----------------------------------------------------------------------- ; Copyright (C) 1995 ; Associated Universities, Inc. Washington DC, USA. ; ; This program is free software; you can redistribute it and/or ; modify it under the terms of the GNU General Public License as ; published by the Free Software Foundation; either version 2 of ; the License, or (at your option) any later version. ; ; This program is distributed in the hope that it will be useful, ; but WITHOUT ANY WARRANTY; without even the implied warranty of ; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ; GNU General Public License for more details. ; ; You should have received a copy of the GNU General Public ; License along with this program; if not, write to the Free ; Software Foundation, Inc., 675 Massachusetts Ave, Cambridge, ; MA 02139, USA. ; ; Correspondence concerning AIPS should be addressed as follows: ; Internet email: aipsmail@nrao.edu. ; Postal address: AIPS Project Office ; National Radio Astronomy Observatory ; 520 Edgemont Road ; Charlottesville, VA 22903-2475 USA ----------------------------------------------------------------------- This file must be used to record any changes made to the code on the Charlottesville VAX in the logon area [AIPS]. The format to use is: line 1: Date Program name Your name line 2: What was done line 3: Where has this change been moved (e.g. OLD, the Modcomp) line 4: Blank The details of line 1 are important since programs are used to find specified strings and do conversions on them. It MUST be typed as follows: 1 blank, 4 digits, a period, two blanks, then the Month, day, comma, year, some space, a title, some more space, your name. CHANGES to the 15JAN90 version of AIPS *********************************************************************** Changes while 15JAN90 was TST: *********************************************************************** *********************************************************************** 5239. August 9, 1989 QUACK Bill C Fixed QUACK to mark the FG table as unsorted. This bug would cause entries entered by QUACK to be ignored under some circumstances Moved from Oz to nowhere. 5240. August 9, 1989 GTPAIR Bill C Fixed bug processing LOGICAL keywords. This error caused FITTP to drop logical type keywords from tables resulting in SN tables being unuseable after being read back in. Moved nowhere. 5241. August 9, 1989 SNPLT Phil Added two new options, one (adverb XINC) allows the user to plot every XINC'th point, the other (OPCODE='DOPL') allows the user to plot the doppler offset column (from the CL table) as a function of time, can be a useful diagnostic for VLBI type data. Also new .HLP file Moved from Australia and nowhere else. 5242. August 9, 1989 MX.FOR Phil Fixed two problems: one was that when making cubes of maps MX would fail on the 2nd plane complaining about FREQID -1, this was happening because the work file was being overwritten but its associated FQ table was being extended not removed and re-written. The second was that the FQ table was being copied to the output map because it was not in the list of those to be excluded in the call to ALLTAB, this did no harm but was unnecessary. Moved from Australia and nowhere else. 5243. August 9, 1989 VMS Eric The VMS optimiser gets us 15--20\%\ improvement in performance when the program works. However, any program having INTEGER*2 in it appears to fail. Since this includes all TV and tape program unavoidably, I have changed SYSVMS:OPTIONS.COM to specify /DEBUG/NOOPTIM as the default. On the Charlottesville VAX, we will keep in SYSLOCAL: a version that defaults to optimization for testing. Moved to VAX1 before the recompilation of the quarterly update and to 15OCT89. 5244. August 17, 1989 SOLCL Chris New task to apply system temperature corrections from selected antennae to VLA data observed in solar mode. Moved nowhere. 5245. August 14, 1989 ZWHOMI Eric Changed generic version to suppress any messages when there are no TVs or TKs. Otherwise, poor values for the logicals TVDEV1 and TKDEV1 lead to meaningless error messages. Moved to 15OCT89 this date. 5246. August 17, 1989 MAXFIT Eric Changed AU9 and MAXFIT.HLP to add the adverb PRTLEV to allow the user to suppress the messages giving the fit coordinates and fluxes. Moved nowhere. 5247. August 18, 1989 SOLCL Eric Corrected bad Fortran --- DATA statements must follow all other kinds of declarations. A PARAMETER statement followed a DATA in the initialization subroutine. The SUN compiler cares about this ANSI rule. Moved nowhere. 5248. August 21, 1989 PRTCC Eric Added another digit (0.1 microJy!) to the displays of point components. Moved nowhere. 5249. August 21, 1989 LISTR Chris Inserted scalar directives (INCLUDE 'INCS:ZVD.INC') for the DO 150 IVIS and DO 250 IVIS loops in subroutine LISTAV. This allows LISTR to be compiled at optimization level 2 (vector optimization) with Convex fc 5.0. This version of the Convex FORTRAN compiler has a bug which causes the compiler to abort on encoutering a ``scalar (to the loop) reduction within an IF-THEN construct inside a DO-loop.'' This problem is fixed in revision 5.1 of the compiler and these directives should be removed when we upgrade to this version. No other programs in OPT2.LIS are affected by this bug. Moved to 15OCT89 this date. 5250. August 21, 1989 TVFLG Eric Changed help file --- there is a 10-second default value used for DPARM(6). Corrected error message when AN file missing or faulty in Fortran. Moved to 15OCT89 this date. 5251. August 21, 1989 PRTCC Eric Changed it to read either MA or UV cataloged files. It was restricted to maps. Moved nowhere. 5252. August 21, 1989 PRTIM Eric Raised allowed ranges of XINC, YINC, and FACTOR in the help file. Changed Fortran a bit to handle 2-axis images better --- the old code wiped out the addresses in a later call to DIETSK. Moved to 15OCT89 this date, nowhere else. 5253. August 22, 1989 UVLOD Chris UVLOD now removes any WEIGHT and/or SCALE random parameters that existed on the FITS tape when DOUVCOMP is true. Previously, UVLOD appended the WEIGHT and SCALE parameters needed for compressed data to the end of the parameter list regardless of whether they already existed. This caused problems for AIPS tasks which read the compressed data. The problem showed up in FITS tapes from the old DEC-10 system (gripe #3740) but would also affect data which had had dummy WEIGHT and SCALE parameters added by FILLM. The problem with DEC-10 tapes arose from the use of a highly unconventional scheme for storing {\it uv\/} data with the integration weight carried in a random parameter. Moved to 15OCT89 this date. 5254. August 23, 1989 UVCMP Chris. I left out the statement that restored the definition of variable WTOFF after adding a new WEIGHT random parameter in a previous modification to UVCMP resulting in UVCMP forgetting the location of the WEIGHT and SCALE random parameters. This statement has now been restored. The error was discovered while reusing the same code in UVLOD (see above) but was responsible for at least one reported gripe (#3782). Moved to 15OCT89 this date. 5255. August 23, 1989 ISPEC Phil Minor modification to protect users from themselves. Ensured that number of planes was never = 1 in input adverbs. If is = 1 then will plot all frequency planes. Moved from Australia to 15OCT89 and 15JAN90. 5256. August 23, 1989 PRTUV Phil Modified PRTUV to deal more sensibly with multi-source data. It will now inform the user when the source changes, the user also now has the ability to select sources to be printed. Also modified .HLP file and moved code to APGNOT:. Moved from Australia to nowhere. 5257. August 23, 1989 PRTUV Eric The program never reported its objection to compressed data --- it just died unnaturally. Added a call to MSGWRT. Moved fix only to 15OCT89 this date. 5258. August 24, 1989 TABIO Eric Corrected some poor code forms due to the overhaul. Added tests for keyword/value pairs and data selection strings to make sure that the record number did not exceed that allowed to fit in the file. Moved nowhere. 5259. August 24, 1989 help files Eric Created help files for OUTFILE, INFILE, and IN2FILE. These had been neglected for some reason. Moved to 15OCT89 this date. 5260. August 24, 1989 GRDROP Eric Corrected AUC to manage the user numbers when updating the gripe directory record. Moved to 15OCT89 this date. 5261. August 24, 1989 TKPL Eric Dropped INTYPE from help and Fortran. This adverb is obsolete for most --- not all --- tasks and was supposed to be dropped in the overhaul. Moved to 15OCT89 this date. 5262. August 24, 1989 FILLM Chris Corrected a DATA statement that was causing FILLM to write out WEIGHT and SCALE parameters in uncompressed data. This made a loop that was added to initialize the random parameter array so that these parameters did not contain invalid numbers redundant. This loop has been removed. It turned out that this loop was responsible for bad entries (times of 0/00:00:00.0) being entered in the index tables at the end of a FILLM run (gripe #3779). During the close-down phase, FLMDAT is called with dummy variables in place of the arrays used to hold the data in other calls to FLMDAT; the dummy variables were not arrays and so the initialization loop overwrote variables following them in memory, including the time for the last index record. Moved to 15OCT89 this date. 5263. August 25, 1989 GRIPR, BATER Eric Corrected overhaul errors in parsing the adverb = value inputs in GRIPR and BATER. Corrected errors reading in gripe name, address, and phone number involving mixed character and hollerith. Brought the GRDROP code up to date to match the much more powerful one inside AIPS itself. Also corrected a bug in BAMODIFY in AUB and some typing. Cleaned up BATER's code for new BQ file format and for new rules on QUEUES listing (0 means all queues). Moved to 15OCT89 this date. 5264. August 25, 1989 TAFFY Jeff Mangum/Eric SUBROUTINE DIDDLE was missing the DDCH.INC include, so FBLANK was not in fact declared or initialized. Moved to 15OCT89 this date. 5265. August 25, 1989 FARAD Chris Corrected a number of typos associated with the reading of TEC data. Also added code to close the TEC data file correctly and changed the handling of bad TEC data. Moved to 15OCT89 this date. 5266. August 25, 1989 COMB Eric Fixed (1) a bad branch that lead to an strange message when using OPCODE = 'MEAN' and (2) an old error in the call sequence to HICLOS that lead the Vax anyway to mess up the output HI file. Moved to 15OCT89 this date. 5267. August 25, 1989 COMB Eric Changed default for BPARM(6) for 'POLA' to 360. degrees in the help and Fortran. Moved nowhere. 5268. August 26, 1989 EDIT, MODIFY, DELETE Eric Added messages to clarify error conditions reading in the arguments to these pseudoverbs in subroutine EDITOR. Moved to 15OCT89 this date. 5269. August 26, 1989 PRTGA Eric Added extra messages to indicate continuing without the AN file. Moved to 15OCT89 this date. 5270. August 26, 1989 EXTLIST Eric Changed AU8A to skip the polarization vector parameters in PCNTR when DOVECT was false. Moved to 15OCT89 this date. 5271. August 26, 1989 APCLN, SDCLN Eric Changed to ignore BOXES adverb above the number given by NBOXES. Moved to 15OCT89 this date. 5272. August 26, 1989 IMVAL Eric Changed help files for IMVAL and QIMVAL to allow 0 for PIXXY since axes > 2 may require that value. Moved to 15OCT89 this date. 5273. August 26, 1989 IMVAL, QIMVAL Eric Changed CUBINT to use a correct formula for cubic-interpolation. The old one had wrong linear and cubic terms! Moved to 15OCT89 this date. 5274. August 27, 1989 STOKES Eric Rewrote this help file, replacing old incorrect information with something that is likely to be correct at least some of the time. Moved to 15OCT89 this date. 5275. August 27, 1989 BLSUM Eric Added more display of the units being used. Moved nowhere. 5276. August 27, 1989 PUTHEAD, GETHEAD Eric Corrected AU7A. It had an addressing problem due to missing parentheses in a Fortran numeric computation. Moved to 15OCT89 this date. 5277. August 28, 1989 GRIPE Eric Added information to help file to tell users that the AIPS version and system string are automatically recorded. Changed AUC and GRIPR to display the system release date and id string before asking for the gripe. Moved nowhere. 5278. August 28, 1989 GET Eric Changed error message to use when trying to GET a non-existant file. Moved to 15OCT89 this date. 5279. August 28, 1989 CLCAL Eric Changed typing in help file to remove tabs, improve appearance and correct misstatement regarding CALCODE test. Moved to 15OCT89 this date. 5280. August 28, 1989 plotting Eric Changed message at end of plots to give only the file size used since the guessed no longer is of any real interest and changed the message level to 2 so that it will appear in the message file as a warning about excessive file sizes. Moved nowhere. 5281. August 28, 1989 GETJY Chris In the subroutine GJYWSN, which modifies the SN tables after determining the fluxes of the secondary calibration sources, GETJY was using column 12 as the weight for polarization one rather than column 13. Column 12 holds the residual phase delay rate which is normally zero for VLA observations. GETJY therefore usually read the weight of the gain solution for polarization one to be zero and did not modify it (gripe #3748). The erroneous EQUIVALENCE has now been corrected. Moved to 15OCT89 this date. 5282. August 29, 1989 DBCON Eric Corrected help file. The explain section said that DOPOS(2,1) required that the two frequencies agree to 0.1\%\ without saying of what. It is the first freqency. Moved to 15OCT89 this date. 5283. August 29, 1989 SU tables Ralph Gaume/Eric Corrected dimension statements for number of columns in SU files in FNDSOU and in GETJY (2 places) and in SOURNU. This could have caused problems in SETJY, GETJY, UVFLG, and other tasks. Moved to 15OCT89 this date. 5284. August 29, 1989 ZPRI2 Ralph Gaume/Eric Corrected the APLUNIX: and APLBERK: versions of ZPRI2.C. They both referred to {\tt op[5]}, where 4 is the maximum subscript. This caused batch jobs to run at too high a priority. All of the tasks must be relinked. Moved to 15OCT89 this date. 5285. August 29, 1989 IMHEADER, TPHEAD Eric Corrected subroutines LSTHDR and MSGHDR to put the leading zeros in the correct columns for observed coordinates. Moved to 15OCT89 this date. 5286. August 30, 1989 J2000 precession Bill C New set of routines to precess, nutate etc. positions according to the "J2000" conventions. The top level routine is JPRECS which is similar to the B1950 routine PRECES but the call sequence is different. The new routines are: JPRECS - Top level J2000 precession routine. JPRENU - Compute rotation matrix for precession and nutation JABER - Compute vectors for aberation and GR light bending JPOLAR - Correct position for polar motion SLAEVP - Earth ephemeris JNUT - J2000 nutation series. Moved nowhere. 5287. September 5, 1989 FARAD Chris Corrected the starting index of the loop that performed the integration over the altitude profile. The error caused only a negligable error in the result. Moved to 15OCT89 this date. 5288. September 6, 1989 IMFIT Bill C Corrected bug in format statment left from the overhaul. This bug caused the display of the RA and Dec reference pixel to be trash and appears to terminate in output if sent to a QMS printer. Moved to 15OCT89 this date 5289. September 8, 1989 CALIB Chris CALIB now ignores the adverbs CALSOUR, QUAL, CALCODE, SELBAND, SELFREQ, FREQID, TIMERANG, BCHAN, ECHAN, and ANTENNAS if used in self-cal mode on a single-source file, as documented in the HELP file. The application of the ANTENNAS adverb to single- source data was associated with a significant problem in this mode of operation. The particular problem arose from the the ill-behaved nature of the library subroutine UVGET with regard to its inputs; that is the fact that UVGET may modify its inputs when called with OPCODE='INIT'. In particular, if the user requested that antenna $n$ be excluded, UVGET was called with ${\tt ANTENS) = ( -n, 0, \ldots)$ to initialize the first (solution) pass through the data but returned with ${\tt ANTENS) = ( +n, 0, \ldots)$. The new value of ANTENS was passed to UVGET when UVGET was called to initialize the second (copy and correction) pass through the data. The result was that the two passes were made with contradictory selection criteria. This came to light when a user tried to exclude an antenna that did not exist in the data: all data were rejected by the selection criteria on the second pass. The route of choosing to ignore the above adverbs rather than patching over the immediate problem was taken since it is not clear how these adverbs apply to single-source files; while it may be desirable to limit solutions to some subset of the data it is probably not desiable to restrict the output data set to that subset of the data for which solutions were found. In addition 'FQ' was added to the list of excluded table types used by ALLCOP (call from CLBHIS) to prevent duplication of the FQ table in the output file. Moved to 15OCT89 this date. 5290. September 11, 1989 LISTR Phil Changed format statement when ch. sep printed out for optype = 'scan'. Moved to 15OCT89 this date. 5291. September 12, 1989 GRIPES.ML Nancy Modified DOCGRIP:GRIPES.ML gripe file cleanup routine to include changing " = " to "=", the only feature not contained in the almost identical GRIPEMODE.ML file which is autoloaded by emacs when a gripe-type file is detected. (See NRAO$ROOT:[EMACSV2.NRAOMACS]GRIPEMODE.ML.) On CVAX, I have eliminated the duplication by making the autoloaded file empty except for a command to point to the DOCGRIP:GRIPES.ML file, since it is an AIPS-specific feature of emacs. Since we do not port the DOCGRIP: directory to nrao1 at this point, I have edited the autoloaded file on nrao1 (/nrao/emacsv2/nraomacs/gripemode.ml) to be the same as the gripes.ml file. Moved nowhere. 5292. September 12, 1989 FILLM Bill C/Phil Some fairly major changes to overcome the problems of the last few months: (1) Changed band code '7' to '4' for 75MHz data (2) Don't check the bandwidths of different files when looking for a match in FLMFND, AIPS data files can now have different bandwidths. (3) Reset the I/O sizes and counters when the new output file does not match the current set known about internally. (4) Reset the channel selection when switching output files, all files were appearing with the number of channels found in the first data set. (5) Ensure that there are separate source lists for each data stream. (6) Ensure that the correct FQ numbers are written to all the files, some things were not being initialized correctly when opening new files and recovering the FQ information from files already created. (7) Ensure that each new line uv file also has a new channel 0 file, this is done by attaching a keyword/value pair to the catalogue headers which FILLM can then read to determine if new data should be attached to this file or not. Was not a problem with line data but was with channel 0. (8) The selection by observing mode was broken, fixed this. Moved to 15OCT89 soon. 5293. September 14, 1989 FILLM Phil I missed a couple of things, when FILLM started reading a new tape the reference frequency it got was incorrect, this would occasionally cause incorrect FQ numbers to be assigned. Also the reading of keyword/value pairs didn't work because I had not told CATKEY what keyword to read. Added a message informing the user of the FQ entry tolerance being used. Moved nowhere, will move to 15OCT89 soon. 5294. September 15, 1989 LISTR Phil One of the frequency labels printed in the 'list' and 'matrix' modes was incorrect due to the IF offset being subtracted twice. Fixed. Moved to 15OCT89 this date. 5295. September 15, 1989 Faraday rotation Chris Added Faraday rotation corrections to the calibration routines in the UVGET package. The ionospheric Faraday rotation measure is read from the CL table (as inserted by FARAD) and applied to the cross-polarization data if DOPOL is set to true. Changes were made to the following files. DSEL.INC --- added the arrays LAMSQ (a look up table of wavelengths squared, indexed by channel number and IF), IFRTAB (the IFR values as read from the CL table) and IFR (the interpolated IFR values) to COMMON /CGNINF/. GAININ.FOR --- added code to initialize LAMSQ. CSLGET.FOR --- added code to read values into IFRTAB. CGASET.FOR --- added code to interpolate IFR values. DATPOL.FOR --- added code to make correction (rotation is subtracted from RL and added to LR). Recompiled all subroutines accessing DSEL.INC and recompiled SPLIT, LISTR, CALIB, PRTUV and UVPLT to pick up changes. Moved nowhere. 5296. September 17, 1989 FQMATC Phil Modified so that the default value for freqid (-1 or 0) does something sensible, it now takes the first FQ number (and tells the user). Also check that if freqid is set it is a legal value. Recompiled all routines that call fqmatc. Moved to 15OCT89 on Sept 22. 5297. September 17, 1989 POSSM Phil Changed slightly the error handling after FQMATC. Moved to 15OCT89 on Sept 22. 5298. September 17, 1989 ZUVPAK Phil When generating the scale factor for a spectrum the wrong element of an array was being examined to see if it was zero. Fixed. Moved to 15OCT89 on Sept 22. 5299. September 18, 1989 CALIB Chris Globally replaced instances of IFR by IFRM to avoid a conflict with DSEL.INC. Moved nowhere. 5300. September 18, 1989 SOLCL Chris Replaced the PARAMETER statement that defined the LUN used to read the FQ table (this had previously been removed since it was incorrectly positioned after a DATA statement but had not been replaced in a legal position). Also converted the value of SELFREQ to a DOUBLE PRECISION variable before calling FQMATC. Moved to 15OCT89 this date. 5301. September 18, 1989 ZTPMI2 Bill J./Chris Initialisation of FCB(1+FCBSP2) prior to tape i/o check in ZTPWA2 was incorrect. This meant that FITTP failed. Moved to 15OCT89 this date. 5302. September 18, 1989 FQMATC Phil Under some circumstances the FQ table was not being closed upon exiting the subroutine. Moved to 15OCT89 on Sept 22. 5303. September 18, 1989 FREQID programs Phil Improved the error handling after FQMATC in all programs that call it. They now die in a more sensible manner. Moved to 15OCT89 on Sept 22. 5304. September 18, 1989 FQTOL Phil New adverb; specifies the tolerance with which two frequencies are compared when deciding what FQ number to assign them. Will be used in DBCON. Also .HLP file. Moved to 15OCT89. 5305. September 19, 1989 RSTOR Patrick Leahy/(by Bill C) New task to convolve a set of clean components with a Gaussian and add them to an image. Also RSTOR.HLP. Moved nowhere. 5306. September 19, 1989 History merging Patrick Leahy/(by Bill C) Added new history routines HIMERG and HIADDN to be used when several history files are to be concatenated and one or more history files are to be produced. Multiple copies of the same history file will NOT be written to the output history file(s). Moved nowhere. 5307. September 19, 1989 ASCOR Patrick Leahy/(by Bill C) Modified ASCOR to use new history merging routine HIMERG. Moved nowhere. 5308. September 19, 1989 .OPT files for QPGM Bill C Added QPGMOPT.OPT and QPGMOPT1.OPT to SYSVMS to allow RSTOR to link. Moved nowhere. 5309. September 19, 1989 D/VAPT.INC Bill C Added RSTOR to list of AP tasks. Moved nowhere. 5310. September 19, 1989 PCAL Chris. Added support for ionospheric Faraday rotation to PCAL. PCAL calculates the ionospheric Faraday rotation for each telescope and visibility when copying the (optionally calibrated) data to a scratch file in PCLCOP using data maintained by UVGET. The rotation is taken into account when subtracting the model polarization from the observed cross-polarization streams in LPCALC (approximate solution for circular feeds) and XYCALC (approximate solution for AT type linear feeds). Moved nowhere. 5311. September 19, 1989 LIBR.DAT for UNIX Systems Kerry Added link list for programs stored in QPGM, a formerly empty directory. Moved nowhere. 5312. September 22, 1989 SOURNU Bill C Cleaned up handling of deselected sources. Also now returns NID = 0 when all sources are selected. Changes declarations of SUNUMV and SUKOLS to use the parameter MAXSUC from PUVD.INC; the previous declarations were too small. Moved nowhere. 5313. September 22, 1989 FREQID, SELBAND, SELFREQ Phil Modified the help files slightly to state what the default values of these adverbs signify to the tasks. Moved to 15OCT89 this date. 5314. September 23, 1989 FILLM Phil Yet another round of fixes, we'll get there. (1) When reading in a second tape and concatenating files the file was being doubled in size instead of incremented by 10000 vis. due to an incorrect index. (2) If the number of polzns or IF's changed between scans FLMCLS would have problems because the values had been changed too quickly. Modified to update various parms in the commons after previous scan has been fully dealt with. (3) In 'LP' mode all of the 'P' band data was being written into seperate files (i.e. one per scan), confusion of what polzn was being read from the tape in this mode. (4) Perley couldn't live without obselete correlator mode 4A, his life is now safe. (5) Occasional Nan's would be present in the nominal antenna sensitivities, if data is solar data these are now checked for and the user is warned. Moved nowhere yet. 5315. September 23, 1989 GETFQ Phil New subroutine, examines the FQ table to obtain information for a specific FQ entry. Put into aplnot. Moved nowhere. 5316. September 23, 1989 POSSM, UVPLT Phil Modified so that the plot label now contains information about the freq/bandwidth for data with a FREQSEL random parameter. Moved nowhere. 5317. September 24, 1989 VLAPROCS.001 Bill J/Phil Removed the freqid kludge that had been inserted to overcome the FQMATC shortcomings. Moved to 15OCT89 this date. 5318. September 24, 1989 VLACLCAL.HLP Phil Added FREQID to the helpfile, it is an adverb that users need to see. Moved to 15OCT89 this date. 5319. September 24, 1989 MAKMAP Phil Minor problem, when the doband option was switched on makmap tried to remove the FQ table from the wrong scratch file and so was getting a non-catastrophic ZCREAT error message. It now does the correct thing. Moved to 15OCT89 this date. 5320. September 25, 1989 VMS Q Routines Glen VMS PSAP version of UVMAP failed because of changes from F-66 to F-77 in subroutines QVMS:QPUT,QGET and QVMOV, although old logic did not appear correct. Incorrect block transfers corrupted map, placing stripes in BEAM. Variable redeclared to I*2 and block transfer logic clarified. Moved to 15OCT89. 5321. September 25, 1989 UVPGET Chris Changed an erroneous KICTPN to KIPTPN. The error meant that UVPGET could not find existing random parameters after the seventh random parameter. This showed up since data read from the DEC-10/ISIS variant of FITS had enough junk in the random parameter list to push the SOURCE parameter beyond slot seven when converting to a multi-source file with MULTI. Moved to 15OCT89 this date. 5322. September 25, 1989 SOLCL Phil Added missing square root operations. Moved nowhere. 5323. September 25, 1989 REVERSE Mark Calabretta/Glen Declared variable J. Moved to 15OCT89. 5324. September 25, 1989 GREYS Eric Corrected bug in axis-information display for LTYPES 4 and 5. The rotation was omitted and the center coordinates were shown twice, but only on rotated images. Moved to 15OCT89 this date. 5325. September 26, 1989 UVIMG Mark Calabretta/Glen UVIMG calls to CATDIR and MAPOPN with contant string parameter were changed to variable parameter PTYPE because these subroutines modify the value. Moved to 15OCT89 this date. 5326. September 26, 1989 TXPL Glen TXPL aborted when writing to a history file because a format statement had the wrong number of characters. Moved to 15OCT89 this date. 5327. September 26, 1989 VLACALIB.HLP, VLAPROCS.001 Phil After user pressure, modified the procedures so that the VLACALIB proc would apply any calibration previously generated. Also removed the printing of the inputs from VLACLCAL because they cover more than one page and require the user to be present and raise a finger to press a button. Moved to 15OCT89 this date. 5328. September 30, 1989 HORUS Phil Added the QUAL adverb to the data selection. We are going to need a qualifier array for general usage. Modified .HLP also. Moved to 15OCT89 this date. 5329. September 30, 1989 ISPEC Eric Rewrote the program. Previously it was limited to the simplest spectral-line cubes, used a variety of non-standard defaults, mislabled the velocity/channel axis, and had a confusing choice between PIXXY and BLC/TRC to set the area averaged for the plot. Changed the Fortran and help file to be more general, to use standard defaults, to use standard (slice) labeling routines, to use only BLC/TRC, and to offer a choice of plot types. Also changed standard subroutines LABINI (to spell {\tt CENTER AT} as {\tt Center at}), CLAB1 (to null out the Y axis also on LTYPE = 2), and CTICS (to define input parameters correctly after the overhaul). Moved nowhere. 5330. September 30, 1989 EXTLIST displays Eric Corrected SNPLT --- the antennas used on each page were not correctly entered in the plot header and the wrong plot type was used for 'SUM ' plots. Changed KNTR to plot type 18, UVHGM to type 19, and ISPEC to type 20. Corrected AU8A for SNPLT --- a call to MSGWRT was omitted and the address to OPTYPE had changed due to the addition of an input parameter. Changed AU8A to differentiate KNTR and CNTR, displaying the planes used for KNTR, and to add displays for UVHGM and ISPEC. Changed AU8A to check all ``unknown'' extension types to see if they are standard tables and to give a useful display if they are. Moved nowhere. 5331. October 1, 1989 MCUBE Eric Corrected WRPLAN --- it was executing a MSGWRT call even on a normal termination. Added the branch around the call. Moved to 15OCT89 this date. 5332. October 2, 1989 IMEAN Glen Add option to write to a log file for later user processing, when file name is specified with additional parameter OUTFIL. Moved nowhere. 5333. October 3, 1989 FQMATC Phil Added ability to deal with Ch tables, just sets FRQSEL = 1. Moved to 15OCT89 this date. 5334. October 3, 1989 TACOP.HLP Bill J. Added some notes to HELP file for VLBI processing. There was some confusion about the use of TACOP for AN files. Users were generating versions of the AN file > 1 for a data set with only one subarray. This caused some problems when UVFIX was run. (This has to be done for data from the CALTECH Mk2 correlator.) AN file diddling is a common requirement for VLBI data sets since the file often lacks essential information. Moved to 15OCT89 this date. 5335. October 3, 1989 GREYS Eric Corrected oversight --- it was placing step-wedge labels to the right of the plot when DOWEDGE <= 0 since there was no test for not doing so. Moved to 15OCT89 this date. 5336. October 4, 1989 CALIB.FOR Phil Removed a spurious msgwrt which was annoying Mk3 vlb users. Moved to 15OCT89 this date. 5337. October 4, 1989 PTIME.FOR Phil Little routine which when fed the time in radians or days produces output which can be easily printed in standard format. Moved to 15OCT89 this date. 5338. October 4, 1989 TXPL Glen TXPL was hanging in batch jobs because of non-standard IO, so these routines were replaced by calls to MSGWRT and PRTLIN. Moved to 15OCT89 this date. 5339. October 5, 1989 DDT Eric Changed DDTSAVE.HLP, adding adverb NCOMP to be saved. Changed DDTLOAD.001 by making UVDIF a bit less demanding, adding adverb NCOMP for each source to specify the number of merged clean components used in ASCAL, changed destruction of any pre-existing ``master'' images plus loading of new ones to be done one type at a time based on TMASK, added CCMRG for all cases ahead of ASCAL, improved final displays adding a PRTMSG of PRIORITY 5 and above as ``answers''. Moved nowhere. 5340. October 6, 1989 Copyright notices Bill C The copyright statments were replaced by another statment in all Fortran, C, Macro and Help files. Mostly this was done outside of the checkout system (since the changes were made only in comments). The copyright notice displayed by ZDCHIN for some stand alone programs was changed to the new statment and comments in example code in IMPFIT were changed. The new statment is: ``This software is the subject of a User agreement and is confidential in nature. It shall not be sold or otherwise made available or disclosed to third parties.'' Moved from 15OCT89 this date. 5341. October 6, 1989 PRTUV Phil Minor problem in the source selction in prtuv, under some circumstances all sources would be disqualified and it would print nothing. Moved to 15OCT89 later. 5342. October 6, 1989 MULTI Phil Multi would write an FRQSEL r.p. even if no CH/FQ tables were present. Fixed. Moved to 15OCT89 later. 5343. October 6, 1989 UVFIX Bill J. For Earth-centred arrays, the arguments to ATAN2 in the evaluation of position for diurnal aberration calculations were both 0.0. ATAN2 is undefined in this situation. Added elementary test to trap this. Moved to 15OCT89 this date. 5344. October 6, 1989 PRTIM Eric Corrected overhaul bug --- the image name and class were all run together. Moved from 15OCT89 this date. 5345. October 9, 1989 DDTLOAD.001 Kerry Fixed bug that caused the ASCAL step to fail if DDISK and MDISK were not the same. Moved nowhere. 5346. October 10, 1989 FARAD Chris Changed the value of the constant used to calculate the Faraday rotation measure from 2.22E-17 to the correct value of 2.62E-17. The error was probably a typo. Moved to 15oct89 this date. 5347. October 10, 1989 TABINI Bill C. Modified to deal with "forgotten" tables; i.e. tables whose existance is not indicated in the catalog header. On reads an old but uncataloged file is cataloged and opened. On write an old but uncatalogued file is deleted before creating a new one. In either case a message is printed. Thus "lost" tables which still exist may be recovered by reading them with, for example, PRTAB. Moved nowhere. 5348. October 11, 1989 VLBIN Phil Modified the way the data weights were written so they reflect the number of bits actually correlated, previously a weight of 1.0 was given to every data point. The new weights are normalized such that they are close to 1.0 and will in fact be almost always a similar number for each data point. Also rationalised the way in which the infile, in2file adverbs are given, previously the user had to use a : on a Vax and a / on a Convex, now use : always. Modified the .HLP file to reflect this. Moved to 15OCT89 this date. 5349. October 11, 1989 FITTP.HLP Mark Calabretta/Bill C Improved the description of the use of adverb DOTABLE. Moved to 15OCT89 this date. 5350. October 11, 1989 TRANS Eric Dropped all handling of integer input images since that is obsolete. Corrected buffer size to 32768, rather than 32767. This caused severe errors in disk-based x-y transposes since the disk IO had a different idea of the buffer size than did the transposition algorithm. Later changed the buffers to be four times larger in the hope of speeding things up some. Intermittant failures on VMS were fixed by the corrections to MINIT and MDISK. Moved to 15OCT89 this date. 5351. October 11, 1989 MINIT, MDISK Eric The fundamental disk IO routines failed to wait for the next to last buffer write to complete. This could cause bad data to reach disk if that next-to-last buffer was reused fast enough. TRANS was able to do this on disk-based x-y transposes. Corrected MINIT and MDISK to record in the FTAB which buffers still need a wait to be performed and to do any pending ones when the FINI code is received. Moved to 15OCT89 this date. 5352. October 12, 1989 MX.HLP Bill C. The HELP section incorrectly stated that EIF=0 meant to grid only 1 IF. EIF=0 means grid up to the highest IF number. Also cleaned up the description of imaging spectral line data. Moved to 15OCT89 this date. 5353. October 12, 1989 BPASS,UVGET,DATGET,DSEL.INC Phil Modified BPASS so that users can now specify an external file to use as the channel 0 when normalising by channel 0. The IN3NAME set of adverbs are used to define the 'channel 0' file. BPASS insists that the line and channel 0 files be identical in terms of sort order, times, baselines etc. i.e. the data in the two files must line up. This modification necessiated changes in UVGET, DATGET and DSEL.INC so all the calibration package has to be relinked. Also updated BPASS.HLP. Moved nowhere. 5354. October 12, 1989 ZCMPR2, ZCREA2, ZDEST2 & ZRENA2 for UNIX Kerry The call to the UNIX "sync" function has been eliminated in the Z-routines used for compressing, creating, destroying and renaming files. This should make these functions considerably faster on systems configured with many file systems (e.g., NRAO1, CHOLLA & YUCCA). For example, deleting files on NRAO1 appears to be about a factor of 4 faster. Moved mowhere. 5355. October 13, 1989 UVMTH Bill New task. Time averages the visibilities in on uv data set on each baseline and adds/subtracts/multiplies/divides them to/from/by/into the visibilities in another data set. Also UVMTH.HLP Moved nowhere. 5356. October 13, 1989 DGETSD, SDGET, TVFLG Phil After modifying DSEL.INC I discovered the three routines above had the same internal names for one of the variables I had inserted into the include. Modified so this was not the case. Moved nowhere. 5357. October 13, 1989 UVFLG Bill J. Fixed this so that single antenna could be deleted throughout specified time range. Previously, the baseline selection routine in UVFTAB screwed up resulting in search for baselines x-x. Moved 15OCT89 this date. 5358. October 13, 1989 BLFIT Bill C. New task. Currently solves for source and/or antenna positions from residual phase measurments in a CL or SN table. Also BLFIT.HLP. Moved nowhere. 5359. October 14, 1989 MK3IN Bill C. Numerous modifications to allow reading polarization observations data from Haystack MKIII correlator tapes. Also MK3IN.HLP. Moved nowhere. 5360. October 15, 1989 POSSM Phil Spedup considerably in case of multiple IF plotting. Previously POSSM would go through the whole calibration loop for all IF's in order to plot just one, then would repeat the process for the next IF. Now does all calibration one time and puts arrays to be plotted in array dimensioned by IF number. Moved to 15OCT89 this date. 5361. October 16, 1989 CLCAL Phil Made default opcode='cali'; previously task insisted users specify an opcode. Modified .HLP file also. Moved nowhere. 5362. October 16, 1989 LISTR Phil Added two new abilities; (1) when doing a matrix listing of amplitude and rms the default is to now use the scaling factor of the amplitude matrix to display the rms, if a separate scale is required for the two matrices that is now an option; (2) another option was inserted to provide DEC-10 like gain listings, the old style listings are still available if required. Also inserted the option to display amplitude and phase in one run for optyp='gain'. Modified .HLP file to describe these changes. Moved nowhere. 5363. October 16, 1989 VLBIN Phil Typo in an equivalence resulted in the polar motion x coord being inserted into the AN table as the frequency. Corrected. Moved to 15OCT89 this date. 5364. October 16, 1989 EXTLIST Eric Changed AU8A to correct call sequences to H2CHR used to display UVHGM plot file information. Moved nowhere. 5365. October 17, 1989 INTYPE Eric Removed INTYPE adverb from EXTDEST help file and subroutine AU8. Removed non-existant OUTTYPE from the CLRONAME help file. Corrected TRANS help file --- INTYPE has not been used for some time. Changed PRTPL and SLFIT and their help files to remove INTYPE. Moved nowhere. 5366. October 17, 1989 DCODEF Phil Was screwed up in the overhaul, the result was that AN tables (at the very least) read from a uvfits tape were wrong in that the antenna coordinates were incorrect by a factor of 10**10. Moved to 15OCT89 this date. 5367. October 17, 1989 ZLASC2 for UNIX Greg Hennessy/Kerry The second argument to the access function was missing in the UNIX version of this routine and is now supplied. Moved to 15OCT89 same date. 5368. October 18, 1989 FILLM Phil The way frequencies were being dealt with in the new era of FQ tables was slightly inconsistent in that for 2 IF data the IF offset was being added to the SU table as well as the FQ table. Now keep track of the total frequency at all times and the offsets are correct. Also changed the default value of the FQ tolerance, now all data will have the same FQ number by default (unless the bandwidth changes), this is the option that will keep most users happy most of the time. Also modified .HLP file. Moved to 15OCT89 this date. 5369. October 18, 1989 SETJY Phil Inserted a call to FQMATC to ensure that the correct FQ number is set when the user specifies the default. Moved to 15OCT89 this date. 5370. October 19, 1989 TABED Bill J. Character array KEYDUM in subroutine TABUP was incorrectly declared. This resulted in truncated character strings for keywords being inserted into the new table file. Moved to 15OCT89 this date. 5371. October 19, 1989 TVFLG Chris Removed a superfluous division by 10 in the upper bound for the smoothing time. This was causing TVFLG to enter an infinite loop if the user attempted to reset the smoothing time when there were less than 10 pixels on the time axis of the baseline-time map. Moved to 15OCT89 this date. 5372. October 19, 1989 Location common Eric Changed DLOC.INC to allow axis type strings to be up to 20 characters long. This is not likely to be desirable most of the time, but 8 characters is too limiting. Changed as a result: AXSTRN Limit this display to first 10 real characters of axis type --- fewer than 8 is better. CLAB1 Corrected string lengths, calls to TRIM, and format to allow longer CTYP. Cleaned up typing. CLAB2 Corrected string lengths, calls to TRIM, and format to allow longer CTYP. Cleaned up typing. IAXIS1 Corrected format and calls to TRIM to allow longer CTYP. LABINI Corrected axis label spelling now that more characters are allowed. SETLOC Increased length of temporary string and simplified typing. TKLAB Corrected format and calls to TRIM to allow longer CTYP. Moved nowhere. 5373. October 20, 1989 UVHGM Eric Changed the program around a bit, correcting a variety of minor errors (wrong list of allowed Stokes, wrong plot ranges) and changing it to use more standard labeling routines. Changed help file, adding advrebs BIF to control the IF selected and DOALL to control whether flagged data are considered at all (they are counted separately in any case). Changed AU8A (verb EXTLIST) to be more suitable to the changed inputs and stored parameters. Moved nowhere. 5374. October 20, 1989 UVFIX PHIL Modified to deal with compressed data. Moved nowhere. 5375. October 20, 1989 TBIN/TBOUT Glen New Table input and output routines which read/write to/from human and editor compatable text files. Should handle all aips tables of any length. Moved nowhere 5376. October 20, 1989 IDWCRD Glen FITS format card reader, parser derived from GTWCRD, but has input of a variable length CARD. It does not depend on input from FITS blocks of 2880 cards, so it can be used with both TAPES and TEXT files. Moved nowhere 5377. October 21, 1989 XSUM Eric Doubled the buffer size to allow better IO on large rows. Moved nowhere. 5378. October 21, 1989 Documentation Eric Changed documentation of file formats for files of types AC, BA, BQ, CA, CB, HI, IC, MS, SG, TC, TD, and TS. The CB and TC ones are new. All are stored as DOCTXT:MV2C06xx., where xx is the first two characters of the file name. Most of the changes are for changes brought about by the code overhaul, including some that make the formats more efficient. Moved to 15OCT89 this date. 5379. October 23, 1989 VLBIN Phil Always assumed the maximum number of antennas possible was 50, this screwed up the end of the AN table. Fixed so now takes MAXANT from PUVD.INC. Moved to 15OCT89 this date. 5380. October 24, 1989 TRANS Eric Made two additional corrections: (1) In the row swapping stage, one of the addressing indices was incorrectly used. I guess we got away with this by only doing things with the first three axes. (2) In the disk-based transpose, the scratch file was able to be absurdly large when the input plane was also large. Fixed the code and IO calls to allow more than one group per sector when that is relevant. Moved to 15OCT89 this date. 5381. October 24, 1989 FILLM Phil An equivalence statement was erroneously plasced after a data statement, was spotted by the Sun compiler. Moved to 15OCT89 this date. 5382. October 24, 1989 BPASS Phil/Bill C A problem occurred when writing bandpass entries for the 2nd of two polarizations when the 1st had been flagged. Corrected to ensure that all entries written. Moved to 15OCT89 this date. 5383. October 24, 1989 LISTR Phil Modified so that when listing gain solutions in the DEC-10 style the scale factor derived from the first record is applied to all. Added new adverb, FACTOR, to enable the users to change the scale used if desired. Also tidied up the optyp='list' format slightly, and added a line to tell the users which BP table is applied if requested. Modifed .HLP to reflect these changes. Moved nowhere. 5384. October 24, 1989 PRTAB Chris Corrected the section of code that outputs numeric or logical data while checking for indefinite values. PRTAB was checking a REAL quantity equivalenced to the first four bytes of the value against FBLANK. This is unsafe since the bit patterns representing valid integers or logicals may not represent valid floating-point values; on some machines (eg. Convexes) this may result in floating-point exceptions. In addition, there are no indefinite values for integers or logical values. Moved to 15OCT89 this date. 5385. October 24, 1989 Copyright notices again Bill C. Changed the copyright notices to mush in the distributed RUN files in RUNSYS and all the *.VFC routines in QFPS, Q120B, Q5000 and QPSAP. Moved to 15OCT89 this date. 5386. October 24, 1989 Overhaul Bill C. Overhauled QFPS32:QTYP.FOR and QFPS:APPEEK.FOR which had previously been missed. Moved to 15OCT89 this date. 5387. October 25, 1989 BPASS, LISTR Phil Fixed a couple of typos which only the Sun compiler caught. Moved nowhere. 5388. October 25, 1989 Removal of unoverhauled code Bill C. Source code which was not overhauled and which will cause problems in the installation procedures was removed from the distributed directories and put under LOCAL:[DEAD.]. If any of the routines are required in the future they can be overhauled, tested and reinstituted into the standard directories. These included: AIPGVMS:FIXFIL, PRNTMN, ZVERPL, VLAMBX. APLVMS:ZDOPR*.* APGVMS:CITCC,COPY,IMGPL,OUTPLT,PRFPL,PRIMG,TOVLB,TVCUB,TVSLD, KONTR,VBCC,VBCIT,WSLOD and corresponding *.HLP files. YARGS:*.* YLEXC:*.* YV20:*.* YNOT:IMIO,TVDISP,YCAMER Moved to 15OCT89 this date. 5389. October 25, 1989 TXPL glen Fixed another bug keeping TXPL from working in batch mode. Moved to 15OCT89 this date. 5400. October 25, 1989 PUTCOL glen Created twin of GETCOL which puts an column of a table row (i.e. an element of a table) into an already opened table. Works with any table type. Moved to nowhere. 5401. October 25, 1989 TABPAK glen Created a routine which packs Table Keys into form used by TABKEY. Moved to nowhere. 5402. October 25, 1989 TABSPC glen Combined TABFRM and TABF3D routines which decode character strings in FITS headers. The strings discribe the ascii format of columns of FITS Tables. TABSPC performs the functions of both routines. Moved to nowhere. 5403. October 25, 1989 TABHDK, TABAXI glen Routines which read FITS cards and decode them. More general than TABHDR and EXTREQ because they work with either TEXT or TAPE files. Will later replace these routines. Moved to nowhere. 5404. October 25, 1989 TBOUT glen Added lines which mark begin and end of data section of output TEXT file. Moved to 15OCT89 5405. October 25, 1989 TBIN glen Allow variable number of input Table Rows and read Rows in any order. Moved to nowhere. 5406. October 26, 1989 XEDIT1.STR Bill C. Removed copyright stuff from the list of XEDIT commands. Moved to 15OCT89 this date. 5407. October 26, 1989 PBCOR Eric Made it work. The Fortran constant 1.0D0 must have been compiled as 0.9999999999... to make a test for illegal $\cos^{-1}$ argument fail when the argument was exactly 1.0. Changed the test to adjust the argument rather than report an error. Found an unexplained --- and totally wrong --- use of the {\it absolute value} of the axis increment in finding the pointing position pixel. Any observed or input (GPOS) pointing position off the reference pixel was used, but used incorrectly as a result. This applied only to the ``fast'', semi-linear mode of angle computation. Removed the alteration of an input argument by an internal subroutine, which could cause aborts on Convexes and Vaxes. Changed the task to prohibit the copying of CC files; without applying the beam corrections to the components, we cannot regard them as any sort of model of the PBCOR output image. Also removed computations from the inner loop, achieving a 20\%\ reduction in the CPU times of the accurate computations. Moved to 15OCT89 this date. 5408. October 26, 1989 TACOP Chris Modified TACOP to take the input version number into account when calculating the number of tables to copy when NCOUNT is zero. This prevents TACOP from attempting to copy tables with version numbers higher than the maximum present (this was harmless, but the error messages were worrying some users). Moved to 15OCT89 this date. 5409. October 26, 1989 More copyright stuff Bill C. Removed copyright notices from QVEX:ISAMAX.s, WHNALT.s Moved to 15OCT89 this date. 5410. October 27, 1989 CALIB Phil Added the source name to the closure error message. Moved nowhere. 5411. October 30, 1989 BPASET Phil BPASET was not dealing correctly with bandpass entries that were flagged in one polarization. Moved to 15OCT89 this date. 5412. October 30, 1989 UVCOP Phil Added ability to copy either AC or XC data to new file. Default is still to do both. Modified .HLP file to reflect change. Moved nowhere. 5413. October 31, 1989 BL tables Phil Modified the BL tables so they would cope with FQ numbers. This needed one new routine, BLREFM, to reformat from old to new BL tables if necessary. Also BLGET, BLINI, GAININ and TABBL were modified. Also modified BLCAL and its .HLP file to take the 3 FQ adverbs and do the appropriate things with them. Moved to 15OCT89 this date. 5414. October 31, 1989 UVCOP Phil Added a check to ensure that AC data would be written when an antenna was specified in the ANTENNAS adverb. Moved nowhere. 5415. October 31, 1989 APLNOT/SN2CL.FOR M.Calabretta/B.Junor (1) Two incorrect branches to statement 650 rather than 700 fixed. (2) Fixed an error in counting the logical record number CLORNO of the output CL file. If COUNT=0, loop 700 was entered with the value of CLORNO returned by CALINI near statement 40. CLORNO then was being incremented before the call to TABIO with erroneous results. (3) Fixed a problem when the input SN table had duplicate entries. Moved to 15OCT89 this date. 5416. November 1, 1989 CVEL Phil Several changes: Now reads and writes compressed data. Removed the interpolation option in calculating the frequency shift, this could be in error by several % over a scan. Fixed up a couple of bugs left over from the overhaul. Gives more informative messages after each scan Moved to 15OCT89 this date. 5417. November 1, 1989 UVPLT glen Added check and warning message to UVPLT when DOCAL is true but no SN table was present. UVPLT now continues when no SN table is found, does no calibration and warns user. Moved nowhere 5418. November 1, 1989 UVFIX Chris Added a missing cosine declination term to the calculation of the rotation of the coordinate frame from apparent RA and declination to mean RA and declination (gripe \# 3346). Also modified UVFIX.HLP to clarify how SHIFT is applied to the tangent position, to explain how APARMs 3 to 8 inclusive are used to generate BX, BY and BZ (gripe \#2460) and to give the sign of the clock error correction (gripe \#2460 again). Moved nowhere. 5419. November 1, 1989 ACFIT Phil New task. Uses the 'template' method to calibrate spectral line data (mainly useful for VLBI). Performs a least squares fit of a high quality, calibrated total-power spectrum to a series of other total-power spectra as a function of time for all antennas in the array. Writes the gain solutions in an SN table. Also a .HLP file. This is still an experimental program, use with caution. Moved nowhere. 5420. November 2, 1989 UVCOP Phil Modified so that the data type (AC/XC) is the first check that is made when deciding to copy a visibility record. Moved nowhere. *********************************************************************** Changes while 15JAN90 was NEW: *********************************************************************** *********************************************************************** 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 from 15APR90 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 from 15APR90 this date. 5423. 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 from 15APR90 this date. 5424. 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 from 15APR90 this date. 5425. 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 messages. Moved to 15APR90 this date. 5426. November 13, 1989 UVFND Bill J. After testing for compressed data, the program failed to warn the user if it found same. The program died uninformatively due to a missing call to MSGWRT. Moved from 15APR90 this date. 5427. November 13, 1989 POPSDAT Eric Removed a second declaration of RESTFREQ, it had caused an addressing error of one location. Moved this fix from 15APR90, nowhere else. 5428. 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 from 15APR90 this date. 5429. November 14, 1989 SPLIT Phil The alternate reference pixel written to the output catalogue header was incorrect by 1. Fixed. Moved from 15APR90 this date. 5430. November 14, 1989 Miscellaneous Eric 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 fixes from 15APR90 this date. 5431. November 14, 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. 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 from 15APR90 this date. 5432. 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 from 15APR90 this date. 5433. November 14, 1989 ALGSUB/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 congigation when it should have been performing AP I/O. Moved from 15APR90 this date. 5434. November 14, 1989 VLBIN Phil Removed array declerations and a couple of intrinsic function calls that violated ANSI standard rules. Moved from 15APR90 this date. 5435. 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 from 15APR90 this date. 5436. 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 from 15APR90 this date. 5437. 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. 5438. November 15, 1989 ALGSUB Glen Removed all complex congigation near U=0 form ALGSUB. Grided 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 from 15APR90 this date. 5439. November 15, 1989 UVMAP Glen Increased buffer to allow 4096*2048 REAL maps, not just I*2 MAPS. Moved from 15APR90 this date. 5440. 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. 5441. 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 from 15APR90 this date. 5442. November 16, 1989 SCREAT Eric Corrected error message to give ``too large'' disk sizes in 8-bit rather than ``AIPS'' bytes. Moved from 15APR90 this date. 5443. 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 from 15APR90 this date. 5444. 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 from 15APR90 this date, nowhere else. 5445. 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 from 15APR90 this date. 5446. November 17, 1989 UVFLG Phil In entry #5428 I missed one occurence of the setting of the number of sources. Fixed. Moved from 15APR90 this date. 5447. 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 from 15APR90 this date. 5448. November 20, 1989 TVFLG Eric Corrected writing of history and flag files: (1) Corrected overhaul bugs in the handling of strings giving the ``REASON'' for the flag and history files. (2) 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. Moved history corrections only from 15APR90, nowhere else. 5449. 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 from 15APR90 this date. 5450. 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 from 15APR90 this date. 5451. November 21, 1989 APLNOT:BPASET.FOR Bill J. This subroutine didn't handle the case when both polarisations had been flagged. When the table was normalised, some divide by zeros were demanded. This affected POSSM, HORUS and SPLIT amongst others. Moved from 15APR90 this date. 5452. 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 from 15APR90 this date. 5453. November 22, 1989 BLREFM Phil The opcode in the call to TABBL was wrong when reformatting the tables. Moved from 15APR90 this date. 5454. 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 from 15APR90 this date. 5455. November 28, 1989 ACFIT Phil Was writing the number of subarrays into the # antennas slot in the SN table header. Moved from 15APR90 this date, 5456. November 28, 1989 FILLM Glen Fixed Corruption of LP band observation of the Sun by subroutine MCSOLR. Moved from 15APR90 this date. 5457. 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 from 15APR90 this date. 5458. November 28, 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. Moved fix from 15APR90, nothing else. 5459. November 29, 1989 UVFIX Phil An index was wrong in the writing of compressed data. Moved from 15APR90 this date. 5460. November 29, 1989 TABBL Phil Wrong index in array lead to incorrect number of IF's being written/read into/from BL tables. Moved from 15APR90 this date. 5461. 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 from 15APR90 this date. 5462. 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. 5463. 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. 5464. 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. 5465. 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 15OCT89 and from 15APR90 this date. 5466. December 8, 1989 GPOS Kerry/Bill C. Fixed zero divide problem. Moved from 15APR90 this date. 5467. December 8, 1989 SNPLT Kerry/Bill C. Fixed zero divide problem. Moved from 15APR90 this date. 5468. 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 from 15APR90 this date. 5469. 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. 5470. 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. 5471. 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 from 15APR90 since this is what users expected. 5472. 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 from 15APR90 this date. 5473. 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 15APR90 same date. 5474. 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 from 15JAN90 and to 15OCT89 this date. 5475. 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 HITIME from TST to NEW to accomodate the history and user-parameter handling changes. Moved to 15JAN90 this date. 5476. December 13, 1989 FILLR Bill J. The LOCAL INCLUDE 'FILLR.INC' had not been changed to reflect the addition of the FREQID column to the NX table. This caused the source (SU) table to have source ids of 1 for all sources. Change in 15APR90 this date. 5477. December 14, 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. 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 from 15APR90 and to 15OCT89 this date. 5478. December 14, 1989 UVPLT Eric Corrected bug positioning the time-date string on the plot. It was one character too high. This upset TVPL at least. Moved to 15JAN90 and 15OCT89 this date. 5479. 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 typo introduced during the conversion of an ``IF ... GO TO ...'' construct to an ``IF ... THEN ... ELSE ..."). Changed in 15APR90 same date, 15OCT89 December 22. 5480. 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 from 15APR90 this date. 5481. 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 from 15APR90 this date. 5526. 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 typo introduced during the conversion of an ``IF ... GO TO ...'' construct to an ``IF ... THEN ... ELSE ..."). Changed in 15APR90 same date, 15OCT89 December 22. 5482. 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 15APR90 same date. 5483. 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 15APR90 same date. 5484. December 28, 1989 APLBELL:ZLOCK Colin/Kerry The check for unavailability from a lockf() error return should be against EACCES, not EACESS. Changed in 15APR90 same date. 5485. December 28, 1989 APLNOT:TABBP Colin/Kerry GO TO 50 should be GO TO 60. Changed in 15APR90 same date. 5486. 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 15APR90 same date. 5487. December 28, 1989 APLUNIX:ZFRE2 Colin/Kerry Replaced the call to index(), which is Berkeley UNIX specific, with a trivial, portable loop. Changed in 15APR90 same date. 5488. 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 only from 15APR90 this date. 5489. January 2, 1990 GTBWRT Eric Corrected bug in addressing in GTBWRT used by TVFLG and UVIMG. It was addressing too many locations in the TVFLG mode and not determining the max and min correctly in all cases in which more than one buffer was used. Moved this correction from 15APR90. 5490. 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 15APR90 same date. 5491. January 2, 1990 APLBELL:ZABOR2 Colin/Kerry This is 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 15APR90 same date. 5492. 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 from 15APR90 this date. 5493. 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 from 15APR90 this date. 5494. January 5, 1990 TABBP Bill C. Cleaned up typing and some declarations. Moved from 15APR90 this date. 5495. 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 cleaned up typing and declarations. Moved from 15APR90 this date. 5496. 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 15APR90 this date. 5497. January 9, 1990 ANCAL Chris Corrected ANCAL to check whether values in the CL table are blanked before attempting to multiply them by the gain correction factor. Failure to do this was giving rise to large, spurious values in the CL table. Changed in 15APR90 this date. 5498. 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 from 15APR90 this date. 5499. 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. 5500. 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. 5501. January 11, 1990 UVSUB Phil Increased the buffer sizes to cope with 4096 square images. Moved from 15APR90 this date. 5502. January 15, 1990 BPASS Bill C. Fixed arguments in call to BPADJ. There was a bug here which caused the routine to fail and die with an error message. Moved from 15APR90 this date. 5503. January 15, 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. Changed FILAI2 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 from 15APR90 today. 5504. 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 from 15APR90 this date. 5505. 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 from 15APR90 this date. 5506. January 19, 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 from 15APR90 this date. 5507. 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 from 15APR90 this date. 5508. January 19, 1990 GRIDTB Eric Corrected the number of buffers actually being handled at any one time. Moved nowhere --- this fix was done in 15APR() some time ago. 5509. January 19, 1990 QPUT Glen Added wait after QPUT in FPS version Moved from 15APR90 this date. 5510. 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 from 15APR90 this date. 5511. January 21, 1990 FILLM.FOR Bill J./Bill C. Added Bill C.'s fixes (1) corrected format statement so that it wouldn't blow up when the FQ tolerance was large. (2) made FILLM do shadow flagging correctly. Moved from 15APR90 this date.