----------------------------------------------------------------------- ; 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. 15OCT90) 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 15JAN91 *********************************************************************** Changes while 15JAN91 was TST: *********************************************************************** *********************************************************************** 6065. June 25, 1990 FITTP Bill J. Minimum value for BLOCKING adverb in HLP file was set at 1.0 which was at variance with HLP file text. Gripe #4300. Moved 15OCT90 this date. 6066. June 25, 1990 POSSM Phil Added new feature in the text output option. A sensible header is now printed in the text file, and the spectrum is listed in columnar form as channel#, vel, amp, phase. A flagged channel is marked as flagged. Also corrected a couple of minor problems with determining the scaling of the plots and the labelling, these were remnants of the recent graphics changes. Moved nowhere yet. 6067. June 25, 1990 UVCOP Phil When selecting by FREQID the channel bandwidth in the catalogue header was not being updated correctly, added a new subroutine (FQSCOP) to ensure that this is all done correctly. Moved nowhere yet. 6068. June 25, 1990 GETJY Phil/Neil The FREQID variable was not passed in common. Moved to 15OCT90 this date. 6069. June 26, 1990 LSERCH Chris ZTPOPN and ZTPCLS assign unique LUNs in the range 51 to 51+NTAPED to the lock files which control access to the tape drives. LUNs above 50 appear to be used nowhere else. However LSERCH was using 50+MAX(3, NVOL) as the maximum valid LUN, causing problems if there were more than 3 tape drives and less (logical) disks than tape drives. LSERCH now uses 50+MAX(3, NTAPED) as the maximum valid LUN. Moved to 15OCT90 this date. 6070. June 26, 1990 PRTCC Bill J. It is sometimes useful to know the sum of components up to (but not including) the first negative delta. PRTCC now outputs this information. Moved nowhere. 6071. June 26, 1990 UVCOP Bill J. A modification for gripe # 4277. UVCOP now has the option (UVCOPPRM(4)>0.0) to keep the user informed of the progress of copying. This is often useful for large files. Currently, it lists every 10,000th visibility copied and the asociated timestamp. This is fine for 'TB' data but looks a little strange for 'XY'. Also fixed a trivial (but often confusing to the user) bug which printed message that data had been "copied from the beginning to 0\ 0 0 0.0" in the case of copying over the FULL time range with selection. Moved nowhere. 6072. June 26, 1990 CALIB Phil / Bill C Fixed array declarations of WT1 and WTB in GASOLV, CLBPA. Fixed numerous errors in GASOLV which were encountered with multi polarization fringe fitting when each section of data had to be read from disk multiple times. Also fixed a number of problems in CLBDR which occured when it was trying to do a fit with insufficient data. Fixed an error in GASOLV setting the value of INCJJJ to the no. of channels rather than the number of IFs as the increment in IF. Cleaned up a few problems with processing IQUV data; it now only does solutions for I rather than I and Q. Closure defects are now labeled Ipol rather than RPOL. Now CALIB fill in the list of sources passed as both the source and calibrator list for SOUFIL. This should fix selection of calibrators by CALCODE. Moved nowhere. 6073. June 26, 1990 MWFBSM Bill C. Modified to be sure to replace blanked values if any valid values were in the input array. Moved nowhere. 6074. June 26, 1990 SN2CL Bill C. Fixed indexing problem; if a blanked IF was found in a SN table entry, that and all following IFs were blanked. Moved nowhere. 6075. June 26, 1990 SWPOL Bill C. Fixed error in indexing visibility record. Moved nowhere. 6076. June 26, 1990 MX Bill C. Cleaned up the handling of frequency channels; also fixed bug which caused too many channels to be processed if CHINC>1. Changed names and useage of variables to more closely reflect channel selection as viewed by the user. Moved nowhere 6077. June 26, 1990 GAPLT Bill J. Call to subroutine GINIT was nonsense --- wrong sequence of arguments and insufficient arguments. Moved to 15OCT90 this date. 6078. June 26, 1990 SNPLT Neil/Phil The data statement initializing QUAL to /30*-1/ should be /30*0/. The tables routines expect QUAL of 0 to mean any qualifier in contrast to the calibration routines that expect -1. Moved to 15OCT90 this date. 6079. June 27, 1990 CVEL.HLP Phil I forgot to outbck the helpfile which reflected the latest CVEL changes, so the program did not know about the new adverbs. Moved to 15OCT90 this date. 6080. June 28, 1990 TAPIO.FOR Bill J. Suppressed the messages about block sizes read being at variance with those expected. This causes so much distress with users that these essentially debugging messages needed to be removed. Prompted to do so by the same question about these messages twice in five minutes! Moved to 15OCT90 this date. 6081. June 28, 1990 DATFLG Eric Corrected gross errors: the offset to the first spectral channel to be flagged was not included when computing the index to negate the weights. Then a flag channel 12-20 acted as a flag channel 1-9. The offset to BCHAN was not used when examining if there were any unflagged channels in BCHAN-ECHAN, making still more trouble. Moved to 15OCT90 this date, 15JUL90 later from OZ. 6082. June 28, 1990 Unix Terminal I/O Chris Changed ZTTOPN in APLUNIX and added a Unix specific version of ZTTYIO. ZTTYIO and ZTTOPN now use the designated input-output unit (*) rather than relying on units 0, 5 and 6 being preconnected to standard error, standard input and standard output respectively (I assume this was calcified FORTRAN 66). In addition, explicit opens of {\tt /dev/tty} have been dropped from ZTTOPN. This provides the necessary hooks for interposing a more friendly interface between the user and AIPS in the Unix environment by allowing AIPS to be run in a pipe. Also removed Convex specific version of ZTTOPN which is no longer necessary. Relinked AIPS and ZSTRTA for now: will slowly permeate other AIPS programs. Moved nowhere. 6083. June 28, 1990 ZLPCLS (APLGEN, APLVMS) Bill J. Changed the logic in the effective handling of OUTPRINT. If OUTPRINT = ' 'then a temporary file is created for the line- printer output and this is deleted after printing. On the other hand, if the user specifies a suitably valid OUTPRINT file to receive the output, this is created, opened and closed properly, but NOT printed. (It has seemed to me that this logic is to be much preferred over the original especially when some users like to print a lot into a file to be examined with an editor elsewhere.) This fix answers Gripes 3921, 4069 and 4128. APLGEN:ZLPOPN.FOR had its output messages changed slightly to remove explicit reference to "lineprinter files". PRTUV had been modified before to have this logic. That program however uses different calls. Moved nowhere. 6084. June 29, 1990 CALIB Neil Killeen/Bill C Fixed bugs in CLBHIS which caused it to blow up if there were more than 35 antennas in the AN table. Moved nowhere. 6085. 30 June 1990 PRTAB Eric Corrected character and bit-field lengths so that the data length controls what is displayed even if the column header is longer. Corrected type in setting the field lengths and starting character numbers which could make one or more of the first columns into blanks fields. Changed to center character fields under the headers using the maximum actual character data length rather than the column width. Improved the centering of column numbers over the columns as well. Changed it to remove nulls from character strings in all places they might occur inside the table. Also had it check for lengths needed for times and floating numbers (i.e. are all positive). Moved from Oz to 15JAN91 and 15OCT90. 6086. 30 June 1990 IMCCLR Eric Raised the lowest level of the standard 8-color contour used by TVFIDDLE among others. It was too dark for standard SUN monitors. Moved from Oz to 15JAN91 and 15OCT90. 6087. 30 June, 1990 ILNCLR, IENHNS Eric Corrected ILNCLR to avoid writing to non-positive array subscripts. It was blithly writing to negative array subscripts and wiping out other local parameters in TVPSEUDO bits of AIPS, SPFLG, and TVFLG. The bug hit unimportant variables only in AIPS itself. Corrected IENHNS to leave the graphics on or off as it was on input. Previously, it could get it wrong. Also corrected the plotting of the transfer function to account for zoom. Before, the plot went to the upper right of the memory, not the visible screen. Moved from Oz to 15JAN91 and 15OCT90. 6088. 30 June 1990 PCNTR Eric Corrected test on IYINC - it used a YINC of 1 except for YINC=1 where it used a YINC of infinity I think. Moved from Oz to 15JAN91 and 15OCT90. 6089. 30 June 1990 MCREAT, GRIDTB and GRIDTC Eric Changed two message levels in MCREAT from 8 to 7 so that disk too full and disk not allowed messages may be suppressed. Changed GRIDTB and GRIDTC to honor BADDISK even when creating a cataloged master grid file. Changed TVFLG/SPFLG help files to reflect this improvement. Added GRIDTC to OPT2.LIS in SYSCVEX:, SYSNRAO1:, and SYSVLAC1:. Moved from Oz to 15JAN91 and 15OCT90. 6090. 30 June 1990 TVFLG & SPFLG Eric Changed them to turn off the menu and restore the zoom only when that is useful. This speeds it up a lot. Corrected the handling of bad sort order in the input data --- it continued on erroneously due to a missing error code. Changed to turn on the TV cursor sooner and to clean up the line/box displays in the flagging routine. Fixed to turn off the graphics planes used to show the window when it is set with the cursor. Changed format statement 1010 in subroutine TVFCLP and SPFCLP so that internal write does not overflow recipient character variable (typed in clip levels). Changed $INC DTVF.INC and DSPF.INC to add source number to the FC table and to pass the source name around more. Changed SPFLG and TVFLG to allow the user to choose whether the current flags are applied to all sources or just to those explicitly visible at the moment on the TV. Made the status line(s) display smarter, allowing 3 lines for narrow windows. Corrected setting of time ranges and areas --- if button B or C was hit before any TRC was set then a bad entry was made in the FC table which could delete a lot of data later but not on the TV. Changed it to make B and C mean A when no TRC is set and to make an entry in the FC table only if some pixel in TVFLG/SPFLG was flagged by the command. Corrected history entries so that SOURCES get a close quote and names are not repeated. Changed so that the wedge when displayed is at least 128 pixels --- some wedges were too narrow to see. Changed help files to mention the new SET SOURCE FLAG. Changed fortrans and help files to exclude autocorrelation by default. This change was in the CV version of TVFLG but was not documented in their CHANGE.DOC. Moved from Oz to 15JAN91 and 15OCT90. 6091. 30 June, 1990 YSTV:WINDOW.C Eric Added more initializations. Previously the window width and height variables were unknown until the screen was redrawn (as for example by opening the icon). So, YWINDO got bad answers when SSS had never been opened. One can write to the icon now. Moved from Oz to 15JAN91 and 15OCT90. 6091. 30 June 1990 PRTAC & Text file reading Eric Corrected text file reading routines to ignore all commented lines. Changed: PRTAC Ignore lines starting with ';' and '*' in user list. AU1A Added line to ignore comment (';') when looking for 2nd ignored line in Inputs section of help file. AU2 Added line to ignore comment (';') when looking for 2nd ignored line in Inputs section of help file. AU2A Added line to ignore comment (';') when looking for 2nd ignored line in Inputs section of help file. Fixed up PRTAC - it was overhauled incorrectly, not taking the new format into account. Someone fixed it partially, but not completely or correctly. Moved from Oz to 15JAN91 and 15OCT90. 6092. 1 July, 1990 FNDEOT, AVTP, FITTP Eric Changed call sequence to FNDEOT to have the calling routine pass in the IO buffer and its size. Some tape drivers require that the buffer always be larger or equal to the largest tape record. Moved from Oz to 15JAN91 and 15OCT90. 6093. 1 July, 1990 tapes Eric Changed a number of routines to test for non-standard and other pseudo-FITS tapes (e.g., RPFITS). This means checking the record length on tapes and checking the value assigned to SIMPLE. Changed: APLSUB/TAPIO Corrected a bug in which the FTAB pointer was used to count logical records in a physical record. Corrected it to return an error code (10) when the logical and physical lengths are incompatible. It was blithly proceding when the physical was < one logical. Added code to mask message about logical and physical length incompatibilities when MSGSUP > 999. BillJ had commented out these messages which is a very bad idea; lengths must be tested. APLSUB/TPHEAD Changed to handle the error 10 from TAPIO and to check the value assigned to SIMPLE. APLNOT/TPIOHD Changed to look for 2880-byte logical records, to handle the error 10 from TAPIO and to check the value assigned to SIMPLE. AIPSUB/AU4 Changed to handle the code for non-standard FITS-like data (-1) now returned by TPHEAD. APGNOT/IMLOD Changed to handle the code for non-standard FITS-like data (-1) now returned by TPIOHD. APGNOT/UVLOD Changed to handle the error 10 from TAPIO and to check the value assigned to SIMPLE. Changed to check tape errors and close tape properly as well in subroutine UVPARM. APLPGM/PRTTP Changed to handle the error 10 from TAPIO and to check the value assigned to SIMPLE. Also simple removed error 10 returns for VLA on-line data tapes. APGNOT/FILLM Changed to ignore error 10 from TAPIO and to mask length incompatibility messages. APGNOT/FILLR Changed to ignore error 10 from TAPIO and to mask length incompatibility messages. APGNOT/MK3IN Changed to ignore error 10 from TAPIO and to mask length incompatibility messages. APGNOT/MK3TX Changed to ignore error 10 from TAPIO. Moved from Oz to 15JAN91 and 15OCT90 - Bill has MK3IN so told him. 6094. 1 July 1990 SETPAR Eric Changed it to test limits on all parameters read from the user (an AIPS Manager). In so doing, relaxed the tests on TIMDEST limits to allow for even shorter time limits. Moved from Oz to 15JAN91 and 15OCT90. 6095. 1 July 1990 IBM discovered Eric Corrected GO TO's with an address inside a DO loop (usually the end thereof) in CLCOR, TABED, UNCAL, SETJY, and TABEX. Corrected two EQUIVALENCE statements which equivalenced a variable to itself as well as to other variables in GETJY. The CLCOR error caused a real error on all machines. Moved from Oz to 15JAN91 and 15OCT90. Moved to CV from Oz (based on version fetched today from CV. 6096. July 1, 1990 IBM discovered Eric UVFND had a logic error due to an error in indentation leading to an incorrectly placed END IF. There were branches into the IF block which the IBM RISC Series 6000 compiler detected. Moved from Oz to 15JAN91 and 15OCT90. 6097. July 3, 1990 CLCOR Bill C Corrected error indexing input parameter array for OPCODE='PCAL' (MkIII VLBI manual phase cal). The wrong phases were being inserted in the table. Moved to 15OCT90 this date. 6098. July 3, 1990 UVFIX Bill J. Didn't report missing AN table(s) --- instead "ends succesfully" leaving user confused. Added user error messages and failure signals. In response to Gripe 4326. Moved to 15OCT90 this date. 6099. July 3, 1990 GETJY Bill J. Bumped up number of secondary calibrators allowed to cope with survey experiments by red-haired people. Moved nowhere. 6070. July 5, 1990 IMLOD Bill J. I had added a trap in IMPARS to cope with "rogue" history records i.e. those which were greater than 80 characters. For example, the end of one record (up to character 80) was "BPA = -" --- this caused the parser grief, and tried to assign a number greater than allowed in R*4 to a real. Bad history records shouldn't cause failure to load image so the trap tested for these and substituted FBLANK for the real. Unfortunately, the test often trimmed the bandwidth to FBLANK. Changed this to something more appropriate. I'm surprised that Gripe 4299 was the only complaint! Moved to 15OCT90 today. 6071. July 7, 1990 LISTR/LISTR.HLP Glen Added Cal Code Data selection to LISTR and LISTR.HLP. Also Shortened standard OPTYPE='SCAN' output for special cases of frequencies the same for all sources. Added 'LR' and 'RL' option to OPTYPE='LIST'. Added a few columns of precision when extra room was available. Stopped auto rescale of OPTYPE='LIST' for small decreases in scale. Removed 60 sec and 60 mins from time print. Added two subroutines ISORT and IMERGE to simplify antenna selection. Moved Nowhere. 6072. July 7, 1990 MX Glen Added further check for channel averaging. Case of averaging several channels, when in the last plane there are not enough channels left to average, the remaining channels are averaged. Moved Nowhere. 6073. July 7, 1990 DDTLOAD.001 Glen Replaced NCOMP=? with NCOMP=0 so ASCAL will pass the DDT test. Moved to 15JUL90 and 15OCT90 this date. 6074. July 5, 1990 OFMDIR, AU5E Eric Moved OFMDIR from AIPSUB to APLSUB, added an error return to its call sequence removing the POPS error handling, and removed any writing of the OFM to the TV. Changed AU5E to account for the new call sequence of OFMDIR, to do the writing of the OFM to the TV on OFMGET, and cleaned up the error handling. Moved from Oz to 15OCT90 and 15JAN91. 6075. July 5, 1990 FILINI and FILAIP Eric Went through FILINI cleaning up the GO TO's into IF-THEN-ELSE and correcting the code. Whoever overhauled it, ignored all the changes in file formats. Corrected file size computation for public CA files in FILAIP (these are never used anyway). Moved from Oz to 15OCT90 and 15JAN91. 6076. July 5, 1990 AU4 Eric Corrected call sequence to FNDEOT to pass in the tape buffer. Fixed precursor remarks to FNDEOT as well. Moved from Oz to 15OCT90 and 15JAN91. 6077. July 8, 1990 DATBND Phil Was not applying the bandpass corrections to cross-polzn spectra (i.e. RL, LR) due to an indexing error. Moved to 15OCT90 today. 6078. July 8, 1990 SMOSP Phil Never smoothed anything other than the first polarization (usually RR) since it only ever though there was one polzn in the data, it now finds out for itself how many there are and does them all. Moved to 15OCT90 today. 6079. July 12, 1990 DATGET Bill C. Added compiler directives includes in front of DO loops 150 and 180 to suppress vectorization. When these loops were short they occasionally produced erratic effects when compiled using the vectorizer on Convexes. Moved to 15OCT90 this date. 6080. July 12, 1990 UVFLG Phil Added the option to delete just the cross-hand polarizations if so desired. Also changed UVFLG.HLP to reflect this. Moved nowhere. 6081. July 13, 1990 SPLIT Phil When the CHNSEL adverb was used the channel bandwidth of the output file was incorrect. Moved to 15OCT90 this date. 6082. July 13, 1990 AVSPC Phil AVSPC had been broken for 6 months, this was 'common knowledge' at the AOC - it wasn't so bloody 'common' to me. The problem was that the number of channels in the spectrum was being extracted from the output catalogue header not the input, so it was always 1. In the course of fixing this I found the same problem as was occurring in SPLIT ie. the channel bandwidth of the output file was incorrect. Fixed. Moved to 15OCT90 and 15JUL90 this date. 6083. July 16, 1990 CALIB,FRING et al. Bill C. Changes mostly affecting multi IF fringe fitting. Modified to honor user specified ANTENNAS for single source data. Changed to blank solutions for which there was insufficient valid data or, in CLBDR, if the determined SNR ratio was below the minimum. Fixed problem in CLBDR that caused IFs to be ignored if a lower IF number failed to converge. CLBDR now sets SNRs for the reference ANTENNA to SNRMIN+1 if the solution converges. Modified CLBDR to use frequency offsets inside of a given IF to do the solutions and then refer the phase to the reference frequency; this allows using the uncertainty of the phase solution to estimate the SNR. Split the amplitude-phase and fringe fitting functions into CALIB and FRING. Split the host of least squares routines into separate files and made local include CALIB2.INC into INC:GAIN.INC. In both CALIB and FRING a new adverb SNVER allows the user to specify if he wants a new SN table or to add to an old one. In the latter case solutions in the old table that are going to be redone are flagged with a call to CALSEL. In both CALIB and FRING, most of the weirdness to allow both fringe fitting and amplitude and phase calibration was cleaned up and the unnecessary adverbs were removed. Also changed CALIB.HLP, added FRING.HLP. Added an option in CALIB to pass solutions that failed or had too little data. The effect is not to change the prior calibration and may be useful in certain self calibration applications. CALIB now reports as failed solutions only those for which the solution failed or the SNR was too low and reports the solutions that had too little data separately. CALIB now prints a header for closure error listings even if no data exceeded the specified limits. Modified RUNSYS:VLAPROCS.001 to reflect the changes to CALIB. Moved nowhere. 6084. July 16, 1990 CALREF Bill C. Corrected to ignore flagged table entries rather than quit. Moved nowhere. 6085. July 16, 1990 New adverbs Bill C. Added new adverbs SNVER and SOUCODE. Changed POPSDAT.HLP and NEWPARMS.001. Added SNVER.HLP and SOUCODE.HLP. Moved nowhere. 6086. July 16, 1990 CALSEL Bill C. New routine to flag selected entries in an SN or CL table. Moved nowhere. 6087. July 17, 1990 UVTBUN glen Difference between UVTBUN and UVUNIF was that UVTBUN doubled row U=0 while UVUNIF conjugated row U=0. Modified UVTBUN to conjugate U=0. Moved nowhere. 6088. July 19, 1990 AIPSSHAR Chris Placed a new shell script, AIPSSHAR, in SYSUNIX. This is a simplified Unix shell archiver: it packages text files into a shell script that recreates the files when executed. Instruction may be found in the header of AIPSSHAR. This script is intended for packaging AIPS related programs, such as the various AIPS screen servers, that do not need the AIPS libraries so that they may be more easily installed on machines without AIPS. Moved to 15OCT90 this date. 6089. July 19, 1990 CLCAL Bill C. Major overhaul. Added adverbs SOUCODE to select the source calibrator code and SNVER to select a single SN table to be applied. CLUPDA was changed from being in a library to incorporated into CLCAL. Now the input SN table are copied to a temporary SN table, smoothed if requested and applied to the input CL table and then deleted. The input SN tables are left unchanged. The output CL table is purged of entries to be replaced. There was a general clean up of the tables handling. Now uses SOUFND and ANTFND to determine selected sources, calibrators and antennas. Removed use of DSEL.INC. Also cleaned up the CLCAL.HLP file. Moved nowhere. 6090. July 19, 1990 ANTFND,SOUFND Bill C. Yet another set of routines to check lists of antennas and sources. SOUFND selects sources from the SU table on the basis of names, qualifiers and "Cal" codes. Both routines handle excluded antennas or sources. Moved nowhere. 6091. July 19, 1990 CL and SN tables Bill C. Modified handling of table column pointers. Added two new parameter includes PSNTAB.INC and PCLTAB.INC with pointers into SNKOLS and CLKOLS from SNINI and CALINI to point to the beginning of a column entry in the record returned from TABIO. These INCLUDE files should be synchronized with any changes made in the the table columns made in SNINI and CALINI. Affected routines: SNINI, TABSN, CALINI, TABCAL. Moved nowhere. 6092. July 19, 1990 SN2CL Bill C. Now selects calibrator solutions by antenna. Cleaned up handling of table column pointers and modified to use logical function SLCTD to determine is a given source, calibrator or antenna was selected. Moved nowhere. 6093. July 19, 1990 SNAPP Bill C. Modified no to delete the input table and cleaned up the handling of column pointers. Moved nowhere. 6094. July 19, 1990 CALADJ Bill C. Changed call sequence. No longer uses arrays in DSEL.INC or even includes DSEL.INC. General cleanup of table column pointers. Moved nowhere. 6095. July 19, 1990 CALREF Bill C. Bug fix. ALL table entries were being ignored due to mishandling a return code of 0 from TABIO (it was treated as a flagged entry). This will cause all referencing of antenna phases to fail. Moved nowhere. 6096. July 20, 1990 TABSRT Bill C. Corrected not to fail on empty tables. Now it closes the table and returns. Moved nowhere. 6097. July 20, 1990 UNSHR Chris Added a standalone FORTRAN program, UNSHR.FOR, to directory YSERV. This program allows AIPS shell archives created with AIPSSHAR to be unpacked on machines that do not have a Bourne shell. Also moved AIPSSHAR to YSERV and created a README file to explain what is happening in this directory. Moved to 15OCT90 this date. 6098. July 23, 1990 GM Phil After the major surgery on CALIB/FRING GM still called DDOT as though it was a function instead of a subroutine. Moved nowhere. 6099. July 23, 1990 COMRPL in SYSUNIX Chris COMRPL in SYSUNIX now excludes any environmental variable that ends in an upper- or lower-case variation of CWD or PWD from consideration when performing the reverse look-up of directory name to AIPS `logical' where it previously only exluded case variations of CWD and PWD as a complete environmental variable name. This prevents COMRPL being confused by the KornShell OLDPWD variable, and most other schemes that keep track of old working directories by storing them in environmental variables. Moved to 15OCT90 this date. 6100. July 23, 1990 FILLM Gareth This is a new version of FILLM which has the capability to create more than 4 uv-files in parallel. The previous version could only manage 4. However, with the implementation of all of the correlator modes in the VLA correlator, it is necessary to be able to fill 8 uv-files at the same time in the worst case (channel 0 and spectra for each of the 4 IFs). The maximum number of simultaneous uv-files is now defined by MXSTRM in FILLM.INC and MXSTRE in MC2.INC. These two MUST be kept identical. This version does not yet implement more than 4 uv-files at once. However, the parameterization meant changing so many lines of code that I think it would be wise to see if I have made any blunders before going further. I have myself tested this version on two spectral line observations. I will let the users provide me with more tests! If this works, then the majority of the further changes have to be made in the routine MCINI. Moved nowhere (surprise, surprise!). 6101. July 25, 1990 AN10RS Chris Reimplemented AN10RS to deal with the case of excluded antennae. AN10RS was producing nonsensical output if any element of ANTENS was negative while all elements of BASELN were zero. This lead to UVCOP, UVFLG and VBPLT not producing output if the user had tried to exclude specific antennae. The code for the original version was sufficiently opaque to make writing a new version easier than locating the bug in the old version. However, inspection of the old code showed up a number of defects in addition to the problem of generating incorrect output for excluded antennae: 1 - the header comments were misleading regarding the output if ANTENS was all zero or positive and BASELN was all zero; 2 - if there were any antenna numbers following a zero in ANTENS or BASELN these would be ignored; 3 - if an antenna number was repeated in ANTENS all subsequent entries in the array would be ignored (BASELN was saved from this by an obvious typing error but the number of unique entries in BASELN would not have been counted correctly; and 4 - only the first NANT antennas in each of ANTENS and BASELN were checked, where NANT was the maximum antenna number in the selected subarray (tasks using the ANTENNA and BASELINE adverbs should function correctly even if the user specifies a number of non-existent telescopes. These defects are all fixed in the new version 6102. July 26 1990 UVTBUN Glen Semi-major change allowing UVTBUN to uniform weight all sizes of maps. Now scales Uniform grid to fit in available "AP" memory if grid is too large. (Unchanged for big APs.) Will move to 15OCT90 after a few days of testing. 6103. July 26 1990 ALGSTB Glen New gridded unsorted subtraction. Will handle any size map. Functions identically to ALGSUB. Moved nowhere. 6104. July 26 1990 UVGRTB Glen New grid without sorting subroutine. Handels any size map. Will also handle Zerospacing with a user defined size. ZEROSP(2) is major axis FWHM in arc seconds. ZEROSP(3) is minor axis FWHM in arc seconds. ZEROSP(4) is position angle on the sky. Uses new Zero spacing subroutine UVSETZ. Moved nowhere. 6105. July 26 1990 UVUNIF Glen Minor changes to make UVUNIF more similar to UVTBUN. Moved nowhere. 6106. July 26 1990 MX/GRDSUB Glen Allow unsorted data to used gridded model subtraction. Both task and subroutine use variable ISORT in common DUVH to choose whether to use subroutines UVTBUN, UVGRTB and ALGSTB or the equivalent (older) set UVUNIF, UVGRID and ALGSUB. Moved nowhere. 6107. July 27, 1990 CLIP Dan Briggs/Phil Modification to CLIP to allow the user just to clip data on specified baselines, or all baselines to an antenna. Also modified help file to reflect these changes. Moved nowhere. 6108. July 27, 1990 VLBIN Phil Added the ability to print out the average clipper bias determined from the autoocorrelation functions. These are printed out for every antenna for every scan when AC data are present. The corrections for quantization and fringe-rotation are now applied to the xc data in VLBIN for spectral line data, they could not be absorbed into the b-factor because that would have to be different for xc and ac data. Added section to help file describing this. Moved nowhere. 6109. July 30, 1990 IBLED Phil So many changes I don't know where to begin. - changed the whole philosophy of the program, instead of gridding the data to a scratch file and having to use the MDISK routines to access it ( which was very slow for this type of editing), I now load up arrays in memory with the amplitude and phase and flagging information and from this I load the TV and do all the editing. This has given IBLED a speed up factor of order 20. - Now use graphics planes for the editing instead of the image plane as before. This means I can use IMVECT instead of YIMGIO, this speeded up the loading and reloading of the screen by a factor of 5 or so. One problem with this occurs on Suns where I sometimes have only one graphics plane and two image planes, in this I load the data plus its labelling to the one image plane and the menu etc goes in the graphics plane. On IIS and IVAS devices this problem does not occur. - Added new flagging and display options, there is now the ability to perform an interactive clip. The user can plot his error bars. Can display any of the STOKES parameters in his data. Has the option to flag all data to a given antenna. - Tidied up a lot of the subroutines to make program maintenance a lot easier. - Rewrote the help file to reflect all the new changes. Moved to 15OCT90 this date. 6110. August 1, 1990 UVGRID glen Removed some of the debugging messages and reformated others. Also allowed points slightly closer to the edge of the V max of the grid to be used. (In earlier versions, some points were excluded.) Moved to 15OCT90 Yesterday 6111. August 1, 1990 ALGSTB glen Removed some debugging messages and spare variables. Moved nowhere. 6112. August 3, 1990 POSSM Phil Someone had screwed up the ability of the user to supply his/her own plotting limits. Fixed so that the program once again uses these if specified. Moved nowhere. 6113. August 6, 1990 AN10RS Chris Fixed a small coding error which caused the first baseline of an explicit set of baselines to be lost (I don't understand why I didn't see this when I tested this case). Moved to 15OCT90 this date. 6114. August 7, 1990 DAYFX Bill C. New task to patch up times screwed up by FILLM at the beginning of a new tape. It looks for a day number that is last than the last good day number and replaces it with the last good day number. With any luck this task will not live long. Moved to 15OCT90 this date. 6115. August 8, 1990 AREAS.DAT Bill C Added definition of logical YXAS to point to Y.XAS as the directory for the NRL X windows version of the AIPS TV. Moved to 15OCT90 this date. 6116. August 8, 1990 XAS Z and Y routines Chris Placed the Y-routines and Z-routines necessary to support NRL's XAS screen server into the system. Some changes to the Z routines were necessary: addition of standard AIPS headers; removal of INTEGER*2 variables; removal of common header file from level-2 Z-routines; and removal of VMS support. The latter change was necessary since we have no mechanisms for compiling C under VMS at present and VMS support seemed incomplete. At present the Z-routines have only been placed in APLBERK since they rely on sockets. Moved to 15OCT90 on August 15, 1990. 6117. August 8, 1990 VLBIN Phil For line data the XC spectra were not being corrected by the fringe-rotation factor (1.28), this is normally part of the b-factor for continuum data but cannot be for line because it applies only to xc not ac data. Moved nowhere yet. 6118. August 8, 1990 IBLED Phil The phase was being incorrectly calculated due to a mis- placed parenthesis, fixed that. Also stopped the screen being refreshed (ngray planes + 1) times every time the YZOOMC is called. YZOOMC is now only called if the magnif. factor has changed. Moed nowhere yet, more changes pending. 6119. August 9, 1990 APLNOT:DATPOL Gareth This routine is internally documented to work only when KNCOR=4. It turns out that this was accurate, but it is nevertheless called with other values! I have changed it so that it does nothing unless KNCOR=2 or 4. In the case of KNCOR=2, it will apply only the mutual coupling between the parallel hands. It is also clear that this only formally went through the code overhaul; it needs major surgery. I will do this in the next week if I have time - otherwise I leave this as a reminder for the next person! Moved nowhere yet. 6120. August 9, 1990 AIPSPUBL:COOKn.TEX,AIPSPUBL:COOKm.MAC Bill J. Where n=1 to 15, m=82, 82G, G, Y and Z. The revision of the Cookbook is now complete. Reordered the complete volume. Added Appendix Y on tape and disk file sizes. Serious revision of all chapters and appendices. Note "Chapter 99" is now "Chapter 4". Moved nowhere. 6121. August 9, 1990 APLPGM:PRTTP.FOR Bill J. A new error code returned by TAPIO (error 10) was not handled correctly in subroutine PRTPHD in the case where PRTTP reads Modcomp archive tapes. Changed the logic to accommodate this. Tested on archive tape and FITS tape. Moved 15OCT90 today. 6122. August 13, 1990 AIPPGM:SETPAR.FOR Bill J. SETPAR wouldn't allow the setting of "-1" in the disk assignments (-1 ==> scratch disk). Minor error in code. Moved 15OCT90 today. 6123. August 14, 1990 XAS Chris Placed a self-unpacking shell archive (XAS.SHR) containing the source code and Makefile for the NRL X-AIPS server in area YSERV. The only change from the NRL release is the inclusion of standard AIPS headers. Documentation will follow. Moved to 15OCT90 this date. 6124. August 14, 1990 AIPSSHAR Chris AIPSSHAR is no longer forced to run under the Bourne shell (it appears to require bash or the KornShell) and added the braces necessary to cope with the case where there are more than 10 files to archive. Moved to 15OCT90 this date. 6125. August 14, 1990 AREAS.DAT in SYSAIPS Chris Moved the declaration of YXAS to its correct place in the Y-routine declarations. Unfortunately the actual directory is also in the wrong place (it should be a subdirectory of YGEN): this will be painful to fix. Moved to 15OCT90 this date. 6126. August 15, 1990 GAL Eric Corrected the plotting aspects of the task: (1) logical variables were only tested for value 0.0, now <= 0.0, (2) MAPOPN used a constant for an in/out variable, (3) plotting of character "o" for data points led to odd shifts of location, not draws an x with GPOS/GVEC, and (4) changed to correct input variables for the model computation routine. Moved from Oz. 6127. August 15, 1990 DFTPL Mark/Eric Fixed a bug in ACDFT related to skipping visibilities. The logic was all wrong resulting in a bad pointer computation, and it could never have skipped anything anyway. Excised everything relating to variable ISKIP. This was a hold-over from its early days as UVPLT. Corrected handling of top strings in plot; changed from 200 to 500 maximum time bins, made sensible defaults for UVRANGE (0,0 was believed!), get default TIMERANGE from the data, set default time increment to 0.01 of the time range, switch to hours on the x axis, and plot the UVRANGE when it is used. Also changed adverb names in the Help file. Moved from Oz. 6128. August 15, 1990 UVPLT Eric In working on DFTPL, I noticed that UVPLT had its handling of the top lines in the plot messed up yet again. The location given is the BLC of the first character! Changed the line number at the top of the plot given to GINIT and altered the line spacing. Also added a digit to the bandwidth. Changed the time display to use TODHMS and corrected a format. Moved from Oz. 6129. August 15, 1990 Printing Eric Began the clean up of printing which will lead in the end to a cleaner interface to the "line printer" which can be one of a scratch file to be disposed to a real printer, a disk file to be saved but not printed, or a terminal. Because of WRITEs to unit 1 in some of the code, the middle of these cannot be properly implemented as yet. As of now, do not use direct writes to unit LPLUN (1) and use LPOPEN and LPCLOS rather than ZOPEN, ZCLOSE, BATPRT, and ZENDPG. At present, these still just call ZOPEN and ZCLOSE as we have been doing. A final switch can be done later after we change things slowly to correct the situation: PRTLIN Changed to handle blank title lines without putting numerous blanks on the output. This will allow special forms to be used on page 1. Also fixed it to watch for nulls in the strings. PRTALN As PRTLIN. LPOPEN New routine to open the full (old) range of line printer options. Also calls BATPRT for batch jobs on real line printers. LPCLOS New routine to close all possible line printer devices and clear them from the FTAB. Handles the final page and batch remarks as well (ZENDPG and BATPRT). IMERGE Cleaned up the typing --- people who do not follow AIPS typing standards should never put code in standard areas. ISORT Cleaned up typing and corrected it to make the algorithm work. Above remarks go double when the routine does not even work! TODHMS Removed from LISTR - converts time to 4 integers with no 60's and no 24's. Even checks the sign. Moved from Oz to 15JAN91 and 15OCT90. 6130. August 15, 1990 Printing tasks Eric PRTIM Removed all direct writing to devices and cleaned up the typing. Made the page titles a bit more legible. Corrected initializing page headers for non-standard types of images. Changed to use LPOPEN and LPCLOS. PRTUV As PRTIM. Also removed the junk referring to ZTX... files. That belongs elsewhere and will eventually be put there. Add support for linear polarizations. Fixed default line count and changed to use TODHMS. LISTR As PRTIM. Removed TODHMS to be an separate subroutine in APLSUB. New CV help file also installed. PRTSD As PRTIM and changed to use TODHMS. UVDIF As PRTIM, plus corrected default line count and changed to use TODHMS. UVFND As PRTIM, plus corrected default line count and changed to use TODHMS. Moved from Oz to 15JAN91 and 15OCT90 (except LISTR). 6131. August 15, 1990 Other printing Eric Changed the following to use LPOPEN and LPCLOS: AU1A AU7 PRTMSG BLSUM CORER PRTAB PRTAC PRTAN PRTCC PRTGA PRTTP TXPL (1) Also installed Mike Kestevan's corrections to RENAME in AU7. He noticed that RENAME did not check all disks and otherwise did not use the tools already available in CATDIR. (2) Changed PRTGA to use TODHMS. Also changed it to print all antennas in sets of as many as possible at a time. Dropped adverb BCHAN from code and help file. (3) Corrected VLA display in PRTAN - a station was listed twice in the mistaken idea that there are 25 stations/arm. (4) Changed TXPL to use the XYRATIO in the plot file and to represent grey-scales better. NOTE: all are based on the 15JAN91 version in CV, so a few improvements have arrived beyond those done in Oz. They include summing the flux before the first negative in PRTCC and handling labeled tapes better in PRTTP. Moved from Oz to 15JAN91 and 15OCT90. 6132. August 15, 1990 IMFIT, JMFIT Eric Mike Kestevan noticed a long time ago that IMFIT/JMFIT did not handle the case of non-square grids correctly. He made a partial change in the RP/AT version. Changed the NRAO versions to use LPOPEN and LPCLOS for the line printer and implemented a full handling of the coordinate increments. Added a sanitised version of ELIPSQ to APLSUB. This routine converts the parameters of an ellipse to different scalings. Note that the AT version had several errors in ELIPSQ and in handling the negative axis increment normal on the RA axis. Moved from Oz to 15JAN91 and 15OCT90. 6133. August 15, 1990 XVSS Chris Repackaged into a self-unpacking archive and move to YSERV. This should make installation easier. Moved to 15OCT90 this date. 6134. August 15, 1990 VLBIN John Benson/Phil Fixed up the FBS correction to work properly when data were correlated with DLYBYREC true. The bad data at the bit shift boundary do not now occur. Also the fringe rotator correction for line data was incorrect, the square root of the true value was being applied. Moved nowhere yet. 6135. August 16, 1990 SSS Chris SSS, like XVSS and XAS is now packaged into a shell archive for easier installation and has been moved to YSERV. Moved to 15OCT90 this date. 6136. August 16, 1990 AIPSSHAR Chris Changed the argument order so that the archive name comes first. AIPSSHAR now works under the Bourne shell again. Moved to 15OCT90 this date. 6137. August 16, 1990 XASVMS.SHR Chris Archived the original versions of the XAS Z-routines from NRL and placed them in YSERV. This allows sites to get at the VMS specific code. This is tucked away like this since it is unsurported code: we do not have the equipment to test this in NRAO. Moved to 15OCT90 this date. 6138. August 16, 1990 PRTUV Phil The logic for controlling how many channels could be printed on a line when docrt = -1 (or 132) assumed that the ampl, phase and weight occupied 15 columns, in fact after recent changes they now occupy 19 columns. Moved to 15OCT90 this date. 6139. August 16, 1990 README in YSERV Chris Revised the README file in YSERV to update the index of archived source files. Moved to 15OCT90 this date. 6140. August 16, 1990 XVSS Chris Minor modifications for release 2.0 of XView. - Added (void) cast to xv_init, which now returns a pointer to the screen object. - Set the XV_VISUAL_CLASS of the frame to PseudoColor. - Removed the explicit call to repaint_canvas() from resize_canvas(). This was necessary in previous versions because the CANVAS_FIXED_IMAGE property was ignored so that the bit gravity was always set to NorthWestGravity. This meant that shrinking the window did not cause an automatic redraw of the canvas. The redraw is necessary to recentre the image. Removal of the call improves performance since the canvas is no longer redrawn twice if the size of the window increases. - XV_X and XV_Y are no longer set in resize_pressed(). The new version of olwm is sufficiently literal-minded that it was placing the origin of the canvas at (0,0) on the screen so that the title bar and the control panel were above the top edge of the screen. Moved to 15OCT90 this date (on the assumption that the customers will have XView 2.0) by October. 6141. August 16, 1990 APLNOT:TABSN Gareth Single character bug fix (.GT. becomes .GE.) to allow the second polarization to be written. Correction to new version made by Bill C - (6091. on July 19, 1990). Moved nowhere. 6142. August 17, 1990 CVEL Phil Modified code and helpfile to deal properly with subarrays. There was a buffer problem if a flag table was being applied to the data while shifting. This was fixed. Also fixed once again the problem of trashing the HI file. Moved to 15OCT90 on Aug. 21 6143. 18 August 1990 Printing Eric I have attempted to regularize printing. Do not use WRITEs to unit 1 for line printers or bother with BATPRT, ZENDPG, or ZOPEN or ZCLOSE for printing. Instead, use LPOPEN and LPCLOS described below for open/close and PRTLIN to print. The printing to terminal, line printer, or text file will be handled. Text files are opened with append so the user can add to an existing file. Batch jobs now print to to PRTFIL:BATCHjjj.nnn for job jjj user nnn unless an OUTPRINT adverb is given. Changed: ZOPEN (APLGEN:) dropped the opening of the line printer, cleaned up the coding. ZCLOSE (APLGEN:) dropped the closing of the line printer. ZLPOPN (APLGEN:) dropped the PNAME argument since it does only real line printers now, added LUN and FIND arguments, and added code to make an open in the FTAB. ZOPEN (APLCRI:) ZFORMF (APLGEN:) NEW - returns a form-feed character LPOPEN Revised to open using ZTTOPN, ZLPOPN, and ZTXOPN (for text file output with APPEND true). Dropped call to BATPRT. Added code to set the default name for batch print files to PRTFIL:BATCHjjj.nnn. Added code to detect whether the text file is new or not. New ones should not get a line feed at the beginning (conveyed via common to ZFORMF). LPCLOS Dropped call to BATPRT. Changed ZCLOSE call to calls to ZLPCLS, ZTXCLS, and ZTTCLS and code to clear up the FTAB. BATPRT Deleted. PRTLIN Added calls to ZTXIO when LUN = 3 (disk text file). PRTALN Added calls to ZTXIO when LUN = 3 (disk text file). ASSNBASIC (SYSVMS:) added definition of PRTFIL. ASSNLOCAL (SYSVMS:) changed LPRNTR to point at PRTFIL. LOGIN.CSH (SYSUNIX:) added definition of PRTFIL. LOGIN.SH (SYSUNIX:) added definition of PRTFIL. Moved from Oz to 15JAN91. 6144. 18 August, 1990 OFMLIST Eric Forgot earlier to change this routine to use LPOPEN and LPCLOS. Moved from Oz to 15JAN91 and 15OCT90. 6145. 18 August 1990 Help files Eric Upgraded the information about OUTPRINT in the following help files: BATCHJOB BLSUM CORER EXPLAIN HELP IMFIT JMFIT LISTR OFMLIST OUTPRINT PRTAB PRTAC PRTAN PRTCC PRTGA PRTHI PRTIM PRTMSG PRTSD PRTTP PRTUV TXPL UVDIF UVFND Moved from Oz to 15JAN91. 6146. 19 August 1990 DDT Eric Changed help files DDT and DDTSAVE: (1) Added TERSE --- otherwise the adverb is useless. (2) Added OUTPRINT --- allows the print to be directed to a disk file, rather than be printed. Changed DDTLOAD.001: (1) Made TERSE apply to TPHEAD listings (2) Made UV flux difference cutoff far less sensitive; with fluxes like 16 Jy, tiny phase differences make more 15 milliJy amplitude differences. Raised it to 50 millis. (3) Made sure DOCRT = -1 on UVDIF runs, PRTMSGs, and PRTAC; DDT is intended to run stand-alone, not with someone at a terminal to hit carriage returns. (4) Restored PRTAC --- it is one of 2 purposes for running the whole thing. Also restored the PRIORITY 5 PRTMSG. (5) Changed to do PRTHED only if TERSE <= 3 and some of the PRTCAT only if TERSE <= 2. Moved from Oz to 15OCT90 and 15JAN91. 6147. 20 August 1990 TVFLG, SPFLG Eric Changed them to use TODHMS in the display of times on the TV and in the history file. This will avoid odd 60's and 24's. Changed how the times of the rows in the master grid are found. Before, odd beating of the recorded times with the times of the rows could occur. Now, the first row begins 1 or 5 seconds (VLA and AT Compact Array) before the first sample. Every other row with data begins $n \times \Delta_t$ thereafter. Moved from Oz to 15JAN91. 6148. August 20, 1990 AN10RS Chris Removed a block on autocorrelations. Moved to 15OCT90 this date. 6149. August 21, 1990 GAPLT Eric Bill J "fixed" the call to GINIT which he called "rubbish" by getting 2 arguments out of order (they were fine) and by adding CATBLK (which was missing). Corrected. Moved from Oz to 15OCT90 and 15JAN91. 6150. August 21, 1990 APLVMS:ZLWOP.FOR Harry Payne/Bill C During the installation of 15APR90 for VMS, LWPLA failed to work, basically because the output spool file was opened with "FORM='UNFORMATTED'" in APLVMS:ZLWOP.FOR, but APLGEN:ZLWIO.FOR tries to write formatted data to it. The file needs to be opened as a formatted file with logical records long enough to hold 256 CHAR*4's. Moved nowhere. 6151. August 21, 1990 CVEL Phil When there are many sources in the datafile the interpolation of velocity shifts was being screwed up because it was not being reset at a source change. Fixed. Moved to 15OCT90 this date. 6152. August 22, 1990 SN2CL Bill C. Removed a bug I recently inserted which caused CLCAL to go into an infinite loop whenever the last calibrator in the SN table was not in the list of desired calibrators. Moved nowhere. 6153. August 22, 1990 TVOVER.TEX Chris Added TVOVER.TEX, a document giving an overview of the various workstation servers, together with installation instructions, to YSERV. Also updated the README file in YSERV and deleted the now obsolete TVINSTALL.TEX in DOCTXT. Moved to 15OCT90 this date. 6154. August 22, 1990 SNINI Phil The call to FNDCOL near the end of the routine was being passed MAXCLC as the number of columns to find, this caused various variables to be overwritten, changed to pass NCOL which is determined just prior to the FNDCOL call. Moved nowhere. 6155. August 22, 1990 ACFIT Phil Problem with FQ ID's, the input adverb was being reset to -1 when the template spectrum is read, fixed this. Also fixed an incorrect format statement which attempted to report the above error. Moved to 15OCT90 this date. 6156. August 22, 1990 APLBERK:ZDATE.C,ZTIME.C Brian These two routines broke when NRAO1 was upgraded to Convex OS 8.1. Times that were declared int should instead be declared type time_t. I checked some other Berkeley systems (SunOS [34].x, Alliant, Convex 7.1, Ultrix) to make sure this wouldn't break anything. Will be moved to 15OCT90 when I figure out how. 6157. August 22, 1990 Unix Directory reading Brian The following all broke at Convex OS 8.1 for the same reason. APLBERK:ZTACT2,ZTKILL,ZTXMA2 APLCVEX:ZSPACE Convex went to a more modern directory reading scheme. A different include file was required, and a struct direct becomes a struct dirent. Also some code that checks for null directory entries was excised since it was non-portable and made backwards compatibility more difficult (this test is only necessary if you have to read the raw directory). The source is ifdef'd on the posix source keyword to try to distinguish when it should try to use the new format. Will be moved to 15OCT90 when I figure out how. 6158. August 22, 1990 APLCVEX:ZTPOP2 Brian This broke at Convex OS 8.1 because the define FASIO became _FASIO to avoid compliance breaking namespace pollution. I ifdefed on posix source so it should still run under pre 8.1 convex OS. Will be moved to 15OCT90 when I figure out how. 6159. August 22, 1990 AIPSPUBL:COOK4.TEX Bill J. Revision of the "Calibration" chapter prior to going to the printers --- to change references to the Ken Sowinski conversion program for pre-Jan 1988 archive tapes. It appears that I was a little premature --- it won't be ready till the year's end. Moved nowhere. 6160. August 22, 1990 ACFIT Phil Made the default value of the subarray adverb mean subarray 1. Moved to 15OCT90 this date. 6161. August 22, 1990 BAKTP Mark Corrected call sequence of FNDEOT in APLUNIX:ZBKTP1.FOR. Moved from Oz to 15JAN91 and 15OCT90. 6162. August 23, 1990 POSSM & VBPLT Phil Both POSSM and VBPLT assumed that the row number = antenna number, in general that was not the case. Changed so that assumption is not made. Moved to 15OCT90 this date. 6163. August 24, 1990 PSNTAB.INC Bill C The pointers to the delay and rate columns were reversed. This caused TABSN to do the wrong thing in FRING. Moved nowhere. 6164. August 24, 1990 SOURNU Phil There was an interesting piece of code in here, on one line it was stated: equal = source(j).eq.sounam on the next: equal = source(j).eq.' ' which of course negated the first line. I removed the second line. Moved to 15OCT90 this date. 6165. August 25, 1990 UVCOP Gareth Two mistakes in a single IF statement in ESTSIZ led to an underestimate of the size of the output data set. Fixed. Moved nowhere. 6166. August 27, 1990 SNPLT Phil Added the qualifier adverb to SNPLT, a data statement inside the code forced all qualifiers to be 0, which of course is not true in general. Moved nowhere. 6167. August 27, 1990 APLUNIX:ZM70X2.C Brian The M70 has been hanging on NRAO1 in Charlottesville. Arnold Rots suggested a kludge. When a read or write to the device fails we lie and say that it works. We then keep throwing out read/write requests until we have a header again. If we have a lot of failures we really exit. This is ugly but might be an acceptable alternative. Moved nowhere. 6168. August 28, 1990 APLNOT:UVGET.FOR Bill J. UVGET was not handling the case of LL Stokes only. This was calling DGINIT with Stokes character code "HALF" and creating an output data set with the single dimension Stokes axis having a value of -1 i.e. = "RR". In most cases, this shouldn't hurt but we should get it correct. Added a subroutine FNDPOL to (1) find the input polarisation specifically if the given code is blank and to copy the input polarisation to the output and (2) given a valid Stokes code, check that the input can support this. This attempts to cover all cases but there may be a problem with XY data (AT) since there is no Stokes code string equivalent to "RRLL" = "HALF" --- for XX, YY only. In this case the subroutine passes " " in the hope that this will be o.k. Moved nowhere. 6169. August 29, 1990 APLVMS:ZTOPE2.FOR Bill C. Added "READONLY" to OPEN statment. This was causing one AIPS using HELP, EXPLAIN or INPUT to block another AIPS from accessing the .HLP file causing HELP, EXPLAIN, INPUT, GO, TPUT and TGET to fail. This was especially a problem for batch jobs. Moved to 15OCT90 this date. 6170. August 29, 1990 DATBND Phil In the case where the data had only one polzn but it was LL not RR the routine did not apply the correct column from the BP table. Moved nowhere yet. 6171. August 29, 1990 POSSM Phil Would not plot the LL bandpass function if there was only one polzn in the database and it was LL not RR. Moved nowhere yet. 6172. August 29, 1990 BPASS Phil Once again, in the case of only LL in the data there was a bug. Fixed BPASS to write both polarization columns to the BP table in this case because there is an implicit assumption that the first polzn is R, the second L in the BP table. Also made a minor fix for the AT case of IQUV; in BPUPDT the NCOR variable was being passed for the purposes of initializing BP table arrays, what should have been passed was NPOLBP, i.e. the number of polzns the BP table was initialized for. Moved nowhere yet. 6173. August 29, 1990 UVCOP Phil When a FREQID is selcted the UVCOP the output tables were supposed to have their FQ ID changed to 1 like the data. The code I had inserted several months ago intended to do that but a typo screwed it up. Moved to 15OCT90 this date. 6174. August 29, 1990 CVEL Phil There was an infrequently used call to CHNDAT that asked for CH/FQ version 0, CHNDAT must be passed the version number actually needed. Changed to ask for version 1. Moved to 15OCT90 this date. 6175. August 29, 1990 APLNOT:FNDPOL.FOR Bill J. Removed format error. Added some combinations of requested Stokes/actual Stokes that are semi-legal and are required by HORUS amongst others e.g. RR (Stokes value = -1) is equivalent to I (Stokes value = 1) Moved nowhere. 6176. August 30, 1990 APGNOT:FETCH.FOR Bill J. Unreasonably small limit on row dimension --- was 100 --- changed to 1024. FETCH did not warn user that the actual row read was larger than the character buffer and truncated the image. Moved nowhere. 6177. August 31, 1990 BPASS Phil When fixing up BPASS for stokes LL I forgot to deal with the case of the user asking for LL alone. Moved nowhere yet. 6178. September 3, 1990 TVMOVIE Eric Changed TVMOVIE to allow it to use zoom factors lower than the number of frames across the screen when it is run on workstations. Then it can run faster and still show all the pixels in the frame. Files changed: SSS.SHR Changed CURSOR.C portion to compute the cursor positions using the displayed center correctly rather than assuming the center to be the middle of the image memory. Corrected a subscript error when forcing a desired window size and changed it to make sure the desired window is fully displayed. TVMOVI.FOR Changed to support workstation windows and zooms of lower value than those suggested by the image frames. AU5D.FOR Changed to pass the new parameter to TVMOVI based on new user adverbs. TVMOVIE.HLP Added new adverbs: DOCENTER to select this new option and FACTOR to control the desired display zoom. TVCUBE.HLP As TVMOVIE.HLP. REMOVIE.HLP As TVMOVIE.HLP. The whole AIPS model of zoom and scroll should be changed away from the old IIS Model 70 (strange) conventions, but I can't do that without IIS Model 70's and IVAS's to test as well as the workstations. So that will have to wait. Moved from Oz to 15JAN91. 6179. September 4, 1990 ANCAL Bill C Fixed a number of bugs and general cleanup. The geometry calculation was changed to use SOUELV to compute the source apparent geometry at each antenna. This lead to the removal of a number of routines. The declaration of UT (time) in subroutine TWERP was REAL rather then DOUBLE PRECISION causing an error interpolating antenna temperatures. Moved nowhere. 6180. September 4, 1990 GETANT Bill C. Modified to return antenna coordinates in the same system for VLBI data as for VLA data. This should fix problems for VLBI data in LISTR and CLCOR. Moved nowhere. 6181. September 5, 1990 TVMOVIE Eric Changed (1) TVMOVI to set the zoom to 0 when using no zoom in the display, (2) YSSS:YZOOMC to allow zoom centers from -MAXXTV to 2 * MAXXTV to allow the movie business to work, and (3) APLBERK:ZSSSX2.C to branch to exit on error (before if SSS died AIPS went in to infinite loops of error messages). Moved from Oz, ZSSSX2 also to 15OCT90. 6182. September 5, 1990 PCAL Bill C. Improved checking of error conditions. LPCALC did not check to see if an adequate total intensity of polarization model was passed. If the total intensity was 0 (a likely failure mode) then the fitted model had 0 for all parameters but PCAL otherwise appeared to work. LPCALC now complains and dies in this case. Also the handling of the case where an inappropriate IN2NAME etc. was provided was cleaned up. Dimension declarations were cleaned up in a number of routines. Moved to 15OCT90 this date. 6183. September 7, 1990 DFTPL Eric Corrected display of channel number. The number used for data selection was correct, but the number displayed was one too low. Moved from Oz to 15OCT90 and 15JAN91. 6184. September 9, 1990 YSSS routines Chris Moved YZOOMC from the now obsolete YSSS directory to YSS. Moved nowhere (but midnight job kicked in both NEW and TST since the migration from YSSS to YSS had not happened on Charlottesville UNIX machines). 6185. September 9, 1990 ZMOUN2 in APLCVEX Chris The Convex version of ZMOUN2 now uses the new (Convex OS 8.0 and later) tape interface ({\tt tpmount} and {\tt tpunmount} if it is available rather than the older {\tt tpalloc} interface. This allows AIPS to bypass ANSI labelling. This is done by testing whether {\tt /usr/convex/tpmount} exists and is executable and branching according to the result. Note that testing for the preprocessor string _POSIX_SOURCE is an unsafe method of distinguishing revisions of the Convex OS since the C compiler does not predefine it in all modes (in particular, it is undefined if the compiler is in strict ANSI mode). Moved to 15OCT90 this date. 6186. September 10, 1990 CLAB1,CLAB2,CTICS,TICINC Phil Modified to accept two more types of label: HA and LST, these are now associated with LABTYP 8 and 9 respectively. This was necessary to accomplish the modifications to SNPLT described below. Moved nowhere. 6187. September 10, 1990 SNPLT Phil Added two new abilities to SNPLT, it can now plot quantities against IAT time, antenna elevation, source hour angle and LST. The choice of which is determined by adverb XTYPE. Also it can now optionally plot error bars, this option is toggled by YTYPE. Modified the help file to reflect these changes. Moved nowhere. 6188. September 10, 1990 APLNOT:FNDPOL.FOR Bill J. Fixed error where I had ignored the possibility of conversion between RR, LL, RL, LR <---> I, Q, U, V. Seems to work o.k. now on LISTR and friends. Moved nowhere. 6189. September 10, 1990 APGNOT:LISTR.FOR Bill J. Replaced PJD changes and EWG changes. Eric's changes are minor --- formatting for AT. Moved nowhere. 6190. September 10, 1990 CCOPTS.SH Chris Changed CCOPTS.SH to define the preprocessor symbol _POSIX_SOURCE on the C compiler command line. This should be removed if using Convex OS 7.1 or earlier. Moved to 15OCT90 this date. 6191. September 10, 1990 ZMOUN2 in APLCVEX Chris It turns out that early versions (7.1 and earlier) of the Convex OS do not have the include file required by later systems. This is now protected by a preprocessor conditional keyed on _POSIX_SOURCE (hence the previous change). Moved to 15OCT90 this date. 6192. September 10, 1990 POSSM.FOR,.HLP Phil After a vigorous campaign mounted by E.Brinks I finally gave in and found the time to insert a positional shift into POSSM. The user can now request that his/her data be shifted in RA/Dec before plotting. Also updated the helpfile. Moved nowhere. 6193. September 11, 1990 ANCAL Bill C When the ratio of antenna to system temperature was given rather that the system temperature (usually the case for the phased VLA) the antenna temperature correction was effectively being done twice causing the corrected flux densities to be low by about the square root of the source flux density. Moved nowhere 6194. September 11, 1990 XAS.SHR Chris A number of minor changes. - IBM support: is included by {\tt xas.c} if the preprocessor symbol AIX is defined (hopefully this is defined by the RS/6000 C compiler. - Incorporated Eric's cursor changes from SSS. - Corrected indexing problems in returning the window size (corrections made to SSS by Eric). - Removed calls to XSetInput on EnterNotify and LeaveNotify in {\tt xas.c} events so that XAS complies with the ICCCM with regard to input focus. Moved to 15OCT90 this date. 6195. September 11, 1990 XVSS.SHR Chris Incorporated changes to {\tt cursor.c} from SSS. Moved to 15OCT90 this date. 6196. September 12, 1990 UVDIF Bill C Modified to accept compressed data. It now also keeps statistics on the correlation values that it checks (which is only what it can print). The ratio of the peak and RMS sine/cosine differences to the peak sine/cosine value is reported and put in the catalog header of the first input file as keywords 'RATMAX' and 'RATRMS'. Moved nowhere. 6197. September 12, 1990 YZOOMC.FOR in YXAS Chris Incorporated recent changes to version in YSS. Moved nowhere. 6198. September 12, 1990 APGNOT:LWPLA Bill J. Modified this to write the text "%!" and "%% (origins)" in the file before the shell script takes it over. "%!" tells the printer that what's coming IS PostScript. This avoids problems when users create a disk file and attempt to plot it later. Chris thinks that it was done the other way for some reason but I can't see why. Let's see what happens. Changed the associated script SYSUNIX:ZLASCL.PS to do the correct thing. Moved to 15OCT90 today. 6199. September 13, 1990 PCAL Bill C. Fixed problem with PCLSEL rereading the source fluxes for each IF. This was preventing the fitted polarization fluxes from the first IFs from being recorded in the SU table. Also cleaned up some array declarations using values for table arrays from PUVD.INC. Also cleaned up handling of AN table in PUTANT to make it relatively immune to changes in the AN table structure. Changed the message level of the messages describing the solutions in LPCALC from 8 to 4. Moved nowhere. 6299. September 13, 1990 GETJY Phil GETJY was complaining that it could only calibrate 130 sources when the user was only trying to calibrate 10, this was because a logic error in one routine caused GETJY to try to calibrate all sources in the SU table regardless of what the user wanted. In the course of fixing this I uncovered another problem; the helpfile claims that the adverb CALCODE only refers to sources in the CALSOU list, this was not the case it was checking all sources for the specified CALCODE, fixed this. Moved to 15OCT90 this date. 6300. September 14, 1990 (AEST) APLPGM:TAFFY.FOR Mark In routine SENDMA, the test for deferred output was logically opposite to that required. Moved nowhere. 6301. September 14, 1990 VLBIN Phil After weeks of investigation I rediscovered the wheel. Apparently the cross-power amplitudes from the MkII correlator are a factor of 2 too high, this may be only true for correlator mode 0, we have to test this. The Cfa group have known this for years but neglected to write it down or tell anybody at NRAO. Effectively what that means is that the quantization corrections for line data should be: s(true) = s(corr) * 1.57 (Van-Vleck) * 1.28 (fringe-rot) ------------------------------------- 2.0 (bugger factor) By a (what I regard as uncannily weird) concidence, 1.57 * 1.28 / 2 = 1.005, i.e. YOU DON'T HAVE TO CORRECT THE DATA AT ALL! Fixed up VLBIN to do these corrections under user control, also modified the helpfile to explain all this. Moved to 15OCT90 this date. 6302. September 14, 1990 ACFIT Phil The weights from the fitting were being written into the SN table incorrectly. The inverse is what should have been written - fixed. Moved to 15OCT90 this date. 6303. September 14, 1990 New task: UVGLU Huib/Phil A new task submitted by Huib Jan van Langevelde. This will glue together matching frequency blocks of spectral line data, i.e. if you had to split your 32 channel database into 4*8 channel files you can glue them back together with UVGLU. Moved to 15OCT90 this date. 6304. September 16, 1990 VBPLT Phil Finally (after how many years?) added the ability to plot error bars associated with the data. The default (YTYPE= 0) is still to not plot error bars. Modified helpfile also. Moved nowhere. *********************************************************************** Changes while 15JAN91 was NEW: *********************************************************************** *********************************************************************** 6305. September 24,1990 MK3IN Athol/Bill C Extensive improvements and enhancements. 1. Extensions for Spectral Line Data. Some changes in the Fourier transform conventions were neccessary in handling multiple correlator module lag functions. In particular the lag order convention needed to be reversed to concatenate the output of the modules. This produced several other changes in the Fourier transform of the lag functions to cross-power spectra which are discussed in a separate memo. A difference in sign convention between FOURG and QUIKFT was resolved. The cross-power spectra are reversed in frequency and have opposite sign of phase with respect to the original MK3IN. The frequency order is consistent with the known single dish spectra and the sign convention is consistent with the SAO program MK3MK2 (which claims to use the NRAO convention). AC functions are treated separately and are corrected for bias and clipping effects before being transformed to AC spectra. Auto-correlation functions are typically generated by cross-correlating half the XC lag range. Individual correlator modules can however be placed in auto-correlation mode and this may be used in the future. Some changes in the code were neccessary to deal with the two methods of generating AC functions. Correlation functions which do not have the same number of lags as the output file are skipped and a warning is printed. This is neccessary to avoid 8-lag fringe finding scans which are often included in the output tape. A warning message is also printed in the centre lag of the correlation function is non-zero. Magic value blanking is used in the lag domain to avoid transforming lag functions with missing lags. 2. Multiple Observing Bands. The new MK3IN produces an FQ table and allows multiple observing bands. 3. FBS Corrections Numerous test were carried out to determine the FBS model exactly. The new version should now do this correctly. 4. Checks on A-tape File Structure. Some checks were introduced to ensure that the A-tape file structure is as expected by MK3IN. The A-tape file headers are decoded and the MK3 file parameters are extracted. A warning is printed if the tape is a member of a multi-volume SAVEM set, as this may cause loss of data. The type 50 records within each extent are required to be in the expected order and only complete scan-baseline headers are used. These checks should diminish the effect of tape errors which may cause loss of synchronization and will also alert the user to any future changes in the MK3 data format. 5. Data Rejection Criteria The data rejection criteria used by FRNGE at the Haystack correlator were implemented in the new MK3IN version. A breakdown of the error statistics is available for each scan which can optionally be indexed on correlator module serial number. This feature was requested by the Bonn correlator group for spectral line data. 6. Baseline Indexing The MK3 tape format is baseline-based and consequently there is a high degree of redundancy in source- and antenna-based parameters which is the format that AIPS requires. The new version of MK3IN assumes that source and antenna parameters in files for different baselines may not be consistent and carries out several checks in this regard. The data are not rejected or changed in any way but a warning message is printed. The correlator model input parameters (or closely associated quantities) are also monitored throughout the run. A summary of the correlator model is printed at the end of the run. These consistency checks should help to isolate problems in the data set caused by changes in the correlation strategy or when merging data correlated at different times or different correlators. These checks may also alert the user to any unnanounced changes in the MK3 output tape format. 7. New Input Parameters Some minor changes in the input parameters were unavoidable but these were kept to a minimum. The APARM parameters are now: APARM(1) CL table increment (min). APARM(2) Maximum parity error rate (def: 0.01) (previously PE count). APARM(3) Maximum dropout rate (def 0.5). Correlator modules are rejected if the number of bits correlated is less than this fraction of the expected correlator count. APARM(4) ge. 1 => No FBS (previous APARM(3)) APARM(5) Max no of files on tape to process. (0 = until EOF). This is useful for short test runs where it is important that task terminates normally and writes the full complement of tables. APARM(9) Print level (0,1,2,3). See below. APARM(10) Debug flag (as before). Experiment file parameters: STATIONS Station names (as before). TIMEOFF IAT-UTC (as before). NO_LAGS Accept only lag functions of this length (if not present this will be set by the first lag function encountered). This parameter was introduced to avoid problems when the first scan is a fringe finding scan. NO_IF Number of AIPS IF's in output file. Introduced for similar reasons to NO_LAGS. Optional, if not given then this is determined from the first scan. SIDEBAND Force double sideband structure in AIPS output file (if this parameter is 'DOUBLE'). Optional, the sidebands will be concatenated if both USB and LSB data are present. STOKES from set {'RR','LL','RL','LR'}. Define the Stokes range of the output file. The largest contiguous range within the parameters is used to define the output file. Optional, if not given this will be determined from the first scan. FREQCODE MK3 character codes for {RR,LL,RL,LR} in that order. These codes are not very standardised. A warning is given if an unidentified MK3 code is encountered. Previously these data were assigned the lowest Stokes value. (def: 'R','L','r','l'). The optional parameters defining the AIPS output file structure were introduced to deal with tapes containing test scans as the first files (eg AC tests) and to force consistency. 8. Print Levels The print levels available are: 0: Print only a source header for each scan. 1: Include a one line error summary for the scan. 2: Print the MK3 file names and extent numbers. 3: Produce a error summary for each correlator module. 9. Miscellaneous The time indexing of each data record required modification due to a few minor typographical errors and rounding errors. It was also discovered that data records may not have time marks at integral correlator accumulation periods for the nominal scan start time. A scan start time offset is accumulated for each scan and the time mark is written for the AP midpoint. The HP number conversion routines were checked against known HP number reprresentations. A minor error in the conversion of single precision reals was found. A correlator index is generated from the 20XX module cross-reference table before processing each new type 51 data extent. This allows video converters to be multiply assigned in a given scan and should allow XC/AC mixing. The delay offsets can be in arbitrary order in the cross-ref table. 10. Code Consolidation The large common blocks were broken up into smaller units to reduce considerable compiler overheads (debug tables) and the preamble comments were updated and consolidated. Moved from 15APR91 and to 15OCT90 this date. 6306. September 24, 1990 VLA cont. calib. test Bill C. Added a new suite of POPS scripts to time and verify the correctness of the AIPS continuum calibration software especially as used for VLA data. These tests are patterened after the DDT tests and consist of two RUN files: VLACLOAD.001 and VLACEXEC.001; the first to read in procedures and the second to execute the tests. The two HELP files are VLAC.HLP and VLACSAVE.HLP. The test data and the master copy of the answers are on a FITS tape and were computed on a Sun 3/60. The test will run in 15JAN91 and later releases of AIPS. The tasks tested in some modes are: UVLOD, IMLOD, FITTP, UVCOP, TABED, UVFLG, SETJY, CALIB, TASRT, TASAV, GETJY, CLCAL, PCAL, TACOP, CLCOR, LISTR, SPLIT, UVDIF, HORUS and COMB. Moved from 15APR91 this date. 6307. September 24, 1990 AVSPC Phil After much negotiation with Neil we have finally settled on what AVSPC should do with the reference frequency, it leaves it alone, so the reference pixel is updated to reflect freq. changes due to channel selection. Also the bandwidth in the output file is the input bandwidth * the number of channels averaged. Moved from 15APR91 this date. 6308. September 24, 1990 BPASS Phil There were two problems remaining with LL polarized line data: (1) If the user had a dual polzn database and requested solutions for LL alone, an indexing error caused the channels to be shifted to the right by one, this meant that the last channel in the internal buffer was occupied by the u, v, w of the next record, this caused large closure error messages to be printed. (2) Also in a dual polzn database, when dividing by an external channel 0, another indexing error caused the LL line data to be divided by the RR continuum. I discovered that I had also broken the ability to generate seperate BPASS entries for multiple bandpass calibrator scans. Moved from 15APR91 this date. 6309. September 24, 1990 ACFIT Phil I added the ability to use the SNVER adverb, by incorporating a call to the CALSEL subroutine. In the course of debugging this I fixed a couple of other problems: (1) the last entry for a particular run of ACFIT had the incorrect source number; (2) if the user just wanted to solve for a single antenna rather than all antennas, there were logic problems, although it still worked; (3) if there was no valid data at all for a particular solution the program printed confusing error messages, but again still worked. Moved from 15APR91 today. 6310. September 24, 1990 New adverbs Phil Two new task-specific adverbs, XAXIS is used in SNPLT to specifiy which type of axis is to be plotted on the abscissa; DOERBAR is used by VBPLT and SNPLT to specify the use of error bars on the plots. Modified POPSDAT.HLP and NEWPARMS.001. Moved from 15APR91 this date. 6311. September 24, 1990 UVBAS Juan Uson/Bill C Fixed a problem with flagged channels. If a channel was flagged, no higher channel numbers were processed in averaging the "continuum" channels. Moved from 15APR91 this date. 6312. September 24, 1990 MAKMAP/UVPREP/HORUS Glen MAKMAP was modified to call UVGRTB which will grid any size problem (up to 4096). Unfortuately an assumption in a subroutine UVPREP was not correct for HORUS (although it was correct for MX.) UVPREP fixed and HORUS relinked. Moved from 15APR91 this date. 6313. September 24, 1990 MX Glen Surprise. A long standing bug was uncovered. MX did not scale U,V,W to second IF for data selection. This caused some data to be outside ALGSTB, ALGSUB limits in gridded subtraction. Particularily frequent when working with multi-array data. Modified only MX subroutine MXSEL. Moved from 15APR91 this date. 6314. September 25, 1990 AVSPC Phil A philosophical argument with Neil was finally resolved (in Neil's favour unfortunately), and the calculation of the new reference pixel was changed. Moved from 15APR91 today. 6315. September 30, 1990 APGNOT:FILLM.FOR Gareth This new version of FILLM represents several months' reworking and debugging. The major changes are as follows: 1. The keywords associated with the multi-source files have been changed from LINECHAN and CORRMODE to OBSCHANS, CORRMODE, SELCHANS, and VLAIFS. This is to permit more sophisticated selection criteria for appending data to existing datasets. OBSCHANS number of channels on VLA archive tape CORRMODE the VLA correlator mode SELCHANS channels selected from OBSCHANS VLAIFS the names of the VLA IFs used (OBSCHANS and SELCHANS are not used for continuum data.) A multi-source file is checked for all of these parameters as well as the Band, # polarizations, 1st Stokes' parameter, and # IFs to determine whether data should be appended to the file by FILLM. This change will mean that this version of FILLM cannot append data to databases created by previous versions. 2. The naming convention for output files has been simplified. The classes are '? BAND' for continuum, and 'LINE' and 'CH 0' for spectral line. Classes such as 'CH 0 1' are obsolete. 3. A new source (NX) entry is now forced whenever the Start LST of the observation changes. This is how the VLA on-line system marks a new scan. This forces a new source in some cases where the older versions did not. It has a two main advantages: it reliably detects a source scan change without checking for the almost infinite possible changes in the set-up of the observations; and it allows QUACK to be run subsequently for each change of the on-line system configuration. The only disadvantage, that logically contiguous data are now in separate scans, can be easily addressed by running INDXR post facto. 4. The subroutine MCWANT, which controls the selection of data and the changes in configuration, has been completely recast. Selection criteria that are always in effect are timerange (and date), program name, and VLA subarray. Selection criteria that may be overridden by DOALL=1 (true) are observing mode, number of channels per baseline, frequency, and band. 5. All possible spectral line observations with the VLA are now fully supported. The flexibility now built in to the VLA on-line system is reflected in the complexity of the subroutine MCINI. There are 26 different ways (3 continuum and 23 spectral line) of writing the data into up to 8 multi-source files in parallel. (An additional 8 ways, for mixed continuum and spectral line data, are supported here but not yet provided by the VLA). 6. The derivation of the number of channels per IF has been moved from subroutine MCINI to MCH1, where most other control information from the VLA archive tape is decoded. Bugs fixed include: 1. FILLM did not always correctly respond to a configuration change, such as FQ number or correlator mode, when averaging data. This was helped by the simpler and more rigorous scan change criterion (see above). A change introduced in February 1990 in the 15APR90 version had this unreliability as an undesirable side effect. 2. A baseline record that was split between to physical tape records had a jump in phase. The subroutine MCGBRC miscounted by 1 the number of data points (real and imaginary being counted as 2) left at the end of the first record. This change was also introduced in February 1990 and was not noticed for several months except by two careful observers. It did not affect continuum data at all. Since the VLA has been in the larger (A and B) configurations since that time, there has not been much use of the spectral line modes. Moved from 15APR91 this date. 6316. October 3, 1990 New directories/logicals Bill C. Added several new directories and corresponding logicals in AREAS.DAT. ...APL/DEV/UNIX/BERK/IBM (APLIBM) IBM RS/6000 Z routines ...APL/DEV/UNIX/BERK/DEC (APLDEC) Decstation Z routines ...Q/DEV/PSAP/IBM (QIBM) IBM RS/6000 Q routines ...Q/DEV/PSAP/DEC (QDEC) Decstation Q routines ...INC/NOTST/IBM (INCIBM) IBM RS/6000 includes ...INC/NOTST/DEC (INCDEC) Decstation includes ...SYSTEM/UNIX/IBM (SYSIBM) IBM RS/6000 shell scripts etc. ...SYSTEM/UNIX/DEC (SYSDEC) Decstation shell scripts etc. ...APL/CONTRIB (APLCONTR) is a directory for contributed software that is not to be included in the standard installation but is available for users to compile, link and use. Source and HELP files go in this directory. Moved from 15APR91 this date. 6317. October 4, 1990 POSSM Phil Minor logic error caused the freq. labelling of the x-axis to fail. Fixed. Moved from 15APR91 this date. 6318. October 11, 1990 POSSM Eric Changed the setting of plot scales to avoid the 360-degree phase ambiguity when possible and to give some slight borders around the data. Also corrected an error message. Used the latest version from CV to return to CV, but did not keep it here. It claims to do phase shifting for position offsets but does not support sprectral-line data - and this in a spectral-line program! Moved from Oz this date to 15APR91 and 15JAN91. 6319. October 11, 1990 SPFLG, TVFLG, Calibration Eric Changed: SOUFIL It did not close the SU table on error and did not examine the kind of error from ISTAB but simply assumed that any ISTAB error meant non-existant SU table (a non-error condition). Now it will close the table and treats most ISTAB errors as real errors. OTBSRT Changed it to write a scratch table as the output and then to rename it to the desired output name only if everything has worked. Otherwise, when a file was sorted onto itself it could be destroyed by any error. TABSRT Removed close of input table - it is always needed now to go to scratch table. Corrected a recent addition to the code --- it is okay to exit quickly when there are no records, but only if the OUTVER is the same as the INVER. SPFLG Changed to call SOUFIL only when it is relevant and to honor its error return. Before it called it always and ignored the error return. TVFLG Changed to call SOUFIL only when it is relevant and to honor its error return. Before it called it always and ignored the error return. Moved from Oz this date to 15APR91 and 15JAN91. 6320. October 11, 1990 SPFLG, TVFLG Eric Changed them so that when they apply the flags to the main FG table they remove them from the master grid file. This means changing the FC table to zero entries and going through the master grid file and magic-value blanking all SPFLG'd or TVFLG'd pixels. Moved from Oz this date to 15JAN91 and 15APR91. 6321. October 11, 1990 LISTR Eric Changed SCAN display to show number of visibilities for each scan in the NX table list whenever the display is wider than 79 characters. Moved from Oz this date to 15JAN91 and 15APR91. 6322. October 11, 1990 SPFLG, TVFLG, TVFIDL Eric Changed SPFLG to display time and channel in all flagging modes. Otherwise, the flux displayed doesn't mean much. Also changed SPFLG to force TVPSEUDO to display some color initially. And changed TVFIDL to do have better initial values and hence not zoom further when it starts zoomed and to do black and white first even if a zoom was done ahead of all coloring. Changed TVFLG as SPFLG. Moved from Oz this date to 15APR91 and 15JUL91. 6323. October 11, 1990 SCHOLD Eric Corrected SCHOLD to move 4 characters to upper case so that "quit" would be accepted as such rather than as an (illegal) AIPS verb. Moved from Oz this date to 15APR91 and 15JAN91. 6324. October 11, 1990 UVAVG Bill C. Modified not to take weighted sum of source and FQ id parameters. This putting garbage into these variables. Moved from 15APR91 this date. 6325. October 11, 1990 APLIBM:* SYSIBM:* INCIBM:* QIBM:* Brian Checked Kerry's IBM RS/6000 modifications into the system. The files were: APLIBM: D2IBM.FOR ZCREA2.C ZGNAME.FOR ZTAP2.C ZTPWAD.C D2IEEE.FOR ZDATE.C ZLPCL2.C ZTIME.C ZTQSP2.C S2IBM.FOR ZDCHI2.FOR ZR32RL.FOR ZTKILL.C ZTTOPN.FOR S2IEEE.FOR ZDCHIC.C ZR64RL.FOR ZTPCLD.C ZTTYIO.FOR ZABOR2.C ZDEAXF.FOR ZRLR32.FOR ZTPMID.C ZTXMA2.C ZABORS.C ZFRE2.C ZRLR64.FOR ZTPMIO.FOR ZGNAM2.C ZTACT2.C ZTPOPD.C (ZACTV9 is checked out by Chris - I'll put it in later). INCIBM: DAPC.INC DBUF.INC DSEL.INC QIBM: QCFFT.FOR QINIT.FOR QRLSE.FOR QCLNSU.FOR QMULCL.FOR WHNALT.FOR SYSIBM: AIPS ASSNLOCAL.SH FCOPTS.SH LIBR.DAT ZLPCL2 AIPXAS CCOPTS.SH INCS.SH LINK ZTRLOG.C ASOPTS.SH CDVER.CSH LDOPTS.SH OPT2.LIS ASSNLOCAL.CSH CDVER.SH LIBR PP Moved from 15APR91 this date. 6326. October 15, 1990 MK3IN.FOR Athol/Bill C. Moved latest bug fixes to this version. Moved from 15JAN91 and 15APR91 this date. 6327. October 16, 1990 IBLED Phil Several new features and some bug fixes: (1) IBLED now has a mode in which it can create catalogued uv work files to which you can return for an editing session. (I call this its 'fire-and-forget' mode). (2) It now creates all its necessary scratch files at the start of the session to avoid disc space problems later. (3) The antennas, baselines adverbs are now used to select antennas to be edited. (4) The handling of the FC tables has been changed, previously when an FC table existed reading it and applying its contents was a painfully slow business. This table is now sorted to time order and the application is much more efficient. (5) IBLED could not apply the flagging information to single source compressed files, fixed this. (6) When averaging in time there was an error which caused one more visibility to be averaged each time, i.e. you asked for 2 minutes averaging time, you got three. (7) The times in the FG table were equal to time +/- interval/2, this was wrong it is now time +/- interval. (8) A display buffer was not being cleared so that when a new baseline or frame was displayed with a smaller number of points the old data from the previous buffer load was still visible. Also changed IBLED.HLP Moved from 15APR91 and to 15OCT90 this date. 6328. October 17, 1990 CVEL Phil I ran a profile on cvel and discovered that 28% of the cpu time was spent in TABIO, this was due to the fact that the CL table was read from the beginning for each visibility until the doppler freq. offset for a particular time was found. I changed this to keep an index of where I was in the CL table for each antenna and CL time and interval, this reduces the amount of TABIOing needed dramtically. Moved from 15OCT90 this date. 6329. October 17, 1990 IMLOD/UVLOD/FITTP Brian To work around a bug in the Sun f77 compiler, the number -2147483648 was changed to -2147483647 - 1. (In 2s complement most_positive_int < abs(most_negative int)). Moved from 15JAN91 this date. 6330. October 19, 1990 XVSS Chris XVSS may now optionally use the MIT shared memory extension to the X Windows System. This should speed up XVSS on machines with large amounts of physical memory. In order to use the shared memory extension you must have a version of Unix that supports System V shared memory: SunOS supports this as an option (the kernel may have to be rebuilt on some systems). You must build XVSS with SHMOPT set to -DUSE_SHM in the Makefile and add the line: {\tt xvss.useSharedMemory: True} to the X resources database before starting XVSS (the X resources database is normally initialised from {\tt .Xdefaults}). If this line is missing or {\tt xvss.useSharedMemory} is set to {\tt False} shared memory is disabled. This is recommended for machines with smaller memories since the use of large amounts of shared memory may cause severe paging problems. Also replaced the Makefile, which was missing from the .SHR file and made the following minor changes. - Excessive blank space has been removed from the control panel. - XVSS will now read the colour of the graphics plane from the {\tt xvss.graphicsColor} (for Americans who can't spell) resource in the X resource database. The value of {\tt xvss.graphicsColor} may be a named colour from the X colour database or a hexadecimal specification in the normal X11 #RRGGBB format. The default colour is now green (the previous purple colour has a poor contrast against the blue dominated AIPS spectral colourmaps, particularly on slides); users with the old style Sun greyscale monitors that run off the red gun may wish to set {\tt graphicsColour} to {\tt red} or {\tt white}. - XVSS now has OPEN LOOK help available for all components. This is limited at the moment since the current version of XView does not support scrolling help windows but it does provide an on-line reminder of the keyboard accelerator bindings. The help text is provided in a seperate uuencoded file, xvss.uu, because the bloody checkin system on CVAX won't accept the long lines of text. Moved from 15APR91 this date (because of missing Makefile). 6331. October 19, 1990 APGNOT:FILLM.FOR Gareth A bug in routine FLMDAT was causing the CH0 and LINE datasets to differ in the number of visibility data points. This annoyed BPASS considerably. The bug only occurred when the last visibility of a given time-stamp was flagged bad. The logic error had also caused the first good visibility after a flagged visibility in a given time-stamp to be rejected! This had not been noticed. The bug only affected spectral line data, but the missing visibility occurred in continuum data as well. Moved to 15JAN91 and 15APR91 this date. 6332. October 22, 1990 APGNOT:FILLM.FOR Gareth Fixed two tests in the selection criteria that were reversed (qualifier and channel code) and led to the exclusion of the requested data rather than the inclusion. Moved to 15JAN91 and 15APR91 this date. 6333. October 23, 1990 POSSM Phil A minor bug, when the user wished to plot the BP table and specified a source name then that was compared with the sourcename in the table, if they didn't match the data were not selected for plotting. I had missed the case of the source number in the table being 0, in which case it should be plotted. Fixed that. Moved from 15APR91 this date. 6334. October 25, 1990 APLSUB:CATKEY Brian There was an off by one error so that an illegal type (of 0) was being accepted. Moved from 15APR91 this date. 6335. October 25, 1990 APLBERK:ZTQSP2 Brian Use a better heuristic to cut down on the number of aips only processes (something at the end of a line that has capital letters and ends in a number). Moved from 15APR91 this date. 6336. October 26, 1990 SYSIBM APLIBM INCIBM Brian SYSIBM:AIPS - source rather than execute AIPXAS SYSIBM:AIPXAS - allow XAS display on remote hosts. APLIBM:ZACTV9 - ZACTV9 was out and couldn't be checked in earlier. APLIBM:ZFRE2 - AIXized so that free now works. APLIBM:ZTQSP2 - Spy now works APLIBM:ZMOUN2 - Mount issues a rewind and selects the correct device for 1600 or 6250 density. In general this will need to be changed from site to site, althout most machines likely won't have a 1/2" tape drive. INCIBM:DSEL.INC - No longer needed by xlf. INCIBM:DBUF.INC - No longer needed by xlf. Move from 15APR91. 6337. October 29, 1990 APLIBM:ZLPCL2.C Brian Removed this stubbed version so we can print from the ibm. Also removed from 15APR91 this date. 6338. October 31, 1990 DOCTXT:USERNO.LIS Gareth The file USERNO.LIS is used by various site-specific routines to display the local system use. It contains the NRAO user number list. This has not heretofore been part of the standard distribution for reasons that I do not understand. It is now. Moved from 15APR91 this date 6339. November 1, 1990 CLBSNR Phil In the self-cal mode of CALIB, and when no closure errors are requested to be printed, CALIB changes the no closure errors to huge closure errors just to catch ridiculous values. This is OK but the user then gets a message on his/her screen for every integration. Changed this to print out the closure error message header only when there is an error. Moved from 15APR91 this date. 6340. November 1, 1990 APLNOT:NCALC Phil NCALC assumed that routine DNRM2 was still a function but it was changed to a subroutine months ago. I'm surprised no one tripped over this before, I suppose it shows how often the GCON option in CALIB is used. Moved from 15APR91 this date. 6341. November 2, 1990 SPFLG Eric Corrected error which caused an incorrect upper channel number to be used for actual flagging when in the one-channel mode. Moved from Oz to 15JAN91 and 15APR91. 6342. November 2, 1990 SGLOCA Eric Corrected bad code for POPS error handling. Moved from Oz to 15JAN91 and 15APR91. 6343. November 2, 1990 AIPS Eric Corrected to actually abort on an error in INIT and not to say "ABORT!!!" when the Sg LASTEXIT file is too old. Moved from Oz to 15JAN91 and 15APR91. 6344. November 2, 1990 SQASH Phil Apparently SQASH has never worked when a sub-image in the first two coordinates was requested. The code looped from 1 -> NY (where NY = # output pixels in y direction) whereas it should have looped from BLC(2) to TRC(2). The loop in the x-direction was OK. Moved from 15APR91 today. 6345. November 2, 1990 APLPGM:UVDIF.FOR Glen Has been broken for some time for UN-Compressed UVdata. (At least uncompressed UVdata in the DDTs.) Recent fix to allow Compressed data caused it to die for Uncompressed, so Compressed data checks are skipped for uncompressed data. Moved from 15APR91 this date. 6346. November 6, 1990 APGNOT:FITTP Bill J. Subroutine EXTWRT was not moving flagged table rows to the output buffer for transfer to tape. This caused a mismatch between the IAXIS2 value recorded on the tape and what actually was on the tape. This caused PRTTP etc some grief. Now moves all rows to tape. Also corrected a typo: "ISORTORD =" --> "ISORTORD=" to conform to standards. Moved from 15APR91 today. 6347. November 7, 1990 APGNOT:FITTP Bill J. Saving the flagged rows to tape effectively unflags them when the table is read back in again (see 6346). Changed this so that EXTWRT first finds how many unflagged rows there are, writes this number to NAXIS2, then writes only the unflagged rows to the tape. This is at variancei with our usual philosophy since the tape file and the disk file are not copies of each other. This should change when the FITS writer/reader is modified for the VLBA. Moved from 15APR91 today. 6348. November 7, 1990 CLBSNR Phil Moved the printing of the closure phase error header message down 2 lines, in its current position it printed the header > 1 times before printing the closuer errors. Moved from 15APR91 this date. 6349. November 8, 1990 APLNOT:TABFQ, TABNDX Gareth Minor changes in the declarations of EQUIVALENCEd arrays in TABFQ and TABNDX so that they can be used, in debugging, with array bounds checking turned on. Moved to 15APR91 and 15JAN91 this date. 6350. November 9, 1990 APLIBM:ZDELA2 Tom Pauls/Brian Tom Pauls has supplied a nice ZDELA2 (based on sleep/usleep) that actually works. I put in a paranoid error check. Moved from NRL to 15APR91 and 15JAN91 this date. 6351. November 13, 1990 TVFLG.FOR TVFLG.HLP DTVF.INC Brian Modified TVFLG to write/read its grid file on the disk pointed at by IN2DISK. Moved from 15APR91 this date. 6352. November 13, 1990 LISTR Bill J. A bug in the indexing when the data base contained only LL data and solutions for the same caused GAINUV and GAIUV1 subroutines to die uninformatively. Moved from 15APR91 today. 6353. November 14, 1990 APGNOT:FILLM Gareth Two major changes: 1. if the user is silly enough to try to force FILLM to append to a file which has incompatible data, FILLM now stops; 2. all character fields read from the data tape are forced to upper case. Also fixed a pedantic complaint: DOCONCAT is now true only if >0. Moved from 15APR91 this date. 6354. November 14, 1990 MX Glen Slight modification to correct Max, Min UVrange to allow second IF to be included in UV data if first IF fits in allowed range. Moved from 15APR91 6355. November 14, 1990 ALGSTB/UVGRTB/UVGRID/UVAPAS Glen Corrections to routines to correctly handle multi-if data for case when not all map fits in "AP". Essentially requires multi-channel passes for two IFs widely separated. Single IF and case of all map in "AP" are not affected by the changes. Moved from 15APR91 6356. November 15, 1990 APLSUB:IMERGE Brian LISTR was not properly selecting baselines because of a bug in IMERGE.FOR - unique antennas were being rejected as duplicates. Besides being buggy, the code was quadratic although this wasn't necessary. Moved from 15APR91 this date. 6357. November 15, 1990 APGNOT:IMMOD Brian The conversion of degrees to radians for BPA was done in exactly the wrong sense. Moved from 15APR91 this date. 6358. November 15, 1990 CLBSNR Phil Finally fixed up CLBSNR to write the closure error message only onece each time the routine is called. Moved from 15APR91 today. 6359. November 16, 1990 CVEL Phil The AC functions were being shifted in the opposite direction to the XC functions, this didn't matter too much for VLBA stations but did for intercontinental experiments. Moved from 15APR91 today. 6360. November 20, 1990 APGNOT:CLCAL.FOR Bill C. Changed call sequence to SNSMO. Also changed to allow OPCODE='SMOO' for multisource data. Moved from 15APR91 this date. 6361. November 20, 1990 APLNOT:SN2CL.FOR Bill C. Removed the use of the weights in the SN table in determining the interpolation in the SN table. This practice could cause serious problems if the SN table weights (SNRs) were not in fact related to the data quality. An example of this is when, in one scan an antenna is used as the reference antenna in fringe fitting and thus has a low "SNR" = min acceptable + 1.0 and in the next scan was not the reference antenna and has a much higher weight="SNR". Moved from 15APR91 this date. 6362. November 20, 1990 APGNOT:DECOR.FOR Bill C. New task to compute the decorrelation amoung a set of selected correlations. The output values have a real part that is the ratio of the scalar amplitude average of the selected correlations to the vector average. This progam is especially useful for looking for coherence problems in multi IF and/or multi channel data. This task allows application of calibration and/or editing and will write a new multi source file. Moved from 15APR91 this date. Also DECOR.HLP 6363. November 20, 1990 APGNOT:SNCOR.FOR Bill C. New task to apply corrections to SN tables. The corrent version only contains the option to force the residual fringe rates to the average of the selected IF/polarization. This option appears necessary to maintain coherence between the polarizations for MkIII VLBA data. Moved from 15APR91 this date. 6364. November 20, 1990 APLPGM:DIFRL.FOR Bill C. new task to divide the RR correlations by the LL correlations. This task is useful for looking for coherence and amplitude calibration problems between RR and LL correlations. It should be generalized to allow referencing phases to an arbitrary correlation. Moved from 15APR91 this date. 6365. November 20, 1990 QPGNOT: FRING.FOR Bill C. Fixed number of words of pops adverbs to ask GTPARM for (it was about 30 too few). Also fixed indexing and numerous other problem that occured when averaging frequencies in the output of a single source compressed data file. Moved from 15APR91 this date. 6366. November 20, 1990 APGNOT:SHOUV.FOR Bill C. New task. A clone of LISTR that does column listings of closure phases for selected triplets or listings with different IF/channels in different columns. These displays are very useful for detecting and analysing problems in multi IF/channel data sets. Also SHOUV.HLP. Moved to 15 JAN91 this date. 6467. November 21, 1990 TVFIDL Eric Correct to use the correct zoom factor when restarting in an already zoomed mode. Moved from Oz to 15JAN91 and 15APR91. 6468. November 32, 2990 APGNOT:FETCH FETCH.HLP Frank Ghigo/Brian A bug was fixed that only allowed square images to be input. The number of characters allowed in a line was bumped up to 16k to allow larger images to be Fetched. Moved to 15APR91 and 15JAN91 20-Nov-1990. 6469. November 25, 1990 APGNOT:DECOR Bill J. The count of NPARMs was low by 1 which meant that OUTDISK was ignored. Also fixed some typos in the HELP file. Moved from 15APR91 today. 6450. November 27, 1990 UVDIF.FOR Bill C. The data being printed was read from the wrong buffer for compressed data causing very erratic results. Added a second decompression buffer and now data to be printed are always copied to and used from the decompression buffers. Moved to 15JAN91 this date. 6451. November 27, 1990 APLUNIX:ZBYTF2 Brian ZBYTF2 was determining if the input and output arrays were the same incorrectly. Also, the case of no byteflip was not being handled properly (but ZBYTFL short-circuits this anyway). Moved from 15APR91 this date. 6452. November 27, 1990 APLGEN:ZR32RL Brian ZR32RL was doing byteflips but not word flips which caused the Decstation to turn perfectly good IEEE numbers into garbage. The other 32 and 64 bit real conversion routines will need to be similarly changed (32 bit ints are wordflipped!). Moved to 15JAN91 this date. 6453. November 28, 1990 HORUS Robert Braun/Bill C. HOURS was only applying uniform weighting corrections to the first of each group of images. Move from 15APR91 this date. 6454. November 29, 1990 ZR32RL ZRLR32 ZR64RL ZRLR64 Brian None of these worked for the decstation because although the bytes were properly flipped, the words weren't. Moreover the blank and zero checks were wrong as well. Moved to 15JAN91 this date. 6455. November 30, 1990 FILLM.FOR, FILLM.HLP Gareth 1. Fixed bug (denkfehler) introduced when the multi-colored VLA spectral line options were added. Two VLA IF spectra from different polarizations (e.g. A & D) were put into the same AIPS IF even if the frequencies were different. They are now correctly separated into different multi-source files. 2. Allow a timerange as short as 10 seconds for debugging purposes. 3. Added code to deal with VLA holography measurements. Moved to 15APR91 and 15JAN91 this date. 6456. December 1, 1990 INC:DAPC.INC, QYPGNOT/MX.FOR Glen Have increased default AP size from 64K words to 256K words = 1 Megabyte. This allows 4Kx4K MXs to run. Also moved all Multi-IF corrections to NEW. Moved from 15APR91. 6457. December 3, 1990 RUNSYS:DDTLOAD.001 Glen Yet another adjustment in TERSE so that some results are always printed. Moved from 15APR91 this date. 6458. December APLIBM:ZABOR2.C Brian I changed every ZABOR2.C, i.e. in APLUNIX, APLBERK, APLBELL, APLSUN, APLCRI, APLIBM. Instead of handing the signal handler a function we have been typically giving it an uninitialized external function pointer which through a quirk of the linker has normally been resolved correctly - everywhere but on the IBM and Decstation. I am checking everything into TST and the IBM version only into New. Moved APLIBM:ZABOR2.C (Only) from 15APR91 this date. 6459. December4, 1990 APLDEC INCDEC QDEC SYSDEC Brian The routines to make the Decstation work under AIPS. QDEC is currently empty, i.e. pure QPSAP. APLDEC: (Notable: byte problem in ZXASX2, Dec doesn't use the unix convention for recordlength in ZTPOPD.FOR) ZABOR2.C ZCREA2.C ZDCHIC.C ZTPWAD.FOR ZXASX2.C ZABORS.C ZDCHI2.FOR ZTPOPD.FOR ZTTOPN.FOR INCDEC: (1MW=4MB PSAP) DAPC.INC SYSDEC: (The usual) AIPXAS CCOPTS.SH INCS.SH QMSLAND.HDR ZLPCL2.QMS ASSNLOCAL.CSH CDVER.CSH LDOPTS.SH QMSPORT.HDR ASSNLOCAL.SH FCOPTS.SH LIBR.DAT ZLPCL2 Moved to 15APR91 and 15JAN91 this date. 6460. December 7, 1990 APLNOT:GCALC1.FOR Bill C. A typo in the phase only solution portion caused the soluton to become unstable and sometimes give bad solutions. It's pretty amazing that it ever worked at all. This was the L1 solution routine for CALIB. Moved from 15APR91 this date. 6461. December 7, 1990 QPGNOT:CALIB.FOR Bill C. In it's computation of the mean gain modulus for the gain stabilization CALIB was only checking that the weight of a solution was positive. Routine CLSNOU was changing the weights of flagged reference antenna data to positive values so the blanking value was added to the sum. This caused a very large mean gain modulus and thus very small amplitudes of data corrected using this mean gain modulus. Moved from 15APR91 this date. 6462. December 8, 1990 IBLED Phil Migrated the TST version here, list the changes below. More new features and bug fixes. (1) Users had complained about the way IBLED switched between the editing and menu screens, this caused a lot of flashing on the IIS and IVAS and was painfully slow on the Sun's. Changed so that now the menu's appear on the same screen as the editing frame, they are always visible. Makes the process much more efficient. (2) Added a new menu option: SELECT FRAME - which enables the user to select his/her own frame to be edited by positioning the cursor on the total vis. fn at the top and define a start and stop position., this is then displayed in the main editing frame. (3) Added an on-line help facility, when users point to a menu item and hit button D a help message is printed on the terminal. (4) The user can now specify the IN2DISK to which the catalogued uv workfile goes, 0 => lets IBLED choose. (5) Added some checks to prevent users hitting the wrong buttons in some of the editing modes. (6) Discovered that when a user came to re-edit a file using an already existing catalogued file his/her previous editing appeared not to have been applied. Actually the editing info was present in the FG table but when reading a catalogued uv-file the FG table is not applied. So when finishing a session the FC table is copied to the uv work file before being translated to an FG table and deleted from the master file, so when a new session is started the FC table is copied back to the master file and applied before any data is displayed. Also updated the helpfile. (7) IBLED, like other tasks, was not correctly dealing with LL only data when reading the FC table. It now does. ALso when a user asked for all data to one antenna to be flagged that information was written correctly into the FC/FG table but internally the program was unaware of it and so it appeared as if the editing had not been done when a subsequent baseline was displayed. Moved from 15APR91 today. 6463. December 10, 1990 HLPFIL:UVAVG.HLP Bill J. The helpfile was rather uncertain whether or not both 'TB' and 'BT' data could be handled. Moved from 15APR91 today. 6464. December 10, 1990 ZLOCK in APLBERK Chris Added a branch (keyed on the _POSIX_SOURCE feature test macro) to use the POSIX.1 fcntl() interface. These appears to solve most of the phantom lock problems on Convex machines. Moved from 15APR91 this date. 6465. December 14, 1990 FILLM Gareth 1. Changed the method of determining whether antennas are present or not. The previous method was not valid for copies of old (pre-1988) tapes. 2. Two fixes to correct bugs in FQ table handling. One was a "mad bomber" - four arrays whose dimensions were too small. The second was the use of control information from VLA IFs A and B for IFs C and D. This had been necessary in the past. Since C and D are now, in general, independent of A and B, this assumption is not longer valid. It was causing incorrect frequencies, bandwidths, channel separations, etc. for experiments where independent tuning is used. 3. The velocity and reference frequency are now written into the SU table for channel 0 as well as the line data. 4. Changed FLMBC, which determines the VLA band code, to use the most up-to-date frequency limits. Other code to determine the band code has been replaced by a call to this subroutine. 5. Fix to ignore shadowing for raster scan (holography) data. 6. Problems with the VLA multiple-IF spectral line modes. The previous fix (November 30, 1990) went too far - now corrected. Fixes made in 15JAN91 and 15APR91 yesterday 6466. December 17, 1990 INC:PUVD.INC Bill C. Changed MAXCHA (number of channels) to 2050. Moved from 15APR91. 6467. December 17, 1990 CVEL Phil Mark Claussen discovered that CVEL did not work for data whose velocity type was heliocentric. That is not strictly true, the first data point was fine, after which the logical specifying if the data were heliocentric was reset to .false.. Fixed this. Moved from 15APR91 today. 6468. January 22, 1991 $AIPPGM Glen Accidently moved $AIPPGM/AIPLAS.FOR to $APLCONTR, then decided to leave it there. Next moved $AIPPGM/AIPMAN.FOR to $APLCONTR. Now consistent with 15APR91 and 15JUL91. Moved and then restorted to original state several other programs in $AIPPGM. Similar changes in 15APR91 and 15JUL91. 6469. January 25, 1991 $APLPGM/UVDIF.FOR Glen UVDIF.FOR was aborting with a segmentation violation once in 10 executions on the SUN-4 (Mandrill). Never completely understood the problem, but have re-arranged some code and it seems to be OK. Will move to 15APR91 and 15JUL91 6470. February 6, 1991 $APLNOT/SOUELV.FOR Glen SOUELV was returning hour angles with range greater than +/-Pi. This cause ANCAL errors for the 140' telescope. Moved from 15APR91 this date.