----------------------------------------------------------------------- ; Copyright (C) 1995 ; Associated Universities, Inc. Washington DC, USA. ; ; This program is free software; you can redistribute it and/or ; modify it under the terms of the GNU General Public License as ; published by the Free Software Foundation; either version 2 of ; the License, or (at your option) any later version. ; ; This program is distributed in the hope that it will be useful, ; but WITHOUT ANY WARRANTY; without even the implied warranty of ; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ; GNU General Public License for more details. ; ; You should have received a copy of the GNU General Public ; License along with this program; if not, write to the Free ; Software Foundation, Inc., 675 Massachusetts Ave, Cambridge, ; MA 02139, USA. ; ; Correspondence concerning AIPS should be addressed as follows: ; Internet email: aipsmail@nrao.edu. ; Postal address: AIPS Project Office ; National Radio Astronomy Observatory ; 520 Edgemont Road ; Charlottesville, VA 22903-2475 USA ----------------------------------------------------------------------- This file must be used to record any changes made to the code on the Charlottesville VAX in the logon area [AIPS]. The format to use is: line 1: Date Program name Your name line 2: What was done line 3: Where has this change been moved (e.g. OLD, the Modcomp) line 4: Blank The details of line 1 are important since programs are used to find specified strings and do conversions on them. It MUST be typed as follows: 1 blank, 4 digits, a period, two blanks, then the Month, day, comma, year, some space, a title, some more space, your name. *********************************************************************** Changes while 15APR90 was TST: *********************************************************************** *********************************************************************** 5421. November 7, 1989 SPLIT Phil SPLIT was screwing up the writing of compressed data. A variable KLOCWT was used to describe the random parameter offset in the output data, this was set to # r.p's + 1. However this same variable was also used to define the offset in the input data, so the input data was screwed. Redefined KLOCWT to LLOCWT. Moved to 15JAN90 this date. 5422. November 7, 1989 VLBIN Phil The lobe rotation correction on station 'B' autocorrelation spectra was being done incorrectly. The ra/dec passed to the geometry calculation routine were passed as degrees rather than radians, also the natural fringe rate calculated was then applied with the wrong sign. Moved to 15JAN90 this date. 5423. November 9, 1989 Print files Eric Added adverb OUTPRINT to all printing tasks and verbs to allow the user to specify a file name outside the regular AIPS file-naming conventions in which to save the printed output. This file must be a new file and is likely to be deleted if the task terminates prematurely. The low-level Z routines have allowed this since the overhaul. It was a matter of adding the adverb and passing it down to them. Changed help files and fortran code for tasks BLSUM, CORER, IMFIT, JMFIT, LISTR, PRTAB, PRTAC, PRTAN, PRTCC, PRTGA, PRTIM, PRTSD, PRTTP, PRTUV, TXPL, UVDIF, and UVFND. Changed help files for verbs EXPLAIN, HELP, PRTHI, and PRTMSG. Other specific changes were: DAPL.INC Added adverb OUTPRT as a HOLLERITH of dimension 12. POPSDAT.HLP Added adverb OUTPRINT as STRING*48. OUTPRINT.HLP New help file for new adverb. NEWPARMS.001 Added adverb OUTPRINT as STRING*48. ZOPEN (APLGEN:) Changed to pass PHNAME through to LPOPEN rather than sending blanks. AU1 Changed call to PRTMSG to get OUTPRT in character form for verb PRTMSG. AU1A Convert OUTPRT to character for ZOPEN call for HELP and EXPLAIN. AU7 Convert OUTPRT to character for ZOPEN call for PRTHI. PRTMSG Added OUTPRINT as character to call sequence and changed ZOPEN call to pass it on. BATER Added OUTPRINT to symbol list and to call to PRTMSG. GRIPR Added OUTPRINT to symbol list and to call to PRTMSG; corrected addressing to character string adverbs. PRTAC Added output file name option to stand-alone mode of this task as well, so long as the print file has not yet been opened. Moved that opening to the last possible moment. TXPL Corrected typing of code some, clarified usage of DOCRT so that it can give width of wider terminals, corrected it to resume AIPS when it writes to the history or message files. The user may just RUN NEWPARMS to get this adverb for his tasks, but to have it recognized for the verbs he/she must do RESTORE 0. Moved nowhere. 5424. November 9, 1989 APCLN Chris APCLN now creates scratch files large enough to hold the larger of an $ny$ by $nx+1$ image or a $2ny$ by $nx/2 + 1$ image, where $nx$ and $ny$ are the numbers of pixels on the x- and y-axis of the input map. Previously, $nx$ and $ny$ had been interchanged, causing too little space to be allocated for the scratch file if $ny > nx$. Moved nowhere. 5425. November 9, 1989 MK3TX Chris New task. MK3TX allows the user to interactively select text files from a MKIII ("A" or "B") archive tape and convert them to disk-based text files acceptable to the local system. Some of these files contain information that may be useful in editting and calibrating the data read from the tape with MK3IN. Moved nowhere. 5426. November 10, 1989 VLBIN Phil Removed the ability to perform the correction for natural fringe rate differences between the two stations of a baseline. This was supposedly needed to line up the frequencies for the later running of CVEL on autocorrelation data. However I hadn't appreciated the fact that my routines in CVEL calulcate the shift for each station exactly and so implicitly do the natural fringe-rate correction anyway. Also changed the .HLP file. Moved to 15JAN90 this date. 5427. November 12, 1989 CVEL Phil There was a problem with the direction of the 2 FFT routines used, the AIPS PSAP has a different direction to FOURG. Modified so they both assume the same sideband now. Also tidied up the informative error messages. Moved to 15JAN90 this date. 5428. November 13, 1989 BPASS Phil Fixed bug put in earlier. When initializing BP table entries was not checking number of polarizations being written. If there was only one some random variables were being clobbered. Moved to 15JAN90 this date. 5429. November 13, 1989 UVFND Bill J. Missing call to MSGWRT when compressed data was encountered let program die uninformatively. Moved to 15JAN90 this date. 5430. November 13, 1989 Printer bugs Eric Corrected bugs in Friday's submission: AU1 Corrected spelling of variable name - was wrong type. Corrected second call to PRTMSG for EXIT. PRTMSG Added close and reopen of message file surrounding the open of the line printer since that uses MSGWRT. ZLPOPN (Generic) improved error message, corrected to return an error code, added message giving file name when it works correctly also. ZLASIO (Generic) corrected message levels. TXPL Corrected some more typing and wrong type of variable in call to PRTLIN. POPSDAT.HLP Removed a second declaration of RESTFREQ, it had caused an addressing error of one location. Moved nowhere. 5431. November 14, 1989 UVFLG Phil Would not work under the circumstance that no source name was specified. If there were none an essential variable was left at 0 and since that was the loop index in the flagging do loop it didn't leave a lot for the task to do. Moved to 15JAN90 this date. 5432. November 14, 1989 Fortran problems (SUN 4) Bill C Fixed a number of problems in the task source code that were uncovered by the SUN 4 Compiler. Arrainged by problem they were: - Lower case c as comment delimiter: COMB, FITTP, GREYS. - Declare an array passed as call argument as dimension (1) rather than (*): CALIB, CLIP, MCUBE, PCAL, XMOM. - Common variable target of DATA statment: CANPL, GAL, QMSPL, SOLCL, IMPFIT. - ANCAL needed to declare a type for DFLOAT - BSMAP had a SAVE statment after a DATA statment. - NINER used a 1-D subscript for WIND33 in an EQUIVALENCE when it was declared 2-D. In a number of cases the Sun compiler was out to lunch; in some cases a change was made to make the compiler complaint go away. - The most common of these was use of an array dimensioned (1) and equivalenced to a common array; the VAX compiler disallows use of the dimension (*) in these cases so the dimensions were changed to a large value: SDTUV, USUBA. - In JMFIT and IMFIT in the history routine RPARM was dimensioned (99) (it's true value) but the compiler was getting confused so the declaration was changed to (*). Moved nowhere. 5433. November 14, 1989 SPLIT Phil The alternate reference pixel written to the output catalogue header was incorrect by 1. Fixed. Moved to 15JAN90 this date. 5434. November 14, 1989 Miscellaneous Eric Corrected yet another call sequence to PRTMSG for CLRMSG in AU1. Corrected lower case ``C'' in column one in Fortran files AU1A, AU5D, ZTPCLD (APLGEN:), CATIO, CHR2H, FILCR, PASWRD, SETVIS, VERMAT, PLNPUT, YGYHDR (YGEN:), and GRBOXS. Corrected use of dummy array declaration length from (1) to (*) for call arguments in BDN and ITBSRT and to real lengths for local variables in GAINI and ITBSRT. Changed intrinsic DFLOAT to REAL in ZDHPRL. Changed DATA of common variables in BMSHP to set the parameters at every entrance to the subroutine through equates. Changed a local COMMON to SAVE variables to allow the use of DATA for TABHDK and moved a SAVE statement in front of DATA statements in DATPOL. All of these errors were found with an ANSI switch on the SUN 4/110; the VMS ANSI switch did not report these violations of Fortran standards. Moved the AU1 fix to NRAO1, rest to 15JAN90 this date. 5435. November 14, 1989 BPASS Phil Modified so that the interpolation across the bandpasses is now a user option, the default is not to do it. Changed .HLP file as well. Also changed the array declarations which were (1) to (*) as spotted by the Sun 4 compiler. Moved to 15JAN90 this date. 5436. November 14, 1989 AIPS/SGLOCA Glen A user found that Save/Get does not find a match when more than 10 names have the same first few characters. SGLOCA now records up to 80 matchs and always detects an exact match to a requested Save/Get area. Expanded error messages. Moved to 15JAN90 this date. 5437. November 14, 1989 ALGSUB in MX, UVSUB P. Leahy/Glen In GRIDDED execution of MX and UVSUB, for the last several months, a series of horziontal stripes were added to output images, because at U=0 ALGSUB was performing a complex conjugation when it should have been performing AP I/O. Moved to 15JAN90 this date. 5438. November 14, 1989 VLBIN Phil Removed array declarations and a couple of intrinsic function calls that violated ANSI standard rules. Moved nowhere yet. 5439. November 14, 1989 VLBIN Phil When FFT'ing autocorrelation functions from the MPI correlator there was an error in calculating the ACF bias since the MPI software wrote the number of padded channels in the function not the number of channels containing data. Modified subroutine ACTRAN to run back through the data array to determine when the padding started and then calculate the bias using that. Moved to 15JAN90 this date. 5440. November 14, 1989 TV routines Eric Changed IMANOT, IAXIS1, and IENHNS to use YSLECT rather than YGRAPH to turn on/off graphics planes. The former uses YGRAPH but also sets the common variable TVLIMG which is used in TVFLG among other places. Moved to 15JAN90 this date. 5441. November 15, 1989 Number parsing Eric Changed GETNUM to require that there be a numeric character to be parsed in the specified string and to return an AIPS indefinite value (DBLANK) if there is not or if there is a floating overflow. Changed INQINT, INQFLT, and INQGEN to check for DBLANK and to return a negative error code if it occurs. This will help detect user format errors in interactive tasks and service programs. Moved to 15JAN90 this date. 5442. November 15, 1989 ALGSUB Glen Removed all complex conjugation near U=0 from ALGSUB. Gridded and DFT MX using ALGSUB now agree to 1 part in 10 million for the test image of a 4 Jy calibrator with 256*256 pixels. Moved to 15JAN90 this date. 5443. November 15, 1989 UVMAP Glen Increased buffer to allow 4096*2048 REAL maps, not just I*2 MAPS. Moved to 15JAN90 this date. 5444. November 16, 1989 Free-format inputs Eric Changed all programs that use free-format input through the INQGEN, INQINT, and INQFLT subroutines to test and use the new error return codes. Changed AIPLAS, AIPMAN, DELSG, FILAIP, FILINI, FIXCAT, FIXUSR, GR2TEX, GRITP, POPSGN, RDFITS, RECAT, SETPAR, SETTVP, and UPDAT. TVFLG remains out for other fixes. Moved to 15JAN90 this date. 5445. November 16, 1989 VLBIN Phil In calculating the AC bias and removing it the output array was not being zeroed properly and some crap was coming through. Fixed. Moved to 15JAN90 this date. 5446. November 16, 1989 SCREAT Eric Corrected error message to give ``too large'' disk sizes in 8-bit rather than ``AIPS'' bytes. Moved to 15JAN90 this date. 5447. November 17, 1989 ITICS Eric Added a call to TICCOR to correct tick mark lengths for declination. Lowered limits on lengths for small subimages. Moved to 15JAN90 this date. 5448. November 17, 1989 TVFLG Eric (1) Corrected bugs related to having fewer than 4 graphics overlay planes. It was possible for the menu to be erased without the menu routine being told about it. The correction to IMANOT, made earlier, fixes a bug in which the graphics plane was not always turned on when needed. (2) Corrected handling of interactive window setting --- made it accept ``error'' returns for cursor off the image as at the outermost pixel of the current image. I thought I had done this a year ago. (3) Added a dummy read of the cursor position to force the TV buffers to be flushed to the device. Otherwise, graphics planes were left on when the program did things like prompt for term,inal input. (4) Changed an initialization so that the current menu selection gets highlighted again when the user returns to the menu mode. (5) Changed to handle format errors reading terminal input using the new error codes of INQINT et al. This provides new exit methods from a couple of places. (6) Corrected TVFOAD to initialize the image catalog when it zeros the memory. (7) Corrected GRIDTB to compute the subarray number correctly; machine accuracy of 1 part in 10**8 or so was assumed. The data for subarrays > 1 were not gridded correctly. NOTE: this error is probably common throughout AIPS. (8) Corrected GRIDTB to set IRET error code on errors from GTBWRT; such errors went undetected by higher-level routines previously. Moved to 15JAN90 this date, nowhere else. 5449. November 17, 1989 ACFIT Phil Debugged. It now will write an SN table containing the full time variable amplitude calibration factors for all antennas selected, will print various useful messages upon request, and also update the HI file. Also wrote the .HLP file. Moved to 15JAN90 this date. 5450. November 17, 1989 UVFLG Phil In entry #5431 I missed one occurence of the setting of the number of sources. Fixed. Moved to 15JAN90 this date. 5451. November 19, 1989 TVFLG Eric Corrected bugs in TVFOAD that caused the load by baseline length to fail. When TVFOAD was rewritten for vectorization, the array indices for the second and third summing arrays were not modified for the baseline length, only the first was. Also changed TVFLG to display the subarray number in the CURVALU-like display while flagging, to use a faster alogorithm to convert from length number to baseline number, and to display current selection of length vs baseline order with an asterisk if the next load will switch that selection. Moved to 15JAN90 this date. 5452. November 20, 1989 TVFLG Eric A variety of changes: (1) Changed DOCAT so that a value of 2 with IN2SEQ = 0, means to return quickly, write the master grid file, and exit. This frees the user terminal during the lengthy creation of the restartable grid. (2) Added DOHIST option to control how much, if any, of the history is written. (3) Changed history times to sexagesimal format and limited the times in the history and flag tables to the start and stop times of the master grid. (4) Corrected overhaul bugs in the handling of strings giving the ``REASON'' for the flag and history files. (5) Corrected missing arguments to WRITE for some history cards and added a test on NPOINT greater than zero. The latter made a confusing message on restarts. (6) Reordered operations in the menu-selection code to give the user control of the cursor before the menu is written. Modified the help file accordingly and the include DTVF.INC. Moved history corrections only to 15JAN90, nowhere. 5453. November 20, 1989 ACFIT Phil When removing a polynomial baseline, there was no trap to determine if the whole spectrum was flagged. Also there was an error in the printing out of the gain factor, the value written into the SN table was fine though. Moved to 15JAN90 this date. 5454. November 20, 1989 Directory changes Bill C. Created a new branch of the AIPS directory structure for non-release specific text files: VMS directory Logical Purpose [AIPS.TEXT.IONS] AIPSIONS Ionospheric monitoring data [AIPS.TEXT.PUBL] AIPSPUBL old DOCPUBL directory [AIPS.TEXT.GRIP] AIPSGRIP old DOCGRIP directory [AIPS.TEXT.WHO] AIPSWHO old DOCWHO directory As a part of this change the contents of TST:DOCPUBL, DOCGRIP and DOCWHO were moved to the new directories. Also created was a directory AIPS_VERSION:[INC.LOCAL] given logical name INCLOC. TVMON was moved from AIPS_VERSION:YVTVPGM to a new directory AIPS_VERSION:YPGVDEV ( [Y.PGM.VDEV]) Moved to 15JAN90 this date. 5455. November 20, 1989 AIPSC Glen Modified AIPSC.FOR subroutine CHSTOP which called PRTMSG with the wrong number of arguments. Added string ' ' in call to PRTMSG which is the File name to save line printer output. Moved nowhere. 5455. November 21, 1989 Service programs Eric Changed GRKEY to offer the option of limiting the output to one or more keywords. Changed GTTEX.COM to give an extra blank line for this option to GRKEY and to specify AIPSGRIP: rather then DOCGRIP:. Changed in AIPGVMS: GRKEY, GETTBC, and GRSORT to specify AIPSGRIP: rather then DOCGRIP:. Changed in AIPGVMS: ASORT, DIRECT, HOK, MAKDIF, and SURVEY to specify AIPSWHO: rather then DOCWHO:. There are plenty of other places these changes should be made. However, Bill Cotton does not give a shit so I am doing no more work on this. Moved nowhere. 5456. November 21, 1989 APLNOT:BPASET.FOR Bill J. This routine didn't trap the case when both polarisations were flagged. On normalisation of the table, divide by zeros were demanded. This screwed up POSSM, HORUS and SPLIT amongst others. Moved to 15JAN90 this date. 5457. November 21, 1989 Midnight job directory changes Bill C. Changed UPDVMS:SRTUNQ.COM not to copy AIPSGRIP and AIPSWHO rather than DOCGRIP and DOCWHO. Moved to 15JAN90 this date. 5458. November 22, 1989 BLREFM Phil The opcode in the call to TABBL was wrong when reformatting the tables. Moved to 15JAN90 this date. 5459. November 22, 1989 TVFLG Eric Changed Fortran and Help: (1) To make FLAGVER default to 1, rather than the highest while reading the data and the highest+1 while writing the flags. (2) To write FLAGVER in the history file for all multi-source files; it used to do it only if calibration was applied. (3) Changed instructions in interactive flagging: buttons A and/or B cause a flag operation followed by a loop for more flagging choices and button D exits with no more flagging. (4) Changed interactive flagging so that button C means to do the currently indicated flagging and then return to the menu mode. The CURVALU-like display will always be used. (5) Changed interactive clipping to use buttons A and B only to switch between max and min, button C to do the clipping before returning to the menu, and button D to return to the menu with no clipping. (6) Added a counter of number of master grid cells newly flagged by a command. This is printed as a warning if 0 or as information if greater than zero. Moved nowhere. 5460. November 22, 1989 Subarrays Phil There was a potential problem in the way in which we were determining the antenna numbers and subarray number. If significant rounding errors were occurring the wrong value would be picked up from the random parameter array. The following subroutines and tasks were changed: DATFLG, DATGET, DGETSD CLIP, CORER, FUDGE, UVFND, UVMTH, ACFIT, ASCOR, AVER, AVSPC, BLOAT, BSMAP, FILLM, INDXR, LISTR, MULTI, POSSM, USUBA, UVAVG, UVCOP, UVDGP, UVFIT, UVFIX, UVIMG, UVMOD, UVPLT, VBCAL, VBMRG, WTMOD, ASCAL, BLCAL, BPASS, CALIB, CVEL, PCAL, VSCAL Moved to 15JAN90 this date. 5461. November 22, 1989 MK3TX Chris Now recognizes filenames beginning with `[' to be text files (log files are stored with filenames of the form `[LOG??'). Also increased the width of the numeric format in the summary line to cope with the fact that MK3 logs typically contain more than 1000 lines. In addition MK3TX no longer treats it as a fatal error if a file header does not contain the expected filename but asks the user whether he still wants to extract it since a few files from Bonn appear to have garbage in the file name field. Finally corrected to priority of an informational message which was being output at priority 8. Moved nowhere. 5462. November 24, 1989 YPGVDEV Eric Renamed the link-editor file SYSVMS:YVTVPGMOPT.OPT to YPGVDEVOPT.OPT. This is required to allow the task to link under VMS. Moved to 15JAN90 this date. 5463. November 24, 1989 TV with Windows Eric Changed AIPS TV model to allow for there to be a rectangular window as the view port into the (larger) TV memory. Changed DTVC.INC Added TVWIND(4), TVDUM2(4), reduced to YBUFF(160). YTVCIN (YGEN) Init new parameters, changed YBUFF fill. Also YDEA:, YM70:, YM75:, YIVAS:, and YSSS: versions. YWINDO (YGEN) New routine to READ the window (and we hope later to write it too). Generic just returns the image memory size. YWINDO (YVTV) New routine to read/write over the wire. TVMON Added interface for new YWINDO routine. AU6 Changed TVZOOM to center in window and reread window size on button pushes. Changed TVSCROL to center in the window. Changed TVPSEUDO to use the window center and to scale everything by the window size and to allow the window to change on each button push. AU6A Changed TVSPLIT and TVBLINK to start the split screen in the middle of the window. AU6B Changed to display the text based on the window at the start of the CURVALUE algorithm. BLANK Changed to start interactive bits from center of window. Image loading is based on TV image memory size as always. BLSUM Changed to start interactions at center of window. GRBOXS Changed to start each box BLC at the center of window. GRLUTS Scale the plot to the window, not the TV memory. GRPOLY Start the cursor in the middle of the window. HIENH Start the cursor in the middle of the window. HILUT Changed the LUT slope and offset to be based on the cursor position within the window. IENHNS Changed the LUT slope and offset to be based on the cursor position within the window. Changed the plot to be placed inside the current window. TVBLNK Changed to base blink rate on current window. TVFIDL Changed to do zoom, LUT, and OFM work relative to the current window. Button hits reread window size. TVMOVI Changed to base rate, enhancement, frame selection relative to the current window. Unable to decide what to do with TVROAM, TVHXF, and TVHLD. None of these work on current windowing machines, so just left them alone. Dropped unused include of DTVC.INC from SET1DG, ZARGXF, ZM70XF, ZV20XF, TKPL, XBASL, XGAUS, XPLOT, LTESS, STESS, TXPL, TKSLIN, and IMERG. Cleaned up typing in TVWIND. Changed DISPTV, APCLN, APGS, APVC, MX, SDCLN, and STEER to put cursor up at center of window rather than at center of image memory. Moved nowhere. 5464. November 25, 1989 SSS windows Eric Changed the SSS TV implementation to support the windowing scheme. Changed: SSS.TXT (SYSSUN:) Added OPCODE 14 to read/write window. HEADER.H (YSVU:) Dropped declaration of cursor scaling parms and interrupt handler. Added WINDO (14) opcode. SSS.C (YSVU:) Dropped init of cursor scaling parms. CURSOR.C (YSVU:) Dropped cursor scaling and {\tt cursor_choice} interrupt handler. Added a routine to force and to return the window size. Added checks on excess window size, forcing it back to coded maximum. WINDOW.C (YSVU:) Dropped creation of panel buttons for cursor scaling. SOCKXF.C (YSVU:) Added handling for new opcode. Moved from MANDRILL this date after tests. 5465. November 25, 1989 Bugs Eric Corrected minor bugs in AU6B, GRBOXS, and TVMOVI affecting things when the window is smaller than the full screen. Also fixed error handling in YTVOPN (YGEN:) --- and error in YWINDO would leave the device partly open. Moved from MANDRILL this date. 5466. November 26, 1989 TVFLG Eric Changed TVFLG to use the YWINDO parameters as well in the various interactive enhancements, the CURVALU-like displays, and even in the placement of the main menu. Improved the flexibility in the dynamic range of the interactive clip. Moved nowhere. 5467. November 27, 1989 SNPLT Phil Changed the tolerance slightly in defining the min and max delays to be plotted when there is no variation. The value used caused a fl.pt exception when the delay was constant and non-zero. Moved nowhere. 5468. November 28, 1989 ACFIT Phil Was writing the number of subarrays into the # antennas slot in the SN table header. Moved to 15JAN90 this date, 5469. November 28, 1989 FILLM Glen Fixed Corruption of LP band observation of the Sun by subroutine MCSOLR. Moved to 15JAN90 this date. 5470. November 28, 1989 FITTP Phil Removed a 'feature'; when FITTP read the FQ table it looked to see if there was only one row, if so it translated it to a CH table so older versions of AIPS could read it. The translation routine checked to see if there was only one IF and if so wrote nothing. Normally this is not a problem, but for some line experiments it means a frequency offset is lost. Modified so that a CH table is now always written. Moved to 15JAN90 this date. 5471. November 28, 1989 TIMDEST Bill J. Incorrect arguments in call to PRTMSG in AIPSUB:AU3A.FOR caused AIPS to crash when TIMDEST was exercised. The incorrect call arguments had not reflected the addition of an output file name for listing in subroutine PRTMSG. Moved nowhere. 5472. November 29, 1989 UVFIX Phil An index was wrong in the writing of compressed data. Moved to 15JAN90 this date. 5473. November 29, 1989 VLBIN Phil Increased the sizes of some arrays to cope with 288 channel autocorrelation spectra from the MkII correlator. Moved nowhere. 5474. November 29, 1989 DDTTEST.TEXT Bill C. Added a documentation file to DOCTXT to be sent with the DDT tapes. The current version refers to 15APR89. Moved nowhere. 5475. November 29, 1989 TABBL Phil Wrong index in array lead to incorrect number of IF's being written/read into/from BL tables. Moved to 15JAn90 this date. 5476. November 29, 1989 TACOP Bill J. If the 'OUT' file specification was set to defaults including OUTDISK=0, then the table to be copied was copied to the first disk with a suitable file - usually the wrong file! Trapped this and forced write to input file. Moved nowhere. 5477. November 30, 1989 DCMP.INC,VISUNP,VISPCK Phil New include and two simple utility routines. Include contains information needed for compression and decompression of vis. data. VISUNP will decompress data from one buffer to another, VISPCK will compress it in the same manner. Moved nowhere. 5478. November 30, 1989 YWINDO Eric Changed the generic YWINDO to return IERR = -1 on WRIT operations, meaning that writes cannot be done on the present device. Returns IERR = 2 and a message, on bad opcodes. Moved nowhere. 5479. December 1, 1989 TVFLG Eric Corrected GRIDTB: when it was converted for vectorization, I suspect, the option to put the data in both as Ant1-Ant2 and as Ant2-Ant1 was subverted with no data going in the latter. Changed GRIDTB, TVFOAD, and TVFLG to keep time in the data row, changed TVFOAD to make a list of the actual times by row number in the TV disk file, and changed TVFLG to use that list in displays and in setting flag times. Changed GRIDTB and TVFLG to keep a sort of ``weight'' with each datum in the master grid. This weight is zero for good points. Positive ``weights'' give the flag sequence number which deletes this datum. Changed to accept up to two related Stokes parameters in the same master grid and to allow switching between them. Acceptable STOKES are all single ones (I, Q, U, V, RR, LL, RL, and LR) and double ones IV, QU, HALF, RLLR. Both FULL and RRLL are taken to be HALF also. Fixed help file for the Stokes change, including the change in menus. Fixed TVFLG to keep a separate Stokes flag pattern for each flagging command, rather than a single one for all commands. Moved GRIDTB fix to 15JAN90, nothing else. 5480. December 1, 1989 Misc Eric Changed AIPINI to avoid disk access when called by FILAI2. Changed APLVMS:ZMKTMP to do a modulo 1000000 on the process ID. Some VMS 5.n systems can have very high process IDs. Moved to 15JAN90 this date. 5481. December 4, 1989 Makefile for SSS Chris Added a makefile for SSS to directory YSVU. The Sun Screen Server may be rebuilt by typing make -f MAKEFILE new which recompiles all the SSS modules, link-edits SSS and moves it to LOAD. If `new' is omitted then only those modules that have been changed since the last rebuild will be recompiled. Macros must be defined in the MAKEFILE to supply the floating-point option for the linker (FPOPT) and the location of the AIPS CC script (CCDIR). The latter should not be necessary but the Sun make utility seems to lose the value of the command-search path (environmental variable PATH) --- I don't know whether this is a bug or a feature (Sun's make is non-standard). Also changed step (5) of the installation procedure in SSS.TXT (in directory SYSSUN) to reflect the existence of a makefile. Moved from the AOC Suns this date. 5482. December 4, 1989 TVFLG Eric Corrected serious bugs in new version: Changed dimension of time list in DTVF.INC and added check on that dimension in TVFOAD to avoid exceeding the array limits. Corrected TVFOAD to leave a pointer fixed when averaging data; it updated it on every included row, making for bad times and bad data locations. Changed TVFLG itself to give more indications of when current parameters will change on the next LOAD in the status line(s) given at the bottom of the menu. Moved from the VLA this date. 5483. December 5, 1989 ZTKCL2 in APLBERK Kerry At some point in the evolution of SunOS 3.x, the disabling of exclusive use upon closing a terminal device quit working as advertised (i.e., closing the device should be sufficient). This meant that the second execution of a Tektronix oriented task or verb would still get a "device busy" error. This has been averted now by adding an {\tt ioctl} call with TIOCNXCL on the device to explicitly relinquish exclusive before closing it. This should not be necessary, didn't used to be necessary and, at least under Convex OS 7.1, still isn't necessary, but should not cause a problem either. Change in 15JAN90 same date. 5484. December 5, 1989 SL2PL, plotting Eric Corrected LABINI, SLBINI to use the passed length of their input character strings (as *(*)) rather than declare 80. This messed up SL2PL which had used length 132 strings. Cleaned up typing in SL2PL and removed improper declarations of dimensions as (1). Moved to 15JAN90 this date. 5485. December 5, 1989 OPCODE.HLP Eric Changed adverb help file to describe every task and verb which uses it including listing all allowed values of OPCODE. Moved to 15JAN90 this date. 5486. December 6, 1989 ZMOUNT, ZMOUN2 Bill J. Added argument NLUSER in call to ZMOUN2 from ZMOUNT specifically so that identification of tape allocation by AIPS user number could be made on CONVEXes. Moved nowhere. 5487. December 5--7, 1989 Times in history Eric Changed the recording of the TIMERANG adverb in history files. All current code assumed that only the fourth and eighth parts of TIMERANG were non-integer, and most have odd ways of recording in history anyway. So changed CORER and UVCOP to make more sensible displays. Wrote a new subroutine HITIME to format TIMERANG to HI cards in both man- and machine-readable forms. Changed ACFIT, BLCAL, BPASS, CALIB, CLCAL, CLCOR, CSCOR, CVEL, FARAD, FILLM, GETJY, HORUS, MK3IN, PCAL, QUACK, SDCAL, SOLCL, SPLIT, TABED, TVFLG, USUBA, UVIMG, and VLBIN to call HITIME. Sent the fixes to CLCOR to Bill Cotton, the fixes to CVEL and UVIMG to Phil Diamond, and held back TVFLG for the moment. Put the rest back. Moved nowhere. 5488. December 7, 1989 CGASET Bill C. Fixed an indexing problem for computing baseline dependent corrections for fringe rate errors. This would cause problems for MKIII VLBI data only. Moved to 15JAN90 and 15OCT89 this date. 5489. December 7, 1989 UVMAP.HLP Bill J. Tweaked HELP file to remove an infelicity concerning UVBOX. Moved nowhere. 5490. December 7, 1989 CLCOR.FOR, CLCOR.HLP Bill C. Added an option to correct CL table for an antenna position error. Delay, rate and phase are corrected. Moved nowhere. 5491. December 7, 1989 ZSPACE.C Bill J. This subroutine calculated size of users' files in kiloblocks using 'st_blocks' assuming that this is returned in units of 1024 byte blocks. It isn't! This makes for nonsense disk usage in the utility "SPACE". Modified this to use 'st_size' which is in units of bytes so that we really know where we are. Moved nowhere. 5492. December 7, 1989 AU3A Phil Modified a call sequence to PRTMSG: DOCRT was being set to false, but the routine expected a real, this caused a bus error on the Convex. Moved nowhere. 5493. December 7, 1989 TVFLG Eric Corrected the addressing to the TV memory and to the master grid when flagging data with the clip option. Moved from the VLA. 5494. December 8, 1989 GPOS Kerry/Bill C. Fixed zero divide problem. Moved to 15JAN90 this date. 5495. December 8, 1989 SNPLT Kerry/Bill C. Fixed zero divide problem. Moved to 15JAN90 this date. 5496. December 8, 1989 Documentation stuff Bill PUTBCK the files used for the AIPSletter: AIPSPUBL:let89d.tex, al82.mac, order.tex,userform.tex(the user agreement form), and letform.tex (template AIPSLetter). Also various files related to the AIPSSERV distribution: - DOC:README. General description of the files destributed by AIPSSERV. - AIPSPUBL:CHANGED.89D CHANGE.DOC for the 15OCT89 release. - DOC:VMSNOTES.89D VMS installation problem notes for 15OCT89. Moved nowhere. 5497. December 8, 1989 PRTAN Glen PRTAN now produces a shorter output if antenna mount, feed polarization and axis offsets are the same for all antennas. Also produces a VLA "Y" plot if all antennas are VLA antennas Moved nowhere. 5498. December 7--9, 1989 IF and channel numbers Eric Corrected sloppy/wrong handling of IF numbers supplied by the user in ACFIT, BPASS, CLIP, CSCOR, GETJY, HORUS, LISTR, MX, PCAL, POSSM, SDCAL, SETJY, SNPLT, SOLCL, SPLIT, TVFLG, UFDIF, UVFIT, UVFLG, UVFND. Corrected sloppy/wrong handling of channel numbers supplied by the user in ACFIT, AVSPC, BPASS, CALIB, CLIP, HORUS, LISTR, POSSM, PRTSD, SDCAL, SPLIT, TVFLG, UVCOP, UVDIF, UVFIT, UVFLG, UVFND. Corrected typo in handling the IF for history writing in GRIDR. Passed on changes of this sort in UVSUB and UVIMG to Phil who has them checked out and in VBPLT to Bill Junor for the same reason. Moved nowhere. 5499. December 8, 1989 GREYS Eric Corrected bug affecting plots where the two images did not identical header coordinates and dimensions. The contour image header was used for the axis labeling which would produce a shift of the labels with respect to the plot. A call to SETLOC in this case could also produce trouble --- so saved not only the header but the whole location common and restored them at the end of GLAB. Moved to 15JAN90 this date. 5500. December 9, 1989 Plot jobs Eric Corrected CNTR, PCNTR, and PROFL to restore a standard location common before calling the star plotting routine. Non-default labeling options would have made the stars be plotted incorrectly. Also made this change in KNTR, but, since it loops over planes, also had to restore the labeling version of the common after the call to STARPL. Moved to 15JAN90 this date. 5501. December 9, 1989 Midnight job Eric Changed UPDVMS:COPY.COM to handle the new directories called AIPS_ROOT:[AIPS.TEXT...]. Before, it only did names with AIPS_VERSION correctly. Moved from VAX1 this date to 15JAN90 and 15APR90. 5502. December 9, 1989 IMHEAD, QHEAD Eric Changed LSTHDR and KWIKHD to make it clearer that the header extension file number if the maximum version number. The old wording suggested that all lower version numbers existed which is not a requirement. Moved nowhere. 5503. December 9, 1989 PRTIM Eric Removed all limits on FACTOR from the help file. Moved nowhere. 5504. December 9, 1989 WTMOD Eric Cleaned up the typing some and changed it to avoid unflagging data. Previously negative weights are altered, but remain negative. Changed help file to clarify this. Moved to 15JAN90 since this is what users expected. 5505. December 9, 1989 BLANK Eric The TV loading function type was left out of the input parameters common causing BLANK to use the function type as an X increment in TV loads. The function never was converted from HOLLERITH to CHARACTER either although the CHARACTER form was given to TVLOAD via the image header. Moved this fix only to 15JAN90 this date. 5506. December 10, 1989 ZMOUN2.C Bill J. I made a error in the filename in the DISMOUNT command which caused the failure of "DISMOUNT". Expansion of filename in 15APR90 only so moved nowhere. 5507. December 11, 1989 LOGIN.CSH and LOGIN.SH Kerry Added the environment variable definitions for the new non-version specific directories (i.e., AIPSIONS, AIPSGRIP, AIPSPUBL and AIPSWHO). Added {\tt limit coredumpsize 0} to LOGIN.CSH so that {\tt core} files are never generated (don't know how to do this in LOGIN.SH for Bourne and Korn shells). Also ``included'' environment variable definitions for local devices, however, this may saturate the environment on some systems. Change in 15JAN90 same date. 5508. December 11, 1989 $APLCVEX/ZMOUN2.C Bill J. Tweaked tape logical name to be AIPSn.USERNOxxxx.INTAPEm so that the forced dismount on exit from AIPS made more sense. Moved nowhere. 5509. December 11, 1989 UVFND Eric Added two new options: (1) WTUV to give the maximum allowed weight, with visibilities haveing greater weights always printed. And (2), APARM(5) > 0 means to check also the Hermitian box in 'UVBX' operation. Corrected ``bug'' in the VCLP operation: APARM(3) was being used as the lower V-flux limit. When a user said APARM = x to AIPS, then all samples were printed (being either greater than x or less than x). Moved nowhere. 5510. December 12, 1989 DBCON Phil Minor fix in DBCON. Under some circumstances it could blow up when attempting to determine the source number for single source compressed data if the array element it was looking in was a Nan. Changed so it only determines source numbers if it has to. Moved to 15JAN90 and 15OCT89 this date. 5511. December 12, 19899 $APLCVEX/ZMOUN2.C Bill J. Forgot to switch the variables around appropriately. This is what happens when fixes are put it at midnight! Moved nowhere. 5512. December 12, 1989 UVFIT Glen Incresed Vis buffers for UVFIT to 25000. Moved nowhere. 5513. December 12, 1989 UVIMG, GRIDTB Eric Changed UVIMG and GRIDTB to set the error code to 0 before calling DATGET. There is some way in which the error code is sometimes not reset by DATGET and this was causing UVIMG to quit when the first point not on the grid was encountered. Moved to 15JAN90 this date. 5514. December 12, 1989 TVFLG Eric Corrected addressing error in recording the baseline number in the data arrays for later flagging. The divide by 2 for 2 values / cell had to become 3 now that weights are saved. Moved nowhere. 5515. December 13, 1989 CALREF Glen Bad data caused divide by zero. Additional tests of data inserted. Moved nowhere. 5516. December 13, 1989 LISTR Phil Now explicitly checks to see if NX and SU tables are present when doing 'SCAN' listing. Moved nowhere. 5517. December 13, 1989 SPLIT Phil When copying the FQ table to the output file the offset for IF's was not being adjusted correctly, except for the 1st IF. This was causing problems later since the u,v,w's for IF's > 1 would then be calculated wrongly. Moved to 15JAN90 and 15OCT89 this date. 5518. December 13, 1989 FILLR Bill J. The LOCAL INCLUDE 'FILLR.INC' had not been modified to reflect the inclusion of the FREQID column in the NX table. This caused FILLR to generate a SU table with SOURCE ID = 1 for all sources in the file. Moved to 15JAN90 this date. 5519. December 14, 1989 SWPOL Chris Rewritten to take care of the general case, where polarizations are swapped at a subset of the antennae in an array. SWPOL now handles compressed data and linear polarization. It also optionally rearranges calibration data in SN and CL tables. SWPOL is no longer capable of overwriting the input file and will no longer handle the case where two parallel hands are present but only one cross-hand (this is rather messy in the general case). Moved to area APGNOT this date. 5520. December 14, 1989 SPLIT Phil For some inexplicable reason SPLIT was not using the freq. offset from the SU table when writing the output file header and rescaling the u,v,w's. The frequency is now updated correctly for each source in SPLCOP. Moved to 15JAN90 and 15OCT89 this date. 5521. December 14, 1989 UVPLT Eric Corrected bug positioning the time-date string on the plot. It was one character too high. Moved to 15JAN90 and 15OCT89 this date. 5522. December 14, 1989 SSS Eric Changed YSVU: HEADER.H, CURSOR.C, and WINDOW.C to make the ``Resize'' panel button do something useful. It will now pos between a full screen (actually a bit less, leaving some text window at the bottom) and the currently set smaller window including both size and location. Moved to SUN4 and SUN3 in CV, nowhere else. 5523. December 15, 1989 TBIN Glen TBIN now handles flagged data correctly. Added subroutine TABREQ to APLNOT and modified TABAXI and TABHDK. These subroutines do not pass arguments in common and are usable in other Programs. Moved to 15JAN90 this date. 5524. December 18, 1989 SSS Eric Changed WINDOW.C to create the ``Resize'' button first. That way it will still be visible on very tiny TV windows. Moved nowhere so far --- SSS will need rebuilding. 5525. December 19, 1989 FIXED.DOC Kerry Fixed UNIX documentation regarding setting area logicals. Moved to 15OCT89 December 20, 1989. 5526. December 20, 1989 TEC data Chris Checked in the Boulder TEC data from 1980 to mid-1989. Files are kept in the directory AIPSIONS and follow the naming scheme TECB.yy, where yy is the last two digits of the year. Also described these files in EXPLAIN FARAD. Moved nowhere. 5527. December 21, 1989 YGYHDR for IIS model 75 Claire Russell/Kerry Changed ``IF (IANGLE.EQ.1) THEN ...'' to ``IF (IANGLE.EQ.0) THEN ...'' in the logic that determines which bits to turn on in the TV subunit header (apparently a error introduced during the conversion of an ``IF ... GO TO ...'' construct to an ``IF ... THEN ... ELSE ..."). Changed in 15JAN90 same date, 15OCT89 December 22. 5528. December 22, 1989 UVPGET Bill C Fixed problem in format 1030. This problem was causing programs to blow up if there was a problem with the random parameters in a data file. Moved to 15JAN90 this date. 5529. December 22, 1989 UVFIX Bill C A logic error caused the UT1-UTC and IAT-UTC corrections set by the user to be ignored. Moved to 15JAN90 this date. 5530. December 27, 1989 TVFLG Eric Changed DTVF.INC to add FC tables, GRIDTB to correct time computations for planes greater than one, TVFOAD to correct time setting for rows with no data, and GTBWRT to remove a blank line. In TVFLG itself: (1) Added FC table to grid file. This saves all Flag Commands to protect against aborts, TV failures, power failures, and the like. Also, it allows deferring the application of the commands to data until a later execution (and allows applying the commands to more than one UV file). (2) Added options to LIST, UNDO, and REDO the flag commands. The last is needed to make sure that flagging of the master grid is consistent after some intermediate flag commands have been UNDOne. (3) Broke the CLIP flagging into a separate subroutine, which allows it to be REDOne as well. (4) Changed to pick up and use the master grid times in an array in memory. This anticipates non-regular grids. (5) Changed to check current ``Stokes flag'' and to flag all planes in the master grid file that are affected. When setting the Stokes flag, it checks to see if the current data (to be loaded next) is included. Removes data from the TV and the TV data file only if the appropriate Stokes value is included in the Stokes flag. (6) Corrected flagging of the master grid file when doing boxes. (7) Corrected to reinitialize the graphics plane after a TVTRANSF operation. Changed TVFLG.HLP to reflect the various improvements. Moved from MANDRILL to CVAX and NRAO1 this date. 5531. December 27, 1989 TVFLG Eric Changed TVFLG to use IMCHAR to clear characters in the highlight graphics plane rather than YZERO. The latter is slow on the IVAS. Also fixed it to reset the graphics plane 2 string before exiting the menu. Corrected GRIDTB to check for input uv data sets with only one Stokes and to use a correct Stokes for the grid file. Corrected TVTRANSF display to zero the graphic planes used and to set a logical to recompute the menu only when there is just one graphics plane. Added code to set BCHAN, ECHAN, BIF, EIF on restarts. Corrected GRIDTB, it was trying to loop over too many channels and IFs. Moved from MANDRILL, test area on NRAO1. 5532. December 27, 1989 Type Declarations for EXTERNAL Routines Kerry In addition to being declared EXTERNAL, such routines must also be given a (dummy) type declaration. Otherwise, ``debug'' compilations (i.e., ``-g'') with the old ``bundled'' Sun f77 compiler will fail with the declaration error ``can't get dbx basetype information''. Sun's new ``unbundled'' FORTRAN 1.x does not have this problem. In the APLPGM program XGAUS, an INTEGER declaration has been added for the EXTERNAL subroutine FCN (passed as an argument to the routine LMSTR1). Also, in the APLSUB routines LMDER and LMDER1, an INTEGER declaration has been added for the routine argument FCN. Changed in 15JAN90 same date. 5533. December 28, 1989 AIPSUB:AU8A Colin/Kerry In the section for listing info on plot extension files from VBPLT, the GO TO 303 should be GO TO 302. Otherwise, missing antenna pair data causes loop 302 to terminate, perhaps prematurely. Changed in 15JAN90 same date. 5534. December 28, 1989 APLBELL:ZLOCK Colin/Kerry The check for unavailability from a lockf() error return should be against EACCES, not EACESS. Changed in 15JAN90 same date. 5535. December 28, 1989 APLNOT:TABBP Colin/Kerry GO TO 50 should be GO TO 60. Changed in 15JAN90 same date. 5536. December 28, 1989 APLUNIX:ZABORS Colin/Kerry The generic UNIX version of the abort handler was trying to use the external character array "sys_siglist" to extract the meaning of the caught signal, however, Bell UNIX systems have no such list. If they do it's got another name. Changed to simply report the number code of the caught signal. Also changed the messages issued by the routine to say "ZABORS:" instead of "ZABOR2:" (i.e., the pre-overhaul name of this routine). Also replaced the call to abort() with a call to kill() such that the process sends an illegal instruction to itself. There is no standard signal issued by the abort() function. For example, Sun issues SIGIOT whereas most other implementations issue SIGILL. IOT and EMT are PDP-11 instruction mnemonics and therefore not issued on Suns, so Sun chose to use it for this purpose. This is not totally screwy. POSIX will probably replace SIGIOT with SIGABRT. Changed 15JAN90 same date. 5537. December 28, 1989 APLUNIX:ZFRE2 Colin/Kerry Replaced the call to index(), which is Berkeley UNIX specific, with a trivial, portable loop. Changed in 15JAN90 same date. 5538. December 28, 1989 ZCREAT Eric Cleaned up the error messages in APLGEN: version of ZCREAT. Moved nowhere. 5539. December 30, 1989 UVDIF Chris Corrected declaration of LPNAME to be 48 characters. Moved to 15JAN90 this date 5540. January 1, 1990 TIMDEST, ZOPEN Eric Changed AU3A to send a blank character string rather than an integer 0 as the task name to PRTMSG to delete messages. The Convex anyway, took it to be some (nonexistant) task name. Changed ZOPEN to declare the file name as character *(*) rather than *48. This is important for dummy names and the like. Moved changes to 15JAN90 this date. 5541. January 2, 1990 PRTAB Bill C. Increased the maximum number of keyword/value pairs to 1024 from 128. Moved nowhere. 5542. January 2, 1990 TABED Bill C. Increase the maximum number of Keyword/value pairs allowed when creating a new table. Also now checks that the maximum is not exceeded. Moved nowhere. 5543. January 2, 1990 APLUNIX:ZCMPR2 Colin/Kerry Fixed error in null terminating the translated logical part of the physical file name (i.e., changed "xlated[tlen]" to "xlated[xlnb]"). Also, changed PMODE to 0666 (i.e., owner, group & world readable & writeable), replaced obsolete reference to the former name of this routine (ZQTRUN), removed calls to "sync" (synchronizes the super- block) and changed the calculation of the number of bytes in an AIPS block to "256 * sizeof (int)". Changed in 15OCT89 and 15JAN90 same date. 5544. January 2, 1990 APLBELL:ZABOR2 Colin/Kerry This is a new, Bell UNIX specific version. It differs in that the signal SIGCLD (death of a child) is ignored. Under Bell UNIX, ignoring SIGCLD prevents terminated child processes from becoming zombies that are then erroneously reported by the task query mechanism (i.e., ZTACT2) as still active. The only problem with this is that, although it is still supported on many System V-based systems (e.g., HP-UX, Cray UNICOS, Masscomp RTU), SIGCLD is not part of the SVID (System V Interface Definition) and programmers are discouraged from using it in new products. The POSIX draft says "applications should set the action for this signal to SIG_DFL in order to ignore it"??? Moved to 15OCT89 and 15JAN90 same date. 5545. January 4, 1990 FQMATC Eric Changed it to give its ``setting FQ id number'' message only when there is more than one possibility. Moved nowhere. 5546. January 4, 1989 NXTFLG Eric Corrected error. To decide whether to include a given FG table entry it was searching a list of 30 included source numbers if any had been specified (via SOURCES, CALCODE, or however). It did not check for more sources if there were more and it did not check for the flag being source independent. Changed to check source independence first and then to test the actual number of source IDs in the list when the flag is source dependent. The full calibration system must be relinked. Moved to 15JAN90 this date. 5547. January 4, 1990 TVFLG's friends Eric Changed and corrected: DTVF.INC Added new adverb for CALCODE, changed to allow 16384 rather than 8192 times, added DOWEDG, added list of source numbers by master-grid row and put the max source number in common. GRIDTB Added list of begin times of each row to call sequence and option to compute the list at regular intervals inside GRIDTB. Changed to store 2 times per row. Changed to use time list rather than regular time intervals for gridding. Changed calls to GTBWRT. GTBWRT It was averaging over much too large a write buffer and thereby writing over a lot of core. It took the Convex with 8 spectral-line channels to reveal this. Added list of start times for each row to call sequence. Changed to store two times from this list in each row rather than averaging sample times. Corrected to find max and min over all buffers rather than testing buffer 1 Nbuffers times. TVFOAD Changed to put correct units in the TV file header. Changed to compute a list of start times for each TV-file row based on the integration time and avoiding any averaging of different sources. Changed to use 2 times per row in the master and TV files. Added call to load optional step wedge. UVIMG Added time array to call to GRIDTB and code to have GRIDTB compute it as a regular array. Moved nowhere. 5548. January 4, 1989 TVFLG Eric Changed: (1) Added CALCODE adverb to allow selection by calibrator code including '-CAL' for no calibrators. (2) Changed it to grid the master grid file at as regular grid intervals as possible, but to allow it to be irregular to avoid running sources together or having many consecutive empty rows. Changed the TV gridding to do the same. (3) Added an option step wedge in the display. (4) Added ``CLIP BY FORM'' operation. This allows you to apply the ``form'' of a clip operation, i.e., the averaging time, type of data clipped against (e.g., rms), and the clip limits, to a selected set of other channels and IFs using the data of those channels and IFs. Code changes: added code to compute times for master grid before calling GRIDTB. Corrected seting of UV file parameters to be before the call to SOUFIL. Improved the times used to be in more pleasing intervals if possible (limits set at integer 1 or 10 seconds). Changed to 2 times and source number in each row of master grid and 2 times in each row of TV file. Added some times for rows after the data --- all times are now based on semi-regular intervals rather than an average of the sample times. Added TVFWED subroutine to do step wedge writing and erasing. Added subroutine TVFBOX to replace the more general and unforgiving GRBOXS. TVFBOX limits the box display and parms to the range actually on the TV, allowing easy graphical setting of windows. Added double check and warning when exiting TVFLG without applying the flags while the master grid is a scratch file. Corrected display at bottom of master grid --- the wrong asterisk was checked for baseline ordering. Corrected CURVALU-like display to blank source name when the row has no source name and to display a line only if it has changed (making the action peppier). Cleaned up unused variables and statement labels. Improved error messages. Changed meaning of times to beginning time of row, thereby using averaging of this row and next as the average time. Corrected CLIP limit read to use the required double-precision argument to INQFLT. Changed LIST FLAGS and UNDO FLAGS to inquire about flag number only when there is a choice. Moved nowhere. 5549. January 4, 1990 DOC:UNIXNOTES.89D Kerry Created file containing solutions to several problems encountered by sites while installing the 15OCT89 release on UNIX systems (mostly Sun VTV, abort handler, Tektronix and tape interface issues). Made a hard copy for Ernie to include in any remaining or future orders for 15OCT89. Moved to 15OCT89 same date. 5550. January 5, 1990 TABBP Bill C. Cleaned up typing and some declarations. Moved to 15JAN90 this date. 5551. January 5, 1990 BLGET Bill C. Cleaned up a problem with nested DO loops with a common termination and GO TOs to loop terminator. This caused trouble on HP compilers. Also cleaded up typing and declarations. Moved to 15JAN90 this date. 5552. January 5, 1990 EXTINI, TABINI Eric When a file's catalog slot is renumbered, the file names stored in the headers of the old and new table format files are not changed. Thus, when TABINI and EXTINI open an old table, they need to force the name in the header to match the current name. Changed it to do so. Relink the system. Moved to 15JAN90 this date. 5553. January 5, 1990 Logical version ID Eric Changed things to determine the logical version identification (TST, NEW, OLD) of the current execution and to record that in the ACcounting file. Changed: DMSG.INC Renamed a dummy variable MSGVER. ACOUNT Changed to call ZVERSN, put MSGVER in ACcount file, display value as string on task termination. ZVERSN APLGEN: version to determine the logical (NEW, OLD or TST) matching the RLSNAM string. PRTAC Changed to support free format inputs, a staff list of user numbers, sum and loop over versions, and numerous minor fix ups. PRTAC.HLP Added DOALL to choose between lumping all versions together and displaying them separately. STAFF.001 Put a list of NRAO staff user numbers in RUNSYS for use by PRTAC (in its stand-alone mode). Need to relink system as for NXTFLG and TABINI/EXTINI anyway. Moved nowhere. 5554. January 8, 1990 TVFLG Eric Corrected bugs that: (1) caused the ANTENNA-DT mode to clear the wrong TV pixels although the correct pixels were cleared in the TV and master files, (2) fail to flag anything in the master file on the PIXEL mode (telling you that it failed for a GETIME error), and (3) displayed only the begin time for PIXEL flags in the SHOW FLAGS function. Moved nowhere. 5555. January 8, 1990 SNPLT Chris Fixed a bug in the time axis scaling: the difference between the maximum and minimum on the y-axis was being used instead of the difference between the maximum and minimum on the x-axis. Changed in 15JAN90 this date. 5556. January 9, 1990 ANCAL Chris ANCAL has been modified so that it can calibrate more than one IF in a single run thus making the calibration of MkIII VLBI data much less tedious. Two new keywords, BIF and EIF, have been added to the control card to give the first and last IF for which calibration information is present; IFNO is retained for compatibility with older calibration files (IFNO=n is equivalent to BIF=n EIF=n). T-sys entries now consist of a day number, a time and NIF T-sys values where NIF is the number of IFs to be calibrated: the first T-sys value applies to IF BIF, the second to IF BIF+1 and so on. In addition ANCAL now checks whether gain entries in the CL table are invalid (equal to FBLANK) before multiplying them by the calibration factor. Files changed are APGNOT:ANCAL.FOR APLNOT:TYINI.FOR APLNOT:TABTY.FOR HLPFIL:ANCAL.HLP Moved nowhere. FBLANK bug fixed in 15JAN90 this date. 5557. January 9, 1990 PRTUV Bill C. Fixed problem that occured when more than 30 sources were present in a uv data file. Now will work on compressed data. Also restructured the task and divided it up into a number of subroutines. Moved nowhere. 5558. January 9, 1990 AIPSUB:GETFLD Bill C. Moved error trap before processing output variable from GETNUM. In the case that a character string starting with "." was typed in (specifically ".false.") GETNUM was detecting a bad value and setting the output to DBLANK which was overflowing a single precision word on IEEE machines; causing an abort. Moved to 15JAN90 this date. 5559. January 9, 1990 TVFLG Eric Changed DTVF.INC and TVFLG to produce correct prompt messages for entering channels and IFs. They are supposed to be between BCHAN and ECHAN and between BIF and EIF, resp. Corrected numerous places to take into account that the reference value for baseline number may not be one (when a SUBARR greater than 1 is specified). Moved from Mandrill, nowhere else. 5560. January 10, 1990 RSTOR Bill C. Fixed problem in computing size for scratch files when the two dimensions of the image were different. Moved to 15JAN90 this date. 5561. January 10, 1990 TVSLV Bill C. Copied TVSLD to [aips.local.dead.pgm] and .hlp] and removed from both NEW and TST. This task apparently survived the last purge of software that didn't survive the overhaul. Moved to 15JAN90 this date. 5562. January 10, 1990 TVFLG Eric Found errors in GRIDTB addressing more difficult Stokes types and when dividing by flux. I doubt that these caused any real trouble. Moved nowhere. 5563. January 11, 1990 UVFLG Bill J. Added a 'mask' option to UVFLG so that users could flag any combination of RR, LL, RL and LR that they like. There are situations where the user would like to flag all cross-hand fringes but this was not possible in UVFLG --- STOKES 'RL' and 'LR' flagged 'RL' AND 'RR' and 'LR' AND 'LL' respectively. Specifying a suitable mask in the STOKES adverb e.g. '0011' will now flag appropriately. Also modified UVFLG.HLP to reflect these changes and explain existing options a little better. Moved nowhere. 5564. January 11, 1990 UVSUB Phil Increased the buffer sizes to cope with 4096 square images. Moved to 15JAN90 this date. 5565. January 11, 1990 COOK10.TEX Phil Rewrote the VLBI chapter of the cookbook. It now provides user recipes for the reduction of VLBI data using the calibration package. It also includes a section on spectral line VLBI data reduction. Full polarization and Mk3 analysis will follow soon. Moved nowhere. 5566. January 11, 1990 TVFLG Eric Corrected handling of the baseline order parameters. The were confused and partly wrong, including in the recorded baseline in the FC table. Changed TVFOAD as well as TVFLG. Also improved the prompt message on the Flag/Confirm function. Moved nowhere. 5567. January 12, 1990 FILAI2, ZDIR Eric Corrected FILAI2, it used the wrong arguments to a WRITE statement, thereby blowing a FORMAT. Brought the initialization statements for the accounting file up to date. Changed ZDIR (APLVMS: version) to append a period to the file names from the root area ('RT'). Otherwise, FILAI2 was seeking SYSPARM.DAT. Moved to 15JAN90 today. 5568. January 13, 1990 DOC:VMSNOTES.89D Bill C. Included documentation about problems with FILAI2 and printers. Moved nowhere. 5569. January 14, 1990 TVFLG, GRIDTB, GTBWRT Eric Changed call sequence to GTBWRT to pass the first dimension of the summing arrays as a parameter and changed that call in GRIDTB. Corrected two errors in TVFLG --- to display the flagged time-range on mode 'TIME' when listing flags and to compute even more times in the master time list to avoid hitting the end of the list. Moved nowhere. 5570. January 14, 1990 SPFLG Eric Files: SPFLG.FOR, SPFLG.HLP, GRIDTC.FOR and SPFOAD.FOR (last 2 may may vectorize). New task to display channel on the x axis and time on the y axis and then to edit the data much like TVFLG. So far, the task will work on only one baseline at a time. Moved nowhere. 5571. January 15, 1990 BPASS Bill C. Changed arguments in call to BPADJ. There was a bug which caused the routine to fail and die with an error message. BPADJ is called when there are multiple reference antennas used i.e. when REFANT was not specified. Moved to 15JAN90. 5572. January 15, 1990 FILLM Bill C Fixed several problems with Frequency accountability. There was some difficulty associated with using the frequency of the first channel written to AIPS as the reference frequency. The frequency offset of the first IF and source were half the bandwidth since the frequency of the observations used was the band center frequency. However, in computing the source and Doppler frequency offsets the sky frequency rather than the reference frequency was used. FILLM now uses the frequency channel corresponding to the center of the observed band to be used as the reference pixel in the output files and the first encountered "sky" frequency as the reference frequency. Also there were numerous errors handling FQ identifiers. Use of STFREQ as the stream reference frequency was changed to use the value in the catalog header since STFREQ may contain a different value if processing of data into a given file is restarted. A major restructuring of the way SU, FQ and CL tables are processed was made. These tables are now written by a new routine FLMTAB which is normally called from FLMUV but for the last entries in the CL table it is called from MCHEAD. Revised the was averaging was handled; usually data was not being averaged as requested. Also fixed a bug in the handling of shadowing; the default was no flagging rather than 25 m. Now makes the frequency reference pixel the velocity, alternate axis reference pixel. Moved nowhere. 5573. January 15, 1990 MX.FOR, MX.HLP Bill C. Fixed error in passing the alternate reference axis pixel. It was not correcting for selection by channel. Also changed the Adverbs for specifying line channels to a more standard useage (BCHAN and ECHAN select channels). Moved nowhere. 5574. January 15, 1990 UVCOP Bill C. Changed how frequency labeling was handled. The same reference frequency as is input file is now used and the reference pixel changed. Moved nowhere. 5575. January 15, 1990 SPLIT Bill C. Fixed error in assigning alternate reference pixel for the velocity axis. Now assumes that the velocity reference pixel is the same as the frequency reference pixel. Now no correction is made for source dependent frequency offsets in the reference frequency but it is done in DGHEAD. Source dependent frequency offsets are put into the FQ table. Moved nowhere. 5576. January 15, 1990 DGHEAD Bill C. Changed the handling of the frequency reference pixel. The same channel is now used in the input and output data. The old scheme was to shift the reference frequency but the reference pixel was not changed. Also, now the source frequency offset is added to the the Reference frequency if the output is for a single source and there is an input SU table. Moved nowhere. 5577. January 15, 1990 HORUS Bill C. Cleaned up handling of the frequency reference pixel. It is now set only in IMCREA and ignores the value set by DGHEAD. Moved nowhere. 5578. January 15, 1990 FILAI2 Eric Changed it to use LUN 6 for messages since it is run in batch where LUN 5 is only an input device. There is no interactivity in this program. Moved to 15JAN90 this date. 5579. January 16, 1990 SPFLG, TVFLG Eric Fixed some messages which were too long in both SPFLG and TVFLG and corrected the wording of some messages (``channel'' should be ``baseline'') in SPFLG. Changed SPFLG.HLP to explain the current, nonstandard use of ANTENNAS and BASELINE. Changed GRIDTC and SPFLG to do two baselines at a time (where the baseline axis can remain regular). Moved nowhere. 5580. January 16, 1990 TBAVG Bill C. New task. TBAVG averages uv data over time and baseline to produce a new uv data file. This is useful for measuring flux densities of time variable point sources. Also TBAVG.HLP. Moved nowhere. 5581. January 17, 1990 UVFLG Bill C. Modified so the channel range written to an FG table were the values passed from AIPS. This is to allow flagging a "channel 0" data file with entries suitable for copying to a line data base without further editing. Also cleaned up declarations. Moved nowhere. 5582. January 18, 1990 TVFLG Eric Added to the help file and code an option to allow autocorrelation data. Before, it was included on the baseline axis, but disallowed by the defaults in the calibration package. Also corrected a bug in setting the internal stop time. When a TIMERANGE was used, the master grid would usually stop prematurely until this fix. Changed most of the numeric inquiry functions to return to the menu on non-numeric input, rather than repeating the question. This allows the user to bail out. Moved nowhere. 5583. January 18, 1990 SPFLG Eric Corrected bug in setting stop times and added option to help file and code to allow autocorrelation data. Changed to support ANTENNAS and BASELINE in the usual fashion and to have as many baselines in the master grid as you may desire. They are put on an irregular axis whose coordinates are stored in a TB extension table. Only allowed baselines with data are included. Added a display to show the correspondence between this ``random'' baseline number and real antenna numbers. Moved nowhere. 5584. January 18, 1990 VLBIN Dan Briggs/Phil Dan spotted an error in the fbs correction routine that remained from the overhaul, it used to be a pseudo I*4 section which was erroneosly erased and caused the fbs correction to fail. Also tidied up the autocorrelation transform routine. Moved to 15JAN90 this date. 5585. January 18, 1990 UVGRID Glen Set zero spacing flux to 1. for creating a beam. Added new general subroutines which handel bandwidth synthesis averageing: UVZERO sets a vis point to zero, except U value. UVZRSP sets zero spacing flux. UVCONJ performs complex conjigate of one vis point UVSETB sets values for creating synthesized beam from VIS. Moved to 15JAN90 on January 19, 1990. 5586. January 19, 1990 UVHGM.FOR,UVHGM.HLP Mark Calabretta/Bill C. Fixed a number of bugs and added the option to plot frequency (FQ) id. Moved to 15JAN90 this date. 5587. January 19, 1990 QPUT Glen Added wait after QPUT in FPS version Moved to 15JAN90 this date. 5588. January 19, 1990 HITIME Eric A bad declaration --- some variable should have been integer, but were listed as real --- caused a formatting error (as well as poor history cards). Moved to 15JAN90 this date. 5589. January 20, 1990 UVGRID/UVCONJ Glen Modifed UVCONJ to conjugate all stokes as well as all frequency channels, if requested. UVGRID call to UVCONJ modified. Moved to 15JAN90 this date. 5590. January 20, 1990 ALGSUB Glen Replaced some GO TOs by THEN, END IFs and added calls to UVCONJ Moved nowhere. 5591. January 20, 1990 FUDGE Bill C. Modified to process compresed data; if the input data is compressed then the output will be compressed. Also modified to make it easier to write different size output records from the input. Moved nowhere. 5592. January 20, 1990 TBAVG.FOR,TBAVG.HLP Bill C. Modified to process compressed data and fixed a couple of bugs in the handling of flagged input data. Fixed a description of the task in TBAVG.HLP. Moved nowhere. 5593. January 20, 1990 UVBAS.FOR,UVBAS.HLP Bill C/Huib van Langevelde New task. Determines a "continuum" visibility from the average of a specified set of channels and subtracts it from all spectral channels. This task is useful for removing continuum emission from line data. It works properly only if the visibility of the source does not change significantly across the bandpass. Moved nowhere. 5594. January 21, 1990 FILLM.FOR Bill J./Bill C. Added Bill C.'s fixes (1) to make FILLM do shadow flagging correctly and (2) to correct blowing up of format when FQ tolerance is large. Moved to 15JAN90 this date. *********************************************************************** Changes while 15APR90 was NEW: *********************************************************************** *********************************************************************** 5595. January 22, 1990 UNICOS Area Logicals Kerry Added area logicals to $SYSAIPS/AREAS.DAT for the implementation of AIPS under UNICOS. These include $APLCRI (UNICOS Z-routines), $INCCRI (UNICOS INCLUDE files), $QCRI (UNICOS Q-routines) & $SYSCRI (UNICOS system files). Eliminated the UNIX area logicals for empty, never used directories, including $APLNUNIX, $QPGUNIX, $QYPGUNIX, $SYSLUNIX and $YPGUNIX. Also cleaned up and embellished the comments. Moved to 15JUL90 same date. To be distributed in 15JAN90 installation kits for UNIX. 5596. January 22, 1990 Source Code Preprocessing for UNICOS Kerry Source code preprocessing for FORTRAN modules under UNICOS is the same as for any other system, however, C modules require some minor massaging, including the removable of the ``_'' (underscore) suffix from entry points and calls to FORTRAN routines. Also, ``cft77'' does NOT convert FORTRAN code to lowercase, so entry points and calls to other AIPS routines must be specified using uppercase. The UNICOS version of PP does this via ``sed'' using the command file C.SED. A special version of INCS.SH has also been created that includes the $INCCRI directory in the preprocessor's search path for INCLUDE files. New files: $SYSCRI/PP $SYSCRI/INCS.SH $SYSCRI/C.SED Moved to 15JUL90 same date. To be distributed in 15JAN90 installation kits for UNIX. 5597. January 22, 1990 Compiler & Loader Options for UNICOS Kerry Created options files specifying rather ``vanilla'' defaults for UNICOS ``cc'' (C) and ``cft77'' (FORTRAN) compilations plus ``segldr'' program loading (linking). New files: $SYSCRI/CCOPTS.SH $SYSCRI/FCOPTS.SH $SYSCRI/LDOPTS.SH Moved to 15JUL90 same date. To be distributed in 15JAN90 installation kits for UNIX. 5598. January 22, 1990 Object Library Maintenance for UNICOS Kerry System V based UNIX platforms do not have ``ranlib'' (eliminated after SYSTEM III). Furthermore, UNICOS uses ``bld'' as an object librarian rather than the standard UNIX ``ar''. New files: $SYSCRI/LIBR Moved to 15JUL90 same date. To be distributed in 15JAN90 installation kits for UNIX. 5599. January 22, 1990 ZLPCL2 Script for UNICOS Kerry Created a special version of the standard UNIX script used to dispose line printer output (i.e., temporary disk files) to a suitable device. It's very similar to the generic UNIX version, except it uses the UNICOS ``asa'' filter to process FORTRAN carriage control characters. New files: $SYSCRI/ZLPCL2 Moved to 15JUL90 same date. To be distributed in 15JAN90 installation kits for UNIX. 5600. January 22, 1990 FORTRAN Compilation Script for UNICOS Kerry The technique used in the generic UNIX version of this script to detect a FORTRAN compilation failure does not work under UNICOS (i.e., ``cft77'' produces a ``.o'' file whether it fails or not), thus a special version of the FC script that tests the return code of the compiler. New file: $SYSCRI/FC Moved to 15JUL90 same date. To be distributed in 15JAN90 installation kits for UNIX. 5601. January 22, 1990 Source & Object Library Mappings for UNICOS Kerry Created a generic version of LIBR.DAT for UNICOS that uses the $QCRI Q-routines, the $YVTV (virtual TV) Y-routines and $APLCRI Z-routines. New file: $SYSCRI/LIBR.DAT Moved to 15JUL90 same date. To be distributed in 15JAN90 installation kits for UNIX. 5602. January 23, 1990 UVBAS Bill C. Corrected indexing error when selection channels. It was selecting data starting at channel 1. Moved from 15JUL90 this date. 5602. January 23, 1990 DeAnza routines Mark Calabretta/Eric Corrected YDEA errors in YZERO (constant used in call where variable is required) and YSPLIT (NE used where EQ was wanted). Moved from 15JUL90 this date. 5603. January 23, 1990 EXPFIT Eric Corrected bad format numbers and added further directory name tests. Changed EXPFIT.COM to include the new text regions. Moved from 15JUL90 this date. 5604. January 24, 1990 SPFLG Eric Corrected the CURVALU-like display in most of the flagging modes. It was not handling BCHAN correctly. Moved from 15JUL90 this date. 5605. January 24, 1990 UVBAS.HLP Huib van Langevelde/Bill C. Added explain section. Moved to 15APR90 this date. 5606. January 25, 1990 BATER Bill J. Two calls to CHR2H in subroutine CUA had incorrect numbers of arguments. Moved from 15JUL90 this date. 5607. January 29, 1990 AIPSHELP.HLP Bill C. Updated Designated AIP schedule. Moved to 15APR90 this date. 5608. January 29, 1990 PLCUB Bill J. The call to NAMEST used a hollerith rather than a character string although the necessary conversion had been performed. Moved from 15JUL90 this date. 5609. January 30, 1990 BATER Bill J. Due to indexing error in subroutine BATPOL, the command interpreter could never have worked for some while! Moved from 15APR90 this date. 5610. January 31, 1990 LISTR.HLP M. Calabretta/Bill C Corrected misstatements about the defaults for frequency identifier. Moved from 15JUL90 this date. 5611. January 31, 1990 APGNOT:UVPLT M. Calabretta/Bill C Fixed a formatting problem in displaying the timerange on the plot; scaled the u,v, and w to the correct frequency when they are plotted and allowed plotting all FQ ids on the same plot. Also documented in UVPLT.HLP. Moved from 15JUL90 this date. 5612. January 31, 1990 DDTs Glen Removed NPOINTS from the DDTSAVE help and from DDTLOAD.001. Updated DDT.HLP. Modified NEWPARMS.001 to include the DDT adverbs DDISK, DDTSIZE, EDGSKP, IOTAPE, MDISK, TCODE, TDISK, TMASK and TMODE. Added help files for the adverbs. Moved from 15JUL90 this date. 5613. February 6, 1990 CALIB Bill Changed call to FRQTAB in GASOLV to determine frequencies from the actual file being read rather than the input data file. Since the data may be reordered by UVGET this is important; under some circumstances this problem causes an infinite loop. Moved from 15JUL90 this date. 5614. February 9, 1990 CALIB Bill C Fixed a couple of MIN functions that were supposed to be MAX in CLBSRC and CLBDR. This error caused CALIIB to abort when doing a common delay-rate solution for all IFs in a MkIII data file. Moved from 15JUL90 this date. 5615. February 9, 1990 MK3IN.FOR,MK3IN.HLP Bill C. Numerous improvments. Any antennas found may be selected by using antenna name 'ANY' in the INFILE. The "Debug" output option was cleaned up to work properly on any machine. CL table entries for antennas for which there was no data are no longer written. The observing date in the AN and HI files are now correct. The time tags associated with the first baseline of the first scan are now correct if REFDATE is not specified. The phase of the phase cal was reversed if the baseline was oriented west to east. This is now fixed. The phase of the phase cal is now adjusted for any offsets in delay. This problem was causing the phase cals on the "remote" stations to be trashed. Also the max and min. peculiar delay offsets are kept on a baseline basis. The delay residual in the output data was wrong by an amount corresponding to one delay lag. Moved from 15JUL90 this date. 5616. February 9, 1990 APLGEN:ZDHPRL.FOR Bill C The use of the intrinsic function "REAL" to convert from a 32 bit integer to double precision was replaced by "DFLOAT". Use of "REAL" was causing a signifigant precision loss in translating HP double precision words on Convexes; DFLOAT should keep the full precision. Moved to 15APR90 this date. 5617. February 10, 1990 CLCOR.FOR, CLCOR.HLP Bill C. Added two new options: 'PCAL' to insert MkIII manual phase cals and 'SBDL' to add corrections to the residual delays without modifying the total model delays. These two options are needed for procesion MkIII VLBI data. Also changed the adverb BPARM to CLPARM which is dimensioned 20. Also CLPARM.HLP. Moved from 15JUL90 this date. 5618. February 10, 1990 POPSDAT.HLP, RUNSYS:NEWPARM.001 Bill C Added the new, task specific adverb CLPARM. In POPSDAT.HLP this was done in a procedure ant the end of the file rather than defining an adverb that POPS knows by name; i.e. DAPL.INC doesn't need to be changed and AIPS etc recompiled and linked. This change required that POPSGN be run by hand on each NRAO computer updated by the midnight job. Moved from 15JUL90 this date. 5619. February 13, 1990 ZFRE2.C in APLSUN Chris Modified version of the generic Unix ZFRE2. Modifications are necessary for Sun OS since the Sun implementation of {\tt df} returns the information for the requested disks in the order in which they occur in {\tt /etc/fstab} irrespective of the order in which they were specified on the command line. Moved from 15JUL90 this date. 5620. February 15, 1990 SDTUV Bill C. Removed restriction that the input data files have 14 data columns. Also fixed a couple of places where zero divides were possible. Also added the Parkes 64 m antenna to the list of antennas that it knows about. Moved from 15JUL90 this date. 5621. February 15, 1990 MK3TX Chris Now behaves more sensibly if there are no text files present in the archive. Moved from 15JUL90 this date. 5622. February 15, 1990 CALIB Bill C Fixed problem with averaging all data in each IF. This problem only occured when fringe fitting MkIII data. Also put the SN table writing stuff in GASOLV into a separate routine. Moved to 15APR90 this date. 5623. February 15, 1990 CLCOR.FOR, CLCOR.HLP Bill C. Modified 'SBDL' option to modify the phase as well as the delay. The phase is assumed to be at the reference frequency if the delay residual correction is non zero. DATCAL corrects to the frequencies of the actual data. Also updated the table of sunspot numbers in the CLCOR.HLP file. Moved to 15APR90 this date. 5624. February 15, 1990 APLNOT:CGASET.FOR Bill C. Removed the phase corrections for channel selection with non zero delay corrections. This correction was also being done in DATCAL and was redundant in CGASET. Moved to 15APR90 this date 5625. February 16, 1990 CALIB Phil There was an indexing error in filling up the frequency arrays in GASOLV, the bchan and bif offsets were not being taken into account. Moved to 15APR90 this date. 5626. February 16, 1990 TABHDR Bill C. Fixed logic error which caused the length of character table entries read from ASCII FITS tables to always be 1. This appears to have had no effect on calibration tables but was discovered at CfA. This affected UVLOD and IMLOD. Moved from 15JUL90 this date. 5627. February 18, 1990 PUTCOL Bill C. Remove a C in column 1 from a line computing the array pointer for a double precision value. Moved from 15JUL90 this date. 5628. February 19, 1990 SLCOL Neil Killeen/Bill C Fixed an error extracting BLC, TRC from the input parameter array. Moved from 15JUL90 this date. 5629. February 19, 1990 PRTSD Bill C Fixed bug in computing time to be displayed; it was multiplying all the times by 2. Moved from 15JUL90 this date. 5629. February 19, 1990 SDTUV Bill C Changed hardcoded number of receivers (14) to the number of columns - 2. Also modified BADRA to accept an RA difference of 360 degrees as 0. Moved from 15JUL90 this date. 5629. February 19,1990 UVCOP Bill C. Fixed precision problem in computing subarray number. Also the 5 day time correction removal when BPARM(2)>0 is only done for single source data and if the uncorrected time is > 5 days. Moved from 15JUL90 this date. 5630. February 19, 1990 UVFIX Bill C. Corrected a number of errors in the handling of compressed data; the output for compressed data was trashed. Also fixed a bug by which the subarray information was stripped off the baseline/subarray code. Also fixed to loop over subarrays if no subarray is requested. Added a SAVE statment in UVWCAL for the values that were computed on the first call. Also changed UVFIX.HLP. Moved to 15APR90 this date. 5631. February 19, 1990 HORUS Bill C. The specified subarray was being ignored. This caused HORUS to fail when DOPOL was true because only 1 subarray at a time can be processed. Moved from 15JUL90 this date. 5632. February 20, 1990 CALIB Bill C. Fixed bug introduced in recent fix of fringe fitting which messed up calibration of VLA like data. The RCP and LCP data were added and called RCP and there were no solutions for LCP data. Also fixed bug which caused the mean gain modulus correction from being computed if requested. Moved from 15JUL90 this date. 5633. February 20, 1990 Mean gain modulus correction Bill C. Bugs in DATCAL and BLSET caused the mean modulus gain correction to fail. DOCAL wasn't applying the correction at all unless delay or rate or baseline corrections were being made. Moved from 15JUL90 this date. 5634. February 20, 1990 TABED Bill C. Modified to ignore flagged table entries. Moved from 15JUL90 this date. 5635. February 20, 1990 ZFRE2 in APLSUN Chris ZFRE2 expected the first part of the pathname of an AIPS ``disk'' to be the mount point of the device on which the logical disk resided. This precondition was not met if any of the AIPS logicals DA01 etc. referred to a symbolic link. This version of ZFRE2 now interrogates each file referred to by the AIPS disk logicals to see if it is a symbolic link and, if so, reads the value of the link. Moved from 15JUL90 this date. 5636. February 21, 1990 TKLAB Bill C. Fixed bad call to NAMEST. It was being passes a Hollerith name/class rather than a character string. This caused an infinite loop (on Suns at least) in verb TKSLICE. Moved from 15JUL90 this date. 5637. February 21, 1990 XMOM Bill C. Fixed bug left over from the overhaul. A numeric array was being passed to ZPHFIL and ZOPEN for the file name rather than a character. This caused failure of the open (on Unix at least - this may work on Vaxes). Also cleaned up some other junk not cleaned up in the overhaul. Moved from 15JUL90 this date. 5638. February 21, 1990 VLBIN Phil The GAST at UT=0 was being written into the AN table as radians rather than degrees. Also the formula used to calculate IAT-UTC was getting progressively wrong as time went on, so I inserted the look up table stolen from Starlink . This is correct up to Jan 1 1990 - we will need to think about how we cope with this - maybe a table distributed by AIPSSERV and let VLBIN read the table? Moved from 15JUL90 this date. 5639. February 21, 1990 AXSTRN glen AXSTRN produces a frequency label in exponential form if the frequency is greater than 1E11. Each internal write statement has an ERR= argument to trap errors. Moved from 15JUL90 this date. 5640. February 22, 1990 ZDEAXF Gustaaf van Moorsel/Bill C Fixed declaration of OPERS(20) to OPERS*4(20). Moved from 15JUL90 this date. 5641. February 23, 1990 NEWPARMS, POPSDAT, DDTs Glen Added help files for two DDT adverbs TNAMF and MAPDIF. Added these adverbs to NEWPARMS and POPSDAT. Copied CLCOR.HLP from TST and modived NEWPARMS to init CLCORPRM instead of CLPARM. Moved from 15JUL90 this date. 5642. February 26, 1990 Q1GRD, QPHSRO, QRECT1 Glen Created QRECT1 which calls sin() and cos() and loads them into the output vector. (Radius is not used and need not be set to 1.) Modifed QPHSRO and Q1GRD to call QRECT1 not QRECT. QRECT1 provides more accurate SIN and COS. Moved from 15JUL90 this date. 5643. February 27, 1990 UVCOP Bill C. Corrected logic error in determining which correlations to keep when selecting compressed data by channel. Moved from 15JUL90 this date. 5644. February 27, 1990 SPFLG, TVFLG Eric Changed DTVF.INC and DSPF.INC to add a variable retaining the source random parameter. Changed GRIDTB and GRIDTC to get the correct source ID number even when the user has specified a single source. Changed SPFLG and TVFLG to check frequency and source ID numbers when making the list of times to be gridded. Also changed them to correct the handling of the interactive (graphical) window setting and to remember the input source random parameter offset. Dropped unused FORMATs. Corrected channel number displayed in SPFLG's interactive modes. It used a correct channel number, but displayed a bad one if the reference channel was not the first channel. Moved from Oz this date to wherever and from 15JUL90. 5645. February 27, 1990 NEWPOS Eric Cleaned up the code a bit. Also changed the formulas used for the SIN and ARC projections. They failed when the angles got very large so that ASIN became ambiguous. Now both use ATAN2. Thanks to Mark Clark (GB) who bugged me about this. Moved from Oz this date to wherever and from 15JUL90. 5646. February 28, 1990?? TVFLG, SPFLG Eric Corrected GRIDTB --- it was regarding an end-of-file condition as unacceptable at the location where the first point of each group was read. At one time, this location was for the first visiblity, but that was changed some time ago to do things in groups for vectorization. Corrected SPFLG and TVFLG --- in the first read through the data to determine the list of times, the tests on subarray, source, and FQ number were branching all the way to the end of the loop. Thereby, the index was not incremented and the same point was retested and re-rejected for the full buffer's worth. This could make the list of times seriously incomplete or in error in some cases. Moved from Oz this date and from 15JUL90. 5647. February 28, 1990 FILLM Phil/Bill J. Migrated the fixes inserted in TST 2 weeks ago, nobody screamed so everything appears to be OK. Several bugs were fixed, one outstanding one and others (re)introduced recently. (1) The problem with the integration just prior to midnight being labelled wrongly was finally found and fixed, the fact that midnight has occurred is now caught and dealt with in MCSUM. (2) When antennas enter the array during an experiment the order of the baselines in the data arrays changes, this was not being trapped. MCWANT now checks to see if the number of antennas and/or the order of antennas has changed and if so it forces a call to MCINI. I fixed this 18 months ago but my fix had been removed even though the variables and arrays I used were still declared. (3) An associated problem is that the AN table is written when a file is first opened so it only contains information for antennas present at the beginning of the observing run. A new subroutine (ANTUPD) now updates the AN table before the file is closed. (4) The bandwidth being written to the SU table was the full bandwidth rather than the channel width. (5) There was a problem occurring when the number of channels changed during the run, this change was not always caught. Added another section to MCWANT to explicitly check for this change and also for a change in correlator mode. When these change the program is now forced to determine if a new file must be opened. (6) The annoying (to the users) feature of having to specify integration time + 0.2 sec to obtain a specified int. time has been disabled. This will need more checking to determine if it really works. Moved from 15JUL90 this date. 5648. March 1, 1990 KEYCOP Bill C New routine to copy any header keywords from one cataloged file to another. Moved from 15JUL90 this date. 5649. March 1, 1990 UVCOP Bill C Modified to call KEYCOP to copy catalog keyword/value pairs. This allows using UVCOP to salvage data from a disrupted FILLM run. Moved from 15JUL90 this date. 5650. March 4, 1990 SNPLT Bill C Fixed several problems with OPTYPE='SUM'. The 2 IF mode (BIF=0) was disabled in the inputs checking and BIF=2 or higher actually did IF=1 due to indexing errors. Moved from 15JUL90 this date. 5651. March 9, 1990 UVPLT Meil Killeen/Bill C When DOCAL=1, UVPLT looked for an SN table rather than CL table for multi-source data bases. It now does not check if the file is a multi-source file to see if an SN table exists. Moved from 15JUL90 this date. 5652. March 9, 1990 KNTR.FOR, KNTR.HLP Marl Calabretta/Bill C Fixed a couple of bugs in KNTR, one relating to annotation of the contour levels, and another in computing the maximum absolute pixel value in KONDRW. Also changed the way that blanked pixels are handled. It used to be based on connecting pixel centres, but this could produce stick-like "contours" for small irregularly shaped clusters of pixels surrounded by blank areas. The algorithm now delineates the pixel edges. There are pros and cons for either method but the latter is less likely to confuse users. Also expanded the explain part of the help file slightly. Moved from 15JUL90 this date. 5653. March 9, 1990 AIPSUB:TVMOVI.FOR Mark Calabretta/Bill C Removed limitation on long cursor moves. Moved from 15JUL90 this date. 5654. March 12, 1990 ZVTVX2 and ZVTVX3 in APLBERK Chris ZVTVX2 and ZVTVX3 were, incorrectly, treating the case where a read from a {\tt SOCK_STREAM} returned less bytes than were requested as an error. This was causing TV-by-wire to fail over long distances (AOC to Charlottesville). Both routines now loop to read the outstanding bytes until either the complete packet is received or a genuine error or end-of-file is detected. Write operations are treated in the same manner, although partial writes are extremely unlikely in the absence of errors. Moved from 15JUN90 this date. 5655. March 12, 1990 BPASET.FOR Mark Calabretta/Phil Ensured that the scratch file written by BPASET was not too small for UVINIT. Moved to 15APR90 this date. 5654. March 12, 1990 AU9A.FOR Mark Calabretta/Bill C Fixed error trap for TKCATL; treats KNTR plots like CNTR. Moved from 15JUL90 this date. 5655. March 12, 1990 TKCATL Mark Calabretta/Bill C Fixed a bug in the logic to see if a pixel is in range. Moved from 15JUL90 this date. 5656. March 14, 1990 APGNOT:MK3IN.FOR Bill C. Modified to open and close SU and CL table whenever they are accessed. This increases the chances that if something goes wrong that the tables will be OK. Moved from 15JUL90 this date. 5857. March 14, 1990 CALIB.FOR Bill C. Fixed logic error in GASOLV in computing the array of frequencies when selecting by channel. This caused later routines to blow up. Moved from 15JUL90 this date. 5858. March 14, 1990 APLPGM:TACOP.FOR Bill J. TACOP didn't behave sensibly if given a blank OUTNAME and OUTCLASS. Default now is to write higher version of extension table attached to input file. Moved from 15JUL90 this date. 5859. March 12, 1990 AIPSUB:AIPINI.FOR Kerry Fixed to take different actions depending on wether the current program is FILIA2 (open logical unit 6 via ZMSGOP) or FILAIP (open logical unit 5 via ZOPEN). Otherwise, non-VMS installations fail. Also if FILAI2 or FILAIP, don't get device characteristics common from disk or make accounting entry. Otherwise, disconcerting error messages are issued when the AC and/or SP files and/or the MS file for user #1 don't yet exist (FILAI2 and FILAIP create these as part of an installation). Moved to 15JAN90 and 15JUL90 this date. 5860. March 15, 1990 GETRLS Bill C. Apparently the 15APR90 release has been identifying itself as the 15JUL90 release. Moved nowhere. 5861. March 15, 1990 QGRIDA/QUVIN Glen Modified QGRIDA and QUVIN to call QRECT1 instead of QRECT to increase precission. Moved from 15JUL90 some time ago. 5862. March 15, 1990 UVMDIV Phil The call sequence to CHNCOP in UVMDIV on CVAX was wrong, it was OK on nrao1 - God knows what was on the other machines. Luckily that piece of code is not used that often. Moved from 15JUL90 this date. 5863. March 15, 1990 QVEX:QPUT, QGET, QPHSRO Glen Removed CVEX QPUT and QGET which were identical to the QPSAP versions. QPHSRO in PSAP uses double precision. Removed this date, consistent with 15JUL90. 5864. March 15, 1990 FILLM Gareth/Bill C Fixed numerous bugs which occured when a scan contained a single averaging interval and the next scan was to go into another file. The control logic was getting hopelessly confused. Errors in the handling of changes in the number or order of antennas was complicating the issue. In this latter case, tests were made using values set in MCANT or MCH1 which had not been called yet. Several of these (number of antennas, correlator mode) are now decoded in MCWANT. All of the changes associated with these problems were in MCWANT. Corrected an indexing problem in MCGBRC which caused and occasional spectrum to be corrupted and an infinite loop under special conditions. The logic error was fixed and a trap was made for the infinite loop although it should not occur if the routine is working correctly. The last baseline in a CDA will be lost if this trap is invoked. Fixed addressing bug for mode 2BC. Inserted correct source and FQ id numbers into the first visibility record of each scan; formerly the source id and FQ id from the previous scan were being used. Modified to accept autocorrelations for revisions 24 and later. Modified to only put data with the same correlator mode in the same file. This involves writing a header keyword giving the correlator mode in each line data file. Changed the error handling process to return continuous EOF conditions from MCREAD if more than PELIM (currently 50) parity errors have previously occurred. The current file(s) being written are now expanded in size by 10,000 visibility records when they are opened, and compressed when they are closed. This minimizes the disk space required when FILLM is running, but could lead to disk fragmentation if the observation forces multiple openings and closings. FILLM used to compress all files that it had appended to when it exited successfully; less fragmentation, but potentially an unnecessary disk hog on abnormal termination(s). Moved from 15JUL90 this date. 5865. March, 19 1990 MK3IN Bill C The source number written into the CL table was always the first source in the source list. Also added call a call to initialize the A tape header info when the end of a scan was reached. Put a trap into AT20XX to ignore a scan if the cross reference table overflowed. This should not happen under normal circumstances but bad (?) type 20?? records can cause this to happen. Moved from 15OCT90 and 15JUL90 this date. 5866. March. March 19, 1990 GETHUT Bill C Added calls to SPFIL to make sure nulls were removed from TABLE info. This was causing FITTP to write nulls in tables headers. Moved from 15OCT90 and 15JUL90 this date 5867. March 19, 1990 FITTP Bill C. Corrected indexing error reading history file. This was adding an extra 8 blanks after the "HISTORY" on the FITS history cards and cause then to occasionally run onto the next card. Moved from 15OCT90 and 15JUL90 this date 5868. March 28, 1990 FILLM Gareth Corrected bug recently introduced in reading the parameters from an existing catalog entry (this is for the case when FILLM will try to append to an existing file.) Corrected the error handling when a logical record is incomplete at the end of tape. Fixed the NX record count when closing a dataset. 5870. April 8, 1990 $APLGEN/ZLWIO.FOR Kerry $APLGEN/ZLWOP.FOR Switched to FORMAT'ed OPEN and WRITEs. Also corrected call sequence to ZLASCL in ZLWIO (error return missing). Moved from 15JUL90/15OCT90 this date. 5871. April 8, 1990 $APGNOT/LWPLA.FOR Kerry Fixed buffering problem and errors introduced by changing from hollerith to CHARACTER constructs during the overhaul. Eliminated EOT character from the beginning of the output PostScript program which causes problems on some laser printers and does not seem to be necessary. Eliminated inappropriate H2CHR call and replaced it with simple CHARACTER substring assignment. Also switched character font from Courier to Helvetica-Bold. Tested CNTR and GREYS plots using the VLBA correlator group's QMS PS810 and at non-NRAO sites using Apple Laserwriter and NEC LC890 laser printers. Moved from 15JUL90/15OCT90 this date. 5872. April 8, 1990 $SYSUNIX/ZLASCL.PS Kerry Created a special version of the ZLASCL script for use with PostScript laser printers that prefixes "%!PS-Adobe-" followed by "%%Creator:AIPS task LWPLA" to the LWPLA output. The first needs to appear at the beginning of the output ("%!... distinguishes PostScript programs from simple text) on a line by itself and this is not possible with the buffering scheme used in LWPLA. ZLASCL.PS should be substituted for ZLASCL at sites with PostScript engines. Some laser printers (e.g., NEC LC890) may also require that the LWPLA output be processed to limit the number of characters per line to, say 132 (e.g., using "fmt -s -132"). Moved from 15JUL90/15OCT90 this date. 5873. April 8, 1990 $APLUNIX/ZABORS.C Kerry $APLBERK/ZABORS.C $APLSUN/ZABORS.C Eliminated "core" file preservation mechanism entirely. Also changed to commit suicide with a self-inflicted illegal instruction rather than by calling "abort" which issues different signals on different systems. Changed in 15JUL90 and 15OCT90 same date. 5874. April 8, 1990 $APLSUN/ZABOR2.C Kerry A new module that differs from the $APLBERK version in that it calls "ieee_handler" to "set" the action for IEEE "common" floating_point exceptions to SIGFPE_ABORT. Otherwise, invalid floating-point results (i.e., NaNs) go undetected. Should abort now. Changed in 15JUL90 and 15OCT90 same date. 5875. April 8, 1990 $APLBELL/ZCREA2.C Kerry $APLBELL/ZACTV9.C New modules. The $APLBELL version of ZCREA2.C differs from the generic UNIX version in terms of header files. The $APLBELL version of ZACTV9.C differs from the generic version in that it calls "fork" instead of "vfork" (a Berkeley UNIXism). Changed in 15JUL90 and 15OCT90 same date. 5876. April 8, 1990 UNICOS Z-routines Kerry All new modules in $APLCRI: ZABOR2.C same as $APLBELL version except abort handler function must be declared differently; also no SIGPWR and SIGCLD action set to SIG_DFL ZABORS.C PUTBCKed then REMOVEd to force REMOVE via midnight jobs on other Crays and to use $APLUNIX version instead ZACTV9.C PUTBCKed then REMOVEd to force REMOVE via midnight jobs on other Crays and to use $APLBELL version instead ZCMPR2.C same as $APLBERK version (Berkeley UNIX "ftruncate" supported, at least under UNICOS 5.0.13) ZCMPRS.FOR same as $APLGEN version except call to CHR2H commented out (4 characters/real won't work on 64-bit machines) and CHARACTER PNAME passed to ZCMPR2 instead of HOLLERITH ANAME ZCPU.FOR calls UNICOS routine SECOND (generic UNIX "times" yileds unreliable under UNICOS 5.0.13) ZCRDIR.C same as $APLUNIX version except calls ISHELL (UNICOS specific) instead of "system" to execute a shell command ZCREA2.C PUTBCKed then REMOVEd to force REMOVE via midnight jobs on other Crays and to use $APLBELL version instead ZCREAT.FOR same as $APLGEN version except call to CHR2H commented out (4 characters/real won't work on 64-bit machines) and CHARACTER PNAME passed to ZCREA2 instead of HOLLERITH ANAME ZDCHI2.FOR same as $APLBELL version except SPFRMT and DPFRMT = 0, and TTYCAR = 0 ZDCHIC.C same as $APLUNIX version except NBBY "defined" as 8 (for some reason not in sys/param.h) ZDCHIN.FOR same as $APLGEN version except NBPS = 1024 and NSPG = 1 ZDESTR.FOR same as $APLGEN version except call to CHR2H commented out (4 characters/real won't work on 64-bit machines) and CHARACTER PNAME passed to ZCREA2 instead of HOLLERITH ANAME ZEXIST.FOR same as $APLGEN version except call to CHR2H commented out (4 characters/real won't work on 64-bit machines) and CHARACTER PNAME passed to ZEXIS2 instead of HOLLERITH ANAME ZFRE2.C same as $APLBELL version except logic (?) to look for continuation line ZGETCH.C PUTBCKed then REMOVEd to force REMOVE via midnight jobs on other Crays and to use $APLUNIX version instead ZI16IL.FOR borrowed from tried & true COS implementation ZI32IL.FOR calls Cray's USICTC to perform conversion ZILI16.FOR borrowed from tried & true COS implementation ZIVSOP.FOR same as $APLGEN version except call to FIVASOPEN commented out (name too long) ZLASC2.C same as $APLUNIX version except calls ISHELL (UNICOS specific) instead of "system" to execute a shell command ZLASOP.FOR same as $APLGEN version except "cft77" objected to RECL=160 without ACCESS='DIRECT' for opcode 'POPN' (used anywhere except under VMS?) ZLOCK.C same as $APLBERK version (Berkeley UNIX "flock" supported, at least under UNICOS 5.0.13) except for additional header file sys/types.h required ZLPCL2.C same as $APLUNIX version except calls ISHELL (UNICOS specific) instead of "system" to execute a shell command ZMSGOP.FOR same as $APLGEN version except call to CHR2H commented out (4 characters/real won't work on 64-bit machines) and CHARACTER PNAME passed to ZDAOPN instead of HOLLERITH ANAME ZMSGWR.FOR restored MSGBUF argument to CHARACTER*80, then set MSGTXT = MSGBUF instead if H2CHR on HOLLERITH MSGBUF(20) (i.e., tried & true pre-overhaul technique) and true pre-overhaul technique) ZMSGXP.FOR same as $APLGEN version except call to CHR2H commented out (4 characters/real won't work on 64-bit machines) and CHARACTER PNAME passed to ZEXIS2 instead of HOLLERITH ANAME ZOPEN.FOR same as $APLGEN version except call to CHR2H commented out (4 characters/real won't work on 64-bit machines) and CHARACTER PNAME passed to ZDAOPN instead of HOLLERITH ANAME ZPUTCH.C PUTBCKed then REMOVEd to force REMOVE via midnight jobs on other Crays and to use $APLUNIX version instead ZR32RL.FOR lots of unpacking as well as bitwise and'ing, or'ing and shift'ing, but it vectorizes; no NaN detection yet, but otherwise pretty well tested ZR64RL.FOR version in a state of flux (not to be trusted) ZRENAM.FOR same as $APLGEN version except call to CHR2H commented out (4 characters/real won't work on 64-bit machines) and CHARACTER OLDNAM and NEWNAM passed to ZRENA2 instead of HOLLERITH OLD and NEW ZSHCMD.C same as $APLUNIX version except calls ISHELL (UNICOS specific) instead of "system" to execute a shell command ZTACT2.C similar to $APLBELL version but checks for zombie'd processes via "grep" on "ps" output to avoid false positives ZTAP2.C in a state of flux, but close to working ZTKILL.C similar to $APLBELL version but checks for zombie'd processes via "grep" on "ps" output to avoid false positives ZTPCLD.C $APLUNIX/ZDACL2.C hacked for pseudo tape (i.e., instead of $APLGEN/ZTPCLD.FOR because direct access FORTRAN I/O under UNICOS 5.0.13 required COS-like machinations); should probably be used by all UNIX platforms ZTPMID.C $APLUNIX/ZMI2.C hacked for pseudo tape (i.e., instead of $APLGEN/ZTPMID.FOR because direct access FORTRAN I/O under UNICOS 5.0.13 required COS-like machinations); should probably be used by all UNIX platforms ZTPMIO.FOR same as $APLGEN version except modified to call C version of ZTPMID; should probably be used by all UNIX platforms; ZTPOP2.C same as $APLUNIX version except additional header file sys/types.h required ZTPOPD.C $APLUNIX/ZDAOPN.C hacked for pseudo tape (i.e., instead of $APLGEN/ZTPOPD.FOR because direct access FORTRAN I/O under UNICOS 5.0.13 required COS-like machinations); should probably be used by all UNIX platforms ZTPOPN.FOR same as $APLGEN version except call to CHR2H commented out (4 characters/real won't work on 64-bit machines) and CHARACTER MTNAM passed to ZDAOPN instead of HOLLERITH ANAME ZTPWAD.C $APLUNIX/ZWAI2.C hacked for pseudo tape (i.e., instead of $APLGEN/ZTPWAD.FOR because direct access FORTRAN I/O under UNICOS 5.0.13 required COS-like machinations); should probably be used by all UNIX platforms ZTQSP2.C same as $APLUNIX version except calls ISHELL (UNICOS specific) instead of "system" to execute a shell command ZTXMAT.FOR same as $APLGEN version except local CHARACTER CNAMES array passed to ZTXMA2 instead of HOLLERITH NAMES (4 characters/real won't work on 64-bit machines) then call CHR2H in loop to convert NNAM CNAMES to NNAM NAMES ZVTVC2.C same as $APLBERK version ($APLBERK not in Z-routine search path) ZVTVC3.C same as $APLBERK version ($APLBERK not in UNICOS Z-routine search path) ZVTVGC.C same as $APLBERK version ($APLBERK not in UNICOS Z-routine search path) ZVTVO2.C same as $APLBERK version ($APLBERK not in UNICOS Z-routine search path); this routine may have trouble getting its arguments correctly from ZVTVOP ZVTVO3.C same as $APLBERK version ($APLBERK not in UNICOS Z-routine search path) ZVTVX2.C same as $APLBERK version ($APLBERK not in UNICOS Z-routine search path) ZVTVX3.C same as $APLBERK version ($APLBERK not in UNICOS Z-routine search path) Changed in 15JUL90 and 15OCT90 same date. 5877. April 8, 1990 Cray INCLUDE Files Kerry All new modules in $INCCRI: DAPC.INC Declarations, dimensions, COMMON definitions and EQUIVALENCE statements for the Cray 256K word Pseudo-AP implementation (same as the old COS version) ZVND.INC Compiler directive to ignore apparent vector dependency (same as old COS version) ZVD.INC Compiler directive to force scalar compilation (same as the old COS version) Changed in 15JUL90 and 15OCT90 same date. 5878. April 8, 1990 Cray Q-routines Kerry All new modules in $QCRI. These are the same as the old COS versions except dynamic memory management is disabled. The former COS technique (CALL MEMORY) conflicts with UNIX memory management. QCFFT.FOR QCLNSU.FOR QINIT.FOR QMAXV.FOR QMEMSZ.FOR QMINV.FOR QMULCL.FOR QPHSRO.FOR QPTDIV.FOR QRFFT.FOR QRFT.FOR QRLSE.FOR QXFOUR.FOR QXXPTS.FOR Changed in 15JUL90 and 15OCT90 same date. 5879. April 8, 1990 $SYSUNIX/AREAS Kerry $SYSUNIX/CREADIR Changed to use $SYSAIPS/AREAS.DAT instead of $SYSUNIX/AREAS.DAT and removed $SYSUNIX/AREAS.DAT. Changed in 15JUL90 and 15OCT90 same date. 5880. April 8, 1990 $SYSUNIX/AREASCSH.SED Kerry $SYSUNIX/AREASSH.SED Added looping to avoid sed'iting matching patterns in comments. Changed in 15JUL90 and 15OCT90 same date. 5881. April 8, 1990 $SYSUNIX/AREAS.CSH Kerry $SYSUNIX/AREAS.SH As generated by $SYSUNIX/AREAS from $SYSAIPS/AREAS.DAT under UNIX. Area logicals for "local" directories should be defined in $SYSLOCAL versions of CDVER.CSH and CDVER.SH now. Changed in 15JUL90 and 15OCT90 same date. 5882. April 8, 1990 $SYSUNIX/LIBR.DAT Kerry $SYSALLN/LIBR.DAT $SYSCRI/LIBR.DAT $SYSNRAO1/LIBR.DAT $SYSSUN/LIBR.DAT Commented better and eliminated references to the defunct areas $APLNUNIX, $QPGUNIX, $QYPGUNIX and $YPGUNIX. Also changed $YVTVPGM references to new logical $YPGVDEV (i.e., new name for $YVTVPGM). Changed in 15JUL90 and 15OCT90 same date. 5883. April 8, 1990 $SYSUNIX/CDVER.CSH Kerry $SYSUNIX/CDVER.SH Added comments (formerly uncommented) and changed to use AREAS.CSH and and AREAS.SH (formerly used AREAS.CSH and AREAS.SH from $SYSLOCAL). Also changed changed the execution search path definition to search the current directory, $SYSLOCAL and $SYSUNIX before all others to avoid command name conflicts (e.g., CC the AIPS C compilation script versus CC the C++ translator commonly installed as a local system utility). This is where "local" area logicals should be defined now (i.e., instead of $SYSLOCAL versions of AREAS.CSH and AREAS.SH). Changed in 15JUL90 and 15OCT90 same date. 5884. April 8, 1990 $SYSNRAO1/CDVER.CSH Kerry $SYSNRAO1/CDVER.SH $SYSNRAO1/AREAS.CSH $SYSNRAO1/AREAS.SH New files (same as $SYSUNIX versions since all local directories for NRAO1 are defined in $SYSAIPS/AREAS.DAT, thus $SYSUNIX/AREAS.CSH and $SYSUNIX/AREAS.SH). Removed obsolete $SYSNRAO1/AREAS.CSH and $SYSNRAO1/AREAS.SH. Changed in 15JUL90 and 15OCT90 same date. 5885. April 8, 1990 $SYSUNIX/ASSNLOCAL.SH Kerry $SYSUNIX/ASSNLOCAL.CSH Eliminated obsolete $AIPSTTn logicals and NRAO1 specific specific logic involving graphics terminal logicals. Added $LPRNTR logical. ASSNLOCAL.CSH is a new file used by LOGIN.CSH to define device logicals at login time. Changed/added in 15JUL90 and 15OCT90 same date. 5886. April 8, 1990 $SYSNRAO1/ASSNLOCAL.CSH Kerry $SYSNRAO1/ASSNLOCAL.SH Changed definition of device logical $DA00 from public scratch file system to file system for exclusive use by AIPS users. Also, removed obsolete $DASSGN and $AIPSTTn environment variables. Changed/added in 15JUL90 and 15OCT90 same date. 5887. April 8, 1990 $SYSUNIX/LOGIN.CSH Kerry $SYSUNIX/LOGIN.SH Changed execution search path definition to search the current working directory, $SYSLOCAL and $SYSUNIX before all others to avoid command name conflicts (e.g., CC). Also changed definitions of $CDOLD, $CDNEW and $CDTST to use CDVER.CSH and CDVER.SH from $SYSLOCAL (formerly they used CDVER.CSH and $CDVER.SH from $SYSUNIX). Also added "source $SYSLOCAL/ASSNLOCAL.CSH" and ". $SYSLOCAL/ASSNLOCAL.SH" to define device logicals at login time. Changed in 15JUL90 and 15OCT90 same date. 5888. April 8, 1990 $SYSUNIX/AIPS Kerry $SYSNRAO/AIPS $SYSSUN/AIPS $SYSUNIX/BATER $SYSNRAO1/DACK $SYSNRAO1/DRUN $SYSUNIX/RUN $SYSNRAO1/SPACE Removed ". $SYSLOCAL/ASSNLOCAL.SH". Device logicals are established at login time now (see LOGIN.CSH and LOGIN.SH entries). Also changed the execution search path definition to search the current directory, $SYSLOCAL and $SYSUNIX before all others in order to avoid command name conflicts (e.g., CC). Changed in 15JUL90 and 15OCT90 same date. 5889. April 8, 1990 $SYSUNIX/COMRPL Kerry Added protection against use on programs which royally screws up object libraries and therefore, installations. Changed in 15JUL90 and 15OCT90 same date. 5890. April 8, 1990 $SYSUNIX/SEARCH Kerry Fixed to correctly remove temporary files and corrected error return code. Changed in 15JUL90 and 15OCT90 same date. 5891. April 8, 1990 $SYSUNIX/ZLASCL.PS Kerry New file for spooling LWPLA output to PostScript printers. Moved from 15JUL90/15OCT90. 5892. April 8, 1990 $SYSALLN/ASSNLOCAL.SH Kerry Removed. Removed from 15JUL90 and 15OCT90 same date. 5893. April 8, 1990 $SYSCRI/FCOPTS.SH Kerry Changed to select static allocation of variables (i.e., like an implicit SAVE statement for each variable). Default is "stack" on all but Cray Y-MP and Cray X-MP EA systems. Changed in 15JUL90 and 15OCT90 same date. 5894. April 8, 1990 $SYSNRAO1/ASOPTS.SH Kerry $SYSNRAO1/CCOPTS.SH $SYSNRAO1/FCOPTS.SH $SYSNRAO1/LDOPTS.SH Changed default action to purge preprocessor output (ASOPTS.SH, CCOPTS.SH, FCOPTS.SH). Changed default C and FORTRAN compilation options (CCOPTS.SH, FCOPTS.SH). Changed FORTRAN compiler command from "oldfc" to "fc" (FCOPTS.SH, LDOPTS.SH). Made $OPT2 optimization option "-O2 -ngr" to avoid Convex global register optimization compiler bug (FCOPTS.SH). Changed in 15JUL90 and 15OCT90 same date. 5895. April 8, 1990 $SYSCVEX/OPT2.LIS Kerry $SYSNRAO1/OPT2.LIS Added new Q-routines to list of modules to be compiled at optimization level "-O2". Changed in 15JUL90 and 15OCT90 same date. 5896. April 8, 1990 $SYSNRAO1/SPACE.FOR Kerry Changed to call EXIT (0) instead of STOP to avoid "at" job error mail. Changed in 15JUL90 and 15OCT90 same date. 5897. April 8, 1990 $SYSNRAO1/ZLPCL2 Kerry Changed to use anonymous ftp for spooling printer files to the CVAX Versatec. Changed in 15JUL90 and 15OCT90 same date. 5898. April 8, 1990 $SYSSUN/SSS.TXT Kerry Changed text referencing ASSNLOCAL.SH to ASSNLOCAL.*. Changed in 15JUL90 and 15OCT90 same date. 5899. April 11, 1990 $APLCRI/ZILI16.FOR Kerry $APLCRI/ZCREAT.FOR Screwed up insertion of user agreement in ZILI16.FOR Screwed up merging of changes to $APLGEN version in ZCREAT.FOR Changed in 15JUL90 and 15OCT90 same date and patched UNIX distribution tape. 5900. April 13, 1990 $QCRI/QCFFT.FOR Kerry Screwed up insertion of user agreement. Changed in 15JUL90 and 15OCT90 same date and patched UNIX distribution tape. 5901.