples was passed to the first call to LMSTR1. It may not have mattered since FCN does not use that parameter. VLBIN Corrected an error in calculating the earth rotation rate and removed unused and erroneous code purporting to set up history info ISPEC Used an otherwise unused variable (SCALX) to set an offset to the plot. Using SCALY as intended will change the offset a lot, but may not affect the plots. IBLED Used a bad subscript into the ERRORB array when plotting error bars for the upper (full vis) plot. Used uninit I rather than IX. PFPL3 Did not have a good value for magic blanking. SPFLG Droped a test against TEMP (=0) since always true and fixed tests in the history writing that left things out when the 1st antenna stayed the same but the 2nd changed. Moved to NEW. 8933. January 14, 1996 BPASET Eric This piece of shit was setting ARRAY(TELNO(I)) without regard for the value of TELNO(I) for i = 1 through 90!. Caused DEC machines to blow up I am tempted to cancel the release when I look at this code. Between BPASET, BPFSEL, BPREFM, and POLVIS (alone) there are 26 arrays dimensioned (MAXCHA,MAXIF) rather than (MAXCIF). The former is 114800, the latter 8192 which may be a little small. Thus these 4 subroutines alone make many tasks (almost all with shared libraries) use about 10 Mbytes more swap than they need. There are many more such misuses of the worst case scenario - in DSEL.INC, DUVV.INC, DBAV.INC etc. BP and Pol cal must be rewritten - but it's not my job anymore. Moved the simple fix to NEW this date. 8934. January 14, 1996 Undefined floats Eric POPS puts V(SP) = C(j1) in a value stack even when this is not appropriate (i.e. j1 = 2) and the C is an integer of the equivalence K. This cannot be changed since the legal/illegal j1's cannot be fully determined. Changed AU2, AU5, AU5C, AU8, AU9C, AUC, and SUBS to check for an array/string name on the stack rather than a scalar. This should reduce the problem. DEC OSF/1 is unforgiving about math with bad values. Moved to NEW this date. 8935. January 15, 1996 Buffers used by ANTAB/GC tables Athol MK3 data require more substantial buffers due to the larger number of IF channels. Increase all gain curve (GC) table I/O buffers from 8192 to 16384. Make this change in $APLNOT/GCINI.FOR, $APLNOT/TABGC.FOR and $APLNOT/GCKEY.FOR. Increase general buffers BUFF1 and BUFF2 in $APGNOT/ANTAB.FOR accordingly. Also increase the ANTAB work buffer in subroutine TYCONS, used when consolidating the system temperature entries in the TY table. This limit breached by a combined VLBA dataset stretching over four days. Moved from TST this date. 8936. January 17, 1996 IMMOD Eric Changed scaling of brightnesses to account for the beam in the input image. Moved to 15JAN96. 8937. January 17, 1996 TVCPS Eric Added OPCODE = 'COLB' which makes blanked and edge pixels 'FFFFFF' rather than '000000'. The latter is black, the former transparent on transparency paper. Option ignored when DOINV TRUE. Moved to 15JAN96. 8938. January 17, 1996 KNTR Eric Needed to use LOCNUM = 2 for calls to PLTBEM - LOCNUM = 1 can be messed up with labeling type choice but 2 is in standard coordinates. Moved to 15JAN96. 8939. January 17, 1996 $APGNOT/SPCAL.FOR & .HLP Athol Upgrade of SPCAL. New features include: i) larger buffers to accomodate VLBA datasets; ii) more careful checks on pre-avg. buffer pointers; iii) new solution type 'RAPR' which assumes that the linearly polarized emission in each channel is a scaled form of total intensity (as in PCAL). Moved from 15OCT96. 8940. January 22, 1996 DOSMTH Eric Changed CALEDIT to init DOSMTH to true - it will be unset if SMOOTH is 0. Changed CALEDIT to declare SMOOTH as an int, matching the DSEL.INC definition. Changed IMAGR and BLING to declare SMOOTH to be integer. Moved nowhere - this is a patch for NEW only. 8941. January 22, 1996 CHINC ERic This was badly messed up too. OUVIMG in QUVUTIL set CHINC to 1 rather than finding out what its value really was and then passed that to IMGMEM which passed it to a header correction routine. Changed QUVUTIL to acquire the real CHINC from the UV object. The header routine had the bizare idea that the frequency increment should be the averaging width if channel are averaged. That makes sense if they are all averaged, but not if just some are. Changed IMHCOR to test the number of points on the freq axis too. Moved to 15JAN96 this date. 8942. January 22, 1996 TRANS Eric The fix of the ALT ref pixel did not consider the sign applied to the new axis - leading to trouble when people reverse a frequency axis (which they should do in some cases). Added a test for sign and the correct fix for reversal. Moved to 15JAN96 this date. 8943.