----------------------------------------------------------------------- ; Copyright (C) 1995 ; 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 code on the Charlottesville VAX in the logon area [AIPS]. The format to use is: line 1: Date Program name Your name line 2: What was done line 3: Where has this change been moved (e.g. OLD, the Modcomp) line 4: 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 15JUL90 *********************************************************************** Changes while 15JUL90 was TST: *********************************************************************** *********************************************************************** 5594. January 22, 1990 UNICOS Area Logicals Kerry Added area logicals to $SYSAIPS/AREAS.DAT for the implementation of AIPS under UNICOS. These include $APLCRI (UNICOS Z-routines), $INCCRI (UNICOS INCLUDE files), $QCRI (UNICOS Q-routines) & $SYSCRI (UNICOS system files). Eliminated the UNIX area logicals for empty, never used directories, including $APLNUNIX, $QPGUNIX, $QYPGUNIX, $SYSLUNIX and $YPGUNIX. Also cleaned up and embellished the comments. Moved to 15APR90 same date. 5595. January 22, 1990 Source Code Preprocessing for UNICOS Kerry Source code preprocessing for FORTRAN modules under UNICOS is the same as for any other system, however, C modules require some minor massaging, including the removable of the ``_'' (underscore) suffix from entry points and calls to FORTRAN routines. Also, ``cft77'' does NOT convert FORTRAN code to lowercase, so entry points and calls to other AIPS routines must be specified using uppercase. The UNICOS version of PP does this via ``sed'' using the command file C.SED. A special version of INCS.SH has also been created that includes the $INCCRI directory in the preprocessor's search path for INCLUDE files. New files: $SYSCRI/PP $SYSCRI/INCS.SH $SYSCRI/C.SED Moved to 15APR90 same date. 5596. January 22, 1990 Compiler & Loader Options for UNICOS Kerry Created options files specifying rather ``vanilla'' defaults for UNICOS ``cc'' (C) and ``cft77'' (FORTRAN) compilations plus ``segldr'' program loading (linking). New files: $SYSCRI/CCOPTS.SH $SYSCRI/FCOPTS.SH $SYSCRI/LDOPTS.SH Moved to 15APR90 same date. 5597. January 22, 1990 Object Library Maintenance for UNICOS Kerry System V based UNIX platforms do not have ``ranlib'' (eliminated after SYSTEM III). Furthermore, UNICOS uses ``bld'' as an object librarian rather than the standard UNIX ``ar''. New files: $SYSCRI/LIBR Moved to 15APR90 same date. 5598. January 22, 1990 ZLPCL2 Script for UNICOS Kerry Created a special version of the standard UNIX script used to dispose line printer output (i.e., temporary disk files) to a suitable device. It's very similar to the generic UNIX version, except it uses the UNICOS ``asa'' filter to process FORTRAN carriage control characters. New files: $SYSCRI/ZLPCL2 Moved to 15APR90 same date. 5599. January 22, 1990 FORTRAN Compilation Script for UNICOS Kerry The technique used in the generic UNIX version of this script to detect a FORTRAN compilation failure does not work under UNICOS (i.e., ``cft77'' produces a ``.o'' file whether it fails or not), thus a special version of the FC script that tests the return code of the compiler. New file: $SYSCRI/FC Moved to 15APR90 same date. 5600. January 22, 1990 Source & Object Library Mappings for UNICOS Kerry Created a generic version of LIBR.DAT for UNICOS that uses the $QCRI Q-routines, the $YVTV (virtual TV) Y-routines and $APLCRI Z-routines. New file: $SYSCRI/LIBR.DAT Moved to 15APR90 same date. 5601. January 23, 1990 UVBAS Bill C. Corrected indexing error when selection channels. It was selecting data starting at channel 1. Moved to 15APR90 this date. 5602. January 23, 1990 CHANGED.90A Bill C. Copied the CHANGE.DOC file for 15JAN90 to HIST:CHANGED.90A. Moved nowhere else. 5603. January 23, 1990 INSVMS:TRANSPRT.COM Bill C. Modified to include the new AIPS PUBL and AIPSIONS directories on the transport tape. Since these are above the release date level they are not automatically included. Moved nowhere. 5603. January 23, 1990 DeAnza routines Mark Calabretta/Eric Corrected YDEA errors in YZERO (constant used in call where variable is required) and YSPLIT (NE used where EQ was wanted). Moved to 15APR90 this date, to 15JAN90 24-Jan. 5604. January 23, 1990 EXPFIT Eric Corrected bad format numbers and added further directory name tests. Changed EXPFIT.COM to include the new text regions. Moved to 15APR90 this date. 5605. January 24, 1990 SPFLG Eric Corrected the CURVALU-like display in most of the flagging modes. It was not handling BCHAN correctly. Moved from 15JUL90 this date. 5606. January 24, 1990 UVBAS.HLP Huib van Langevelde/Bill C. Added explain section. Moved to 15APR90 this date. 5607. January 25, 1990 BATER Bill J. Two calls to CHR2H in subroutine CUA had incorrect numbers of arguments. Moved to 15APR90 this date. 5608. January 29, 1990 AIPSHELP.HLP Bill C. Updated Designated AIP schedule. Moved to 15APR90 this date. 5609. January 29, 1990 PLCUB.FOR Bill J. The call to NAMEST had an incorrect argument - hollerith instead of character string although the conversion had been performed to generate the correct argument. Moved to 15APR90 this date. 5610. January 30, 1990 AIPSSERV documentation Bill C. Updated doc:readme. to drop discussion of bug fixes in 15OCT89. Also added initial versions of DOC:VMSNOTES.90A and UNIXNOTES.90A. Moved nowhere. 5611. January 30, 1990 BATER Bill J. Incorrect indexing in the command interpreter in subroutine BATPOL ensured that this would never work. Moved to 15APR90 this date. 5612. January 30, 1990 BATER Bill J. A trivial problem noted after above fix. Error conditions would not be reported correctly from BATPOL due to another indexing error. Moved nowhere. 5613. January 31, 1990 LISTR.HLP M. Calabretta/Bill C Corrected misstatements about the defaults for frequency identifier. Moved to 15APR90 this date. 5614. January 31, 1990 APGNOT:UVPLT M. Calabretta/Bill C Fixed a formatting problem in displaying the timerange on the plot; scaled the u,v, and w to the correct frequency when they are plotted and allowed plotting all FQ ids on the same plot. Also documented in UVPLT.HLP. Moved to 15APR90 this date. 5615. January 31, 1990 APLNOT:UVGET Bill C. Improved precursor comments especially use of common variables. Moved nowhere. 5616. January 31, 1990 DDTs Glen Removed NPOINTS from the DDTSAVE help and from DDTLOAD.001. Updated DDT.HLP. Modified NEWPARMS.001 to include the DDT adverbs DDISK, DDTSIZE, EDGSKP, IOTAPE, MDISK, TCODE, TDISK, TMASK and TMODE. Added help files for the adverbs. Moved to 15APR90 this date. 5617. January 31,1990 QPHSRO/QRECT Glen Replaced a matrix multiply calculation of the Phase rotation by calls to SIN and COS. Replace QRECT use of Look up table in PSAP with SIN and COS. Removed QPHSRO from the QVEX this improved small DDT by 3 BITs. Moved to nowhere. 5618. February 5, 1990 Going AIPS vol. 1 Bill C. Revised the AIPS Programmer's Manual "Going AIPS" to reflect the changes made in the overhaul. The new edition refers to the 15APR90 version of AIPS. Volume 1 is finished; files in DOCTXT:GAVOL1.TEX, GOAIPS.STY, CHAP1-8.TEX, APPENDIXA.TEX and APPENDIXB.TEX. Also added a file README with instructions for using LaTeX for Going AIPS. Moved nowhere. 5619. February 6, 1990 CALIB Bill Changed call to FRQTAB in GASOLV to determine frequencies from the actual file being read rather than the input data file. Since the data may be reordered by UVGET this is important; under some circumstances this problem causes an infinite loop. Moved to 15APR90 this date. 5620. February 6, 1990 UVFLG Phil When writing to an FG table UVFLG was not inserting any information in the HI file, it now does. Also tells the user which FG table is being written. Moved nowhere. 5621. February 6, 1990 SETJY Phil Modifed the 'calc' option to use the new flux coefficients determined by Perley from VLA data. The default is to now use these values for any of 3C48, 3C138, 3C147 or 3C286. If the user wishes the old Baars et al coefficients can be selected. Also added the QUAL adverb since this is sometimes needed when inserting velocity information for the same source observed at different velocities. Also changed .HLP. Moved nowhere. 5622. February 7, 1990 UVPLT Phil Modified the plot labelling very slightly to remove a gap between the top of the plot and beginning of the labelling. Moved nowhere. 5623. February 9, 1990 GETFQ Phil Modified to look in CH tables for freq info if there is no FQ table. If neither of them are there then it fills in the data from the catalogue header. Moved nowhere. 5624. February 9, 1990 POSSM Phil POSSM was not labelling the plot correctly if the user requested the x-axis be labelled in Hz. It now does a much better job of this, overriding the automatic scaling produced by METSCA if necessary. Moved nowhere. 5625. February 9, 1990 UVPLT Phil Modified the labelling even more, inserting the IF number being plotted, and generally tiding up a little. Moved nowhere. 5626. February 9, 1990 DOUVCOMP Phil Changed the default value of douvcomp to .true.. So modified POPSDAT.HLP and RUNSYS:NEWPARMS.001. Will run POPSGN on Monday. Moved nowhere. 5627. February 9, 1990 SETJY Phil Changed a format statement slightly, and ensured that when no sources are selected that the user gets a sensible error message. Moved nowhere. 5628. February 9, 1990 ANTINI Phil There was an error in the call to CHR2H such that only 8 characters of the column title were written to the table instead of 24, the others were all set to nulls. Normally this did not matter but some FITS readers don't find this very healthy. Moved nowhere. 5629. February 9, 1990 CALIB Bill C Fixed a couple of MIN functions that were supposed to be MAX in CLBSRC and CLBDR. This error caused CALIIB to abort when doing a common delay-rate solution for all IFs in a MkIII data file. Moved to 15APR90 this date. 5630. February 9, 1990 MK3IN.FOR,MK3IN.HLP Bill C. Numerous improvments. Any antennas found may be selected by using antenna name 'ANY' in the INFILE. The "Debug" output option was cleaned up to work properly on any machine. CL table entries for antennas for which there was no data are no longer written. The observing date in the AN and HI files are now correct. The time tags associated with the first baseline of the first scan are now correct if REFDATE is not specified. The phase of the phase cal was reversed if the baseline was oriented west to east. This is now fixed. The phase of the phase cal is now adjusted for any offsets in delay. This problem was causing the phase cals on the "remote" stations to be trashed. Also the max and min. peculiar delay offsets are kept on a baseline basis. The delay residual in the output data was wrong by an amount corresponding to one delay lag. Moved to 15APR90 this date. 5631. February 9, 1990 APLGEN:ZDHPRL.FOR Bill C The use of the intrinsic function "REAL" to convert from a 32 bit integer to double precision was replaced by "DFLOAT". Use of "REAL" was causing a signifigant precision loss in translating HP double precision words on Convexes; DFLOAT should keep the full precision. Apparently this bug was introduced in an attempt to clean up problems found by the SUN 4 compiler. Moved to 15APR90 this date. 5632. February 10, 1990 CLCOR.FOR, CLCOR.HLP Bill C. Added two new options: 'PCAL' to insert MkIII manual phase cals and 'SBDL' to add corrections to the residual delays without modifying the total model delays. These two options are needed for procesion MkIII VLBI data. Also changed the adverb BPARM to CLPARM which is dimensioned 20. Also CLPARM.HLP. Moved to 15APR90 this date. 5633. February 10, 1990 POPSDAT.HLP, RUNSYS:NEWPARM.001 Bill C Added the new, task specific adverb CLPARM. In POPSDAT.HLP this was done in a procedure ant the end of the file rather than defining an adverb that POPS knows by name; i.e. DAPL.INC doesn't need to be changed and AIPS etc recompiled and linked. This change required that POPSGN be run by hand on each NRAO computer updated by the midnight job. Moved to 15APR90 this date. 5634. February 11, 1990 FILLM Phil/Bill J. Several bugs were fixed, one outstanding one and others (re)introduced recently. (1) The problem with the integration just prior to midnight being labelled wrongly was finally found and fixed, the fact that midnight has occurred is now caught and dealt with in MCSUM. (2) When antennas enter the array during an experiment the order of the baselines in the data arrays changes, this was not being trapped. MCWANT now checks to see if the number of antennas and/or the order of antennas has changed and if so it forces a call to MCINI. I fixed this 18 months ago but my fix had been removed even though the variables and arrays I used were still declared. (3) An associated problem is that the AN table is written when a file is first opened so it only contains information for antennas present at the beginning of the observing run. A new subroutine (ANTUPD) now updates the AN table before the file is closed. (4) The bandwidth being written to the SU table was the full bandwidth rather than the channel width. (5) There was a problem occurring when the number of channels changed during the run, this change was not always caught. Added another section to MCWANT to explicitly check for this change and also for a change in correlator mode. When these change the program is now forced to determine if a new file must be opened. (6) The annoying (to the users) feature of having to specify integration time + 0.2 sec to obtain a specified int. time has been disabled. This will need more checking to determine if it really works. Moved nowhere. 5635. February 12, 1990 VLBIN Phil VLBIN was unable to cope with lower sideband data, it just assumed everything was upper sideband. Fixed the labelling routines and the FFT routines. Moved nowhere. 5636. February 13, 1990 DOCTXT:VGUIDE.RNO Bill C/Gareth Revised the VMS installation guide for 15JAN90. Moved nowhere. 5637. February 13, 1990 ZFRE2.C in APLSUN Chris Modified version of the generic Unix ZFRE2. Modifications are necessary for Sun OS since the Sun implementation of {\tt df} returns the information for the requested disks in the order in which they occur in {\tt /etc/fstab} irrespective of the order in which they were specified on the command line. Moved to 15APR90 this date. 5638. February 15, 1990 SDTUV Bill C. Removed restriction that the input data files have 14 data columns. Also fixed a couple of places where zero divides were possible. Also added the Parkes 64 m antenna to the list of antennas that it knows about. Moved to 15APR90 this date. 5639. February 15, 1990 MK3TX Chris Now behaves more sensibly if there are no text files present in the archive. Moved to 15APR90 this date. 5640. February 15, 1990 CALIB Bill C Fixed problem with averaging all data in each IF. This problem only occured when fringe fitting MkIII data. Also put the SN table writing stuff in GASOLV into a separate routine. Moved to 15APR90 this date. 5641. February 15, 1990 CLCOR.FOR, CLCOR.HLP Bill C. Modified 'SBDL' option to modify the phase as well as the delay. The phase is assumed to be at the reference frequency if the delay residual correction is non zero. DATCAL corrects to the frequencies of the actual data. Also updated the table of sunspot numbers in the CLCOR.HLP file. Moved to 15APR90 this date. 5642. February 15, 1990 APLNOT:CGASET.FOR Bill C. Removed the phase corrections for channel selection with non zero delay corrections. This correction was also being done in DATCAL and was redundant in CGASET. Moved to 15APR90 this date 5643. February 15, 1990 POSSM Phil Added an option to plot all IF's in the same frame, essentially adding the second IF onto the end of the first so creating a longer spectrum. Also ensured that flagged channels are not plotted. Updated the .HLP file to reflect the change. Moved nowhere. 5644. February 16, 1990 CALIB Phil There was an indexing error in filling up the frequency arrays in GASOLV, the bchan and bif offsets were not being taken into account. Moved to 15APR90 this date. 5645. February 16, 1990 TABHDR Bill C. Fixed logic error which caused the length of character table entries read from ASCII FITS tables to always be 1. This appears to have had no effect on calibration tables but was discovered at CfA. This affected UVLOD and IMLOD. Moved to 15APR90 this date. 5646. February 18, 1990 PUTCOL Bill C. Remove a C in column 1 from a line computing the array pointer for a double precision value. Moved to 15APR90 this date. 5647. February 19, 1990 SLCOL Neil Killeen/Bill C Fixed an error extracting BLC, TRC from the input parameter array. Moved to 15APR90 this date. 5648. February 19, 1990 AIPSUB:AU7A Bill C Added new verbs to access tables. - GETTHEAD: Reads table header keywords and other info - PUTTHEAD: Writes table header keyword values. - TABGET: Reads values from a table. - TABPUT: writes values to a table. Added new routines to AIPSUB:TBHEAD, TBDATA. Also GETHEAD.HLP, PUTHEAD.HLP, TABGET.HLP, TABPUT.HLP, POPSDAT.HLP. This change requires running POPSGN. Moved nowhere. 5649. February 19, 1990 PRTSD Bill C Fixed bug in computing time to be displayed; it was multiplying all the times by 2. Moved to 15APR90 this date. 5650. February 19, 1990 SDTUV Bill C Changed hardcoded number of receivers (14) to the number of columns - 2. Also modified BADRA to accept an RA difference of 360 degrees as 0. Moved to 15APR90 this date. 5651. February 19,1990 UVCOP Bill C. Fixed precision problem in computing subarray number. Also the 5 day time correction removal when BPARM(2)>0 is only done for single source data and if the uncorrected time is > 5 days. Modified UVCOP.HLP to use new adverb UVCOPARM rather than BPARM. Moved to 15APR90 this date. 5651. February 19, 1990 New adverbs Bill C. Added new "task enrichment parameter" arrays (i.e. ?PARM array). Added UVCOPARM for UVCOP and UVFIPARM for UVFIX. Changed HLPFIL:POPSDAT.HLP and RUNSYS:NEWPARMS.001. Added HLPFIL:UVCOPARM.HLP and UVFIPARM.HLP. This change requires running POPSGN. Moved nowhere. 5652. February 19, 1990 UVFIX Bill C. Corrected a number of errors in the handling of compressed data; the output for compressed data was trashed. Also fixed a bug by which the subarray information was stripped off the baseline/subarray code. Also fixed to loop over subarrays if no subarray is requested. Added a SAVE statment in UVWCAL for the values that were computed on the first call. Also changed UVFIX.HLP. Changed APARM and BPARM to UVFIPARM in UVFIX.HLP. Moved to 15APR90 this date (except for new adverb). 5653. February 19, 1990 HORUS Bill C. The specified subarray was being ignored. This caused HORUS to fail when DOPOL was true because only 1 subarray at a time can be processed. Moved to 15APR90 this date. 5653. February 20, 1990 Task specific parameters Bill C. Changed the name of several task specific adverbs added recently. The new convention for simple cases is task_namePRM. This allows a predictable if long and/or ugly task specific parameter arrays. Those changed are: CLPARM => CLCORPRM (CLCOR) UVCOPARM => UVCOPPRM (UVCOP) UVFIPARM => UVFIXPRM (UVFIX) CLPARM.HLP, UVCOPARM.HLP, UVFIPARM.HLP were removed and CLCORPARM.HLP, UVCOPPRM.HLP and UVFIXPRM.HLP were added; CLCOR.HLP, UVCOP.HLP and UVFIX.HLP were changed as were POPSDAT.HLP and NEWPARMS.001 Moved nowhere. 5654. February 20, 1990 VLBIN Phil In the case of lower sideband data the reference pixel in the header was being incorrectly set. Moved nowhere. 5655. February 20, 1990 CVEL Phil Modified to work on lower sideband data, also now will not copy the BP table to the output file if it was applied. Speeded up the shifting by only calculating the shift every 5 minutes and interpolating for points in between, previously the full shift was calculated for every data point, this lead to a very long run time for the task. Moved nowhere. 5656. February 20, 1990 CALIB Bill C. Fixed bug introduced in recent fix of fringe fitting which messed up calibration of VLA like data. The RCP and LCP data were added and called RCP and there were no solutions for LCP data. Also fixed bug which caused the mean gain modulus correction from being computed if requested. Moved to 15APR90 this date. 5657. February 20, 1990 Mean gain modulus correction Bill C. Bugs in DATCAL and BLSET caused the mean modulus gain correction to fail. DOCAL wasn't applying the correction at all unless delay or rate or baseline corrections were being made. Moved to 15APR90 this date. 5658. February 20, 1990 POSSM Phil The source number was being incorrectly set when determining the velocity of the source, if the source was the last in the dataset. Moved nowhere. 5659. February 20, 1990 TABED Bill C. Modified to ignore flagged table entries. Moved to 15APR90 this date. 5660. February 20, 1990 ZFRE2 in APLSUN Chris ZFRE2 expected the first part of the pathname of an AIPS ``disk'' to be the mount point of the device on which the logical disk resided. This precondition was not met if any of the AIPS logicals DA01 etc. referred to a symbolic link. This version of ZFRE2 now interrogates each file referred to by the AIPS disk logicals to see if it is a symbolic link and, if so, reads the value of the link. Moved to 15APR90 this date. 5661. February 21, 1990 TKLAB Bill C. Fixed bad call to NAMEST. It was being passes a Hollerith name/class rather than a character string. This caused an infinite loop (on Suns at least) in verb TKSLICE. Moved to 15APR90 this date. 5662. February 21, 1990 XMOM Bill C. Fixed bug left over from the overhaul. A numeric array was being passed to ZPHFIL and ZOPEN for the file name rather than a character. This caused failure of the open (on Unix at least - this may work on Vaxes). Also cleaned up some other junk not cleaned up in the overhaul. Moved to 15APR90 this date. 5663. February 22, 1990 BPASS Phil There was an error in setting the row numbers when writing the BP table such that row number 1 was always the first written. This meant that if a user wished to add to an existing BP table he/she could not, the existing table would be overwritten. Also removed a now unused subroutine and changed the APARM array in the helpfile to BPASSPRM, also meant changing POPSDAT and NEWPARMS.001. Moved nowhere. 5664. February 21, 1990 VLBIN Phil The GAST at UT=0 was being written into the AN table as radians rather than degrees. Also the formula used to calculate IAT-UTC was getting progressively wrong as time went on, so I inserted the look up table stolen from Starlink . This is correct up to Jan 1 1990 - we will need to think about how we cope with this - maybe a table distributed by AIPSSERV and let VLBIN read the table? Moved to 15APR90 this date. 5665. February 21, 1990 AXSTRN glen AXSTRN produces a frequency label in exponential form if the frequency is greater than 1E11. Each internal write statement has an ERR= argument to trap errors. Moved to 15APR90 this date. 5666. February 22, 1990 ZDEAXF Gustaaf van Moorsel/Bill C Fixed declaration of OPERS(20) to OPERS*4(20). Moved to 15APR90 this date. 5667. February 22, 1990 PRTAB Phil Modified the code dealing with the DOHMS option, it was relying on a hard coded table to decide which columns had time info. It now examines the UNITS column label and decodes the type of column using GETCOL, so it is now general and does not need to be updated whenever a table format changes. Moved nowhere. 5668. February 23, 1990 NEWPARMS, POPSDAT, DDTs Glen Added help files for two DDT adverbs TNAMF and MAPDIF. Added these adverbs to NEWPARMS and POPSDAT. Moved to 15APR90 this date. 5669. February 23, 1990 FG tables Phil Added the FREQID to FG tables, this of course meant changes in the flagging routines. FGREFM - new routine to reformat FG tables FLGINI, TABFLG, DATFLG, NXTFLG, UVGET, SDGET, FLAGUP all modified to deal with the new column. TABFLG and FLAGUP had their call sequences changed. UVFLG and QUACK had the FREQID adverbs added and suffered moderate changes to reflect the addition of FQ numbers. Also changed their .HLP files. TVFLG and SPFLG had minor changes - just changed the call sequence to FLAGUP. Moved nowhere. 5670. February 24, 1990 MX/UVGRID Glen New UVGRID which does not require sorted UV data. MX modified to allow unsorted data and to determine Max and Min baseline length. New subroutines: UVAPAS calculates number of passes for bandwidth synthesis. APVCHK prints max and min of a vector in the AP. UVPREP creates a UV scratch file name, opens the file, inits the file and reads maximum baseline length. Moved nowhere. 5671. February 26, 1990 Q1GRD Glen Modified to call QRECT1 not QRECT. QRECT1 provides more accurate SIN and COS but does not multiply by radius. Moved to 15APR90 this date 5672. February 26, 1990 M-Z*.HLP Bill C Added the one line descriptions and the categories to the help files for which this had not already been done. Moved nowhere. 5673. February 26, 1990 UVGET Phil When I inserted the FQ ID into the FG tables the column numbers changed, this caused UVGET to fail to sort the FG table on the correct column numbers. Changed the data statement to force the sort keys to be columns 5 and 1 instead of 4 and 1. Moved nowhere. 5674. February 27, 1990 UVCOP Bill C. Corrected logic error in determining which correlations to keep when selecting compressed data by channel. Moved to 15APR90 this date. 5675. February 27, 1990 SPFLG, TVFLG Eric Changed DTVF.INC and DSPF.INC to add a variable retaining the source random parameter. Changed GRIDTB and GRIDTC to get the correct source ID number even when the user has specified a single source. Changed SPFLG and TVFLG to check frequency and source ID numbers when making the list of times to be gridded. Also changed them to correct the handling of the interactive (graphical) window setting and to remember the input source random parameter offset. Dropped unused FORMATs. Corrected channel number displayed in SPFLG's interactive modes. It used a correct channel number, but displayed a bad one if the reference channel was not the first channel. Moved from Oz this date to wherever and to 15APR90. 5676. February 27, 1990 NEWPOS Eric Cleaned up the code a bit. Also changed the formulas used for the SIN and ARC projections. They failed when the angles got very large so that ASIN became ambiguous. Now both use ATAN2. Thanks to Mark Clark (GB) who bugged me about this. Moved from Oz this date to wherever and to 15APR90. 5677. February 27, 1990 TIMDEST Bill J. Minimum value of TIMDEST set in TIMDEST.HLP was 14 days. Changed this to 5 days for the usual situation at the AOC. Moved nowhere. 5678. February 27, 1990 APLSUB:FPARSE.FOR Bill J. On occasion a history card is more than 80 characters long. When PRTTP calls GETNUM to parse the numerical information in this record, it can return a NaN to signify some illegal value. FPARSE didn't trap this appropriately and equivalenced the R*8 DBLANK which is 1.3E45 to an R*4 causing f.p. overflow. Added a trap so that isn't a fatal condition in the case of a "bad" history record. 5679. February 28, 1990 TVFLG, SPFLG Eric Corrected GRIDTB --- it was regarding an end-of-file condition as unacceptable at the location where the first point of each group was read. At one time, this location was for the first visiblity, but that was changed some time ago to do things in groups for vectorization. Corrected SPFLG and TVFLG --- in the first read through the data to determine the list of times, the tests on subarray, source, and FQ number were branching all the way to the end of the loop. Thereby, the index was not incremented and the same point was retested and re-rejected for the full buffer's worth. This could make the list of times seriously incomplete or in error in some cases. Moved from Oz here and to 15APR89. 5680. March 1, 1990 KEYCOP Bill C New routine to copy any header keywords from one cataloged file to another. Moved to 15APR90 this date. 5681. March 1, 1990 UVCOP Bill C Modified to call KEYCOP to copy catalog keyword/value pairs. This allows using UVCOP to salvage data from a disrupted FILLM run. Moved to 15APR90 this date. 5682. March 4, 1990 SNPLT Bill C Fixed several problems with OPTYPE='SUM'. The 2 IF mode (BIF=0) was disabled in the inputs checking and BIF=2 or higher actually did IF=1 due to indexing errors. Moved to 15APR90 this date. 5683. March 5, 1990 Table descriptions. Phil Modified the DOCTXT:MV...nn files to include the descriptions of FREQ ID and I.FAR.ROT. Description files modified were those for BL, NX, FG, SN and CL tables. Moved nowhere. 5864. March 6, 1990 CHNCOP, CHNDAT Phil Modified the precursor comments to fully explain what the default values of FREQID do. Moved nowhere. 5865. March 6, 1990 SOURNU Bill C. Corrected null (selects all) value of QUAL from 0 to -1 as is advertised to the users. Corrected the DATAed values of QUAL passed to SOURNU in BLFIT, QUACK, SNPLT, UVFLG. Moved nowhere. 5866. March 7, 1990 UVCOP Bill C. Added selection by sources and qualifier, FQ id, and IF. Also UVCOP.HLP. When selecting by IF then the reference frequency is changed, the u,v and ws are scaled and the AN, CL, SN and SU tables are suitably modified. Moved nowhere. 5867. March 7, 1990 Table selection by IF Bill C. Added a number of new routines to reformat tables selecting a given range if IF. CLSEL - Selects IFs in a CL table SNSEL - Selects IFs in a SN table SUSEL - Selects IFs in a SU table FGSEL - Renumbers IFs in a FG table Moved nowhere 5868. March 9, 1990 AIPSLetter stuff Bill C. 15JAN90 AIPSLetter (let90a.tex) was put into AIPSPUBL. Also changed macro al82.mac to give better uunet address as per DWells. Moved nowhere. 5869. March 9, 1990 YSVU:window.c,cursor.c Mark Calabretta/Bill C Modified WINDOW.C and CURSOR.C to bind a RESIZE event to the F2 and F7 keys, users can take their pick! Sunview considers pressing a button and releasing it to be two separate events. Added a call to event_is_down() to remove the "bounced key" behaviour of the function keys within SSS. I, also replaced the calls to event_id() by event_action() which supercedes it. Moved to 15APR90 this date. 5870. March 9, 1990 UVPLT Neil Killeen/Bill C When DOCAL=1, UVPLT looked for an SN table rather than CL table for multi-source data bases. It now does not check if the file is a multi-source file to see if an SN table exists. Moved to 15APR90 this date. 5871. March 9, 1990 KNTR.FOR, KNTR.HLP Mark Calabretta/Bill C Fixed a couple of bugs in KNTR, one relating to annotation of the contour levels, and another in computing the maximum absolute pixel value in KONDRW. Also changed the way that blanked pixels are handled. It used to be based on connecting pixel centres, but this could produce stick-like "contours" for small irregularly shaped clusters of pixels surrounded by blank areas. The algorithm now delineates the pixel edges. There are pros and cons for either method but the latter is less likely to confuse users. Also expanded the explain part of the help file slightly. Moved to 15APR90 this date. 5872. March 9, 1990 AIPSUB:TVMOVI.FOR Mark Calabretta/Bill C Removed limitation on long cursor moves. Moved to 15APR90 this date. 5873. March 12, 1990 ZVTVX2 and ZVTVX3 in APLBERK Chris ZVTVX2 and ZVTVX3 were, incorrectly, treating the case where a read from a {\tt SOCK_STREAM} returned less bytes than were requested as an error. This was causing TV-by-wire to fail over long distances (AOC to Charlottesville). Both routines now loop to read the outstanding bytes until either the complete packet is received or a genuine error or end-of-file is detected. Write operations are treated in the same manner, although partial writes are extremely unlikely in the absence of errors. Moved to 15APR90 this date 5874. March 12, 1990 BPASET.FOR Mark Calabretta/Phil Ensured that the scratch file written by BPASET was not too small for UVINIT. Moved to 15APR90 this date. 5875. March 12, 1990 AU9A.FOR Mark Calabretta/Bill C Fixed error trap for TKCATL; treats KNTR plots like CNTR. Moved to 15APR90 this date. 5876. March 12, 1990 TKCATL Mark Calabretta/Bill C Fixed a bug in the logic to see if a pixel is in range. Moved to 15APR90 this date. 5877. March 12, 1990 YSVU:CURSOR.C,WINDOW.C Mark Calabretta/Eric Modified WINDOW.C to accept F2 and F7 as the AIPS TV resize button. Modified CURSOR to surpress erroneous double button hits. Moved nowhere. 5878. March 12, 1990 APLVMS:ZMCACL.FOR Bill J. A redundant call to ZC8CL (to convert compressed ASCII to local characters) was removed. This code was probably a hangover from pre-f77. Affects only VAX version of FILLR. Moved nowhere. 5879. March 12, 1990 APLNOT:CALREF.FOR Bill J. Under some circumstances the arguments for SIN and COS exceeded permissible range for REAL functions. Changed evaluation to use double precision. Moved nowhere. 5880. March 12, 1990 APLPGM:CLIP.FOR Bill J. Currently, CLIP can't handle compressed data. The error message to warn the user advises the use of SPLIT. Changed this to suggest use of UVCMP --- this handles both multi and single source databases --- pending conversion of CLIP to handle packed u-v data. Moved nowhere. 5881. March 14, 1990 APGNOT:MK3IN.FOR Bill C. Modified to open and close SU and CL table whenever they are accessed. This increases the chances that if something goes wrong that the tables will be OK. Moved to 15APR90 this date. 5882. March 14, 1990 CALIB.FOR Bill C. Fixed logic error in GASOLV in computing the array of frequencies when selecting by channel. This caused later routines to blow up. Moved to 15APR90 this date. 5883. March 14, 1990 APLPGM:PRTTP.FOR Bill J. Added a 'terse' option for FITS format tapes (PRTLEV=-2). This gives tape file#, image and filename ONLY. Changed HELP file to reflect this. Moved nowhere. 5884. March 14, 1990 APLPGM:TACOP.FOR Bill J. Blank OUTNAME and OUTCLASS weren't being interpreted usefully i.e. just copy an extension table to higher version attached to the input file. The behaviour of the program was at variance with information in "Chapter 99" and caused confusion. Moved 15APR90 this date. 5885. March 12, 1990 AIPSUB:AIPINI.FOR Kerry Fixed to take different actions depending on wether the current program is FILIA2 (open logical unit 6 via ZMSGOP) or FILAIP (open logical unit 5 via ZOPEN). Otherwise, non-VMS installations fail. Also if FILAI2 or FILAIP, don't get device characteristics common from disk or make accounting entry. Otherwise, disconcerting error messages are issued when the AC and/or SP files and/or the MS file for user #1 don't yet exist (FILAI2 and FILAIP create these as part of an installation). Moved to 15JAN90 and 15APR90 this date. 5886. March 15, 1990 SYSAIPS:AREAS.DAT Kerry Added and/or redefined area logicals for screen server and virtual TV implementations including: Logical Directory Contents YSS ...Y.DEV.SS Screen Server Y-routines (formerly YSSS as ...Y.DEV.SSS) YSERV ...Y.SERVERS Screen servers YSTV ...Y.STV Sunview TV (formerly YSVU as ...Y.DEV.SSS.SVU) YVDEV ...Y.VDEV Virtual TV program (formerly YPGVDEV as ...Y.PGM.VDEV) YXTV ...Y.XTV X Window TV YXTVUNIX ...Y.XTV.UNIX UNIX XTV communications YXTVBELL ...Y.XTV.UNIX.BELL Bell UNIX (streams) YXTVBERK ...Y.XTV.UNIX.BERK Berkeley UNIX (sockets) YXTVVMS ...Y.XTV.VMS VMS (DECnet) The following areas are still defined until they can be phased out entirely: YSSS ...Y.DEV.SSS (now YSS as ...Y.DEV.SS) YSVU ...Y.DEV.SSS.SVU (now YSTV as ...Y.STV) YPGVDEV ...Y.PGM.VDEV (now YVDEV as ...Y.VDEV) Moved nowhere. 5887. March 15, 1990 DDCH.INC, ZDCHIN Bill C. Increased the maximum number of files open of the 3 different types to 20 each. Also increased the size of FTAB to 2048. This should take care of the MX problem that size the overhaul it could not clean 16 fields. Moved nowhere. 5888. March 15, 1990 UVCOP Phil Under some circumstances when no IF selection was done the CL, SN, SU and FG tables were not copied to the output file. Fixed. Moved nowhere. 5889. March 15, 1990 QGRIDA/QUVIN Glen Modified QGRIDA and QUVIN to call QRECT1 instead of QRECT to increase precission. Moved to 15APR90 some time ago. 5890. March 15, 1990 UVGRID and friends Glen Moved sections of UVGRID into new subroutines APLOAD, APSHFT, APSUM, APFNSH and APWRIT. Moved nowhere. 5891. March 15, 1990 UVMDIV Phil The call sequence to CHNCOP in UVMDIV on CVAX was wrong, it was OK on nrao1 - God knows what was on the other machines. Luckily that piece of code is not used that often. 5892. March 15, 1990 QVEX:QPUT, QGET Glen Removed CVEX QPUT and QGET which were identical to the QPSAP versions. Removed to 15APR90 this date. 5893. March 15, 1990 FILLM Gareth/Bill C Fixed numerous bugs which occured when a scan contained a single averaging interval and the next scan was to go into another file. The control logic was getting hopelessly confused. Errors in the handling of changes in the number or order of antennas was complicating the issue. In this latter case, tests were made using values set in MCANT or MCH1 which had not been called yet. Several of these (number of antennas, correlator mode) are now decoded in MCWANT. All of the changes associated with these problems were in MCWANT. Corrected an indexing problem in MCGBRC which caused and occasional spectrum to be corrupted and an infinite loop under special conditions. The logic error was fixed and a trap was made for the infinite loop although it should not occur if the routine is working correctly. The last baseline in a CDA will be lost if this trap is invoked. Fixed addressing bug for mode 2BC. Inserted correct source and FQ id numbers into the first visibility record of each scan; formerly the source id and FQ id from the previous scan were being used. Modified to accept autocorrelations for revisions 24 and later. Modified to only put data with the same correlator mode in the same file. This involves writing a header keyword giving the correlator mode in each line data file. Changed the error handling process to return continuous EOF conditions from MCREAD if more than PELIM (currently 50) parity errors have previously occurred. The current file(s) being written are now expanded in size by 10,000 visibility records when they are opened, and compressed when they are closed. This minimizes the disk space required when FILLM is running, but could lead to disk fragmentation if the observation forces multiple openings and closings. FILLM used to compress all files that it had appended to when it exited successfully; less fragmentation, but potentially an unnecessary disk hog on abnormal termination(s). 5894. March 15, 1990 AVSPC.FOR/.HLP Neil/Phil Modified AVSPC to deal more sensibly with reference pixels and values, now sets the output bandwidth to the bandwidth between the two extreme channels selected. Removed the BCHAN, ECHAN adverbs as they caused confusuion. Added the ability to copy any keyword/value pairs. Moved nowhere. *********************************************************************** Changes while 15JUL90 was NEW: *********************************************************************** *********************************************************************** 5895. March 16, 1990 FPARSE.FOR Bill C. Previous change caused this routine not to compile. Moved from 15OCT90 this date. 5896. March, 19 1990 MK3IN Bill C The source number written into the CL table was always the first source in the source list. Also added call a call to initialize the A tape header info when the end of a scan was reached. Put a trap into AT20XX to ignore a scan if the cross reference table overflowed. This should not happen under normal circumstances but bad (?) type 20?? records can cause this to happen. Moved to 15APR90 and from 15OCT90 this date. 5897. March 19, 1990 GETHUT Bill C Added calls to SPFIL to make sure nulls were removed from TABLE info. This was causing FITTP to write nulls in tables headers. Moved from 15OCT90 and to 15APR90 this date 5900. March 19, 1990 FITTP Bill C. Corrected indexing error reading history file. This was adding an extra 8 blanks after the "HISTORY" on the FITS history cards and cause then to occasionally run onto the next card. Moved from 15OCT90 and to 15APR90 this date 5901. March 19, 1990 Generic ZLWOP and ZLWIO Kerry Switched to FORMAT'ed OPEN and WRITE. Also corrected call sequence to ZLASCL in ZLWIO. Moved to 15OCT90 this date. 5902. March 19, 1990 LWPLA Kerry Fixed buffering problem and errors introduced by changing from hollerith to CHARACTER constructs during the overhaul. Eliminated EOT character from the beginning of the output PostScript program which causes problems on some laser printers and does not seem to be necessary. Eliminated inappropriate H2CHR call and replaced it with simple CHARACTER substring assignment. Also switched character font from Courier to Helvetica-Bold. Tested CNTR and GREYS plots using the VLBA correlator group's QMS PS810 and at non-NRAO sites using Apple Laserwriter and NEC LC890 laser printers. Moved to 15OCT90 this date. 5903. March 19, 1990 ZLASCL.PS in SYSUNIX Kerry Created a special version of the ZLASCL script for use with PostScript laser printers that prefixes "%!PS-Adobe-" followed by "%%Creator:AIPS task LWPLA" to the LWPLA output. The first needs to appear at the beginning of the output ("%!... distinguishes PostScript programs from simple text) on a line by itself and this is not possible with the buffering scheme used in LWPLA. ZLASCL.PS should be substituted for ZLASCL at sites with PostScript engines. Some laser printers (e.g., NEC LC890) may also require that the LWPLA output be processed to limit the number of characters per line to, say 132 (e.g., using "fmt -s -132"). Moved to 15OCT90 this date. 5904. March 20, 1990 CHLTOU Kerry Corrected a 32-bit-ism by changing an integer constant "1" argument in a call to ZI32IL (starting position in units of 32-bit integers) to NP which is assigned NBITWD / 32. Otherwise, the routine fails on 64-bit platforms (e.g., Crays). Moved to 15OCT90 this date (also put on 15APR90 UNIX distribution tape). 5905. March 20, 1990 CATIO Kerry Eliminated a baroque EQUIVALENCE of variables used as opcodes that confused CRI's cft77 and substituted CHARACTER constants in CALLs. Moved to 15OCT90 this date (also put on 15APR90 UNIX distribution tape). 5906. March 20, 1990 MX Bill J. Channel averaging was not handled correctly. In particular, this affected users who had written out their UV databases from ISIS with UVFITS. (UVFITS treats the 2 IFs as 2 FREQUENCY CHANNELS.) Moved from 15OCT90 this date. 5907. March 22, 1990 XVSS Chris XView screen server. This is a port of the SunView screen server to the X Window System using the XView toolkit. The current version should be regarded as a beta test release and only functions with bsd sockets. Files added to the system (in area XTV) are: AIPS.ICON -- aips tv icon COLMAP.C -- colour map handling CURSOR.C -- cursor and buttons HEADER.H -- global definitions IMGIO.C -- image loading/readback MAKEFILE -- Unix makefile SOCKET.C -- miscellaneous socket operations SOCKXF.C -- socket I/O WINDOW.C -- set up windows & event handlers XVSS.C -- main program The current version works with versions of AIPS linked with the same set of Y-routines used by the Sun screen server (SSS). These routines have been moved to the new YSS area and the ZSSSxx routines have been moved from APLSUN to APLGEN (at some stage they will be renamed to something more appropriate). Moved to 15OCT90 this date 5908. March 27, 1990 BPASS Phil When solving for bandpasses on multiple IF data (MkIII VLBI) a frequency index was not being updated every IF loop. This caused the resultant BP entries to degrade as a function of IF. Moved from 15OCT90 this date. 5909. March 28, 1990 FILLM Gareth Corrected bug recently introduced in reading the parameters from an existing catalog entry (this is for the case when FILLM will try to append to an existing file). Corrected the error handling when a logical record is incomplete at the end of tape. Fixed the NX record count when closing a dataset. 5910. March 28, 1990 MX Glen Modified UVGRID to correctly handle averaging two (or more?) IFs Also added error checks in subroutines APWRIT and APFNSH Moved from 15OCT90 this date. 5911. March 28, 1990 APLVMS:ZDCHI2,ZMI2 Glen VMS Asynchronous I/O requires Event Flags for determining when I/O is finished, but only 32 flags are "Unassociated?" New modifications to INC:ZDCH.INC caused more than 32 flags to be used. ZMI2 aborts on event flage outside the range 32 to 63. ZDCHI2 modified to assign only numbers 32 to 63 and ZMI2 to reassign numbers in FTAB if event outside 32 to 63 ever occur again. (ZMI2 code checks should not now be need, but are there just in case...) Moved from 15OCT90 this date. 5912. March 219 1990 TABBP, BPASET Phil When a single channel of a BP table was flagged, TABBP marked the whole spectrum as flagged - this was OK because BPASET did not know what to do if only one channel was flagged bad. They now both deal correctly with flagged channels. Moved from 15OCT89 this date. 5913. March 29, 1990 POSSM Phil Modified so that when plotting BP tables a flagged channel is not plotted. Also extended the option put in a few weeks ago to plot all IFs and channels on the same plot to work on BP tables. Moved from 15OCT89 this date. 5914. March 29, 1990 APLSUB:FPARSE.FOR Bill J. The logic of the trap for bad history records was incorrect and was my fault. This caused PRTTP to fail disastrously on encountering negative coordinates (amongst other things). Moved from 15OCT90 this date. 5915. April 2, 1990 MK3IN Bill C. Fixed bug in AT20XX in which the cross index reference table was messed up. Also fixed a problem with the Fractional Bit Shift correction; It was apparently a factor of 2 too small and with the wrong sign although handeling of the FBS correcion in the correlators is confusing. Also added option (APARM(3)=1) to suppress the FBS correction altogether. Also MK3IN.HLP. Moved from 15OCT90 and to 15APR90 this date. 5916. April 2, 1990 APLGEN:ZRHPRL.FOR Bill C. Corrected error in complimenting mantissa for negative numbers. This had the effect of causing negative FBS corrections on MK3IN to be too small. Moved from 15OCT90 and to 15APR90 this date. 5917. April 4, 1990 UVFIX Glen UVFIX appeared not to work because the default frequency for UVW scaling is taken from the ANTENNA table. This frequency is often zero, making the new UVWs zero. If Antenna frequency is zero, the CATBLK frequency is now used by default. UVFIX is unchanged if user specifies the FREQUENCY. Moved from 15OCT90 this date. 5918. April 5, 1990 PRTAC Bill C. Fixed problems initializing the accounting (AC) file. Incorrect values were being put into the first 4 words and the file was not being compressed. This rendered the AC file unuseable. Moved from 15OCT90 this date. 5919. April 6, 1990 UVAVG Bill C. Modified to process compressed uv data. Also knowns about source and FQ ids and will not average data withdifferent values. Also increased the size of the working arrays. Moved from 15OCT90 this date. 5920. April 9, 1990 TABMRG Bill C. Modified to be smarter about blanked values. Blanked values are not summed and if table entries are being merged and one value is blanked and the other is not then the valid value is written into the output record. Moved from 15OCT90 this date. 5921. April 11, 1990 MK3IN Athol Kemball/Bill C. Fixed bug in logic detecting and end of tape in ATREC. Now if an EOF is encountered while NEWFIL is true then this is treated as an end of tape condition. Moved from 15OCT90 this date. 5922. April 11, 1990 CLCAL Chris Removed the erroneous description of the BOTH option, which smooths amplitude and phase, from the EXPLAIN file. This had claimed, incorrectly that this was a vector average. Moved nowhere (more extensive changes planned for 15OCT90). 5923. April 12, 1990 APGNOT:ASCOR.FOR Bill J. Expanded buffers JBUFSZ from 3200 to 4800 and BUFF1, BUFF2 from 1600 to 2400 to cope with spectral line data. Moved from 15OCT90. 5924. April 8, 1990 $APLUNIX/ZABORS.C Kerry $APLBERK/ZABORS.C $APLSUN/ZABORS.C Eliminated "core" file preservation mechanism entirely. Also changed to commit suicide with a self-inflicted illegal instruction rather than by calling "abort" which issues different signals on different systems. Changed in 15APR90 and 15OCT90 same date. 5925. April 8, 1990 $APLSUN/ZABOR2.C Kerry A new module that differs from the $APLBERK version in that it calls "ieee_handler" to "set" the action for IEEE "common" floating_point exceptions to SIGFPE_ABORT. Otherwise, invalid floating-point results (i.e., NaNs) go undetected. Should abort now. Changed in 15APR90 and 15OCT90 same date. 5926. April 8, 1990 $APLBELL/ZCREA2.C Kerry $APLBELL/ZACTV9.C New modules. The $APLBELL version of ZCREA2.C differs from the generic UNIX version in terms of header files. The $APLBELL version of ZACTV9.C differs from the generic version in that it calls "fork" instead of "vfork" (a Berkeley UNIXism). Changed in 15APR90 and 15OCT90 same date. 5927. April 8, 1990 UNICOS Z-routines Kerry All new modules in $APLCRI: ZABOR2.C same as $APLBELL version except abort handler function must be declared differently; also no SIGPWR and SIGCLD action set to SIG_DFL ZABORS.C PUTBCKed then REMOVEd to force REMOVE via midnight jobs on other Crays and to use $APLUNIX version instead ZACTV9.C PUTBCKed then REMOVEd to force REMOVE via midnight jobs on other Crays and to use $APLBELL version instead ZCMPR2.C same as $APLBERK version (Berkeley UNIX "ftruncate" supported, at least under UNICOS 5.0.13) ZCMPRS.FOR same as $APLGEN version except call to CHR2H commented out (4 characters/real won't work on 64-bit machines) and CHARACTER PNAME passed to ZCMPR2 instead of HOLLERITH ANAME ZCPU.FOR calls UNICOS routine SECOND (generic UNIX "times" yileds unreliable under UNICOS 5.0.13) ZCRDIR.C same as $APLUNIX version except calls ISHELL (UNICOS specific) instead of "system" to execute a shell command ZCREA2.C PUTBCKed then REMOVEd to force REMOVE via midnight jobs on other Crays and to use $APLBELL version instead ZCREAT.FOR same as $APLGEN version except call to CHR2H commented out (4 characters/real won't work on 64-bit machines) and CHARACTER PNAME passed to ZCREA2 instead of HOLLERITH ANAME ZDCHI2.FOR same as $APLBELL version except SPFRMT and DPFRMT = 0, and TTYCAR = 0 ZDCHIC.C same as $APLUNIX version except NBBY "defined" as 8 (for some reason not in sys/param.h) ZDCHIN.FOR same as $APLGEN version except NBPS = 1024 and NSPG = 1 ZDESTR.FOR same as $APLGEN version except call to CHR2H commented out (4 characters/real won't work on 64-bit machines) and CHARACTER PNAME passed to ZCREA2 instead of HOLLERITH ANAME ZEXIST.FOR same as $APLGEN version except call to CHR2H commented out (4 characters/real won't work on 64-bit machines) and CHARACTER PNAME passed to ZEXIS2 instead of HOLLERITH ANAME ZFRE2.C same as $APLBELL version except logic (?) to look for continuation line ZGETCH.C PUTBCKed then REMOVEd to force REMOVE via midnight jobs on other Crays and to use $APLUNIX version instead ZI16IL.FOR borrowed from tried & true COS implementation ZI32IL.FOR calls Cray's USICTC to perform conversion ZILI16.FOR borrowed from tried & true COS implementation ZIVSOP.FOR same as $APLGEN version except call to FIVASOPEN commented out (name too long) ZLASC2.C same as $APLUNIX version except calls ISHELL (UNICOS specific) instead of "system" to execute a shell command ZLASOP.FOR same as $APLGEN version except "cft77" objected to RECL=160 without ACCESS='DIRECT' for opcode 'POPN' (used anywhere except under VMS?) ZLOCK.C same as $APLBERK version (Berkeley UNIX "flock" supported, at least under UNICOS 5.0.13) except for additional header file sys/types.h required ZLPCL2.C same as $APLUNIX version except calls ISHELL (UNICOS specific) instead of "system" to execute a shell command ZMSGOP.FOR same as $APLGEN version except call to CHR2H commented out (4 characters/real won't work on 64-bit machines) and CHARACTER PNAME passed to ZDAOPN instead of HOLLERITH ANAME ZMSGWR.FOR restored MSGBUF argument to CHARACTER*80, then set MSGTXT = MSGBUF instead if H2CHR on HOLLERITH MSGBUF(20) (i.e., tried & true pre-overhaul technique) and true pre-overhaul technique) ZMSGXP.FOR same as $APLGEN version except call to CHR2H commented out (4 characters/real won't work on 64-bit machines) and CHARACTER PNAME passed to ZEXIS2 instead of HOLLERITH ANAME ZOPEN.FOR same as $APLGEN version except call to CHR2H commented out (4 characters/real won't work on 64-bit machines) and CHARACTER PNAME passed to ZDAOPN instead of HOLLERITH ANAME ZPUTCH.C PUTBCKed then REMOVEd to force REMOVE via midnight jobs on other Crays and to use $APLUNIX version instead ZR32RL.FOR lots of unpacking as well as bitwise and'ing, or'ing and shift'ing, but it vectorizes; no NaN detection yet, but otherwise pretty well tested ZR64RL.FOR version in a state of flux (not to be trusted) ZRENAM.FOR same as $APLGEN version except call to CHR2H commented out (4 characters/real won't work on 64-bit machines) and CHARACTER OLDNAM and NEWNAM passed to ZRENA2 instead of HOLLERITH OLD and NEW ZSHCMD.C same as $APLUNIX version except calls ISHELL (UNICOS specific) instead of "system" to execute a shell command ZTACT2.C similar to $APLBELL version but checks for zombie'd processes via "grep" on "ps" output to avoid false positives ZTAP2.C in a state of flux, but close to working ZTKILL.C similar to $APLBELL version but checks for zombie'd processes via "grep" on "ps" output to avoid false positives ZTPCLD.C $APLUNIX/ZDACL2.C hacked for pseudo tape (i.e., instead of $APLGEN/ZTPCLD.FOR because direct access FORTRAN I/O under UNICOS 5.0.13 required COS-like machinations); should probably be used by all UNIX platforms ZTPMID.C $APLUNIX/ZMI2.C hacked for pseudo tape (i.e., instead of $APLGEN/ZTPMID.FOR because direct access FORTRAN I/O under UNICOS 5.0.13 required COS-like machinations); should probably be used by all UNIX platforms ZTPMIO.FOR same as $APLGEN version except modified to call C version of ZTPMID; should probably be used by all UNIX platforms; ZTPOP2.C same as $APLUNIX version except additional header file sys/types.h required ZTPOPD.C $APLUNIX/ZDAOPN.C hacked for pseudo tape (i.e., instead of $APLGEN/ZTPOPD.FOR because direct access FORTRAN I/O under UNICOS 5.0.13 required COS-like machinations); should probably be used by all UNIX platforms ZTPOPN.FOR same as $APLGEN version except call to CHR2H commented out (4 characters/real won't work on 64-bit machines) and CHARACTER MTNAM passed to ZDAOPN instead of HOLLERITH ANAME ZTPWAD.C $APLUNIX/ZWAI2.C hacked for pseudo tape (i.e., instead of $APLGEN/ZTPWAD.FOR because direct access FORTRAN I/O under UNICOS 5.0.13 required COS-like machinations); should probably be used by all UNIX platforms ZTQSP2.C same as $APLUNIX version except calls ISHELL (UNICOS specific) instead of "system" to execute a shell command ZTXMAT.FOR same as $APLGEN version except local CHARACTER CNAMES array passed to ZTXMA2 instead of HOLLERITH NAMES (4 characters/real won't work on 64-bit machines) then call CHR2H in loop to convert NNAM CNAMES to NNAM NAMES ZVTVC2.C same as $APLBERK version ($APLBERK not in Z-routine search path) ZVTVC3.C same as $APLBERK version ($APLBERK not in UNICOS Z-routine search path) ZVTVGC.C same as $APLBERK version ($APLBERK not in UNICOS Z-routine search path) ZVTVO2.C same as $APLBERK version ($APLBERK not in UNICOS Z-routine search path); this routine may have trouble getting its arguments correctly from ZVTVOP ZVTVO3.C same as $APLBERK version ($APLBERK not in UNICOS Z-routine search path) ZVTVX2.C same as $APLBERK version ($APLBERK not in UNICOS Z-routine search path) ZVTVX3.C same as $APLBERK version ($APLBERK not in UNICOS Z-routine search path) Changed in 15APR90 and 15OCT90 same date. 5928. April 8, 1990 Cray INCLUDE Files Kerry All new modules in $INCCRI: DAPC.INC Declarations, dimensions, COMMON definitions and EQUIVALENCE statements for the Cray 256K word Pseudo-AP implementation (same as the old COS version) ZVND.INC Compiler directive to ignore apparent vector dependency (same as old COS version) ZVD.INC Compiler directive to force scalar compilation (same as the old COS version) Changed in 15APR90 and 15OCT90 same date. 5929. April 8, 1990 Cray Q-routines Kerry All new modules in $QCRI. These are the same as the old COS versions except dynamic memory management is disabled. The former COS technique (CALL MEMORY) conflicts with UNIX memory management. QCFFT.FOR QCLNSU.FOR QINIT.FOR QMAXV.FOR QMEMSZ.FOR QMINV.FOR QMULCL.FOR QPHSRO.FOR QPTDIV.FOR QRFFT.FOR QRFT.FOR QRLSE.FOR QXFOUR.FOR QXXPTS.FOR Changed in 15APR90 and 15OCT90 same date. 5930. April 8, 1990 $SYSUNIX/AREAS Kerry $SYSUNIX/CREADIR Changed to use $SYSAIPS/AREAS.DAT instead of $SYSUNIX/AREAS.DAT and removed $SYSUNIX/AREAS.DAT. Changed in 15APR90 and 15OCT90 same date. 5931. April 8, 1990 $SYSUNIX/AREASCSH.SED Kerry $SYSUNIX/AREASSH.SED Added looping to avoid sed'iting matching patterns in comments. Changed in 15APR90 and 15OCT90 same date. 5932. April 8, 1990 $SYSUNIX/AREAS.CSH Kerry $SYSUNIX/AREAS.SH As generated by $SYSUNIX/AREAS from $SYSAIPS/AREAS.DAT under UNIX. Area logicals for "local" directories should be defined in $SYSLOCAL versions of CDVER.CSH and CDVER.SH now. Changed in 15APR90 and 15OCT90 same date. 5933. April 8, 1990 $SYSUNIX/LIBR.DAT Kerry $SYSALLN/LIBR.DAT $SYSCRI/LIBR.DAT $SYSNRAO1/LIBR.DAT $SYSSUN/LIBR.DAT Commented better and eliminated references to the defunct areas $APLNUNIX, $QPGUNIX, $QYPGUNIX and $YPGUNIX. Also changed $YVTVPGM references to new logical $YPGVDEV (i.e., new name for $YVTVPGM). Changed in 15APR90 and 15OCT90 same date. 5934. April 8, 1990 $SYSUNIX/CDVER.CSH Kerry $SYSUNIX/CDVER.SH Added comments (formerly uncommented) and changed to use AREAS.CSH and and AREAS.SH (formerly used AREAS.CSH and AREAS.SH from $SYSLOCAL). Also changed changed the execution search path definition to search the current directory, $SYSLOCAL and $SYSUNIX before all others to avoid command name conflicts (e.g., CC the AIPS C compilation script versus CC the C++ translator commonly installed as a local system utility). This is where "local" area logicals should be defined now (i.e., instead of $SYSLOCAL versions of AREAS.CSH and AREAS.SH). Changed in 15APR90 and 15OCT90 same date. 5935. April 8, 1990 $SYSNRAO1/CDVER.CSH Kerry $SYSNRAO1/CDVER.SH $SYSNRAO1/AREAS.CSH $SYSNRAO1/AREAS.SH New files (same as $SYSUNIX versions since all local directories for NRAO1 are defined in $SYSAIPS/AREAS.DAT, thus $SYSUNIX/AREAS.CSH and $SYSUNIX/AREAS.SH). Removed obsolete $SYSNRAO1/AREAS.CSH and $SYSNRAO1/AREAS.SH. Changed in 15APR90 and 15OCT90 same date. 5936. April 8, 1990 $SYSUNIX/ASSNLOCAL.SH Kerry $SYSUNIX/ASSNLOCAL.CSH Eliminated obsolete $AIPSTTn logicals and NRAO1 specific specific logic involving graphics terminal logicals. Added $LPRNTR logical. ASSNLOCAL.CSH is a new file used by LOGIN.CSH to define device logicals at login time. Changed/added in 15APR90 and 15OCT90 same date. 5937. April 8, 1990 $SYSNRAO1/ASSNLOCAL.CSH Kerry $SYSNRAO1/ASSNLOCAL.SH Changed definition of device logical $DA00 from public scratch file system to file system for exclusive use by AIPS users. Also, removed obsolete $DASSGN and $AIPSTTn environment variables. Changed/added in 15APR90 and 15OCT90 same date. 5938. April 8, 1990 $SYSUNIX/LOGIN.CSH Kerry $SYSUNIX/LOGIN.SH Changed execution search path definition to search the current working directory, $SYSLOCAL and $SYSUNIX before all others to avoid command name conflicts (e.g., CC). Also changed definitions of $CDOLD, $CDNEW and $CDTST to use CDVER.CSH and CDVER.SH from $SYSLOCAL (formerly they used CDVER.CSH and $CDVER.SH from $SYSUNIX). Also added "source $SYSLOCAL/ASSNLOCAL.CSH" and ". $SYSLOCAL/ASSNLOCAL.SH" to define device logicals at login time. Changed in 15APR90 and 15OCT90 same date. 5939. April 8, 1990 $SYSUNIX/AIPS Kerry $SYSNRAO/AIPS $SYSSUN/AIPS $SYSUNIX/BATER $SYSNRAO1/DACK $SYSNRAO1/DRUN $SYSUNIX/RUN $SYSNRAO1/SPACE Removed ". $SYSLOCAL/ASSNLOCAL.SH". Device logicals are established at login time now (see LOGIN.CSH and LOGIN.SH entries). Also changed the execution search path definition to search the current directory, $SYSLOCAL and $SYSUNIX before all others in order to avoid command name conflicts (e.g., CC). Changed in 15APR90 and 15OCT90 same date. 5940. April 8, 1990 $SYSUNIX/COMRPL Kerry Added protection against use on programs which royally screws up object libraries and therefore, installations. Changed in 15APR90 and 15OCT90 same date. 5941. April 8, 1990 $SYSUNIX/SEARCH Kerry Fixed to correctly remove temporary files and corrected error return code. Changed in 15APR90 and 15OCT90 same date. 5942. April 8, 1990 $SYSALLN/ASSNLOCAL.SH Kerry Removed. Removed from 15APR90 and 15OCT90 same date. 5943. April 8, 1990 $SYSCRI/FCOPTS.SH Kerry Changed to select static allocation of variables (i.e., like an implicit SAVE statement for each variable). Default is "stack" on all but Cray Y-MP and Cray X-MP EA systems. Changed in 15APR90 and 15OCT90 same date. 5944. April 8, 1990 $SYSNRAO1/ASOPTS.SH Kerry $SYSNRAO1/CCOPTS.SH $SYSNRAO1/FCOPTS.SH $SYSNRAO1/LDOPTS.SH Changed default action to purge preprocessor output (ASOPTS.SH, CCOPTS.SH, FCOPTS.SH). Changed default C and FORTRAN compilation options (CCOPTS.SH, FCOPTS.SH). Changed FORTRAN compiler command from "oldfc" to "fc" (FCOPTS.SH, LDOPTS.SH). Made $OPT2 optimization option "-O2 -ngr" to avoid Convex global register optimization compiler bug (FCOPTS.SH). Changed in 15APR90 and 15OCT90 same date. 5945. April 8, 1990 $SYSCVEX/OPT2.LIS Kerry $SYSNRAO1/OPT2.LIS Added new Q-routines to list of modules to be compiled at optimization level "-O2". Changed in 15APR90 and 15OCT90 same date. 5946. April 8, 1990 $SYSNRAO1/SPACE.FOR Kerry Changed to call EXIT (0) instead of STOP to avoid "at" job error mail. Changed in 15APR90 and 15OCT90 same date. 5947. April 8, 1990 $SYSNRAO1/ZLPCL2 Kerry Changed to use anonymous ftp for spooling printer files to the CVAX Versatec. Changed in 15APR90 and 15OCT90 same date. 5948. April 8, 1990 $SYSSUN/SSS.TXT Kerry Changed text referencing ASSNLOCAL.SH to ASSNLOCAL.*. Changed in 15APR90 and 15OCT90 same date. 5949. April 10, 1990 $APLUNIX/ZPRI2.C Kerry $APLBERK/ZPRI2.C The hard-coded value for the lowering of batch job execution priority was far too severe (10 on a scale of 0-20). Changed it to "0" (same as interactive), however, to take effect, each task executable must be regenerated. Execution priorities should really be system parameters (i.e., set using SETPAR and based on POPS #). Changed in 15JUL90 same date. 5950. April 11, 1990 $APLCRI/ZILI16.FOR Kerry $APLCRI/ZCREAT.FOR Screwed up insertion of user agreement in ZILI16.FOR Screwed up merging of changes to $APLGEN version in ZCREAT.FOR Changed in 15OCT90 and patched 15APR90 UNIX distribution tape same date. 5951. April 13, 1990 MX glen Uniform check for max frequency channel greater than min. Forces DFT subtraction if UV data is not X* sorted. Moved from 15OCT90 yesterday. 5952. April 13, 1990 $QCRI/QCFFT.FOR Kerry Screwed up insertion of user agreement. Changed in 15APR90 and 15OCT90 same date and patched UNIX distribution tape. 5953. April 16, 1990 IMCREA Bill C. The rotation on the sky was not being written to the catalog header for beams, an undefined (usually 0) value was. Moved from 15OCT90 this date. 5954. April 16, 1990 QNOT:MAKMAP Bill C. MAKMAP was depending on GRDAT to set the necessary gridding parameters. This did not work when the uv data had been rotated. Now NGRDAT is set to .TRUE. and the gridding parameters are explicitly set. This caused HORUS to incorrectly shify rotated data. The IF frequency offsets were being applied twice. Changed the call to CHNCOP to calls to CHNDAT and the frequency offsets are refered to BIF. This error caused HORUS to use incorrect frequencies for scaling u, v and w from the reference frequency to the actual frequency and thus give a misscaled image. Moved from 15OCT90 this date. 5955. April 18, 1990 XVSS Chris A number of miscellaneous changes prior to the release of the beta test version. - Added a makefile (accidentally missed from the original check-in). - Corrected bad C code: get_client was declared static in WINDOW.C and sock_io was declared static in SOCKET.C. This should have caused the Sun loader to fall over but was found while porting XVSS to the DECstation 3100. - Now expands TVDEV to TVDEVn and then expands TVDEVn to find the device name (no longer fixed to TVDEV1). - Now assumes any device name containing a colon is of the form : (no longer fixed to XVSSIN). - Function keys F2 and F7 are now accelerators for the resize operation. - Added code to prevent the apparent function key ``bounce''. - Added network byte order macros to S OCKXF for byte-swap architectures (eg. DECstations). The transmission is already in network byte order since ZSSSXF calls ZILI16 to pack the short integers into the buffer. Files affected are: CURSOR.C in YXTV MAKEFILE in YXTV SOCKET.C in YXTV SOCKXF.C in YXTV WINDOW.C in YXTV Moved from 15OCT90 this date. 5956. April 19, 1990 MX Bill C. Fixed problem with making the output file too large if frequency channels were being averaged or the channel increment was larger than 1. If the number of channels being averaged equals the total number of channels used then bandwidth synthesis is assumed and the number of output channel images is assumed to be 1. Also fixed logic error which caused the dirty image to be made twice if no CLEANing was to be done. Moved from 15OCT90 this date. 5957. April 23, 1990 PRTAB Eric Corrected excess widths for floating columns, put in tests to set widths more sensibly, including reading the full file to see how many digits are really required for the integer columns. Moved from Oz and from 15OCT90 this date. 5958. April 23, 1990 SNPLT Eric Corrected an error --- when PIXRANGE was specified, SNPLT did not count the number of plots correctly and hence did not end the last one properly. Changed it to be smarter about phase so that plots around +180 and -180 come out on a useful scale. Also changed the SUM amplitudes to be self-scaling or scaled by PIXRANG while SUM phases are self-scaling. Inserted CV changes allowing proper handling of IFs when summing (March 4 change). Moved from Oz and from 15OCT90 this date. 5959. April 23, 1990 APLSUB:ALLTAB Bill C. This routine was checking to see if version 1 of a given table type existed and if not then it didn't copy any of that table type. Now an existance test is done for each version of each table and a table is copied to the same version number as the input file. This problem was most noticible in TASAV when SN tables were not copied. Moved from 15OCT90 this date. 5960. April 23, 1990 SUBIM Bill C. Modified to call KEYCOP to copy catalog header keywords. Moved from 15OCT90 this date. 5961. April 24, 1990 Remote TVs Chris Added the following documents (in LaTeX format) to DOCTXT. Both are preliminary versions. TVINSTALL.TEX -- installation notes for TVMON, SSS and XVSS (designed to supplement the Unix installation guide). XVSSUSER.TEX -- user notes for XVSS under OPEN LOOK. Moved from 15OCT90 this date. 5962. May 1, 1990 ZFIO, ZMIO Bill C. The Convex compiler when using -O0 optimization level was generating bad code which caused erratic and sometimes obscure problems. It was modified to reduce the number of go to statments but the compiler bug appears to be associated with having a local variable which is computed and then used only as the subscript of an array argument of a subroutine call. Both of these routines shared this problem. These problems could have affected all AIPS tasks to some degree. Moved from 15OCT90 this date. 5963. May 1, 1990 OPT2.LIS Bill C. Removed CALIB from the lists of routines to compile with level -O2 on Convexs in SYSCVEX, SYSNRAO1, SYSVLAC1. The current compiler used at NRAO incorrectly compiles CALIB at this level. Moved from 15OCT90 this date. 5964. May 3, 1990 FITTP Phil When a uv file had an associated FQ table with only one row, FITTP translated this to a CH table, deleted the FQ table (which caused no harm) and wrote the file. This behaviour was causing considerable alarm to some users, so I have modified FITTP to translate the FQ table to a CH, write the CH table then delete it, leaving the uv-file in its original state. The FQ table is only written if it has multiple rows. Moved from 15OCT90 this date. 5965. May 7, 1990 SETJY Phil Modified to cope correctly with lower sideband data, it was missing the sign of the sideband from one key statement. Moved from 15OCT90 this date. 5966. May 11, 1990 TVFLG, SPFLG Eric Corrected 3 bugs which occurred in both tasks: The UV input file was read from visibility 2 through NVIS+1 in both the initial routine which listed the times and in the routine which applied the flags to single-source files. This could cause an apparent error in the time sequencing. The subroutine SOUFIL changes the list of sources, removing any minus signs. Since SOUFIL is needed for the time determination and then again for the calibration package, the times were found for all sources not in the list and then used for all sources in the list whenever an initial minus sign occurred in the source list. Corrected by restoring the list to its initial form after the first call to SOUFIL. Moved from Oz, should go to 15JUL90 and 15OCT90. 5967. May 15, 1990 XVSS Chris Corrected handling of zoom and scroll while image planes were turned off in SCRWRT.C (migrated Eric's change from SSS). Plugged a hole in the error checking in SOCKET.C that caused XVSS to crash with no error message if TVDEV was not defined. Changed WINDOW.C and CURSOR.C to use ACTION_SELECT events rather than explicit MS_LEFT events in case the mouse keys have been rebound for left-handed users. Moved from 15OCT90 this date. 5968. May 16, 1990 PRTAB Phil Corrected a typo in the new subroutine PRTSCL, which was looping from I to NCOL instead of 1 to NCOL when determining the starting position of the data in the output string. Moved from 15OCT90 this date. 5969. May 16, 1990 IMLOD Bill J. Fixed error in handling of value of keyword/number pair in IMPARS. GETNUM returns DBLANK as value if it can't be interpreted correctly. This in turn was being assigned to an R*4 in some cases with obvious results. Moved from 15OCT90 this date. 5970. May 22, 1990 SNSMO Chris Changed modes of operation so that phase is smoothed using a scalar average rather than deriving the smoothed phase from the smoothed sine and cosine components. Also restructured the code to make the flow of control more obvious, added some new error messages and forced reading of the SN table column numbers from the SN table. Also changed the help/explain file for CLCAL, which is the principal client of SNSMO, to reflect the changes. Moved from 15OCT90 this date. 5971. May 22, 1990 SNSMO Phil SNSMO was always assuming that SN tables had entries for 2 polzns, this is not always true and so FNDCOL was failing looking for the absent entries. Fixed so that the number of columns searched for in FNDCOL is dependant on the number of polzns present. Moved from 15OCT90 this date by Chris. 5972. May 22, 1990 SLCTD Chris New routine to determine whether a given source, calibrator source or antenna is in the set of sources, calibrator sources or antennae selected by the user. Moved from 15OCT90 this date. 5973. May 22, 1990 CLCAL Chris CLCAL now only smoothes calibration data from the requested calibration sources (as expected) when using boxcar smoothing rather than smoothing calibration data from all sources within the boxcar interval. SNSMO in APLNOT now takes additional arguments to specify the requested calibration sources. CLUPDA in APLNOT (the only routine which calls SNSMO) has been modified to provide SNSMO with the required list of calibration sources. Moved from 15OCT90 this date. 5974. May 24, 1990 FILLM Gareth The following two fixes made in 15OCT90 have now been copied to to 15JUL90. April 20, 1990 update in 15OCT90: FLMIN now checks to see that a REFDATE, if given, has the correct syntax. Blank is used if not. Short integrations are now correctly handled. There was some confusion about the IAT for geometry calculations and the IAT time stamp. The IAT time stamp is now used everywhere for determining whether a new integration, new scan, or new catalog entry is needed. The day number was not being incremented in MCWANT. This led to the program terminating at midnight if the stop time was specified and was on the same day as the start time! May 18, 1990 update in 15OCT90: I believe that the FQ entry selection in FLMFQ been causing unpredictable behavior with and without optimization, with and without the debugger on the Convex C1. I have therefore recoded the FQ entry selection loop in a manner that makes in easier for compilers to do local optimization (e.g. Convex -O0). Maybe this will solve some of the FQ table entry problems. Fixed a bug in FLMCLS which did not store the current FQ ID, but the previous one, for the last scan - only a problem when the FQ ID had changed from the previous scan. Moved from 15OCT90 this date. 5975. May 30, 1990 FILLM Gareth May 24, 1990 update in 15OCT90: The auto-correlation values are now stored in the Channel 0 multi-source file. This is so that BPASS will find the same number of visibilities in channel 0 as in the line data. This is not done for continuum data. Moved from 15OCT90 this date. 5976. May 30, 1990 MAKMAP Phil There was a typo in the call sequence to CHNDAT, a 'read' should have been a 'write', I have no idea why this has never bothered us before. Although I have a vague memory of fixing this. Moved from 15OCT90 this date. 5977. May 31, 1990 $AIPGUNIX/ZSTRTA Kerry A WRITE statement had an I/O list but the corresponding FORMAT statement had no format specifiers, which caused problems under UNICOS. Eliminated the I/O list since the info was hard-coded in the FORMAT statement. Changed in 15OCT90 same date. 5978. June 1, 1990 DDTLOAD Glen Added DOUVCOMP=FALSE to Tape read procedure so that UVDIF will work correctly. Moved from 15OCT90 this date 5979. June 4, 1990 UVMDIV Phil Source code was different on cvax and chucca. Cvax was correct. Prodded the code and conrpl'd uvmdiv, then comlnked everything that used it. This fix was missed by the midnight job on March 15. Moved to 15JUL90 this date. 5980. June 7, 1990 BPASET Phil When antennas were deselected by the user the logic in BPASET was wrong so that only the deselected antennas were used. Fixed. Moved from 15OCT90 this date. 5981. June 7, 1990 APLNOT:SNSMO.FOR Bill J. An error occured if SNSMO had already reached the end of the table but there were still antennas to loop over. Added trap to test for this. Moved from 15OCT90 this date. 5982. June 8, 1990 ADDIF Phil ADDIF was missing the H2CHR calls to translate the name and class of the input data set. Another relic of the overhaul. Moved to 15JUL90 this year. 5983. June 12, 1990 AIPSHELP.HLP Bill J. Versions in TST = 15OCT90 and NEW = 15JUL90 were not the same --- NEW was outdated. This caused complaints from local NRAO users. All three versions should be identical. Moved from 15OCT90 this date. 5984. June 12, 1990 PCAL Chris Fixed a conceptual error in the handling of ionspheric corrections that led to array bounds violations if there were more than 10000 pre-averaged visibility points. It also led to errors in applying the ionspheric Faraday rotation correction that would have caused subtle errors in the calculated polarizations if ionospheric Faraday corrections were applied. Ionospheric rotation data is now stored in a scratch file rather than an array. Moved from 15OCT90 this date. 5985. June 13, 1990 Call to TABKEY Chris TABKEY was being called to read or write keywords with character values without converting the value argument to or from Hollerith in a number of places. This would cause problems under VAX/VMS but would not have been noticed on most (if not all) UNIX systems. Fixed in UVFIT, GETANT and PCAL, which appear to be the only routines with defective calls. Moved from 15OCT90 this date. 5986. June 13, 1990 POSSM Phil Added the DOPOL adverb to POSSM and POSSM.HLP. Moved from 15OCT90 this date. 5987. June 13, 1990 PRTTP Bill J. As part of test to see whether tape is Modcomp archive tape or not, the number of antennas in the array had to be greater than 1. This is inappropriate for a single dish subarray for VLBI. Moved from 15OCT90 this date. 5988. June 13, 1990 HORUS.HLP Bill J. The default for OPTYPE was not explained. In fact, OPTYPE = ' ' defaults to OPTYPE = 'LINE'; this caused some grief to continuum users since OPTYPE = 'LINE' uses ONE IF only/ Moved from 15OCT90 this date. 5989. June 14, 1990 NULL pointers Chris ZSSSOP2.C (APLBERK) and ZVTVO2.C and ZVTVO3.C (APLBERK and APLCRI) were erroneously copying a null terminated string from the location returned by getenv() and comparing it against the string "(NULL)" to test for errors. This would cause programs to crash with segmentation violations or bus errors if the appropriate logical names were undefined or had inappropriate contents. All now correctly check the value returned by getenv() against NULL. In addition all now report an error and abort if an Internet host is not specified or has a malformed name rather than trying to connect to lynx.toronto.edu. Also added function declarations for getenv(), index() and strncpy(), removed unused local variables and fixed some dubious casts. These modules now lint cleanly, apart from ignored return values, on Suns and Convexes. Moved from 15OCT90 this date. 5990. June 14, 1990 ZFRE2 Chris Fixed a bad sprintf() format string that caused disk numbers greater than 10 to be translated to lower case hex digits in DA0n logical names. Fixed in APLMASC, APLCRI, APLSUN and APLUNIX. Moved from 15OCT90 this date. 5991. June 15, 1990 CVEL Phil Major overhaul. Now deals with FQ tables, the catblk handling was modified - this removed two problems, the first was that compressed data could not be handled correctly, secondly the HI file was not copied from the input to the output file. Fixed numerous other little bugs along the way. Moved from 15OCT90 this date. 5992. June 16, 1990 FILLM Bill J. Subroutine MCWANT called ZBYMOV to translate the date. This is inappropriate for VAXes --- needs a call to ZI32IL instead. Symptoms of failure on VAX were formatting errors (in diagnostic messages) and a failure to find any valid data. Tested out the repaired code on BOTH VAX and Convex. Moved form 15OCT90 this date.