----------------------------------------------------------------------- ; 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 the master version of AIPS kept in Charlottesville. An entry 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 15JAN94 *********************************************************************** Changes while 15JAN94 was TST: *********************************************************************** *********************************************************************** 7669. July 9, 1993 $HIST/CHANGE.DOC (this file) Pat Initialized this file as part of the quarterly update. Moved nowhere, but also edited NEW (15JUL93) version accordingly. 7670. July 9, 1993 IBM Fortran problems Bill C. Fixed a couple of minor FOrtran problems which gave trouble on the new IBM compilers: $APGOOP/CL2HF.FOR Arguments to MOD function of different type. $APGOOP/TBDIF.FOR Logical comparison of LOGICALs requires .EQV. or .NEQV. Moved to 15JUL93. 7671. July 9, 1993 GAL.FOR Gustaaf Fixed a couple of minor Fortran problems which gave trouble on the new IBM compilers. Moved to 15JUL93. 7672. July 13, 1993 MINSK, UVMAP Eric Changed MINSK to force the buffer size used to be a power of 2 less than or equal to the buffer size given in the call. Wrote a new $APLSUB/POWER2 to compute this. Changed UVMAP to use the standard UVBFSS buffer size again. This will fix some of the mysterious bugs in MX. Moved to 15JUL93. 7673. July 13, 1993 TPMON Eric Found that several of the error messages for checks on the number of words read/written in the handshakes were wrong. Corrected the error messages. Note that 15JUL93 TPMON does not work with 15OCT92. Moved to 15JUL93. 7674. July 13, 1993 COMAP Eric Changed COMAP.001 to set FACTOR=0 ahead of UVPLT calls and changed UVPLT.HLP to allow this value. UVPLT sets all FACTORs under 0.1 to 1 anyway. Moved to 15JUL93. 7675. July 14, 1993 Various Pat Went through the entire TST hierarchy with an automated checker shell script, comparing rcs text of files with what was in the clear text area. Found and fixed the following discrepancies: $AIPSUB/VERSTR.FOR 6 char string, not 8 $APGOOP/TBDIF.FOR use .NEQV. in place of .NE. $APGOOP/CL2HF.FOR use d.p. variable, not single $APLNOT/DVDMIN.FOR Remove obsolete hack for suns $APLSUB/CATIO.FOR No lowercase c-omment (col.1) $APLSUB/GETRLS.FOR Date (15JAN94) $APLSUB/LMDER.FOR Remove obsolete hack for suns $APLSUB/LMSTR.FOR Remove obsolete hack for suns $INSUNIX/INSTEP1 version date $SYSUNIX/AREAS.CSH,SH Binary areas AIPSBIN/AIPSFITS $SYSUNIX/AIPSPATH.CSH,SH OLD/NEW/TST definitions $SYSUNIX/FDEFAULT.SH SUN4: OPT8,9="-O4 -fast" $YPGNOT/GAL.FOR couple minor changes $SYSSUN/CCOPTS.SH PURGE=FALSE $SYSSUN/COMRPL,COMLNK check if -d $transact.. Some of these will be moved to NEW shortly pending an investigation of the automated checker output there. 7676. July 14, 1993 $YPGNOT/GAL Gustaaf See 7671. Due to disk crash an older version was put in TST. Put the same version in TST as is in NEW. Moved nowhere. 7677. July 14, 1993 $APGNOT/FILLM Gustaaf FILLM was treating the case with a blank field for VLAOBS incorrectly. Fixed this. Moved to 15JUL93. 7678. July 15, 1993 $APGNOT/JMFIT,BLFIT Pat Removed an ancient hack -- declaring an external function a second time as an integer -- that apparently was needed in an unknown version of Sun Fortran (only comment I found: "this is for STUPID SUNS"). No longer needed and broke IBM xlf compiler. Moved to 15JUL93. 7679. July 15, 1993 $YPGM/ISPEC Gustaaf Added the option DOCRT=0 which means NO text output at all (i.e. just the plot). This is an easy way out; it might be better in the future to allow OUTPRINT = 'NONE' for such a case in order to maintain the integrity of the DOCRT definition. Also changed the output format somewhat to prevent overflows. Moved to 15JUL93 (today!). 7680. July 16, 1993 $HLPFIL/ISPEC.HLP Gustaaf Added small improvement to ISPEC.HLP. Moved to 15JUL93. 7681. July 16, 1993 IBLED Eric Changed it to spread the main plot out some when the frame size is small (i.e. small amounts of data or a small selected frame). This will make it easier to see things. Made first frame, next frame, last frame support the selected small frame size as well. Changed IBLED to accept negative values of DPARM(5) as initial guesses for the interval which it will then set. Before, if the actual interval was > 50 seconds the program blew up. Also added test to avoid the blow up when the interval is not determined. Changed the help file for this one. Moved to 15JUL93 which is a bit ballsy. 7682. July 16, 1993 $RUNSYS/VLBA.001 Phil Speeded up the plotting of the visibility considerably by forcing a SPLIT/AVSPC on the data before plotting, so the spectral averaging is not done in VBPLT. Added several new features, the user can now decide whether they want the plots on the printer or the Tek window, they can decide whether they want to zap the intermediate files, they must also now specify the source to be dealt with. Moved nowhere. 7683. July 19, 1993 Shift and rotation bugs. Bill C. Undo some of the damage I did in attempting to fix the problem with rotation in the gridded uv data modeling and fix other problems. There were previously numerous bugs affecting the computation of the Fourier transform of set of CLEAN components with shifts AND rotation on the sky. These only affected the gridded-FFT modeling method (CMETH='GRID'). The recent incarnation of this series of bugs affected both deconvolution (MX, WFCLN) and modeling (CALIB, FRING UVSUB etc.). -GRDCRM: Add all of XPOFF, YPOFF to the model positions; if these values are correct (not previously true) then the positions in the CC table plus XPOFF and YPOFF should correspond to grid cells. Added a test to check that component positions are on grid cells and to complain loudly (but continue) if not. -ALGSUB,ALGSTB: corresponding changes to those in GRDCRM; i.e. shift is that given in DXCG, DYCG. -GRDAT: XPOFF, YPOFF were being rotated by MAPROT with the wrong sign. -MX: in MXIN MAPROT was not being converted into radians before being fed to the sine and cosine routines. This affected the ra and dec shifts being put into the catalog header. The DFT model calculation in CALIB/FRING was not sensitive to this error but the gridded-FFT routines were. $APLNOT/IMCREA,IMHCOR: same problem with argument to sin/cos as MX. $QYPGNOT/UVMAP: same problem with argument to sin/cos as MX. Moved to 15JUL93 this date. 7684. July 19, 1993 MX restore problem Bill C. Equivalences of local arrays with task buffers in MXCCRM caused the restore step to fail when more than 2048 components were restored. This routine now checks that these arrays are not over flowed. Moved to 15JUL93 this date. 7685. July 19, 1993 ALGSUB Bill C. In the test to see if a row was finished, a floating point ".LT." was used rather than a ".LE." causing erratic behavior in those rare cases where there was an EXACT floating point match. Symptoms observed are the introduction of a series of ripples into the residual image. Moved to 15JUL93 this date. 7686. July 19, 1993 $QOOP/QCLEAN.FOR Bill C. Now uses BEAM size in determining how big to make imaging scratch files. WFCLN was having problems with beams larger than could be made in memory. Also fixed CLCCRM so that it wouldn't overflow internal arrays for ra, dec etc of components. Moved to 15JUL93 this date. 7687. July 19, 1993 $APLIBM/ZMOUN2, ZTPOP2 Pat Added code that supports Exabyte 8500 drives via the same DENSITY setting as on suns (22500 => high density 5gb mode, anything else => 2gb low density mode). This includes a "hack" that permits the 22500 density to be used on 8200 drives, as some 3rd party 8500 drives (TTI, for example) will identify themselves as 8200's. Also inserted code that will report the pid of the owner of the lock on the lockfile, i.e. the process number of any process that has the drive already mounted. This change should probably be moved to the other versions of ZMOUN2 someday. Moved to 15JUL93 right now. 7688. July 19, 1993 $APLDEC/ZABOR2.C Pat The DEC Fortran compiler doesn't need a signal declaration for SIGFPE exceptions. Commented out. Moved to 15JUL93 right now. 7689. July 19, 1993 MK3TX Chris Added code to deal with a new oddity in HP text files: negative record lengths. Examination showed that large negative record lengths were actually the true record length (in 16-bit words) minus 32 769 while small negative record lengths denote the end of file. Moved to 15JUL93 on this date. 7690. July 19, 1993 $QYPGNOT/VTESS.FOR Gustaaf VTESS would exit with an arithmeitc exception when NOISE was specified as 0. VTESS now checks NOISE and traps illegal values. Moved nowhere. 7691. July 20, 1993 $QOOP/QCLEAN.FOR Bill C. Routines CLSCSUM and CLNCYC now call routine OUSETF (in $QOOP/QUVUTIL.FOR) to copy total and field summed flux densities into the imaging (DGDS.INC) common so that using zero spacing fluxes will work. CLNINI no longer uses the default box if NBOX=0 but box 1 is actually specified. Moved nowhere 7692. July 20, 1993 $QOOP/QUVUTIL Bill C. Added utility routine OUSETF to set the values of TFLUXG and FLUXG (the total and field summed flux densities) into the imaging common (DGDS.INC). Moved nowhere 7693. July 21, 1993 $APLBERK/ZLOCK.C Pat Forced it to use the "posix" [sic] file locking via routine fcntl() instead of the old flock() call; latter can't tell if a NFS daemon has the file locked for a process on a different host. Moved to 15JUL93 this date. 7694. July 21, 1993 shift bug Eric Old versions of UVMAP and MX put the shift parameters incorrectly into the header but did shift the reference pixel correctly. Unfortunately, modeling routines use the shift parameters. Wrote a new routine to compare the 2 header entries and to correct the header shift parameters for the old bug if that gives agreement between the 2 header representations. It is $APLSUB/FXSHFT.FOR. Changed GRDAT to call it. Moved to 15JUL93 this date. 7695. July 22, 1993 priorities Eric Changed $APLUNIX/ZPRI2.C to use a batch "nice" level of 12 rather than 1. Also changed it to attempt to raise priorities on 'UPPP' although it may never have the privilege to do so. Also UPPP was used only in FPS AP120B code - and DOWN was used more than once only there. The new code will drop the priority by 12 only once even if there are several calls for DOWN, which makes the new large increment ok. Remove $APLBERK/ZPRI2.C as redundant. It is too bad POSIX did not address this issue. Moved nowhere. 7696. July 22, 1993 BPGET Phil Ran into a different problem with BPGET, it failed with DOBAND=2 on kiowa (the big IBM) because the LUN defined for the scratch file operations lay within the range 31 -> 30+NTAPED for kiowa only. Changed to use LUN's 48 and 49. Oh for a routine to figure out available LUN's like OOP does. Moved nowhere. 7697. July 26, 1993 FITTP Eric FITTP could write 2 EOFs on a tape if it did not find the requested files to write since TAPIO opened for write will write an EOF on a CLOS call unless there have been FLSH calls. Changed to do a ZTPCLS and avoid TAPIO in the no-image case. Moved nowhere. 7698. July 29, 1993 BLING Chris Replaced the simple-minded Levenberg-Marquardt minimization procedure with Algorithm 717 from ACM Transactions on Mathematical software. This is much more robust than the original procedure which had been found to suffer from numerical stability problems for some data. The new procedure also calculates the model function values and the objective function in one subroutine call rather than using a separate call for each observation making it more amenable to optimization on vector and pipelined machines. Also added a message that prints the version number of the BS table being created. Moved nowhere. 7699. July 29, 1993 DLAMCH in APLNOT Chris Added the LAPACK auxillary routine DLAMCH and its supporting routines (DLAMC1, DLAMC2, DLAMC3, DLAMC4 and DLAMC5) to APLNOT. DLAMCH calculates various parameters that characterize the host machine's implementation of double-precision floating- point arithmetic. Some of these parameters are used in setting convergence criteria for the new fitting algorithm in BLING. DLAMCH will fail if optimized and has been added to OPTIMIZE.LIS in SYSUNIX with optimization level zero. Moved nowhere. 7700. July 29, 1993 History things Eric Added to the verbs handling history files. Changed PRTHI giving it a new adverb HIEND as the last line of print and respelling adverb PRSTART as HISTART for the starting line of print. Changed HINOTE, moving it to subroutine AU7B and changing its verb number and giving it the DOHIST option to control whether or not HINOTE is prepended to lines from text files. Created verb HITEXT to write selected history lines to a plain text file and verb STALIN to delete (with DOCONFIRM option) a range of lines from the history file. Changed DAPL.INC Add the HIEND adverb AU7 Removed HINOTE code, added HISTART and HIEND to PRTHI. AU7B Created to do HINOTE, HITEXT, STALIN. VERBS Added call to AU7B. VERBSB Added call to AU7B. VERBSC Added commented out call to AU7B. KTRIM Version of JTRIM that kill all after a non-printable character. POPSDAT help file: changed and added verbsm HIEND. HISTART New help file replaces PRSTART. HIEND New adverb help file. PRTHI Changed help for new and altered adverbs. HINOTE Changed help for DOHIST adverb. HITEXT New help file. STALIN New help file. Moved nowhere. 7701. July 30, 1993 OOPS object memory managment Bill C. Revised the Object memory managment to use c routines for object specific storage to allow increased number and size of objects. This involves some shuffling of the commons. $APLGEN/ZOINTD.C - package of Fortran callable c language for memory managment of AIPS object memory. The data structures consist of an object directory and zero or more objects each of which contains a name, class, CATBLK and a linked list of arbitrary information. This new scheme allows up to 1000 objects rather than the older limit of 20. The following table summarizes. Function: Routine name init_dir ZOINTD create_object ZOCROB destroy_object ZODEOB find_object ZOFNOB copy_object ZOCPOB find_llelem ZOFNLE info_llelem ZOINLE store_data ZOSTDT fetch_data ZOFEDT store_catblk ZOSTCT fetch_catblk ZOFECT fetch_name ZOFENM In the above llelem refers to a linked list element. $INC:OBJPARM.INC - Removed parameter MAXOBJ; increased number of possible classes to 10. $INC:CLASSIO.INC - All arrays dimensioned by MAXOBJ now dimensioned by MAXIO. Moved LUNUSE from OBJMNGR private common to CLASSIO.INC. Added Hollerith array HBUFF to use for converting character strings to pass to the c routines. Added character array BUFOBJ to give the object names associated with each buffer. Moved nowhere: $APLOOP:UVDATA.FOR - Changed indexing of arrays OBJLUN and OBJFIN to be by I/O stream rather than by object number. $APLOOP/CXIMAGE.FOR - Removed usage of MAXOBJ; the routines now have to lookup the names of the component parts. $APLOOP/OBJMNGR.FOR - Extensive changes to use new c language routines which can dynamically allocate object memory. Moved nowhere 7702. July 30, 1993 $APLOOP/OBJMNGR.FOR Bill C. Fixed bug in OBCLOS which clobbered memory if the file was already closed. Moved nowhere. 7703. August 2, 1993 $QOOP/QCLEAN.FOR Bill C. CLBHIS now calls ARRCWI to clear any windows set on the beam. Moved nowhere. 7704. August 4, 1993 POPS adverb fetch/set Eric Dropped the awful equivalence file DAPL.INC and created two subroutines instead. ADVERB fetches the value of an adverb of any sort, converting to integer if requested for numeric adverbs and to character for all string variables. ADVRBS is the reverse of ADVERB (with the same calling sequence) to set adverb values. Now the order of adverbs in the K array will be irrelevant, as it already is for task adverbs. Changed: ADVERB (New) get adverb values from K array ADVRBS (New) set adverb values in K array PRTALN Fix bug causing blank pages on 2nd and later print jobs from AIPS. POPSDAT New adverb GREMAIL - deliberately stuck in the middle! GRIPE.HLP Add GREMAIL adverb. GREMAIL New help file. VERBS Added code to protect and restore the POPS environment that ADVERB and ADVRBS might disturb. Removed DAPL.INC add added calls to ADVERB and ADVRBS as needed to HELPS, TBDATA, TBHEAD, VERSTR, AU1, AU1A, AU2, AU2A, AU3, AU3A, AU3B, AU4, AU5, AU5A, AU5B, AU5C, AU5D, AU5E, AU6, AU6A, AU6C, AU6D, AU7, AU7A, AU7B, AU8, AU8A, AU9, AU9A, AU9B, AU9C, AUA, AUB, AUC, AUT and AIPSC. Moved nowhere. 7705. August 5, 1993 AIPSC Eric Corrected the call to PRTMSG so that it would actually give a task name to the subroutine: it was sending ' ' to a routine that declared the variable CHARACTER*6. As a consequence it got garbage. Either one send the expected, or one declares CHARACTER*(*). Anyway, what was now needed was 'AIPSC ' anyway. Had this worked as intended, a batch submission would have cleared the whole message file. Now it clears AIPSC messages or prints them on error. Fixed VERBSC and VERBSB to be like VERBS. Moved nowhere. 7706. August 5, 1993 Message levels Eric Corrected message levels in CALIB, ZLPCL2 (Unix, IBM, CRI versions), MCREAT, and UVCREA. Changed them too in UVMSUB and UVMDIV to be level 3 but to honor MSGSUP = 32000 suppression. Message levels mean: 0 message file only(users input) 1 terminal only (interactive instructions) 2 mildly informative (i.e. progress, ..) 3 4 5 the main answers 6 mild warning - suppressed by MSGSUP 7 serious warning - suppressed by MSGSUP 8 fatal error 9 major screw up 10 Call 911 and the fire department now. These should be honored. We print messages based on this in DDT and tell manufacturers and users that > 5 => error. Moved nowhere. 7707. August 5, 1993 GREP Eric New procedure to search for a string in AIPS files. The grammar is GREP [] [] where the directory is optional with default $AIPS_VERSION if defined or `pwd` if not. The file name is optional and the actual name searched is *. The search is case sensitive. File is in $SYSUNIX. Moved nowhere. 7708. August 6, 1993 FITLD Phil Some fairly major changes to enable the seamless joining of FITS files to an AIPS output file. Also tidied up the logic for the choice of an output file - no doubt this area will need more work. FITLD now applies a global scaling to VLBA data, using a parameter that is passed from the correlator. Task now deals with premature end of files correctly. Added a donconcatanate command to the helpfile so the user has control over concatanation. Moved nowhere. 7709. August 7, 1993 PRTLIN, PRTALN, COMPIL Eric Changed the 2 line printer routines to show the host name and AIPS release on top of each page. With the shared printers it is getting harder to figure out what comes from where. Changed COMPIL: it was not compiling IF (I+J) < X THEN correctly. Since IF is a pseudoverb it had too few symbols to regard the + as an addition. The demand that we be able to omit = signs strikes again! Moved nowhere. 7710. August 9, 1993 MK3IN Bill C. Added a warning message in MK3ANT to tell if the reference day has been changes and to advise the user to do it again with the correct reference day. Also strengthened the warnings aginst an incorrect reference day in MK3IN.HLP. MK3IN can put different reference days in the catalog header (used for the data and tables) and in the antenna table (used to compute observing geometry). This can cause signifigant errors in the computation of the apparent observing geometry for purposes of parallactic angles, elevations etc. Moved nowhere. 7711. August 9, 1993 Misc bugs Eric Changed: UVPGET To test for, fix, and report FREQ = 0. AU7A To actually pick up adverb values and use them in PUTHEAD. PRTLIN To use a correct format for output disk files. It was just hanging. PRTALN To use a correct format for output disk files. It was just hanging. Moved nowhere. 7712. August 9, 1993 $APGNOT/FILLM.FOR Gustaaf FILLM until now used the time at the beginning of a scan to see whether a change in source had taken place. This fails for data from the survey, in which the pointing is changed, but the beginning-of-scan time stays the same. FILLM now uses the fact that the SUBMODE changes when the pointing changes. So a new scan is assumed not only when the beginning-of-scan time changes, but also when the submode changes. Moved nowhere. 7713. August 9, 1993 BLING Chris Added code to calculate and display a measure of the degree of coherence. The measure used is the ratio of the amplitude of the peak in delay-rate space to the weighted sum of the gridded amplitudes. This should be 100% if there is no noise and the rates and delays are constant but will be lower if the rates and delays change. Low coherence measures may indicate that the scan length is too long. The coherence measure is also a good diagnostic of errors in the calculation of orbits for space-based antennae. Moved nowhere. 7714. August 10, 1993 $APLPGM/PRTAB Phil PRTAB dies in the scaling subroutine while reading AN tables if one of the AN rows has a blanked antenna position. Fixed so that if any real or d.p variable is blanked it is not used in the scaling check. The printing of these values was already correct. Moved nowhere. 7715. August 10, 1993 $APGOOP/DTSUM Phil Now will not list # pts on baselines that don't exist, i.e. there is an entry in the AN table but no baselines containing that antenna in the data. Also will now list XC and AC integration times separately, these can be very different. Moved nowhere. 7716. August 10, 1993 UVDIF Eric Changed it to buffer the uv data lines in order to detect when the 2 data sets are identical but the order of 2 vis's has been swapped. This is quite common in the DDT, for example. It now swaps such samples, performs correct tests on flux and uv differences, and prints only those samples which have excess differences. The total number of swaps is reported, but those lines are not printed. This also helps make the peak difference and rms displays more meaningfull. They were badly affected by swaps. Moved nowhere. 7717. August 11, 1993 BLING Chris Improved the initial estimates of the width of the peak used in the least-squares fit to the peak. Also added an absolute convergence criterion to stop the fitting process wasting cycles grubbing about in the noise (also prevents many cases of singular and false convergence. Moved nowhere. 7718. August 12, 1993 AVGIF Phil Took the search for flagged channels out of the averaging loop, this provides a substantial speed increase for this routine. Meant changing the call sequence slightly. Moved nowhere. 7719. August 12, 1993 $APLPGM/AVSPC.FOR Phil Changed call seq to AVGIF. Moved nowhere. 7720. August 12, 1993 $APGNOT/SPLIT.FOR Phil Added the option to average data across IF's as well as the as the, already existing, option to average spectral channels within an IF. HI table reflects the type of averaging, the FQ table is also updated. Changed the .HLP file. Moved nowhere. 7721. August 12, 1993 Abort handling Eric Changed all versions of ZABORS.C. They were simply returning on recursive calls which could lead I would think to an infinite loop of abort signals and returns. Changed to enable default handling of SIGILL and then to signal a SIGILL. This did not fix our problem on the IBMs in which stale load modules (running ones replaced by the midnight job) generate a heavy nfs traffic. To get rid of this, one had to change ZABOR2.C to use default handling of SIGBUS. With the change to ZABORS, the verb ABORTTASK stopped working until the same SIGBUS change was made to ZABOR2. This is very odd since SIGBUS is not involved with the aborttask process and messages put in ZABORS never appeared. Changed ZABOR2.C in $SYSUNIX to split out the routine formerly called signel as an AIPSish one called ZIGNAL.C. Changed ZABOR2.C and $APLBEL/ZDELA2.C to call zignal_ rather than signel. The splitting was needed since the latest HP compiler cannot handle the 2 routines in the same file. Also it is cleaner by our usual standards. Fixed $APLBEL/ZDELA2 to renable trapping of SIGALRM after suspending it for the timer use. Some of the other ZDELA2's also have this bug, but not $APLSUN. Moved nowhere. 7722. August 12, 1993 VTESS Eric VTESS was adding n+1 reads to the status of the beam, but only clearing n at the end. Changed the first CATIO call (the excess read) to set REST instead. Moved nowhere. 7723. August 12, 1993 DDT Eric Changed the old DDTLOAD.001 and DDTEXEC.001 to use the name DOT throughout and put them in $RUNSYS as DOTLOAD.001 and DOTEXEC.001. Changed the DDT procedures: (1) Cleaned up typing, dropping excess semicolons, string concatenations, etc. Made them use more uniform symbology rather than typing string values everywhere. (2) Added concept of FITS disk. The files will have names of the form DDTsttttt where s is S, M, L, H and ttttt is the task name or other file class name. They go in an area which the user specifies with an environmental variable in DDTDISK. And it's all triggered by non-blank DDTDISK. (3) Moved the code for huge over from the DDTHLOAAD.001 file. (4) Switched to floating point FITS files. (5) Widened the BOX in the large test - the whole source was not included in the old box. (6) Changed ASCAL into CALIB with appropriate adverbs and removal of the SN file generated. (7) Corrected PRTTP call - the old NFILES = 4 was still there, causing the first 4 files not to be listed. These used to be text files. (8) Cleaned up help files, removing erroneous tape stuff and adding DDTDISK. (9) Made all ALLDEST calls check all disks. Otherwise, if files are moved they fail to clean up properly. Moved all over for testing. 7724. August 12, 1993 $APGOOP/POLSN.FOR Bill C. Flagged entries in the SN table were treated as errors rather than being ignored. Also modified to average phases by sine and cosine part rather than as phases. Moved nowhere. 7725. August 12, 1993 VERB Eric Changed POPSDAT.HLP to have a new pseudoverb called VERB. Changed POLISH to recognize its number (75) and call PSEUDO. Changed PSEUDO to compile a verb assuming the input format VERB = v1 , = v2 , ... where is a verb name and vi is the number to be assigned to it. Note that the Fortran of AIPS (and AIPSC, AIPSB) has to know about this number in order to call the correct routines. But, this will allow programmers to create new verbs and enter them in NEWPARMS as well as POPSDAT. Then users can use the new verb without having to do a RESTORE 0. Changed VERB.HLP to VERBS.HLP and created a new VERB.HLP to describe this pseudoverb. Another use of this would be to allow synonyms - although one would have to make an inputsd/help file too for a true psynonym. Moved nowhere. 7726. August 13, 1993 GETFQ Phil IF FRQSEL was -1 but an FQ table was present - a common occurence for calibrated data then GETFQ would die even if there was only 1 FQ row. Cannot understand why this was never seen before. Moved nowhere. 7727. August 13, 1993 FITTP Eric Writing VD tables discovered a bug: logical data fell straight into the error trap on short integers in EXTWRT. Fixed. Moved nowhere. 7728. August 16, 1993 Last-minute fixes Pat As usual, the installation testing on the various architectures revealed many minor but irritating bugs. These are the last changes made to 15JUL93 prior to release. Any additional changes will have to be implemented as patches. Summary: $APLDEC/ZDCHI2.FOR Changed SYSVER to 'ULTRIX 4' $APLDEC/ZTPOPD.FOR key off SYSVER instead of SYSTYP $INSUNIX/INSTEP1 Many hacks, mostly for convex/ultrix brain-dead bourne shells, uname/convex. Probably lots more too. $SYSUNIX/TPDEVS.SH Remove "unset" $SYSUNIX/XASERVERS Ditto $SYSUNIX/MAKEAT Extra sed filter to remove "ENV" in reverse "logical" lookup $SYSUNIX/COMRPL Ditto $SYSUNIX/START_AIPS watch out for sh5 shell on ultrix, remove "echo -n". $APLNOT/STEPCH.FOR Make '' empty string into ' '. $APLDEC/ZTXMA2.C Need __POSIX and _XOPEN_SOURCE defined $APLDEC/ZTKILL.C Ditto $APLDEC/ZTACT2.C Ditto $SYSUNIX/SYSETUP Remove "unset", also HP fix $SYSUNIX/FC Ditto $QYPGNOT/MX.FOR Typo in equivalence statement $APLSOL/ZDAOPN.C APLUNIX version causes error in SunOS 5.2 $SYSSOL/LIBR Shared Library fixes $SYSSOL/LDOPTS.SH Various $SYSSOL/CCOPTS.SH OPT0 set to blank, not "-O". $QYPGNOT/VTESS.FOR Wrong INPARM count (was 198, correct=195) $SYSSOL/ZLPCL2 Use "postprint" as a filter, also asa. $APLHP/SIGNEL.C Split from ZABOR2.C $APLSOL/ZTPWAD.FOR EOF code differs. Same as Sun4 I think. There are other modules changed that are not worth mentioning in detail: $SYSSUN/COMRPL, $SYSUNIX/OPTIMIZE.LIS, $SYSUNIX/AIPS.L, $SYSSOL/COMRPL, $SYSUNIX/FDEFAULT.SH, $APLUNIX/ZMSSRV.C, $SYSSOL/CDVER.*, $SYSSOL/LIBR.DAT, $SYSSUN/LDOPTS.SH. Moved to 15JUL93 and to the master 15JUL93 tar tapes today. 7729. August 17, 1993 FRING Chris The arrays WTT and TIME were declared too small in FRNFIT causing variables in COMMON to be trashed if the number of channels and IFs was small. Increased their dimension to XASIZE and removed the equivalences between WTT and TIME and BUFF2. No code used the fact that WTT and TIME were equivalenced and the equivalencing may just have been an attempt to save storage. FRING now has an enormous BSS segment: roll on, Fortran 90! Moved nowhere. 7730. August 17, 1993 $APGNOT/SWPOL.FOR Bill C. Fixed incorrect test for blanked SN and CL table entries. Moved nowhere. 7731. August 17, 1993 $RUNSYS/CROSSPOL.001 Bill C. Now initializes values of STOKES, BIF, EIF, BCHAN and ECHAN before running UVCOP. The last user set values of these parameters were being used which could cause problems. Moved nowhere. 7732. August 18, 1993 $APGOOP/DTSUM Phil Made sure that all valid antenna names are printed, not just the first 3. Moved nowhere. 7733. August 18, 1993 $APGNOT/SHOUV.FOR Bill C. Modified to use the PUVD.INC parameter MAXCIF as the maximum # channels * # IF. The old value was 128 which is rather silly for the VLBA. Moved nowhere. 7734. August 19, 1993 $QPGOOP/WFCLN.FOR Bill C. Fixed a bug with averaging several channels. The data selection was set for a single channel and attempts to image several channels (not IFs) would fail. Also now defaults output name to input name rather than dying. Moved nowhere. 7735. August 19, 1993 $APGNOT/FILLM.FOR Gustaaf Fixed a small bug in FILLM which led to a wrong and confusing message about the shadowing limit, if any shadowed antennae were found. Moved nowhere. 7736. August 19, 1993 $APGNOT/CLCOR.FOR Leonid Fixed a bug at opcode PHAS; fixed a bug at opcode SBDL. Installed a new opcode MBDL which added phase difference between the given IF frequency and reference one corresponded to the given multiple band delay. The value of this delay is added to the corresponded value at the modified CL table. The Help update as well. Moved nowhere. 7737. August 20, 1993 $APLPGM/PRTCC.FOR Gustaaf Fixed default behavior for BCOUNT and ECOUNT. It was impossible to select ONE component when specifying BCOUNT = ECOUNT. Moved nowhere. 7738. August 20, 1993 Z's and Solaris Eric Changed $APLSOL/ZFRE2.C to use the base disk block size when it is not 0 rather than the recommended one since the units of the disk blocks are base blocks. This made NFS systems look 8 times too large. Corrected $APLUNIX ZVTPX2.C and ZVTPX3.C. They increased the send and recieve buffers with the wrong variable which Solaris is suddenly setting to 0. It did not do this earlier. This will have slowed all systems with remote tapes however. Moved to ftp patch area for 15JUL93. 7739. August 23, 1993 $SYSAIPS/AREAS.DAT Pat Added two new areas: APLALPHA (boy is _that_ going to challenge all us dyslexic typists!) and SYSALPHA for the new DEC/Alpha architecture. The former is under the $APLDEC architecture as indications are that most of the DEC/Ultrix routines will work unchanged. Also removed $UPDATE as it is no longer used. Moved nowhere. 7740. August 23, 1993 BLING Chris Added an option that will dump the amplitudes of the transformed search grid into a file readable by AVS. Each scan generates a file of ASCII data and a field descriptor file. The data is dumped in column major order and the field descriptor gives the lengths of the axes. This option has been added because I need to inspect the FT'd grid to diagnose some oddities in the TDRSE data. Moved nowhere. 7741. August 23, 1993 $QOOP/QCLEAN.FOR Bill C. Fixed call argument to OUVIMG in CLNUV; it was passing WORK1 as both of the scratch files. This was causing imaging failure for large images where these scratch files are actually used. CLRHIS was setting the BLC, TRC after opening the input array which caused the previous window to be used. CLNINI now checks for a box with all zeroes but a nonzero nboxes and fills in the default box. Moved nowhere. 7742. August 23, 1993 $QNOT/MAKMAP.FOR Bill C. MAKMAP was using UVGET buffer UBUFF as an I/O buffer which wasn't large enough for it to do a 4096x4096 image. It now uses polarization calibration array POLCAL instead. Moved nowhere. 7743. August 23, 1993 $QPGOOP/SCMAP.FOR Bill C. For compressed input data in the self calibration routine the Stokes' parameter is set to ' ' when uncompressing the data before the model division. Now defaults output name to input name rather than dying. Default maximum Imaging uv range now set to 1.0E15 rather than 0; this was causing some problems in the routine deciding the uv range to give full weight in the self calibration solution. Moved nowhere. 7744. August 24, 1993 $QOOP/QCLEAN.FOR Bill C. Modified CLRGRD and CLREST to do a disk based computation of the FT of the restoring Gaussian times the CLEAN components. This may use work object WORK2 for a temporary accumulation. This change should allow restoration of any size image. Moved nowhere. 7745. August 24, 1993 VLBA Antenna Characteristics tables Phil Added the number of digitizer levels for an observation to the antenna characteristcs table - seemed the logical place to put it. Moved nowhere. 7746. August 24, 1993 $APGNOT/FITLD.FOR Phil Deals properly with the new Ant Char table. Also now switches the location of the Rl/LR pixels when the baseline direction has to be switched. Moved the loading up of the output uv buffer into a separate subroutine in order to maintain much better control over it. Moved nowhere. 7747. August 24, 1993 $APLOOP/ARRAY.FOR Bill C. In ARRIO a check in now made to insure that array VALIDX is not overflowed. This case arrises in the FT of 4096 images where the row length is 8192. The fix caused the max and min to be only for the first half of the image but under the circumstances these are meaningless numbers. Moved nowhere. 7748. August 24, 1993 $APGNOT/DBCON.FOR Gustaaf DBCON, when used on multisource data with FG tables, would apply NO correction for the difference in days to the timerange in the second FG table when copying it to the output table. The result of this was that the whole flagging of the combined data set was screwed up. This problem must have been around for a long time and it is hard to see why it has not been discovered before. Maybe most users use SPLIT first, or use DBCON before any flagging. The bug itself was trivial: DBCON was looking for a column called 'TIMERANG' to apply the shift to, but did not find any in the FG table. The FG table has a column called 'TIME RANGE', so I changed DBCON accordingly. Moved nowhere, but will move it to the ftp area. 7749. August 24, 1993 Unused variables Eric Unused variables cause some compilers to abort (!!) when there are too many of them. Therefore, they will NO LONGER BE TOLERATED! Corrected: $AIPSUB/ KWIKHD, OFMLST, OFMMOD, SGLAST $APLSUB/ AVGCHN, AVFCIF, CATIO, IDWSRC, LSTHDR, MAPCOP, MSKIP, PLNGET, TABIO $APLGEN/ ZCREAT, ZMSGCL, ZR64RL, ZRLR64, ZTAPE, ZV20CL, ZV20XF $YSS/ YIMGIO $YSUB/ GPOS $APLNOT/ ACINI, ACIO, ACTAB There are many more to go. Moved nowhere. 7750. August 25, 1993 $SYSUNIX/OPTIMIZE.LIS Pat Somehow the settings here for APLNOT, APLOOP and APGOOP were set to always produce debug code. On most systems this inhibited any optimization from being done. Removed the debug setting. Also commented out the debug/OPT0 setting for APCLN. Moved nowhere yet (probably to HP test site at JPL soon). 7751. August 25, 1993 INPUTS Eric Changed PSEUDO to limit string declarations to 132 characters and changed AU1A to display the inputs for long strings. It was just putting the string value in and beyond the end of MSGTXT. This caused AIPS to abort or quit at some later time. The POPS routines set the offset for TYPE 9 (SUBSTR strings) as 100 * offset. This is too small for 132-character strings. Changed DSMS.INC to make X large enough for 132 chars. Changed ASSGN, CONCAT, EQUIV, and KWICK to store this as 1024 * offset. Requires recompiling a lot. Changed POPSGN to limit to 132 characters also. Fixed KWICK so that PRINT and TYPOE verbs do the right sort of things with long strings - rather than blowing aips out of the water. Moved nowhere. 7752. August 25, 1993 UVTINI Eric Added a blank keyword to the DATA statement for keywords since Phil put back the DUVV.INC with a larger number of allowed keywords. It would not compile otherwise. This is why I do not care for these includes that define purely local variables. Moved nowhere. 7753. August 27, 1993 POSSM Phil When SOLINT > 0, POSSM would stop if the last frame on the plot page was empty, fixed so it carries on now until the end of the timerange is hit. Moved nowhere. 7754. August 27, 1993 $APLOOP/HISTORY.FOR Bill C. OHLIST was not writing logical adverbs to the history file. Moved nowhere. 7755. August 27, 1993 IMLOD Eric Fixed a bug in the routine that reads the oldest type of extension table we once used. There may be more - I do not remember the format. Anyway, the pointer was not initialized and so could fail when IMLOD read a second file. Phil has been asked to make the same change in FITLD, which he has out. Moved nowhere. 7756. August 29, 1993 VLBA interface Phil Tidied up all the includes that define the VLBA tables. These now include all the DATA statements so I get rid of the problems of declared variables not being used. This causes errors on some compilers - moronic if you ask me, but that's life. Moved nowhere. 7757. August 29, 1993 VLBA interface routines Phil Changed the call sequences of all routines that deal with the VLBA tables. The variables are now all passed through commons. Moved nowhere. 7758. August 29, 1993 AIPSPLIT.TAR in APLCONTR Chris AIPSPLIT.TAR is a tar archive containing two perl scripts and their man pages. aipsplit tears an AIPS program apart so that each program unit (local include, main program, subroutine, function or block data subprogram) is in a separate file and creates a makefile and (provided that the program area is specified) .OPT file for the dismantled program. aipsjoin will reverse the process or build a single AIPS program from any collection of .FOR and .INC file. [APLCONTR isn't really an appropriate area for this but we don't have anywhere else for odds and ends that might be useful but don't really want to maintain.] Moved nowhere. 7759. August 30, 1993 UVFIT.HLP Phil Minor change, helpfile claimed 5000 vis was the limit for fitting, actually it is 25000. Moved nowhere. 7760. August 30, 1993 VBPLT.FOR Leonid The program did not work when both DOCALIB and DOAVG existed and number of polarization NUMPOL is >1. The reason was that NUMPOL was destorted at subroutines SCALVB and PLTVB in DOAVG option because the identificator NUMPOL was used as temporal one for subroutine AVGCIF. So NUMPOL has been changed for random identificator NUMTEM at the two places. 7761. August 30, 1993 CGASET.FOR Leonid Specification statement - SAVE DOXDR was after DATA statement. 7762. August 31, 1993 AP2SIZ.FOR Bill C. This routine returns a power of 2 AP size for use in the FFT routines. It was only checking up to 2**12 so the FFT routines would do a disk based 4K FFT no matter how much memory was available. The test now goes up to 2**20. Moved nowhere. 7763. August 31, 1993 FILLM.FOR Gustaaf Allowing survey pointings to be treated as separate scans caused the same thing to happen for holography data. This is fixed now. Also, the user can now choose to have the TY table filled with Tsys values. This used to be the case only for solar observations. Moved nowhere. 7764. September 2, 1993 AP memory Eric Changed: $QPSAP/QINIT to use min (actual 2nd AP, setpar value) $QVEX/QINIT to use min (actual 2nd AP, setpar value) $Q3090/QINIT to use min (actual 2nd AP, setpar value) $INC/DAPC.INC to use 1.25 Mword AP, to get parameters from another include, to drop RWORK. $INC/DAPS.INC New: AP size parameters $INCVEX/DAPS.INC New: AP size parameters Convex version $INCCRI/DAPS.INC New: AP size parameters Cray version $INCALN/DAPS.INC New: AP size parameters Alliant version $INCSUN/DAPC.INC removed $INCIBM/DAPC.INC removed $INC3090/DAPC.INC removed $INCDEC/DAPC.INC removed $QCRI/QXFOUR.FOR to remove erroneous reference to RWORK in precursor comments. $Q3090/QCFFT.FOR to make RWORK local. $QNOT/MAKMAP removed unused reference to KAPWRD. SETPAR to use DAPS.INC to limit the 2nd AP memory and to clarify the first AP memory meaning. SETSP to use DAPS.INC to limit the 2nd AP memory and to clarify the first AP memory meaning. Wholely retyped to standards. FILAIP to use DAPS.INC parameters in recreating SP file. FILAI2 to use DAPS.INC parameters in creating SP file - limiting users' answer to KAPWRD. Moved nowhere. 7765. September 2, 1993 AP memory usage Eric Changed $QNOT routines ALGSUB, ALGSTB, UVUNIF and CCSGRD to use the full AP memory (usually huge) rather than just the basic 64K-word area. Also changed UVMAP similarly. Now only the cleans in APCLN, SDCLN, and MX refer simply to KAPWRD. The limited size in ALGSUB and ALGSTB caused the routines to use a very small support size for the gridded modeling in large cases (NY = 4096) leading to rather inaccurate results (the source was not fully subtracted despite having appropriate components). Moved the ALGSUB, ALGSTB changes to the patch area. 7766. September 2, 1993 MX Eric Corrected the setting of image type in the header: residual was called components. Changed the handling of a restart to make sure that the image is correctly computed. It was making all sorts of assumptions about the image otherwise and really messing up. Moved to the patch area. 7767. September 2, 1993 DDT Eric Changed $RUNSYS DDTLOAD.001 to react to the changes I made earlier to the TYPE verb. Other procs may wish to do similar things. Moved nowhere. 7768. September 3, 1993 $QPGOOP/WFCLN.FOR Bill C. Fixed several problems involved with selecting data by frequency or IF. The initial open of the uv data no longer includes selection by frequency, IF or Stokes' so the underlying data structure will be visible. Moved nowhere. 7769. September 6 1993 Beam Gaussian fitting Bill C. Dan Briggs has pointed out that the beam Gaussian fitting used in AIPS (and copied in SDE) sometimes gives incorrect (but semi-plausible) results. I've corrected the algorithm and implemented it in new routine FITBM which is now called from the places which previously performed thie operation: $QYPGM/APCLN $QYPGNOT/MX, SDCLN $APLNOT/BMSHP (used by $QYPGNOT/UTESS, VTESS) $QOOP/QCLEAN (Used by $QPGOOP/SCMAP, WFCLN) Moved nowhere. 7770. September 6, 1993 $QOOP/QCLEAN.FOR Bill C. Fixed CLNINI not to write to the uvdata object for an image clean. Moved nowhere. 7771. Sepember 6, 1993 $APLNOT/FITBM.FOR Bill C. New routine. This routine does the least squares portion of the beam Gaussian fitting after the matrices have been computed. This routine will replace the many separate versions of this, previously incorrect, logic. 7772. September 6, 1993 $APLNOT/SYMINV.FOR Bill C. New routine to invert symmetric, real matrices. Moved nowhere. 7773. September 7, 1993 $APGNOT/FILLM.FOR Gustaaf Added option in FILLM where the user can ask for ALL correlations in holography mode, not just the ones involving the reference antenna. Since FILLM's CPARMs amd DPARMs were already taken, I made CPARM(2) a bitmap, so that CPARM(2) serves three distinct functions. Though this is not a trivial thing for a user to handle, I prefer this over adding yet another set of parameters. After all, all three functionalities will only be used by very experienced users. Moved nowhere. 7774. September 8, 1993 Location common Eric Changed the location common (DLOC.INC) so that all variables are dimensioned by parameter NUMLOC. Added LOCNUM as a pointer to the currently active subscript in these arrays. Changed $AIPSUB: AU5 Set LOCNUM=1, use subscripts in display. AU6D Set LOCNUM=1, use subscripts in display. AU9 Set LOCNUM=1, use subscripts in display. AU9A Set LOCNUM=1, use subscripts in display. AU9B Drop the include. AU9C Drop the include. $APLOOP POSITION Dropped the copying back and forth of coordinates, and changed it to use LOCNUM 1 and 2. $APLSUB: CHECKL New routine to check LOCNUM value. CHNTIC Call CHECKL, use subscript. DIRCOS Use subscript. DIRDEC Use subscript. DIRRA Use subscript. FNDX Use subscript. FNDY Use subscript. LABINI Call CHECKL, use subscript. LMPIX Use subscript. Corrected GROSS error. It used the maximum of a small positive number and the header axis increment to get the pixel number. But increments are often negative. It is best to test for zero divide, not play this lazy game. Affects IMLOD, FITLD. NEWPOS Use subscript. SETLOC Call CHECKL, use subscript. SLBINI Call CHECKL, use subscript. STCLOS Removed the include. STOPEN Removed the include. STPUT Removed the include. TICCOR use the subscript. TICINC Call CHECKL, use the subscript. TKLAB Call CHECKL, use the subscript. TKTICS use the subscript. XYPIX use the subscript. XYVAL use the subscript. $YSUB: CLAB1 Call CHECKL, use the subscript. CLAB2 Call CHECKL, use the subscript. COMLAB Call CHECKL, use the subscript. CTICS use the subscript. IAXIS1 Call CHECKL, use the subscript. ITICS use the subscript. $YPGM CNTR Set LOCNUM=1, drop call to SETLOC before STARPL. GREYS Set LOCNUM = 1 and 2 as needed, dropped the location common save areas and the call to setloc before STARPL, use the subscript. IMEAN Set LOCNUM values and use subscript. IMVIM Set and use the subscript. ISPEC Set and use the subscript. PCNTR Set and use the subscript. PROFL Set and use the subscript. SL2PL Set and use the subscript. TAPLT Set and use the subscript. TVHUI Removed unused location common TVRGB Removed unused location common UVHGM Set and use the subscript. $YPGNOT CLPLT Set and use the subscript. DFTPL Set and use the subscript. FRPLT Set and use the subscript, drop many unused variables. GAL Set and use the subscript. Corrected the use of list-directed read on a CHARACTER variable. ANSI, in their wisdom, made that illegal. GAPLT Set and use the subscript, remove and rarrange unused variables, correct unreachable couint of good points. GNPLT Set and use the subscript, dropped unused variables. IBLED Set and use the subscript. IRING Set and use the subscript. KNTR Set and use the subscript, drop saving common and use subscript 2 instead. Drop excess call to SETLOC ahead of the STARPL calls. PFPL1 Set and use the subscript. PFPL2 Set and use the subscript. PFPL3 Set and use the subscript. PLCUB Set and use the subscript. Cleaned up long lines and some of the other non-standard typing. PLROW Set and use the subscript, drop unused variables. POSSM Set and use the subscript, drop garbage code that pretended to save the common but never used it, and drop several included DLOC that were not used. SNPLT Set and use the subscript. UVPLT Set and use the subscript. VBPLT Set and use the subscript. $APLPGM PBCOR Set and use the subscript. PRTIM Set and use the subscript, fix handling of PRGNAM. PRTPL Set and use the subscript, remove 2 unused I's. QMSPL Set and use the subscript. SLFIT Set the subscript, remove 1 unused variable. STFND Set the subscript. TKPL Set and use the subscript, remove unused I. XBASL Set the subscript. XGAUS Set the subscript, removed several unused variables, fixed numerous improper (1) buffer declarations. XPLOT Set the subscript, removed 1 unused variable. $APGNOT BAKLD Removed unused include DLOC. CANPL Set and use the subscript, remove unused variable. GSCAT Set and use the subscript. GSTAR Remoce unused include DLOC. HGEOM Set and use 2 values of the subscript rather than swapping 1 Kbyte of common data back and forth all the time. IMFIT Set and use the subscript, remove a few undeclareds. IMLOD Set and use the subscript. JMFIT Set and use the subscript, remove a few undeclareds. LTESS Include DLOC.INC and set LOCNUM = 1 for SETLOC. LWPLA Set and use the subscript. Cleaned up the typing --- why do some people go to considerable trouble to undo standards? Removed all formats given as character strings in the WRITE statement. That is not within the ANSI standard. MOMFT Set and use the subscript. PGEOM Set and use the subscript. REGRD Set and use the subscript. Removed all formats given as character strings in the WRITE statement. That is not within the ANSI standard. SELSD Set the subscript. STARS Drop the unused location common. STESS Include DLOC.INC and set LOCNUM = 1 for SETLOC. STRAN Drop the unused location common. TXPL Set and use the subscript, drop unused variables. UVPRM This one exemplifies lazy coding. The user will never figure out what it is supposed to do from the help file or the code comments. It actually does a useful MSGWRT print out - not the plots it claims. Removed the large section of unused code setting up the location common and cleaned up the typing and message levels some. UVPRT Remove unused location common. $QPGNOT FFT Set and use subscript, remove an unused variable. IM2UV $QYPGNOT UTESS Include DLOC.INC and set LOCNUM = 1 for SETLOC. VTESS Include DLOC.INC and set LOCNUM = 1 for SETLOC. $APLCONTR CCNTR Set and use subscript. Fixed handling of STFACTOR for star plotting, allowing negatives. It will not plot CC files since the subroutine CCPL is missing. SAD Set and use the subscript. Fixed numerous undeclared variables, unused variables, illegal functions (SIND, COSD), long lines, and what not. Maybe it will work. It compiles and links anyway. STARC Set and use subscript, delete unused variables. Moved nowhere. 7775. September 8, 1993 CCNTR Eric Added Jaffe's task to plot CC components along with contours. It required cleaning up, updated call sequences, etc. I made it do multiple symbols like star plotting as well. Cleaned up STARPL and separated out subroutine PLOTWO as well. Moved nowhere. 7776. September 9, 1993 BLING Chris Fixed a bug in BLING which caused it to look for an FQ table whether or not an IF axis was present. Moved nowhere. 7777. September 12, 1993 $APGNOT/FILLM Gustaaf In the case of online filling, FILLM will now check changes in TELL every 10 seconds. For off-line filling this time remains 1 minute. Moved nowhere. 7778. September 13, 1993 $APGNOT/STARS.FOR Gustaaf Under some circumstances STARS would create RA entries which are negative, actually 360 degrees off. Though this does not harm the further handling of ST files, I fixed this for reasons of esthetics. Moved nowhere. 7779. September 16, 1993 $APLOOP/ARRAY.FOR Bill C. Modified ARRSAD to accept a complex scalar to add to a complex array. Also added function ARRSMU to multiply an array by a scalar; again the scalar is assumed to be the same type as the array. Moved nowhere. 7780. September 16, 1993 $APLSUB/CMPCFL.FOR Bill C. New routine that compresses unused entries out of the lists in DFIL.INC commons of the current scratch files and files with marked catalog flags. Moved nowhere. 7781. September 16, 1993 DFIL.INC usage in OOPS Bill C. Both the image and uv data class destructors (IMGZAP, OUVZAP) failed to compress scratch file entries from the DFIL.INC commons when deleting scratch objects. The corresponding entries were marked as unused but were not compressed from the list. Since new entries are always added at the end of the list it could eventually be exhausted. Now OUVZAP and IMGZAP call routine CMPCFL to compress these commons. Also OUCFCL and CFLCLR use CMPCFL to compress these arrays when clearing a catalog status flag. Moved nowhere. 7782. September 16, 1993 $APLOOP/UVUTIL.FOR Bill C. Modified UVCLIP to allow optional specification of the Stokes' value to be clipped. This means it now clips by I rather than individual RR and LL as before. As before, clipping by I will also flag other polarizations if I is flagged. Moved nowhere. 7783. September 16, 1993 $QOOP/QINTER Bill C. Added function IMGIAE to interpolate an al-el image at the positions of the output image. Special projections are used in the az-el image. Moved nowhere. 7784. September 17, 1993 OOP uvdata reference freq. Bill C. There were problems in the OOPS uv data handling routines with the handling of frequencies in the presence of IF selection, especially if IF 1 were not included. The frequencies passed to derived uv data objects were the frequency of the first actual IF passed whereas the u,v and ws and the FQ table entries were unchanged. The effect of this is that any derived images have incorrect frequencies and the u, v and w terms are not scaled appropriately causing a scale error in the derived image. Also in the translation of a uv descriptor to an image descriptor, the correct reference frequency was not being used. There is a new uvdata descriptor (UV_DESC) class public member, REFFREQ, available after the object is opened which gives the actual reference frequency. The UV_DESC copy functions and frequency functions in UVUTIL.FOR now use this value as the output uvdata frequency. Affected functions: $APLOOP/UVDATA.FOR - OUVOPN $APLOOP/UVDESC.FOR - UVDCOP, UVDSCP $APLOOP/UVITIL.FOR - UVREFQ, UVFRQS Moved nowhere. 7785. September 17, 1993 $SYSLOCAL/NOSHARE.LIS Dave A. Added OHGEO.FOR to NOSHARE.LIS. Moved nowhere. 7786. September 17, 1993 $YPGM/GREYS and CNTR Pat/Dave Shone The size of the INPARM (or whatever) common was incorrectly set too large in these programs. On Solaris 2, this causes other stuff to be clobbered, usually the stuff set by ZDCHIC. I wonder how many more of these are lurking... Moved to anonymous FTP patch area today. 7787. September 18, 1993 $APLOOP/IMAGE.FOR Bill C. Modified function IMCLNX to copy catalog header keywords; this should ensure that they get copied under most circumstances. IMCLNX now also opens the input and output images to be sure they're fully instantiated. Moved nowhere. 7788. September 20, 1993 $APLOOP/IMAGE.FOR Bill C. IMGCLN now calls IMCLNX to copy image and array descriptors and catalog header keywords. Moved nowhere. 7789. September 20, 1993 Imaging Eric I have stumbled across several effects in our imaging process. One is the fact that we have been using a very coarse histogram to determine which pixels should be searched for components in each major cycle of Clean. This led in some cases to very few components being found per cycle and a very long clean. A second is that the user has had no control over how many pixels could go into that search in each cycle. The current limit (effectively 20050 most of the time for minpatch 51) is too large for some cases and too small for others. A third is the way in which we handle data near the edges of the UV plane while gridding. There is currently a "guard band" of 6 uv pixels to prevent mathematical errors in convoluttion. But that is not enough to prevent very high frequency problems from occurring in the clean. The fixes are: (1) a finer histogram, (2) give the user adverb control over the clean memory, (3) give the user adverb control over the guard band and display information about the correct cell size to use to avoid data loss and a warning to watch for high frequency effects. Changed are: POPSDAT.HLP Add adverbs GUARD, MAXPIXEL. NEWPARMS.001 Add adverbs GUARD, MAXPIXEL. UVMAP.HLP Add GUARD. HORUS.HLP Add GUARD. MX.HLP Add GUARD, MAXPIXEL. APCLN.HLP Add MAXPIXEL. SDCLN.HLP Add MAXPIXEL. APCLN.FOR Double buffer size, change AP size computation to full size with MAXPIXEL added to limit number of pixels, increase histogram size, remove task name from history message displays. SDCLN.FOR Double buffer size, change AP size computation to full size with MAXPIXEL added to limit number of pixels, increase histogram size. MX.FOR change AP size computation to full size with MAXPIXEL added to limit number of pixels, increase histogram size . Added GUARD, including telling the user when data exceeds the size of the guarded uv plane. Set DMPR.INC parms to 0 to avoid disturbing UVGRID. UVMAP.FOR Added GUARD and information to the user about samples outside a sensible guard and the actual guard. DMPR.INC Add parameters for gridding to return the actual and included maximum U and V and to set a guard band. GRDMEM.FOR Added handling of guard band and return of maximum U and V seen and used. UVGRID.FOR Added handling of guard band and return of maximum U and V seen and used. UVGRTB.FOR Added handling of guard band and return of maximum U and V seen and used. UVTBGD.FOR Added handling of guard band and return of maximum U and V seen and used. UVTBUN.FOR Revise edge test to use MAX minus guard bands. UVTBUW.FOR Revise edge test to use MAX minus guard bands. UVUNIF.FOR Previously paid no attention to points off the edge. Added test so that only weights inside the guard bands are altered. I don't know what the AP makes of points off the edge! IMCREA.FOR Removed a TAB character. Moved nowhere. 7790. September 20, 1993 remove unused variables Eric Continuing to winnow unused variables: $QYPGNOT/ MX $AIPPGM/ FILAIP $APLNOT/ CLBSNR, CRDSET, DATGET, EULROT, FFTV2, FGREFM, MAKGAU, NDXINI, PLNPUT, PTIME, R3DTAB, SELSMG, SOUFIL, SOUINI, SUSEL, TABAXI, TABFQ, TABHDK, TABREQ, TABSRT, TPIOHD, XEULER Also: CRDSET Removed lower-case code in DATA, formats as character strings in a WRITE statement. Both are not allowed in strict ANSI. FFTV2 Fixed typo leading to double declaration of variable, corrected code that was branching into a DO loop (having branched out) just to avoid IF-THEN-ELSE, corrected an un-numbered DO (END DO) structure. Moved nowhere. 7791. September 20, 1993 $SYSUNIX/PROG Pat I got fed up using the brute force WHICH - it uses "find" - to figure out which program directory a given program is in. This shell script does it much faster as it only checks the directories that have "/PGM" in the name - only 20. Moved nowhere. 7792. September 20, 1993 $HLPFIL/POPSDAT.HLP Gustaaf Added the various online FILLM commands as AIPS procedures containing TELL commands. Moved nowhere. 7793. September 20, 1993 $QOOP/QUVUTIL.FOR Bill C. Added optional "GUARDBND" to input uvdata object to specify the guardband in the uv data grid. Moved nowhere. 7794. September 20, 1993 OOP CLEAN class Bill C. Added class public members SCALERES and BMSCLSZ to enable and control scaling of the residuals to the same units as the restored components before the restoration. The scaling is by the ratio of beam areas. The dirty beam area is determined inside a specified box. Add routine CLSCRS to implement scaling and added to both image and uv CLEANs. Changes in both $APLOOP/CLEAN.FOR and $QOOP/QCLEAN.FOR. Added class public member MAXNRES to allow control of the maximum number of residuals in major cycles. This parameter can be adjusted to the size of the problem. The default is the older, hard value of 20,000. Modified CLEAN residual histogram to 8192 levels and the beam histogram to 1030 allowing much larger beam patches. Moved nowhere 7795. September 20, 1993 $APLOOP/UVUTIL.FOR Bill C. Modified UVCOPY and UV2SCR to select by optional U and V limits. Moved nowhere. 7796. September 20, 1993 $QPGOOP/WFCLN.FOR Bill C. Added option to scale the residuals by the ratio of the restoring beam area to the dirty beam area before restoring components. This allows adjusting the residuals to the same physical units as the restored components. The parameters of this correction are the size of the box about the dirty beam to be used to determine its area. Modified to exclude uv data outside of a region of uv space defined by the "guardband" inside the uv grids used. Also fixed a bug which caused the scaling of U,v and w to be incorrect. Implemented new adverbs MAXPIXEL and GUARD to allow user specification of the maximum number of residuals to consider in a CLEAN cycle and the guardband around the outside of the uv data grid. Also added documentation in WFCLN.HLP. Moved nowhere. 7797. September 21, 1993 LISTR.FOR Bill C. Increased PARAMETER MAXSOU in SCANUV (OPTY='SCAN') from 500 to 2500. This allows LISTR to do something sensible for a larger number of sources. Moved nowhere. 7798. September 21, 1993 SETJY.FOR Bill C. Increased the maximum number of sources in the SU table from 300 to 2500. Also parameterized this value and added a test that the arrys were not being overrun. In its previous condition, SETJY would trash memory and sink without a trace if there were too many sources in the SU table. Moved nowhere. 7799. September 21, 1993 FITLD.FOR Phil Several changes: (1) Added the multidimensional DLOC.INC stuff (2) Added the ability to read VLBA archive tapes, these are different from distribution tapes in that they have SQL headers on them, so FITLD needs to know how to skip these headers, this is a temporary fix. It won't last here long. (3) Tidied up the default naming of output files when reading VLBA data. (4) Tidied up all the VLBA type includes and all the VLBA -> AIPS table translation routines. (5) Added the ability to shift the correlator model polynomial from the time tags in the IM table to those in the CL table. (6) Switched the sign of the phases in the data to make the maps come out the right way round, PHASE = -PHASE strikes again! Moved nowhere. 7800. September 21 $APGNOT/FILLM.FOR Gustaaf Solved bug in online FILLM. FILLM stops automatically at the end of the current scan when a) a 'quit' is given, or b) the specified VLAOBS changes. But when a 'quit' was given during the *last* scan of an observation, after which the change in VLAOBS would have stopped FILLM anyway, FILLM became unstoppable! This is fixed now. It took Miller Goss to find this one; others had never bothered to do a 'quit' during the last scan. Moved nowhere. 7801. September 22, 1993 $SYSUNIX/AIPSASSN.(C)SH Dave Added the ONLINE logical to the AOC-specific part of the script. Moved nowhere. 7802. September 26, 1993 PRTLIN, PRTALN Eric Changed the two print subroutines so that the header with DOCRT true has the system and release names also. That info is always useful when there are so many machines/releases that can talk to the same screen. Moved nowhere. 7803. September 26, 1993 routines in wrong place Eric If a subroutine calls a routine in $QNOT, then it must be in $QSUB or $QNOT. If it is placed in $APLNOT, then when a task is linked only with $APLNOT, there will be an unresolved reference. Some computers care about this with shared libraries (Solaris guys - our future). Moved CALRES, MAKCVM, DSKFFT from $APLNOT to $QNOT. IMOPEN, PLOTDA, MODPLT, PLTSET were moved to $YNOT. $APLGEN/ZVTVRX was renamed VTVRX and moved to $YSUB. FITLD could not run otherwise. Moved nowhere. 7804. September 28, 1993 Source list size Eric Changed the size of the source list - formerly 500 most places to 2500 for the sky survey. Changed PUVD.INC moved parameters from DSEL to here and increased the source list size. DSEL.INC removed parameters to PUVD. DTVF.INC add include PUVD, set source list by XSTBSZ. DSPF.INC add include PUVD, set source list by XSTBSZ. IBLED changed hard-coded 500's to XSTBSZ. TVFLG changed hard-coded 500's to XSTBSZ. SPFLG changed hard-coded 500's to XSTBSZ. SPLIT changed hard-coded 30 (!!) to XSTBSZ. SOUNDX changed 30 to XSTBSZ, put error message on list overflow, cleanedf up commenting (IT DOES NOT IMPROVE CLARITY TO PUT AN INANAE COMMENT ON EVERY LINE OF CODE. CVEL Removed a locally declared parameter of the same name and usage (but different value) as used by DSEL, but now moved to PUVD. Moved nowhere. 7805. September 28, 1993 SAD Eric/Walter Walter submitted a very non-standard but lovely little task called SAD for search and destroy. It finds "islands" of flux and fits gaussians to them. I cleaned it up, fixed a small bug which caused it to find sources twice sometimes, and added some other features. It writes a MF model fits file, it is restartable, it uses the line printer, it writes a residual file, etc. etc. It works in only 1 plane at a time. Moved to Leiden today, nowhere else. 7806. September 28, 1993 CCNTR Eric Upgraded to plot MF files too and to use a range of symbols to indicate source strength as well (for CC or MF, not ST). Moved nowhere. 7807. September 28, 1993 $APGNOT/FILLM Gustaaf FILLM, when reading in survey data, did not recognize pointings to the same position as the same source, and wrote an extra entry in the SU table. This is fixed now. FILLM, when appending to an existing file, will now also tell the user what file it is appending to. Moved nowhere. 7808. September 29, 1993 $QPGOOP/OHGEO Bill C. Made it less picky about the error code from OGET call to read catalog header for 'PARANGLE' before deciding to use the value passed as an adverb. OGET was not properly interpreting the return code from CATKEY. Moved nowhere. 7809. September 29, 1993 $APLOOP/OBJMNGR.FOR Bill C. In several places the error codes from CATKEY were returned undigested. Now if the return code from CATKEY is greater than 20 (Keyword(s) not found) a return code of 1 is substituted. This is the usual, not found, code under the circumstances. Moved nowhere. 7810. September 29, 1993 DISPTV Eric The $QNOT routine DISPTV calls both Y and Q routines. It must either appear in a $QYSUB which we have never made, or be copied into each task that needs it. Moved it into VTESS and UTESS - all others that use a DISPTV have their own already. Moved to Solaris to try it. 7811. September 30, 1993 LIBR.DAT (Sun4, Sol) Pat/Eric I have been told that AIPNOT tasks should never reference AIPSUB, as AIPSUB routines can call TV routines. So I have simplified the LIBR.DAT entries -- for Sun and Solaris only -- in an attempt to see if this solves certain other problems with shared libraries on Solaris. Required movements to make this so: AIPINI from AIPSUB to APLSUB to AIPPGM: AIPSC, BATER, BSTRT1, QMNGR, GRIPR, POPSGN, RECAT Removed unused variables from BATER, GRIPR, POPSGN in $AIPPGM FIXCAT, SHOPB, and UPDAT in $AIPNOT - removed unusable subroutines in the last. Moved nowhere. 7812 October 1, 1993 UVGET.FOR Bill C. Added hook in UVGET to allow specifying the initial visibility number to read for a single source (only) file. Changed the following: - DSEL.INC: Added integer variable INITVS as the first vis. no. - SELINI.FOR: Initialize INITVS to 1 - INDXIN.FOR: Set FSTVIS to INITVS in the absence of an NX table - UVGET.FOR: Added documentation describing INITVS. Also modified SELINI to set the default FRQSEL to +1 rather than -1 as UVGET now refuses -1. Moved nowhere. 7813. October 1, 1993 OOP vis. offset on read Bill C. Modified the uvdata class to take advantage of recent changes in UVGET. These changes allow reads to start at an arbitrary visibility number in a single source file. Changed are: UVDATA.FOR (OUVOPN): Set DSEL.INC value INITVS to VISOFF before opening for read. UVDATA.FOR (OUVINI): Set DSEL.INC value INITVS to VISOFF before reinit for read. UVDESC.FOR: Updated class documentation. Moved nowhere. 7814. October 1, 1993 FITLD Phil Fixed a subtle bug; a local common had the same name as a subroutine called by MERGE, took a while to find this one. Added the source and antenna mapping functions, now when a file from tape is being concatanated onto a file on disc the source and antenna tables from the existing file are read into arrays and a mapping function set up so that the incoming sources/antennas get the correct numbers assigned to them. We have to do this because the VLBA correlator can change source/antenna numbers during the correlation of an experiment. Fixed up the shifting of the correlator polynomials onto the CL table, as it was it would stop doing it after the first antenna. Moved nowhere. 7815. October 5, 1993 $APLOOP/PRINTER Bill C. New OOP class to allow access to printers. Standard headers are generated for uvdata of image objects. All of the features of the standard AIPS routine PRTLIN are available. Moved nowhere 7816. October 5, 1993 $APLOOP/GENERIC.FOR Bill C. Added new printer class to most generic routines. Moved nowhere. 7817. October 5, 1993 $APGOOP/RFI.FOR Bill C. New task to look for interference based on high values of amplitude RMS. Also RFI.HLP. Moved nowhere. 7818. October 5, 1993 OPTIMIZE.LIS Phil Added FITLD to the list, it can only be optimized to level 1 on IBMs - otherwise the result is a segmentation violation. Moved nowhere. 7819. October 6, 1993 $APGOOP/RFI.FOR Bill C. Fixed indexing bug in RFIRMS; the bug had no visible symptoms. Changed criterion for picking "worst" correlator; it now must be from the set of correlators marked as bad. Moved nowhere. 7820. October 6, 1993 EPOSWTCH Eric Moved subroutines CRDSET, CRDTRN, EULROT, MATCH STEPCH, and XEULER form $APLNOT to $APLSUB, cleaning them up on the way. These are used for position precesion and conversion. Also cleaned up TXTLEN which was already in APLSUB. Changed AU7 to do switch between 1950 and 2000 coordinates as verb EPOSWTCH. Changed POPSGN.HLP and NEWPARMS.001 to declare the new verb - note that one does not have to do a RESTORE 0 to get this verb, just a RUN NEWPARMS. Also a help file. Moved nowhere. 7821. October 7, 1993 FFT Eric Raised all the buffers. There was one real buffer of 4098 which had to equivalence to a complex buffer of NY+1. This did odd things with NY=4096! Moved nowhere. 7822. October 8, 1993 EPOSW Eric Changed EPOSWTCH to average the output of the CRD... routines and a new subroutine called B2JPOS. The latter attempts to copy the method from The Astronomical Almanac, but has to adopt a new inverse matrix since we always force the proper motion terms to 0. This is all dubious, but gets us closer to the position pairs in the VLA Calibrator book. Moved nowhere. 7823. October 11, 1993 $APLOOP/ARRAY.FOR Bill C. ARRIO was not keeping track of the maximum and minimum pixel values for array planes on the 4th and higher axes. Moved nowhere. 7824. October 12, 1993 $APLNOT/TABHDR.FOR Phil Now recognizes the TAPE STATISTICS (VT) table from the VLBA correlator. John Benson finds this very useful for assessing data quality. Moved nowhere. 7825. October 12, 1993 $APGNOT/GETJY.FOR Gustaaf In correcting the SN table, GETJY did not properly take undefined values into account, with distastrous consequences. Added test for undefined. Moved nowhere. 7826. October 13, 1993 $APGOOP/HF2SV.FOR Pat The Alpha compiler barfed at several non-sequitors in this source code. I removed the keyword FILEOPT='EOF' (sun extension) and replaced it with code to scan the file to find the EOF instead. Also corrected a format statement that had a variable name instead of a format specifier by mistake, and generally cleaned it up (no names > 6 chars, no lowercase, move data statements after all declarations). There are still three problems but they're not fatal: INTEGER*2 definitions, equivalencing character and integer variables, and a GOTO 303 that comes from outside an IF-block. Someday when we've got time (yeah right). Moved nowhere. 7827. October 14, 1993 $QNOT/UVTBUW.FOR Bill C. If the "AP" size was too small to handle the entire grid this routine was attempting to increase the size of the summing box. However, it wasn't doing this correctly and it was giving a totally bogus and misleading message. This routine is used in tasks WFCLN and SCMAP. Moved nowhere. 7828. October 15, 1993 $QOOP/QCLEAN.FOR Bill C. CLNUV was failing to check all fields for previous components on restart. Moved nowhere. 7829. October 15, 1993 $APGNOT/GETJY.FOR Gustaaf See entry 7825. The fix was a little bit too quick, and I had overlooked that the SN table was not updated at all. Now it is. Mover nowhere. 7830. October 19, 1993 $APLNOT/R3DTAB.FOR Phil If a table was of zero size (can happen with VLBA tables) then R3DTAB would read the next tape record anyway which would screw up the reading of subsequent tables. Fixed. Moved nowhere. 7831. October 19, 1993 $APGNOT/SNSMO.FOR Craig/Phil Fixed a bad typo (scaled delays by 10**19 instead of 10**-9), and also adjusted the rate limits when editing rates. Moved nowhere. 7832. October 19, 1993 $YPGNOT/CLPLT & VBPLT Phil This must be the oldest bug that I personally have fixed in AIPS. The routine the calculated the GAST at a given UT for the model plots had an error of ~2 minutes. This showed up when plotting the model vs the data for the VLBA first science data in which we have very high SNR and 2 sources beating against each other producing well defined minima. This must have been there since 1982 when Lars first wrote VBPLT. Moved nowhere. 7833. October 19, 1993 $YPGNOT/VBPLT Leonid The new option 'AUTO' has been installed. Now if you give the value 'auto' to the new input parameter OPTYPE the program plots the points correspoded to auto correlation. The value '' or 'cros' are reserved for cross correlation. Several bugs were discovered at the original program and fixed: 1. The ploting of weight did not work at regim of time averaging of the signal. 2. Timerange limiting did not work at regim of time averaging of the signal. 3. The regim of time averaging of the signal did not work correctly. The number of plotting points was wrong. Moved nowhere. 7834. October 19, 1993 FITTP Eric Changed FITTP, FITTP.HLP, NEWPARMS.001, and POPSDAT.HLP to reset the default for BLOCKING to 10. Moved nowhere. 7835. October 19, 1993 AIPS Eric Corrected gross bugs dealing with substrings being assigned to variables. If the substring had no characters in it and was not at the beginning of the string, then it would copy blanks all over places beyond the end of the string. Ever wonder about adverb values like 1.356316E-19 ? Fixed ASSGN.FOR. Moved nowhere. 7836. October 20, 1993 RUN Eric Fixed one oddity - RUN claimed to find run files locally but then used only ones in RUNFIL or RUNSYS. Corrected HELPS and $APLUNIX/ZDIR to support and allow (resp) local versions. Moved nowhere. 7837. October 20, 1993 $INC/PUVD.INC Dave A. Changed the bseline table size XBTBSZ from 3500 to 7500. After talking to Phil, it became apparent that 3500 would not handle the number of antennas, polarizations, and IFs currently available. Moved nowhere. 7838. October 20, 1993 $APGNOT/DBCON.FOR Gustaaf Put the new DBCON into the system. It allows DOPOS(1,1)=2, in which case position shifts are calculated frequency depen- dent. This makes DBCON more exact, but also slower. For the moment, the new feature only works for input data sets with identical sort order. In other cases the user is warned. Moved nowhere. 7839. October 22, 1993 CLRTEMP Eric New basic verb CLRTEMP to clear the temporary literal area while running procedures. Changed POPSDAT.HLP and KWICK to add the verb. Changed OERROR, KWICK, GETFLD, CONCAT, and HELPS to make a more explicit error message when the temporary area runs out of room. Created new help file. Changed AIPS, AIPSB, AIPSC, and POPSGN to set up the pointers of the temp area. Changed PSEUDO and NEWPARMS.001 to allow and define the new verb. Moved nowhere. 7840. October 22, 1993 RUN again Eric ZTXMAT was automatically concatenating RUNSYS to whatever RUNFIL or VERSION area was being searched for RUN files. This caused a situation in which the same name file was in RUNSYS and RUNFIL (with different user numbers) to refuse to procede. Changed it to do the concatenation only on standard file areas and only when no match is found. Moved nowhere. 7841. October 26, 1993 RFI.HLP Bill C. Now the default solution interval is given. Also a couple of other minor wording changes. Moved nowhere. 7842. October 27, 1993 $APLOOP/PRINTER.FOR Bill C. Added automatic header for tables. Moved nowhere. 7843. November 1, 1993 $APLOOP/UVUTIL.FOR Bill C. Fixed indexing bug in UVTAVG addressing baselines in the accumulation array. Moved nowhere. 7844. November 2, 1993 $QPGOOP/WFCLN.FOR Bill C. The default value for NFIELD was not being enforced; a value of 0 was causing bizzare behavior. Moved nowhere. 7845. November 2, 1993 $APGNOT/SPLIT.FOR Gustaaf When splitting a database with more than 1 IF, the contribution from the SU table to the frequency was removed for all IF's larger than BIF! Fixed. Also, added the option to store the full source name in two keywords. There is only room for the first eight characters in the header. The new option makes it possible to store the full source names for single source data. Moved nowhere. 7846. November 4, 1993 $APLIBM/ZTPOP2.C Pat The command to find out tape attributes (lsattr) did not include the path, so it was failing if the user did not have /etc in the default PATH. Added /etc/ to the string. Moved nowhere. 7847. November 4, 1993 $APLPGM/PRTAB.FOR Dave A. Added the TY table to TIMTAB, and increased NUMTIM to 8. This allows DOHMS=1 to work for TY tables. Moved nowhere. 7848. November 5, 1993 $APGNOT/DBCON.FOR Phil When DBCON changed source numbers on the incoming file it printed out a message each time it did it. Now prints out the message the first time only for each change. Moved nowhere. 7849. November 5, 1993 $APGNOT/USUBA.FOR Phil Added the option to move data to other subarrays by specifying by source name. Also updated .HLP file. Moved nowhere. 7850. October 28-November 5, 1993 tapes Eric Changes for survey and corrections PRTTP Added PRTLEV=-3 to go to text file with 1 short line per file . Changed no movement NFILES from > 1000 to > 10000. Also help file. FITTP Changed it to write begin ends of file only when not doing DOEOT. Added a BAKF (1) at the very end just before the close. This gets the IBM OS to skip writing an extra unwanted EOF at the end of the tape. AU4 Fixed message format. ZTAPE Moved ZTAPE.FOR from $APLBELL to $APLGEN replacing old one with one that sends the count of operations to ZTAP2. Moved the old $APLGEN/ZTAPE.FOR to $APLALLN, $APLCVEX, $APLVAX, $APLDEC, and $APL3090 since I have yet to upgrade those areas. Changed the following to take the count as an argument. Also SUN/ZTAP2.C Changed WEOF to leave out the BAKF and ADVF it did for safety's sake, doing only necessary movements at BOT or on 1/2-inch tapes. Changed BEGW to skip the WEOF at BOT. IBM/ZTAP2.C as for SUN. The following already took count as an argument, so changed HP/ZTAP2.C Changed WEOF to leave out the BAKF/ADVF when possible, changed BEGW to write EOFs only when not at beginning of tape and to leave out the BAKF/ADVF when possible, changed AEOI to use a hardware advance to EOD on DAT tapes even though that means not knowing anything about file numbers, corrected and enhanced debug message sections, corrected ADVF following BAKFs since ierr was 4 when that op had to be done. Moved nowhere. 7851. November 7, 1993 Make room for more basic verbs - we used up all 100 that we once thought were more than enough forever. Changed: POPSDAT.HLP Changed verbs numbers, adding 100 and then rearranging to leave room for growth. Change basic pseudoverb numbers too and rearrange to put similar things together. VERBS Changed branch table to match POPSDAT. VERBSB Changed branch table to match POPSDAT. VERBSC Changed branch table to match POPSDAT. SGLOCA Changed SG version number to 14. AU2A Changed SG version number to 14. AIPS Changed range of verb numbers for call to VERBS. AIPSB Changed range of verb numbers for call to VERBSB. AIPSC Changed range of verb numbers for call to VERBSC. POPSGN Changed range of verb numbers for call to VERBS. NEWPARMS.001 Dropped almost everything since RESTORE 0 will be required and POPSDAT has all that was in NEWPARMS. POLISH Move error testing to PSEUDO, STORES, ..., clean up pseudoverb numbering. PSEUDO Reorder pseudoverbs to match new, coherent numbering and clean up the code a bit. EDITOR As PSEUDO. STORES As PSEUDO. HELPS As PSEUDO, new verb numbers. KWICK Changed verb numbers for WHILE, THEN, ELSE. VERB.HLP Add option to give synonym rather than number. PSEUDOVB.HLP New - like VERB for pseudoverbs. DSMS.HLP Changed variable TYPE to SYTYPE since it seems to be causing the Sun compiler some problem. Changed the following for the spelling of TYPE: ADVERB ADVRBS ASSGN AU1A AU2 COMPIL CONCAT EQUIV GETFLD LTSTOR SUBS SYMBOL AU2A EDITOR HELPS KWICK OERROR POLISH PSEUDO STORES VERBS VERBSB VERBSC Moved nowhere. 7852. November 7, 1993 XAS Eric Changed init.c to check mutiple visuals to pick out the default visual. HP offers 2 pseudocolor visuals in the basic window and the second one is the default! All vislist references became subscripted by visnum. Also fixed a large number of long lines. Moved from the HP 755 test machine. 7853. November 7, 1993 APROPOS Eric New pseudoverb implemented in HELPS, AU1A, POPSDAT.HLP. It reports all 1-line help summaries containg the word or words specified by the user. It is case insensitive and minimum matching and suprisingly fast. Created $AIPNOT programs HLPA to create the needed list of words from the full help area and HLPB to update the list for a specific help file. Crated a help file for APROPOS and the special list help file it uses which is called LSAPROPO.HLP. Unfortunately, putbck is refusing to allow new files to be putbck so I cannot putbck the following: $HLPFIL/CLRTEMP.HLP $HLPFIL/PSEUDOVB.HLP $HLPFIL/LSAPROPO.HLP $HLPFIL/APROPOS.HLP $AIPNOT/HLPA.FOR $AIPNOT/HLPB.FOR Moved nowhere - and the midnight job won't help on these either! 7854. November 8, 1993 BLING Chris If the region around the peak in delay-rate space was flat then BLING would calculate an initial set of model parameters that would cause underflow and loss-of-precision problems on machines that used IEEE-754 arithmetic. These would usually lead to floating-point exceptions during the least-squares fit to the peak and may have been responsible for the wild peak positions that have occaisionally been reported. BLING now uses a fixed oversampling factor of 2 in the FFT to prevent the excessive interpolation that would sometimes result from adjusting the oversampling to fill the AP. In addition BLING checks that all values neighbouring the peak are at least 10% less than the peak before going ahead with the least-squares fit; this should also catch the case where the peak value in the search window is the shoulder of a maximum outside the search window. Moved nowhere. 7855. November 8, 1993 APROPOS Eric Various due to APROPOS: HLPA Changed to include the category words and reject duplicate entries. HLPB As HLPA. AU1A Changed to sort the matches in alphabetical order. Before they were alphabetized by the actual word matched and secondly by the symbol name. BLING.HLP Add VLBI to category and fringe fit to 1-liner. FRING.HLP Add VLBI to category and fringe fit to 1-liner. HF2SV.HLP Fix up pre-cursor remarks which were left unedited by the task's author. UVNOV Task and help removed from system. UVNOU New task and help = the actual UVNOV task, but with the comments and name fixed to describe what it actually does. There will be a lot more of this. Moved nowhere. 7856. November 9, 1993 $APGNOT/FITLD.FOR Phil Fixed up the subroutine that recognizes global keywords so that it initializes them before reading values from the incoming table and also it will now recognize the RDATE keyword that has been added to all VLBA tables. Moved nowhere. 7857. November 10, 1993 $APGNOT/FITLD.FOR Phil FITLD now deals correctly with data that cross the midnight boundary. Did not think I would have to deal with this but the correlator starts a new job at midnight (most times) and since it has no global view of an observation it doesn't know that this is day n>1 of a processing run. Still need some work on some of the tables in this area. FITLD also does the source mapping correctly now, I had all the logic in place but forgot to update the source random parameter. Will now print a visible warning if output file has zero visibilities. Moved nowhere. 7858. November 11, 1993 $YPGNOT/IBLED.FOR Phil Minor logic error resulted in the user not being able to switch between 1-IF and ALL-IF flagging, same for ALL-CHANNEL and 1-CHANNEL. Moved nowhere. 7859. November 11, 1993 2 new adverbs Phil To keep Craig happy - he can walk to my door in < 10 seconds - I have added 2 new adverbs. SYMBOL to allow the user to choose a plotting symbol, will be in VBPLT/CLPLT soon; and PLVER to replace INVERS in those routines that display PL files. Will modify the relevant HLPFIL's soon. Added .HLP files for both adverbs. Moved nowhere. 7860. November 11, 1993 $DOCTXT/PRIMARY.TXT Pat This is a simple file with a list of (currently 8) primary keywords that are allowed in help files. The checkout system will shortly be modified to check this file (and a SECONDARY.TXT that doesn't exist yet) to make sure the help file conforms. If not, the putbck will be rejected. Moved nowhere. 7861. November 11, 1993 Gustaaf's complaints Eric Changed CNTR, PCNTR, GREYS to put back the extra SETLOC call before the call to STARPL. It is needed because the location common is buggered for odd axis labeling types (i.e., relative axes). Changed HLPA, HLPB to add tasks, verbs, ... (i.e. the plurals) to the stop lists. Changed APROPOS.HLP to clarify where the words are found and what is displayed. Moved nowhere. 7862. November 11, 1993 Bill's complaints Eric Fixed the AVEOT message in AVTP to give the correct information on no error rather than incorrect info only on error. Moved nowhere. 7863. November 12, 1993 IBLED Eric Fix some fucked up logic with the handling of single channel and/or single IF flags. Someone changed .OR. to .AND. and then that same someone I suspect decided months later to remove the flags due to this error. Restored to their original state. Moved nowhere. 7864. November 12, 1993 TCOPY Bill C. New task. This utility copies selected contents from one tape to another. It's especially designed to copy many tapes to one tape which is difficult to do properly with unix utilities. Disk FITS files are also supported. Also TCOPY.HLP. Moved nowhere. 7865. November 12, 1993 CLCAL Craig/Phil Putbck a modified version of CLCAL for Craig. This now has the option of fitting a polynomial to the SN values of the residual rate and writing the fit to the CL table. This compiled and worked on the Sun but apparently does not compile on the IBM/Alpha machines. Eric is fixing up the illegal Fortran. Moved nowhere. 7866. November 12, 1993 VBPLT Phil Added the ability to choose the plotting symbol. Only has two choices for now, a '+' or a '|'. Moved nowhere. 7867. November 12, 1993 clean up the shit again Eric CLCAL was just putbck with a bit of stupid and wrong code. It does the silly thing of trying to pass a subroutine name as a call argument. Technically this is legal, but practically it usually causes trouble. In this case however, the routine was declared to be CHARACTER*(*) which it is not. EXTERNAL would probably work, but why play games passing a fixed single argument? Changed it to do the obvious. Corrected numerous long lines of comments and other minor bits. Moved nowhere. 7868. November 12, 1993 Plot display tasks Phil Changed INVERS to PLVER in the plot display tasks TXPL, TKPL, TVPL, QMSPL, PRTPL (is this ever used now?), LWPLA and CANPL. This will remove the annoyances caused by the global nature of INVERS. Moved nowhere. 7869. November 15, 1993 NCCBOX, CCBOX Phil 2 new adverbs for Craig for his forthcoming CC editing task. He wanted this because he uses the new task in a procedure with MX and does not want to override the NBOXES and BOX adverbs. Moved nowhere. 7870. November 15, 1993 New procedure: $RUNSYS/BOX2CC Craig/Phil A new procedure that will convert up to NBOXES BOX values in pixels to CCBOX values in arcseconds. Moved nowhere. 7871. November 15, 1993 $APLSUB/FPARSE.FOR Phil Recognize keyword NAXIS, and set logical if GROUP structure present. Needed for PRTTP. Moved nowhere. 7872. November 15, 1993 $APLPGM/PRTTP.FOR Phil Added the ability ro recognize VLBA binary FITS files and print out the pertinent information. Also fixed up the counting of the # bytes on the tape, PRTTP was missing most of the bytes in the extension tables. Moved nowhere. 7873. November 16, 1993 TCOPY Bill C. Fixed format problem, missing comma. Moved nowhere. 7874. November 16, 1993 HELPS Eric APROPOS moved things around a bit in HELPS.FOR and I forgot to change the part of the code that covers up for the user saying GO to a verb. Fixed it up. Moved nowhere. 7875. November 16, 1993 SLAMCH and DLAMCH in APLNOT Chris Added the single-precision numerical inquiry routine SLAMCH to the system. Also merged the DLAMCn routines back into DLAMCH.FOR. DLAMCH and SLAMCH are now identical to the LAPACK versions apart from the addition of AIPS header comments. Optimization is disabled for both. Moved nowhere. 7876. November 16, 1993 BLAPP Chris BLAPP will now abort if no solutions are found instead of trying to apply an empty SN table (this was the cause of the bogus sort error messages from SN2CL). An incredibly stupid error in calculating the time range has been fixed and the SLAMCH code has been moved out to the AIPS library (see above). Two new adverbs have been added. ANTWT allows the user to give greater weight to data from baselines to certain telescopes and OPCODE controls whether the solutions are applied to the CL table automatically. Moved nowhere. 7877. November 18, 1993 FWRITE Eric Changes to call sequences require programmers to use GREP to find all uses of the routine. FPARSE was changed, but FWRITE was not. So TPHEAD was bust. Fixed. Moved nowhere. 7878. November 18, 1993 UPDCONTROL Pat Modified the copying of the control files so that success of copying is dictated by (a) a successful return status from the actual copy, and (b) a checksum comparison. This involved modifying the checkout software so that it always puts the checksum of the *.UPD files in *.UPD.SUM in addition to the change to UPDCONTROL itself. The latter now tries to copy the .SUM file over if the .UPD file made it, gets the numeric checksum from it, and compares it with the actual checksum of the copy. Moved to various MNJ locations today. 7879. November 18, 1993 BLING Chris Improved the calculation of the degree of coherence. This should now represent the ratio of the vector sum of the amplitudes over the scalar sum of the amplitudes in a scan after fringing and should not go over 100% for strong fringes any more. This change will result in the degree of coherence being about 1/3 of what it was before in most cases. Moved nowhere. 7880. November 22, 1993 New adverb: CLINT Phil No this is not pronounced after whathisname Westwood, but is 'CL int(erval)'. A new adverb for FITLD to enable the user to specify the CL interval the desire. Should maybe also move into INDXR, FILLM, MK3IN and VLBIN. Moved nowhere. 7881. November 22, 1993 New task: CCEDT Craig/Phil Craig has written a task within the OOP package to edit CC extensions based on a cutoff flux and/or their positions. Uses NCCBOX and CCBOX to specify boxes. There is a procedure called BOX2CC in $RUNSYS to convert between a BOX specified in pixels to the CCBOX in arcsec needed by CCEDT. Moved nowhere. 7882. November 22, 1993 BLING Chris Modified the calculation of the coherence to compensate for the fact that the grid is not completely filled due to over- sampling and flagged data and implemented a new method for calculating the SNR. The new method assumes that all power that is not in the immediate neighbourhood of the peak is noise and is evenly distributed over the grid. This will overestimate the noise level if the peak is blurred but this is probably desirable. Moved nowhere. 7883. November 23, 1993 $QNOT/VSCDFT.FOR Bill C. Implemented correction for shift between tangent point and observing position for bandwidth smearing correction. This correction was formerly stubbed and this change should allow UVADC to to this correctly. Also fixed bug in telling the fraction of the work done. Moved nowhere. 7884. November 23, 1993 $QPSAP/QPTADC.FOR Bill C. Fixed a number of fundamental problems in the table lookup. These problems affected relatively large bandwidth smearing corrections. Affected task = UVADC. Moved nowhere. 7885. November 23, 1993 $QPGNOT/CVEL.FOR Phil CVEL now determines if the data comes from the VLBA correlator by examining the ANAME keyword in the AN(tenna) table. If it is 'VLBA' then CVEL assumes that all data have been fringe-rotated to the same reference point (the centre of the Earth) and will determine the velocity correction accordingly. If the VLBA string is missing then CVEL assumes a MkII or MkIII correlator and, although these correlators do the fringe-rotation differently, CVEL knows how to deal with them. A message is printed if CVEL makes the VLBA assumption. Also added text to the helpfile explaining this. Moved nowhere. 7886. November 23, 1993 $YPGM/BLANK.FOR Gustaaf Small change allows user to specify - if DPARM(5) > 0 - by what value the blanks have to be replaced (formerly only 0). Moved nowhere. 7887. November 23, 1993 $HLPFIL/IRING.HLP Gustaaf Changed one-liner since APROPOS RING did not list IRING ! Moved nowhere. 7888. November 27, 1993 AP size Eric If the AP size exceeds some limit, AP addresses may no longer be stored accurately in the floating-point AP core memory. For IEEE single precision, this limit occurs at 2**24, or 16 Mwords. The reason to store integers as floating arose because we wished to roll AP memory as fp buffers to disk and back again in order to share a true AP among users. Problems with integers arose when the host and ap did not have the same fp format so that a translation had to occur. It is our guess that this situation will never again arise. The combination of external device with different format and multiple simultaneous users will not occur. Thus, we can set the roll times to infinity even if the other 2 conditions arise. Therefore, we can now store integers correctly in the AP memory. To do this, changed: QCLNSZ $QPSAP - changed to full AP memory. QCLNSZ $QSUN, QIBM - removed. QRLSE $QVEX - removed. It was overriding KAP2WD unnecessarily. QMAKMS Fixed it to set in 1's for when vector A < C. Before it set in 0 for A < C and A >= C. Fortunately, nothing seems to use this routine. QMULCL Word 0 for each field descriptor changed to integer. APCLN Changed APSIZ to integer, fixed 1 roller call to roll more than the 64K words. Changed to load an integer for the field address. SDCLN As APCLN. MX Changed to enter integer addresses for each field in the AP memory. QCLEAN $QOOP - Changed to enter integer addresses for each field in the AP memory. UVTBUN Changed to send integer addresses into the weight summing and fetching routines. UVTBUW Changed to send integer addresses into the weight summing and fetching routines. QIHIST $QPSAP - new routine does histogram of integers with no scaling. Complains if the range is wrong. QVINDE $QPSAP - changed to use integer-valued gather vector. QWTCNT $QPSAP - changed to use integer-valued gather vector. QINTP $QPSAP - cleaned up index computation hoping to avoid floating point. QINT $QPSAP - as QINTP and changed call sequence to get integer pointers to arrays rather than integer pointers to fp addresses of those arrays. QUVIN $QPSAP - changed to pass the integer array pointers to QINT. GRDSUB Corrected test on rotation which was performed before the rotation parameters were set. The consequence of the bug was the use of ALGSTB under all circumstances, never ALGSUB. ALGSUB Changed to pass in integers to the AP for addresses. ALGSTB Changed to pass in integers to the AP for addresses. Q1GRD $QPSAP - changed to pass integer addresses in to QGRD4 rather than via indirection (and fp) QGRD4 $QPSAP - changed call sequence to expect actual AP addresses rather than fp-valued locations of the addresses. Changed to try to force integer addresses. QGRD1 $QPSAP - changed to try to force integer addresses QGRD2 $QPSAP - changed to try to force integer addresses QGRD3 $QPSAP - changed to try to force integer addresses UVGRID Changed to put integer values of addresses in AP. UVGRTB Changed to put integer values of addresses in AP. UVTBGD Changed to put integer values of addresses in AP. GRDMEM Changed to put integer values of addresses in AP. GRIDR Changed APSIZ to integer throughout. APGS Removed unused reference to APSIZ as KAPWRD * 1024.0 and other unused variables. APVC Removed unused reference to APSIZ as KAPWRD * 1024.0 and other unused variables. STEER Removed unused reference to APSIZ as KAPWRD * 1024.0 and other unused variables. Moved nowhere. 7889. November 29, 1993 ALGSUB Eric In order to avoid 320 buffer copies each of 41000 words, I changed the address of the QVMOV call to reuse the full AP memory. After much pain, got all the +- 1's correct. I did not notice this because all my APs were too large to care. It blew off small APs - e.g. 5 Mbyte ones with 1024 on a side MX. Moved nowhere. 7890. November 30m 1993 HLPB Eric Cleaned up error return at end. The EOF on the input was leading to an assumed actual error. Moved nowhere. 7891. November 30, 1993 more AP stuff Eric (1) Changed the handling of the clean patch in MX, APCLN, SDCLN to allow for much larger beam batches. Changed them to call QCLNSZ to get the correct AP size for cleaning even on vector computers and to use the DDCH.INC variables for gridding and component subtraction. Cleaned up the help files too. (2) Renamed DAPS.INC to PAPC.INC since it is a parameter include for DAPC.INC. Added another parameter to it to define the max AP to use in QCLNSU. Changed DAPC.INC, SETPAR, SETSP, FILAIP, and FILAI2 to use the new include name. Changed $QPSAP QCLNSZ to use the new parameter include and to limit its answer with the new parameter (PKPWDC) in it. Moved nowhere - more to come. 7892. December 1, 1993 $AIPSPUBL/AIPSOOF.TEX, .PS Bill C. Final version of the AIPS OOPS documentation (at least by me). Moved nowhere - no more to come. 7893. December 1, 1993 $AIPSPUBL/AIPSMEMO51.TXT Gareth This general memo by Alan Bridle from 1987 still gives a good outline of the AIPS project. Although, Alan says it was checked in, it was somehow lost. This was scanned from the original by Gene Runion. I checked (most!) of the generated typos. Moved Nowhere 7894. December 1, 1993 AVGCIF in APLSUB Chris AVGCIF was calculating output indices using the index number of the Stokes axis (JNCS) instead of the Stokes axis increment (INCS). This resulted in visibilities being partly overwritten if more than one polarization was present. This has been fixed. CLPLT and VBPLT are affected. Moved nowhere. 7895. December 1, 1993 $QNOT/GRDCRM.FOR Bill C. Increased tolerance from 1% to 5% that a component is allowed to be from a exact grid cell before complaining. Moved nowhere. 7896. December 1, 1993 Vector APs Eric Changed booboos in $QPSAP version: QMULCL one address computation still affected by floating. Changed Convex versions: DAPC.INC to use PAPC.INC PAPC.INC to define parameters including the clean limit which is here not the full AP QINT Changed call sequence, broke up address computations to try to avoid the use of floating, dropped unused variables. QGRD4 Changed call sequence, broke up address computations to try to avoid the use of floating, dropped unused variables. QMULCL Changed pick up of address to integer, dropped unused variable. Changed Alliant versions: DAPC.INC to use PAPC.INC and dropped unused complex arrays which were 1 word longer than the things they were equivalenced to. PAPC.INC to define parameters including the clean limit which is here not the full AP QINT Changed call sequence, broke up address computations to try to avoid the use of floating, dropped unused variables, changed scratch work vector from 7 to 9 since only 9 is unknown to non-QALN. QGRD3 Broke up address computations, dropped unused variables. QGRD4 Changed call sequence, broke up address computations to try to avoid the use of floating, dropped unused variables, changed scratch work vector from 7 to 9 since only 9 is unknown to non-QALN.. QMULCL Changed pick up of address to integer, dropped unused variable. QCFFT Changed scratch work vector from 7 to 9 since only 9 is unknown to non-QALN, dropped unused variable. Moved nowhere. 7897. December 1, 1993 UVBOX Eric Changed the help files for CXPOLN, GRIDR, HORUS, MX, SCMAP, UVMAP, and WFCLN to allow UVBOX values up to 128. The previous limit was 10. Large values of UVBOX apparently help in the beam pattern of the VLA but cost a lot of computing and require a large pseudo AP. Moved nowhere. 7898. December 2, 1993 Vector APs continued Eric Changed CRI versions: DAPC.INC to use PAPC.INC and dropped unused complex arrays. Changed from dimensions of (1) to correct dimensions which stops errors from the compilers. PAPC.INC to define parameters including the clean limit which is here not the full AP QMULCL Changed pick up of address to integer, dropped unused variable, corrected array declaration. QCFFT Changed scratch work vector from 7 to B since only A and B are unknown to non-QCRI, dropped unused variable. QRFFT Changed scratch work vector from 7 to B since only A and B are unknown to non-QCRI, dropped unused variable. QPHSRO Changed scratch work vector from 8 to A since only A and B are unknown to non-QCRI, dropped unused variable. QRFT Changed scratch work vector from 2 to B since only A and B are unknown to non-QCRI, dropped unused variable. Moved nowhere. 7899. December 2, 1993 FITS binary tables Eric Binary extension tables are now called BINTABLE. Added that string to IMLOD, UVLOD and GSCAT lists of synonyms for binary tables. Phil will do FITLD. FITTP needs to be changed too but only after the release date. Moved nowhere. 7900. December 2, 1993 $APGNOT/FILLM.FOR Gustaaf FILLM now recognizes Q-band. Moved nowhere. 7901. December 2, 1993 FITTP Eric Dropped dangerous and mostly ignored adverb DOTABLE. Changed DONEWTAB to refer to calling binary tables BINTABLE vs calling them A3DTABLE. Dropped all ascii table writing and any conversion of FQ to CH and back. Also help file of course. Moved nowhere. 7902. December 3, 1993 $APLOOP/POLSN.FOR Bill C./Kari Leppanen Fixed incorrect logic in determining R-L phase offset. Moved nowhere. 7903. December 3, 1993 ZACTV9 in APLUNIX Chris Added code to explicitly close open file descriptors in the first generation children. Without this, the file positions in the UNIX file table entries used by the parent could be trashed by its children. This was spotted as a partial rewinding of RUN files on some machines and also appears to have been the cause of the line of garbage that appeared when starting a task under the debugger under Sun OS 5. Moved nowhere. 7904. December 3, 1993 ZACTV9 in APLUNIX Chris ZACTV9 no longer closes file descriptors that might refer to the message terminal. Messages should no longer be lost and the message server will no longer be any wierder that it normally is. This also required a modification to ZTTOP2 in APLUNIX so that the file descriptor for the message terminal is stashed when it is not the controlling terminal. Moved nowhere. 7905. December 3, 1993 $APGNOT/FITLD Phil Several changes: (1) Now recognizes BINTABLE as a valid table id, re Eric's change 7899. (2) Deals with VLBA LSB data, reflects the LSB spectrum and negates the phases. (3) When reading VLBA data tells user about latest bug fixes. (4) Contains kludge to deal with the problem of VLBA correlator scre-up of cross hand polzns. This is very temporary. (5) Can now deal with different data mapping functions for different freqids' (6) If there is a 30sec time gap in the data a new NX record is written. (7) Mapping function deals with compressed/uncompressed data. Moved nowhere. 7906. December 5, 1993 $APGNOT/FITLD Phil Missed some 3DTABLE -> BINTABLE strings. Moved nowhere. 7907. December 7, 1993 $APLUNIX/ZLPCL2.C, ZLASC2.C Pat Changed the way in which the temporary file is removed. Unlinking it right away will often remove the file before the spooler has had a chance to get to it. There is now a 300 second delay before printing, implemented via a fork and sleep. Also, replaced the system command with a popen/pread/pclose combination; this will remove the need for an IBM-specific version of these routines (the system call there was returning error 10). This has the side effect of making the output of the ZLASCL and ZLPCL2 shell scripts go through MSGWRT and they get put in the message file too. Looks a little weird due to wrapping. Moved nowhere. 7908. December 9, 1993 UVTBUN Eric The routine had an ill concieved + 0.001 in an integer expression causing it to mess up values of negative array locations. It did this once, not twice, so the weights for points with u=0, v < 0 got too much by rather more than half. Moved nowhere. 7909. December 9, 1993 BLING Chris Made several changes to BLING. - Removed AVS dump option. - Improved error reports so that lost scans are identified. - fixed a bug in the noise calculation that could cause floating-point exceptions if windows were in effect. - Provided access to relative-function convergence and X-convergence criteria through CPARMs. - Removed width parameters from the fit to the peak in order to reduce the chances of singular convergence. - Tidied up HELP and EXPLAIN files. Moved nowhere. 7910. December 14, 1993 BLING Chris Changed CPARMs so that the user can request that scans where interpolation fails are to be retained with reduced weight. This may bail some users out of sticky situations at the cost of allowing them ample rope to hang themselves. Also rewrote the explain file section on interpolation in an attempt to clarify the use of the CPARMs. Moved nowhere. 7911. December 16, 1993 COMRPL Pat Gave this overweight script a thorough going-over. The main concern was that it was not handling errors right when processing an "at" file or list of files. Now it keeps track of whether or not there was any error, tries the other files regardless and at the end gives the ok or error message depending on the "loop" status. I also cleaned up the options checking somewhat. Moved nowhere. 7912. December 16, 1993 POSSM Phil Added a line to the annotation specifying whether calibration was applied when making the plot, and if so which CL or SN table was used. Moved nowhere. 7913. December 17, 1993 $APLOOP/POSITION Bill C. PSNCV3: The radial dependent scaling was on offsets in RA and dec rather than in projected coordinates. This was blowing up near the poles. This change affects task OHGEO. Moved nowhere. 7914. December 17, 1993 COMLNK Pat See entry #7911 for COMRPL. Did same thing here. Moved nowhere. 7915. December 17, 1993 UVNOU Eric Corrected the wording to state correctly what the program does. It is the V axis at which u = 0 at which for all baselines the fringe rate goes through 0 which is why this program is useful. Moved nowhere. 7916. December 17, 1989 GUARD Eric Clearly this option is causing problems. Users do not seem to grasp what it is doing or why one doesn't want data all the way out to (and off) the edge of the UV plane. Changed the limits on GUARD in the help files and code to 0.9 as an upper limit. Changed: (1) Help files for MX, HORUS, WFCLN, and UVMAP to limit GUARD to < 0.9 and not to imply that larger values -> a default. (2) Added test on GUARD so that values outside 0 - 0.9 are taken as 0 in subroutine UVUNIF and changed the 0.6 limit to 0.9 in GRDMEM, UVGRID, UVGRTB, UVTBGD, UVTBUN, and UVTBUW (3) Added upper limit to tests on GUARD in MX, UVMAP, and HORUS and changed the setting of the default in WFCLN to be similar to the others (with no taper term however). Moved nowhere. 7917. December 17, 1993 $QPSAP/QPTAD.FOR Bill C. Fixed logic bug in computing address in lookup table. This error could cause the task to blow up. Used in UVADC. Moved nowhere. 7918. December 17, 1993 $YPGNOT/POSSM Leonid The number of BP table is printed now if the data is coming from the table. The header information at output file has been fixed The statement: 'The outfile is written only if NCOUNT=0' is added at the description of OUTFILE at POSSM.HLP file 7919. December 19, 1993 $QPGNOT/FRING Phil/Leonid At multyband delay solution, the INDE was appearing at SN table for all IF at DELAY and RATE column and only for the first IF at REAL and IMAG column. REAL was equal 1 and IMAG was equal 0 insted of INDE for the rest IF. This bug was fixed at subroutine FRNMS. 7920. December 20, 1993 BLING Chris Fixed an index error in G2MOD. This was causing the Jacobian to be calculated incorrectly for the single-IF case and was responsible for the excessive number of interpolation errors. Also fixed a bug in setting the FORTRAN unit number for DGLG debugging output. Moved nowhere. 7921. December 20, 1993 $SYSAIPS/AREAS.DAT Pat Moved $APLLINUX to the bell side of the tree. It is probably more appropriate there than on the Berkeley side, though it's still somewhat of a toss-up. Moved nowhere, but need to create/delete areas. 7922. December 20, 1993 XAS.SHR Eric The PC port of XAS pointed up a call seq error in our calls to XShmDetach --- the 2nd arg is an address of a structure not its first value. Moved nowhere. 7923. December 20, 1993 $APGNOT/FITLD Phil More in the series of changes aimed at making FITLD more robust as we see different types of VLBA observation pass through. (1) Added the adverb CLINT to .HLP, enables the user to set the CL time interval. Necessitated writing a new subroutine to fill in the values. Ensured that a scan has entries at the beginning and end and then every CLINT minutes. (2) Several changes in the area of file concatanation. FITLD can now run through a tape with ~100 VLBA data files and successfully read and concatanate them. (3) Unfortunately the correlator can occasionally write records out of time order, added logic to skip these records and carry on with data loading. Will allow up to 5% of file to be out of time order, or bad for other reasons, before skipping the rest of the file. (4) Corrected the logic in creating and adding to history records. Also the catalogue header was not being updated after the HI file was copied to it. (5) Sorted out the problems with the write status disappearing from the file after the first file was loaded from tape. (6) Will now deal correctly with frequency changes every scan. AT the end of a FITLD run a routine will run through the SU/CL tables and fill in the freq. offsets correctly. Difficult to do this on the fly. (7) FITLD now prints the latest bug fixes when it first starts up. Temporary situation only. (8) Added useful messages to user as reading progresses. (9) Fixed a bug that caused the task to fail if the number of antennas changed while concatanating. Moved nowhere. 7924. December 20, 1993 BLING Chris Added linear bounds for the parameters of the Gaussian models used for interpolation (which required switching driver routines from DGLG to DGLGB). The peak is now constrained to lie within the bounds of the grid cell in which the maximum amplitude was found. Without these bounds DGLG(B) would sometimes find a route to a local minimum of chi-squared that first reduced the peak value and then shifted the peak by several cells in delay before increasing the peak value again; this would occur if the peak were highly skewed and corresponded to a solution where the data used for the fit corresponded to the shoulder of a fictitious Gaussian located in the direction of the shallowest fall-off. Extremely skewed peaks will now sometimes cause interpolation errors as DGLGB beats its brains out against the bounds but should never cause wildly erroneous delays (whether or not interpolation fails). Also rewrote the interpolation section of the explain document to reflect the fact that end-users shouldn't fiddle with any CPARMs other than CPARM(1) and then only in desperate circumstances. Also added references for the optimization routines used in interpolation. Moved nowhere. 7925. December 21, 1993 VBPLT Chris Changed the HELP-file descriptions of the defaults for BCHAN, ECHAN, BIF and EIF to correspond to what the program really does. Moved nowhere. 7926. December 21, 1993 PROFL Eric Added comments to the help file about using TRANS when the hidden line algorithm has trouble. Changed the dimension of the string labeling variable which was assumed 2 but declared 1. This prevented LTYPE 4,5,6 from working. The strings are suppressed in PROFL in the interest of simplicity, but are computed by the low level routines. Moved nowhere. 7927. December 22, 1993 $APGNOT/UVCRS.FOR Leonid The new task finds the crossing points of UV-ellipses, calculates the average ratios of the amplitudes at the crossings points and findss the solution for the gains of antennas by a least squares approach minimizing the deviation of the ratios from 1. Also UVCRS.HLP. Moved nowhere 7928. December 22, 1993 $APGNOT/ANCAL.FOR Phil/Lincoln Greenhill Added the A22o, A22e and A30 terms of the 140ft spherical harmonic description of its gain curve to the task. Needed to expand the keyin section because they went in the middle - that's always a pain. Updated the .HLP also. Moved nowhere. 7929. December 22, 1993 VBPLT Chris The BASAVG subroutine assumes that the random parameter and visibility buffer are unaltered between calls. PLTVB altered them anyway which lead to points being misplaced in x and y when SOLINT was not zero. This was most obvious when they fell outside the bounds of the plot. Modified SOLINT to save and restore these variables at appropriate times. Moved nowhere. 7930. December 23, 1993 ZFREE Eric Deleted $APLUNIX/ZFREE.FOR - it made too many assumptions about the form of disk names and hence had bad displays under many circumstances. Fixed up $APLGEN/ZFREE to show the last real characters of the disk name up to 22 with a '<' character to show that characters are missing at the left. Moved nowhere. 7931. December 23, 1993 DEC ALPHA Eric Corrected $APLALPHA version of ZFRE2.C. I don't know why anyone thought it was working - it did a multiply when a subtract was correct, passed the address of a string pointer rather than the pointer to strcpy, and expected a return code of 1 rather than the 0 that always means success. Removed the old form of ZTAPE.FOR and put in $APLDEC a rewritten ZTAP2.C that looks a lot like most of the others. Changed ZMOUN2.C to set the density code in the tape name and to report back what tape device was found - again done in $APLDEC. Moved nowhere but pongo. 7932. December 23, 1993 HOLGR Mike Kesteven Changed the holography task and help: 1. The "surface rms" is now the half-path rms, weighted by the illumination amplitude. Its value is recorded in the history file. 2. The gains are estimated at the observing frequency and at K and Q band. 3. Some extension to the input parameters (largely dictated by the VLA requirements) a. INFILE. if set to "ANT", HOLGR will look for a file in the FITSarea with the name HOLOnnSS.EDT where nn is the antenna number, specified in APARM(9) SS is a stokes identifier, in APARM(10) : 1..LL; 2..RR; 3..LR; 4..RL b. APARM(2) : nominally the satellite elevation; if negative, it is assumed that the true holography coordinates, rather than azimuth and elevation offsets from the target. c. BPARM(3,4) : if negative, you get a radial taper, down 50% at the specified |u| d. BPARM(9) : 2000 will disable the focus offset calculations, but retain the pointing. Moved nowhere. 7933. December 27, 1993 UVNOU.HLP Tim C. Changed the strange comments in the help file to reflect why the task is useful. Moved nowhere. 7934. December 29, 1993 $YPGNOT/POSSM Phil 2 minor fixes, one removes the attempt to close the plotfile when it was never filled, that occasionally happens; the second ensures that the average bandpass weight is only computed when data have been averaged. Moved nowhere. 7935. December 29, 1993 $APLUNIX/ZIGNAL.C Pat Checked it out to insert an #ifdef for linux; it has an underscore where SunOS has a hyphen in one of the structures. Not sure if what I did is really kosher but it makes the darn thing work under Linux and other systems. While doing this, noticed a comment "/ Sun OS */" (missing * after the first /) and a missing semicolon in the SYSV #ifdef part. I wonder what will change now. I really shouldn't do this from home... Moved nowhere. 7936. December 29, 1993 $QPGNOT/BPASS.FOR Phil Ensured that, when in total-power mode, the BP table does not contain totally flagged entries. This scres up the DOBAND=2 option, totally flagged entries have never been written for cross-power bandpassses. Moved nowhere. 7937. December 29, 1993 $YPGNOT/GAL.FOR Gustaaf Added the adverb OUTFILE. GAL optionally appends a one-liner containing results from the fit to this OUTFILE. Is useful when applying GAL to many rings. One option is 'redirect' OUTFILE to STARS which allows one to plot the fitted major axis. Moved nowhere. 7938. December 30, 1993 UVINIT Eric Changed it to allow NVIS = 0 on reads only. The first UVDISK call will return no (more) vis's. This should cover cases in which, due to flagging, there are no vis's for a particular source. Moved nowhere. 7939. December 30, 1993 RUN Pat Added more functionality than this poor little script was ever supposed to have. Arguments now include: program name, "DEBUG" or "-d", "LOCAL" (debug forces local), and the already present da= disk selection option. Order is irrelevant (resistance is futile). I put case insensitivity into the argument loop. If the environment variable $DBUGR is not set, you will be prompted for the name of the debugger. It doesn't go through the theatrics that START_AIPS does to check this, so if you type fbx instead of dbx, it'll crash and burn. Moved nowhere. Do I get the last change of 93? 7940. December 30, 1993 $YPGNOT/VBPLT.FOR Phil Sorry Pat, you don't. The line that copied the user specified timerange into the array used by UVGET was commented out so all the data were being read for scaling and plotting, although the timerange was used when placing points into the plot. Was this why VBPLT was so slow? Moved nowhere. 7941. December 30, 1993 $APGNOT/FITLD.FOR Phil More changes: (1) Added data selection by source, qualifier, calcode and timerange. (2) Changed error codes for data out of sequence and bad day number, they could be confused with EOT codes. (3) Tidied up the logic for concatanation and # files to skip/read, depending on whether incoming data is from tape or disc. (4) Ensured that file being written to is marked with write status always, this flag was lost when concatanation occurred. (5) Fixed the antenna renumbering logic, and also when # antennas increases due to new antennas in an incoming file the NO_ANT keyword in the CL table was not being set, fixed. (6) Removed the polzn cross-hand kludge, the data are now correct coming out of the correlator. Also updated help file. Moved nowhere. 7942. January 1, 1994 ABOUT Eric New pseudoverb to display the list-like help files. Changed HELPS and AU1A to do the verb, POPSDAT.HLP and NEWPARMS.001 to declare the pseudoverb, and created ABOUT.HLP to describe the verb. Changed woprding of HELP.HLP to account for ABOUT. Changed HLPA, HLPB to ignore ZZ*.HLP files. Changed VERMAT, TXTMAT, ZTOP0EN, ZTXMAT, and, $APLUNIX/ZDIR to allow the help file name to be somewhat longer than 8 characters (plus the .HLP). Changed call sequence to ZTXMA2 in areas $APLDEC, $APLSUN, $APLUNIX, $APL3090 and changed OFMDIR to match. Also created SHOPH.FOR to create the list-like help files and to maintain them. One runs SHOPA on a list of the help files, sorts the output, and then runs SHOPH. This last copies the precursor and all other lines from the pre-existing list file in $HLPFIL up to the first blank line. Then it compares the list in $HLPFIL with the one it has made from the sorted SHOPA output. If they are different, the user is told of the need to update. Created list files in $HLPFIL: ZZADVERB.HLP ZZIMAGING.HLP ZZRUN.HLP ZZANALYSIS.HLP ZZINFORMAT.HLP ZZSINGLEDI.HLP ZZAP.HLP ZZINTERACT.HLP ZZSPECTRAL.HLP ZZASTROMET.HLP ZZLINE.HLP ZZTABLE.HLP ZZBATCH.HLP ZZMAP-UTIL.HLP ZZTAPE.HLP ZZCALIBRAT.HLP ZZMODELING.HLP ZZTASK.HLP ZZCATALOG.HLP ZZOBSOLETE.HLP ZZTV-APPL.HLP ZZCOORDINA.HLP ZZOOP.HLP ZZTV.HLP ZZCUBE.HLP ZZOPTICAL.HLP ZZUTILITY.HLP ZZDELETE.HLP ZZPARAFORM.HLP ZZUV-UTIL.HLP ZZEXT-APPL.HLP ZZPLOT.HLP ZZUV.HLP ZZFITS.HLP ZZPOLARIZA.HLP ZZVERB.HLP ZZGENERAL.HLP ZZPOPS.HLP ZZVLA.HLP ZZHARDCOPY.HLP ZZPRINTER.HLP ZZVLBI.HLP ZZIMAGE-UT.HLP ZZPROCEDUR.HLP ZZIMAGE.HLP ZZPSEUDOVE.HLP Moved nowhere. 7943. January 3, 1994 WFCLN Eric Changed WFCLN to set some parameters into the clean object(s) before calling the uniform weighting instead of after. Changed QUVUTIL routine OUNFWT to avoid a divide by 0. Both of these will stop a bug in which CHINC = 0 slipped through and caused a 0 divide. Moved nowhere. 7944. January 5, 1994 CLPLT Phil Added the SYMBOL adverb to CLPLT. Also updated the help file. Moved nowhere. 7945. January 5, 1994 PLCUB Eric Made all the buffers much bigger. It was limited to rows of 512 without really checking for it! Moved nowhere. 7946. January 5, 1994 HELP Eric Corrected all versions of ZTXMA2.C to truncate the file name at the last period, not the first. Names like VLAD.OLD.HLP were confusing it. Changed AU1A to put the full name of the help file in the displayed title, not the min match. Also COMRPL'd HELPS which I forgot to do in public areas. Moved nowhere. 7947. January 5, 1994 PRTAC Eric Pat's neat fix to the file deletion problem has an interesting sidelight. By forking from a running process, all file lock connections are inherited by the forked process and hence held for 5 minutes until it departs. This can be nasty. PRTAC called LPCLOS before closing the accounting file which was then locked for 5 minutes while the task shut down waited (bleating forlornly every 30 seconds) in ACOUNT called by DIETSK. We will have to check that all functions doing line printing close their open files before calling LPCLOS. Sigh... Moved nowhere. 7948. January 6, 1993 RTIME Eric New task to test machine timings on loops. Moved nowhere. 7949. January 6, 1993 $APGNOT/AFILE.FOR Phil Two minor changes, first the experiment numbers for MkIII correlations are now > 1000, so had to change the format for reading that number from I3 to I4. Secondly, the logic for selecting by processing date collapsed if an observation was correlated over the New Year, fixed so that it checks the year number also. Moved nowhere. 7950. January 6, 1994 Socket codes Eric Removed the getsockopt calls for tcp max segment size. The answer was not actually used in the tape codes. The TV code used it, so I changed the TV code to be like the tape code, raising buffer sizes should that ever be needed. Changed $APLUNIX versions of ZVTPO2, ZVTPO3, ZVTPX2, ZVTPX3, ZSSSO2, and ZSSSX2. Lots of link edits. Moved nowhere. 7951. January 7, 1994 $APGNOT/MODVF.FOR Gustaaf MODVF stands for MODel Velocity Field. This task will create a model of a galaxy velocity field. As in GAL, information about the orientation of the galaxy with respect to the observer is contained in APARM. Detailed information about the rotation curve and warps are read in from INFILE. The warp is specified in terms of the angular momentum vector change from ring to ring, which makes it far easier to enter specific warp models than previous similar software, where warps had to be given in terms of position angle and inlination. 7952. January 7, 1994 $APLUNIX/ZCPU.C Pat Added an #ifdef for __linux__ so that it gets the CPU units right. Now it won't show 100 cpu seconds for 60 seconds of wallclock time anymore! Moved nowhere. 7953. January 7, 1994 $APLUNIX/ZLPCL2.C Pat Removed the code that did a fork() and had the child process sleep for a while and then delete the scratch file. This was causing all sorts of problems as the child inherited the locks of the parent. Instead, made it do a popen() on a pipe to a shell command "((sleep xxx; rm -f yyy)&)". Thus a grandchild process will do the sleeping and removal, and will not (hopefully) have any locks as the process is a simple shell, not a copy of AIPS or LWPLA or whatever. Moved nowhere, but I have to do the IBM version too... 7954. January 7, 1994 $YSERV/XAS.SHR Pat Mostly editing the comments in the Makefile so that the LIBDIRS and INCDIRS variables can possibly be set automatically via INSTEP1 (wish I had time to learn more on make; it'd be easier I'm sure). Also put in an "install" target so that you have to type "make" to make xas, and "make install" to move it to the destination (usually $LOAD). I think "make install" will do everything if it all needs done. Moved nowhere. Why do I do these things on Fridays? 7955. January 10, 1994 SAD Eric Put back my latest version. It offers the option of trying 2 sources whenever the residuals of a single-source fit are excessive. The MF file has more (!) columns for polarization and for the residual rms, peak, and sum. The help file was improved rather generally. I don't think this is the final version of SAD, but it should be good enough to release. Moved nowhere. 7956. January 10, 1994 SOUELV Leonid The longtitude - STNLON(ANTNO) was multiplied by coefficient 1.002738. But the longtitude need not to be multiplied by a coefficient. So it has been remooved. Moved nowhere. 7957. January 10, 1994 CLCOR Leonid[B The new option 'ANAX' has been added for correction of phase, delay and rate for antennas' axis shifts. The AN table is modified as well. Moved nowhere. 7958. January 11, 1994 LP close Eric Changed JTRIM to terminate strings on all non-printable characters (char < ' ') except "ht" (horizontal TAB) and to not count ht as a printable character when looking for the last non-blank. Created $APLGEN/ZTAB to return a ht character to JTRIM. Changed ZLASC2 and ZLPCL2 to send messages only up to 64 characters and corrected the placing of any remaining characters in a 2nd message line. Simplified the code (JTRIM will fix new-line and other odd characters and messages do not need to be blank filled if they are null or new-line terminated). Probably ought to relink everything for JTRIM, but did not do so yet. Moved nowhere. 7959. January 12, 1994 MSGWRT Eric Changed onset of file too big message from 750 to 5000 and frequency from every 21 messages to every 210. Moved nowhere. 7960. January 13, 1994 $APLNOT/BPASET.FOR Phil Changed a variable defining a buffer size for the sort routine from 4096 to UVBUFSL, this was causing the application of bandpasses from the AT to fail when a large number of channels were used. Moved nowhere. 7961. January 13, 1994 ZACTV8, ZACTV9 Eric Added a ZERROR call message to all possible error completions in ZACTV9 including those from child and grandchild. The latter would not get back to ZACTV8 to call ZERROR and so were omitted or avoided by calling exit rather than return. Many sources of error were then left more mysterious than they needed to be. Dropped the ZERROR call from ZACTV8 as a consequence. Also made this fix to $APL3090/ZACTV9 upgrading it to use zmsgwr. Moved nowhere. 7962. January 13, 1994 PC found Eric/Pat Changed a number of routines for inconsistent call sequences involving character strings. These matter since UNIX systems changed a Fortarn string into 2 arguments, putting them somewhere in the call sequence (time and OS dependent places). In other cases, the call sequences just were missing arguments. CORER A subroutine had too few arguments, so IERR was never set. UVLSF A call to HENCOO had the out class given twice and no out sequence number for the BASFIT file. XBASL Used a floating buffer to avoid having another character variable for a ZPHFIL/ZOPEN pair. XGAUS Used a floating buffer to avoid having another character variable for a ZPHFIL/ZOPEN pair. CANPL Used a dummy 0 for an in/out character buffer in a call to close ZLASIO, left out 2 arguments (used) in calling its own CNTPLT and CANVEC. This could never have worked. DCONV Sent character string 'BLNK' (as if that were the magic value blanking used in AIPS) to LINIO. A real was expected - and used most places in DCONV. Droppped unused variables. FILLR Sent 2 arguments to MCANT where none were declared or used, used a CHAR argument to ZILI16 and ZC8CL where an integer is expected. GSCAT Left out the offset argument in a call to CHFILL. This must have never been used (blanked character column in the GS table). HOLGR The DFT routine called the data retrieval routine without one of its arguments. Changed call to DFT to pass that one through. IMLIN Used a floating buffer to avoid having another character variable for a ZPHFIL/ZOPEN pair. Removed unused variables, corrected list-directed READ from a string with calls to GETNUM. IMLOD Fixed several calls to ZC8CL in the non-FITS sections. JMFIT Changed 1 call to H2WAWA which was passing in a character string where a hollerith was required. MK3IN Changed 1 call to TABIO for clos which used a character string for a dummy argument. SDTUV Changed 1 call to TABIO for clos which used a character string for a dummy argument. UVLOD Changed several calls to TABIO for clos which used a character string for a dummy argument. Changed a call to GETCRD that used a dummy string in place of an integer. UVSIM Changed 4 calls to CHR2H used to attempt to set real character strings to blank. VLBIN Changed 1 call to TABIO for clos which used a character string for a dummy argument. The OOP package and more still need this done. Moved nowhere. 7963. January 14, 1994 Deletions Eric Moved $APGNOT/BLFIT.FOR and $HLPFIL/BLFIT.HLP to $APLCONTR since this routine was not upgraded when the CL file format was changed and is reputed not to have worked properly. It can be rescued if anyone needs it and wants to do the work. Also moved $INC files CATDAT.INC, CATREC.INC, and CONDAT.INC. They are not used (fortunately since they are in DEC PDP-11 specific forms). Moved nowhere. 7964. January 16, 1994 New adverbs Phil Created the new adverb set IN4NAME, IN4CLASS, IN4SEQ, IN4DISK and IN4TYPE. These will be needed for the new VLBA/VLBI task VBGLU that will glue together data from multiple passes through the VLBA correlator. It will have to deal with 28 channel MkIII Mode A data that requires 4 passes, so I will have to deal with up to 4 simultaneous input files. May as well do it properly. Created .HLP files for all the new adverbs. Moved nowhere. 7965. January 16, 1994 More from the PC Eric Corrected: MAPCOP Corrected a dummy argument to CATDIR to be character. The routine got it right 5 lines after getting it wrong which is why it was found. TABINI Corrected a call to ZDESTR using a hollerith file name. It could never have worked for correct error recovery. GRITP One buffer was left character rather than hollerith in calls to ZTPMIO and ZC8CL. It would not have read tapes. RSTOR The parameter OLD was left out of the declaration of FILES causing IERR not to be set and hence the task proceded even if there were problems with the inputs, scratch files, etc. IMVIM One of the clean up on error calls to DELEXT sent in IERR for the version and no address at all for IERR. It would have blown up if called. TVMON Call to AIFM gave an extra, unused buffer. GAPLT Dropped unused arguments to GAPANT. BLING Added IRET to declaration in INITNL. CXCLN Changed a pile of get and put routines to use one of the 2 parameters as a DUMMY or CDUMMY. FRCAL Changed a pile of get and put routines to use one of the 2 parameters as a DUMMY or CDUMMY. OHGEO Changed a pile of get and put routines to use one of the 2 parameters as a DUMMY or CDUMMY. SCMAP Changed a pile of get and put routines to use one of the 2 parameters as a DUMMY or CDUMMY. Corrected a long line that caused an object name to be 3 rather than 32 characters long. WFCLN Changed a pile of get and put routines to use one of the 2 parameters as a DUMMY or CDUMMY. Moved nowhere. 7966. January 17, 1994 BLING Chris Fixed a misplaced end-of-loop that caused multipolarization data not to be written to the BS table correctly. Moved nowhere. 7967. January 18, 1994 SOUELV Leonid Returned back the coefficient 1.002738D0 after STNLON(ANTNO) at line ANTLST. Moved nowhere. 7968. January 21.1994 SNPLT Leonid Fixed two bugs occured at OPCODE 'PLIF'. 1. In the labelling in the header of the plotfile the second line was too long when OPTYPE's represantation requires more than 5 characters: AMP - Gain amp, PHAS - Gain phs, DOPL - doppler offset, etc. As the result the antenna name was trancated. I was especially confusing at the VLBA case when the name were truncated to "VLBA_". The second line has been splitted on two lines. So the header has three lines now. 2. If FRING failed to find solution for a particular antenna, SNPLT stopped plotting at that antenna and closed down after reporting an error even though there may be perfectly valid solutions for the next antennas of the list. The new version prints 'NO DATA SELECTED' for the 'bad' antenna and continues analysis of the next ones. The antenna's number and name is printed now to help understanding of the process. Moved nowhere. 7969. January 24, 1994 FITLD Phil Usual long list of changes: (1) Added code to skip over non FITS files if encountered. (2) Ensured that when multiple files are open for writing that the HI records are piped to the correct one. (3) Ensured that if a file is just created the LSTTIM (time of last record is reset). (4) Changed GET3DT error codes from 2, 3 -> 12, 13 to avoid confusion with error codes from other routines. (5) Fixed error that caused catblk to be copied into wrong save array when multiple files open. (6) Ensured that write flag is set on current output file. (7) Changed some call sequences to ensure that character strings are not passed as dummy non-character variables - for PC-AIPS. (8) Added code to add day number offset to VLBA-style tables on the fly as they are read in, maybe should improve this. (9) Added code to skip over source numbers <= 0 (10) Tidied up the code that counts bad records. (11) At end of UV data have tapio read to eof, rather than skip. If there were only a few unknown records (due to padding) then the skip option would jump over the next file. (12) Change the logic for renumbering sources/antenna during concatanation, now only keep a list of those to be changed rather than all possible ones. This was causing a very subtle bug. (13) Have the data mapping function switch the sign of all sidebands only when cross-power data encountered. This is a very complex area. Moved nowhere. 7970. January 17-21, 1994 OOP Eric Changed all calls to GET and PUT routines in $QPGOOP, $APGOOP, $QOOP, and $APLOOP (and there are a lot of them) so that they use a non-character and a character variable in the correct place. Also changed PUTs and other places to use the parameterized data type codes (OOAINT, OOCAR, etc.). Moved nowhere. 7971. January, 24, 1994 AREAS.DAT, APLULTRIX Pat Made new area $APLDEC/ULTRIX = $APLULTRIX, moved routines ZFRE2, ZDELA2, and ZCREA2 from $APLDEC to $APLULTRIX, moved $APLALPHA versions of same routines to $APLDEC, and modified AREAS.DAT accordingly. Ran AREAS locally; the MNJ _should_ take care of it elsewhere. Moved nowhere. 7972. January 25, 1994 VBPLT Phil When averaging in time VBPLT was losing track of the subarray number in the scaling routine. Fixed. Moved nowhere. 7973. January 25, 1994 SNPLT Leonid I splitted the second line of the header on two lines to exclude possible truncating of IF number when OPTYPE is represented by many characters, for example 'multyband delay' Moved nowhere. *********************************************************************** Changes while 15JAN94 was NEW: *********************************************************************** 7974. January 28, 1994 $HIST/CHANGE.DOC (this file) Pat Initialized this file as part of the quarterly update. Moved nowhere, but also edited TST (15JUL94) version accordingly (i.e. emptied it). 7975. January 30, 1994 AIPS, VERBS Eric Moved the VERBS* subroutines inside the AIPS* program files. VERBSC referred to some routines which are in AIPSC only, causing unresloved external references. The VERBS routines are only used by the single programs anyway. Moved from 15JUL94 this date. 7976. January 30, 1994 ZDCHIN Eric Added the PAPC include and tests on KAP2WD. This parameter was being used before the call to QINIT and hence the software did not have a chance to defend itself from bad settings in the SP file. Now one should set the SP file to the largest for any of the 3 versions and the others should be able to take cae of themselves. Moved to 15JUL94 this date. 7977. January 31, 1994 LIBR Eric Changed most of the LIBR.DAT files - they specified $AIPSUB for $AIPGUNIX which is incorrect. The $SYSUNIX one and a couple of others had already been corrected. Moved to 15JUL94 yesterday. 7978. January 31, 1994 PRTLIN, PRTALN Eric/Dave Changed PRTLIN and PRTALN to call ITRIM rather than JTRIM for the page header line. JTRIM was dropping the line due to the special character (form-feed) we wanted in the line. Moved to 15JUL94 this date. 7979. January 31, 1994 MX, ZEROSP Eric Dropped the option of allowing ZEROSP(2-4) to give the widths of a finite gaussian as the zerospacing. It takes too many cells to represent one for the gridding to handle and messes up the beam. We need to meet this need in other ways - placing extra vis samples in the uv data set. Also shortened long line in MX output. Moved from 15JUL94 this date. 7980. February 1, 1994 FILLM.HLP, BAND.HLP Gustaaf Added the Q-band information to the HLP files. FILLM already handled Q-band for some time, but there was no way for the user to know. Though not in the 'urgent' category, moved it to 15JAN94 nonetheless. 7981. February 1, 1994 TBSUB, TABLE Eric Corrected TABLE.FOR subroutine TABOPN to set EXISTS = STATUS.EQ.'READ'. Otherwise, it set NCOL et al. to zero when trying to create a table (with a default OUTVERS). It then failed of course. Fixed two character declarations in TBSUB, one of which caused the first character of each column to appear as the column label of the first output column, and the other output labels were blank. Moved from 15JUL94 this date. 7982. February 1, 1994 UVCOP Phil UVCOP forced the first IF-offset in the FQ table to be 0.0 even when no IF selection or freq. scaling was being done. Added an IF statement that forces this only when necessary. Moved from 15JUL94 this date. 7983. February 1, 1994 FRING Phil FRING selects/calibrates/flag etc the data to a scratch file and does the fitting from that scratch file. However it sets up its solution intervals based on the NX table from the original input file. When any data selection was done the solution intervals could be wrong because there was no check made to see if the NX record read was a deselected one. Added those checks and we now get the correct solution intervals on short-scan VLBI data. Moved from 15JUL94 this date. 7984. February 2, 1994 $APLOOP/UVUTIL.FOR Bill C. UVCLIP wasn't handling the case of doing Ipol when one of the two parallel hands was already flagged. Now Ipol is computed from either of the two parallel correlators if only one is good. Moved from 15JUL94 this date. 7985. February 2, 1994 $APGNOT/MAPBM Bill C. New task. Uses a VLA raster made in holography mode (with no reference antennas) and makes images of the BEAM in I, Q, U and V with RMS uncertainties in each. The input is a calibrated multisource uv data set and the total and polarized intensities of the source. In Q, U and V the resultant images are fractional polarization and the Q and U images can be used by VLABP to correct widefield snapshot VLA images for off axis instrumental polarization. Also help file. Moved from 15JUL94 this date 7986. February 2, 1994 $QPGNOT/VLABP Bill C. New task. Task to apply widefield instrumental polarization corrections to a snapshot VLA polarization image. The imput images can be either a set of I, Q and U images or a polarization cube containing the three and a set of instrumental polarization images. This task interpolates and rotates the instrumental polarization image to align with the observed image, rotates between Q and U to counter the effects of the on axis calibration of the off axis instrumental response, multiplies times the Ipol image and subtacts from the (Q + iU) image. The parallactic angle of the observation can be passed either as a header keyword or an adverb. This will ONLY work for snapshot images. Also help file. Moved from 15JUL94 this date 7987. February 2, 1994 $QPGNOT/FRING Phil The FRING fix I put in yesterday did not check for the code running off the end of the NX table. Now does. Moved from 15JUL94 this date. 7988. February 2, 1994 FITLD Phil 3 small changes. 1) allow FITLD to look at more than one disc when deciding on concatanation; 2) Benson has moved the correlator history records to the base header (where they should have been all along) - added code to ensure that the string FITLD is appended to these records; 3) when an antenna number is changed during concatantion and as a result a baseline direction has to be changed it was not doing it, fixed now. Moved from 15JUL94 this date. 7989. February 3, 1994 $SYSSOL/LIBR.DAT Pat The $AIPGUNIX area only references two libraries: APLSUB and APLSOL (in this case). It needs to repeat these for the loader to resolve all references, at least when shared libs are not used. Moved from 15JUL94 this date. 7990. February 3, 1994 $SYSSUN/COLOR,PRINT Pat These shell scripts were never part of AIPS formally. Modified them to use the AIPS printing scheme, automatically selecting the first appropriate printer if none is specified, and allowing the PR=n command line printer selection. Moved from 15JUL94 this date. 7991. February 3, 1994 AJAX.FOR Pat As part of the clean-up (sorry, couldn't resist!) I do with the quarterly update, I noticed that this task failed to link on Solaris. It has a call to DESCR which is in AIPSUB, and AIPNOT tasks are not supposed to call AIPSUB routines. So I moved it to AIPPGM. This cures the problem. Moved from 15JUL94 this date. 7992. February 4, 1994 WFCLN.FOR Bill C. Fixed serious bug in WFTAV which averaged the times in the uv data set to determine the average observing geometry. This was being done in solar time which caused serious problems if there were observations over several days. Now it averages in sidereal time modulo 1 day. Moved from 15JUL94 this date. 7993. February 9, 1994 FITTP, UVLOD, FITLD ERIC Changed FITTP to leave the projection type on the U, V, and W random parameter. It does this as 'UU -SIN' for DONEW false and as 'UU---SIN' for DONEW true. Changed UVLOD to recognize either 'UU ' or 'UU--'. Moved from 15JUL94 plus did FITLD myself. 7994. February 11, 1994 $SYSUNIX/TVDEVS.SH Pat In Multi-site configurations (i.e. where there is more than one SITE specified in HOSTS.LIST; this is useful when you want opposite endian architectures to share the same $AIPS_ROOT) the selection of HOSTS, TVSERVS and XTERMS was not discriminating on the SITE name. Fixed via a simple "grep". Moved from 15JUL94 this date. 7995. February