----------------------------------------------------------------------- ; Copyright (C) 1999-2000 ; 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 Master version of AIPS kept in Charlottesville. An entry is: line 1: Date Program name Your name line 2: What was done (may span multiple lines) line n-1: Where has this change been moved (e.g. 15JUL94, nowhere) line n: 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 15OCT99 *********************************************************************** *********************************************************************** Changes while 15OCT99 is TST *********************************************************************** *********************************************************************** 10202. March 26, 1999 Object Manager Chris Fixed the guard for the while loop that establishes the value of len in create_llelem (ZOINTD.C in APLGEN). This incorrectly used a bitwise and operator instead of a Boolean and operator and listed the conjuncts in the wrong order so that there would always be a read from beyond the expected length of dim. Moved to 15APR99 on this date. 10203. March 27, 1999 Freq offset in Source table Eric The source table contains a frequency offset as well as the FQ table. Technically, the CL table has both the source offset added to the time-variable Doppler offset so one really ought to find every time record and make everything go 10 times slower! But the Doppler change is small (0.5 km/s max) while the source offsets can be large (100 Mhz possible). So fix things to at least take the source part into account. Actually, UVGET already does this in the INIT call when there is 1 source out so DSEL.INC Add parameter SFREQS(MAXIF) SELINI Init SFERQS = 0.0d0 DGHEAD Set SFREQS CHNCOP Change call seq to add SFREQS array to call and to add the SFREQS values to the FOFF(if) written to the output FQ table. UVDATA In OUVOPN, save SFREQS from UVGET (INIT) call as SOURFREQ in UVDATA object UVUTIL Changed UVFRQS and UVREFQ to get SOURFREQ from the UV data set and add it to FOFF. Changed UVFCOP to get SOURFREQ and pass it to OCHNCO. Fixed them to use BCHAN and ECHAN - wrong channel frequencies were found when BCHAN and ECHAN were still active i.e. not 1 and N. QUVUTIL Added OUVOPN calls in OUVIMG and OUNFWT. Frequencies are known only after an open. OUVIMG found FREQS and then used FREQG for the list of frequencies. TABLEUTIL Changed call seq of OCHNCO to take source offsets to pass on to CHNCOP. UVMDIV Send 0's for SFREQs to CHNCOP - all single source by this routine. UVPDIV Send 0's for SFREQs to CHNCOP - all single source by this routine. CALIB Send SFREQS from DSEL and UVGET(INIT) to CHNCOP. PCAL Send SFREQS from DSEL and UVGET(INIT) to CHNCOP. LPCAL Send SFREQS from DSEL and UVGET(INIT) to CHNCOP. FRING Send SFREQS from DSEL and UVGET(INIT) to CHNCOP. NOT DONE - Ketan has it. KRING Send SFREQS from DSEL and UVGET(INIT) to CHNCOP. CPASS Send SFREQS from DSEL and UVGET(INIT) to CHNCOP. BPASS Send SFREQS from DSEL and UVGET(INIT) to CHNCOP. BLCAL Send SFREQS from DSEL and UVGET(INIT) to CHNCOP. MX Send SFREQS=0.0D0 to CHNCOP. Moved to 15APR99 April 8 10204. March 27, 1999 CLINV Eric New task to copy an SN or CL table inverting some or all of the calibration contained therein. Moved to 15APR99 April 8. 10205. March 27, 1999 FLGIT Eric Added option to flag on VPOL, fixing up displays and help file. Moved to 15APR99 April 8. 10206. March 27, 1999 FITAB Eric Added more parameters to the main header as HISTORY AIPS since that is what UVLOD uses to build its headers. Moved to 15APR99 April 8. 10207. March 27, 1999 IMAGR Eric Major gridding bug in 3D disk-based images. If there is a gap in the u coverage so that there are rows beyond the end of one swath in which there are no data, GR3DUV was supposed to write rows of 0 to the grid file, but it did not. Instead it wrote the next swath to early in the file and did not re-write the file at rows near 0. This should have done very bad things to the image - in the case under study where the data weights are small (<< 1), the image ended up in the KiloJy range. With more normal weights, horizontal stripes particularly toward the image corners were the result. Also change IMAGR.FOR to force VERSION to the output channel number for line images, so that the correct CC file would be created at the beginning. Restarts failed otherwise even for line (although the help says OUTVERS matters only for continuum). Changed TVMENU to remember the last TV size when it did a menu. Then, if that has changed it will redraw all of the menu. It could get confused otherwise. Changed QCLEAN to wait 5 minutes for the first TV menu to time out and to init the color tables only on the first pass. Moved to 15APR99 April 8. 10208. March 27, 1999 Anticipate Eric Changed: PUVD.INC Added MAXAFL = max fields in AIPS adverbs DGDS.INC Added LIMFLX = lowest flux to subtract in CC files Moved nowhere. 10209. March 30, 1999 IMAGR, SCMAP Eric Changed SCMAP Added tests to control whether the LUTs are initialized by self-cal TV function and added longer TIMLIM for first time around. QCLEAN Fixed TIMLIM setting a bit and CLIINF test on cell size which failed sometimes for large cells. Added test to see if anything is found in OVERLAP 2 since the last time all images were remade. If not, it switches to OVERLAP 1 mode rather than go into its infinite loop. Moved to 15APR99 April 8 10210. March 30, 1999 SPLIT Eric On error in UVCREA, there was a call to UVGET ('CLOS'...) although UVGET ('INIT'...) had not been called since the previous close. This caused a failure because DOCAL was true by that point, but had been false in the earlier quick UVGET open-close. CLBUFF had been used for scratch purposes to see if an SU table existed and so had whatever was in the SU tables's first record but 0's for the data area (CLBUFF(257-512)). So the SU table was closed with erroneous data and thereby effectively destroyed! This only happened if it had been used with LUN 27 previously and successfully - i.e. by a previous SPLIT. Anyway - put in a logical to keep track of whether we should do a UVGET close or not. Moved to 15APR99 April 8 10211. March 30, 1999 CALIB Eric Missed one of the calls to CHNCOP which needed SFREQS added to it. Moved to 15APR99 April 8. 10212. March 30, 1999 XGAUS Chris Fixed the dimensions of the SCRTCH array which was a factor too small. It is hard too see how this was missed since XGAUS blew up in a spectacular fashion. Moved to 15APR99 on this date. 10213. March 31, 1999 LaCook9 Ketan/Eric I've added information on how to process data that has calibration transfer from the correlator. Also, there is now some material on phase-referencing. I took the opportunity to place more emphasis on EDITR. Probably somewhere it should say that EDITR is now the recommended task for initial examination and for editing of VLBA data. Upgraded LaCook0, LaCookI and all of the PostScript copies and the html TofC. Moved to 15APR99 this date. 10214. April 2, 1999 KNTR Eric Added NY adveerb to allow user control of number of panes in Y direction. Changed POPSDAT.HLP, NEWPARMS.001, created NX.HLP, NY,HLP. Moved to 15APR99 this date. 10215. April 2, 1999 GETDEL in APLNOT Chris The loop that minimized the number of records to be checked for for time/source/subarray/frequency matches assumed that the sum of reference time and time interval increased monotonically with increasing row number and could therefore exclude potentially valid IM table records from the validity test. It also assumed that delay polynomials are valid before the reference time of the record. Rewrote the preselection using two simulated while loops to maintain the correct invariant conditions and removed a redundant loop. Moved to 15APR99 on this date. 10216. April 2, 1999 GETMC in APLNOT Chris Fixed a defect that lead to DTIME being calculated relative to a record that was either earlier or later than the one actually being used. Moved to 15APR99 on this date. 10217. April 4, 1999 UVPRT, UVHOL Eric Removed holography from UVPRT and fixed bug affecting data weight scaling. Created new task UVHOL which is like UVPRT. However, it squirrels away all records with matching W (a holography code for pointing/phasing) and then prints or writes out only the last N of them and then only if at least M are in the set. Moved to 15APR99 April 8 10218. April 5, 1999 ZTPOPN Eric ZTPOPN set the common parameter TPNAME(ivol) before finding out if the tape would actually open. It then left it set when things failed. In AIPS, after this failure, the set name would cause the program to try to dismount an uninitialized on exit and to fail on a mount as well. Moved to 15APR99 this date. 10219. April 5, 1999 TABIO Eric Changed to reset write status in the table header to a no-op on CLOS. Otherwise, a header just read by ZFIO will always appear to be in write form and hence confuse a TABIO call. It was this that ultimately caused to destruction of source tables in SPLIT. Moved nowhere. 10220. April 6, 1999 MNJ Eric The min-mnj did not work properly because UPDCONTROL spelled UpdMiniPRG as UpdMiniPrg. Changed spelling and added TPMON to the default list. Moved to Kitt Peak and the 15APR99 this date. 10221. April 7, 1999 LIBR Pat Added AXLINUX (in fact *LINUX) to the list of systems on which "lorder" will not be done for libraries. Moved to 15APR99 this date. 10222. April 8, 1999 BPASS Eric There was an addressing error that caused bad bandpasses to be computed for the first half of a good polarization for an antenna when all channels of the other polarization of that antenna were flagged. The bug has been there since 1996 (in CVX) and has therefore affected both 1998 releases. Moved to 15APR99 April 8. 10223. April 8, 1999 FITLD Chris Added code to pick up the observing date from RDATE keywords in FITS-IDI files. Normally this will be the reference date for the first ARRAY_GEOMETRY file. Previously FITLD only looked for DATE-OBS keywords. This was potentially dangerous since DATE-OBS is not required in any FITS-IDI tables and usually only occurs in UVDATA tables. Added comments to the help file describing circumstances in which the current scheme might fail. Moved to 15APR99 on this date. 10224. April 8, 1999 KRING Ketan Added new SOLTYPE = 'NOFT' to totally disable FFT stage. This allows KRING to perform only the least squares and behave just like CALIB except that KRING can also solve for rates, single-band and multi-band delays. When a solution's SNR falls below the threshold, the solutions are now blanked. The minimum gridding time is now used to better grid the data in time. This should fix the 'two-times found message'. Unfortunately, the user should now take care to actually enter the minimum integration time as accurately as possible [For the VLBA, the minimum integration time should be enterred as the appropriate multiple of 0.131072]. Once the time-gridding routines are overhauled, this nuisance will be abated. The test suite for KRING has now been fully automated. The SNR calculations for the FFT stage and the LS stage have been overhauled to be on the same scale. Two SNRs are now calculated in each stage. The first is labelled SNR and is used for threshold testing against the user-supplied SNR cutoffs. The second is labelled SNR/vis and should be used for comparing SNRs between the FFT stage and the LS stage. Further details supplied upon request. Some refinements for these calculations are forthcoming. Moved to 15APR99 on this date. 10225. April 8, 1999 OMFIT Ketan OMFIT now supports INFILEs with internal versioning - for easier prototyping of models within a single INFILE. The adverb DOFIT has been added to allow the user to supply a starting model from POPS. Any available model-type can be specified using DOFIT. The adverbs have been rearranged. Errors in the calculation of machine precision (resulting in a lax convergence criterion), the estimation of map noise, and construction of a catalogued AIPS file have been fixed. The OMFIT test suite now checks the self-calibration convergence. OMFIT now includes the covariance matrix of the parameters in the OUTFILE. Moved to 15APR99 on this date. 10226. April 9, 1999 SPCAL, USUBA Eric Fixed bad declaration in SPCAL (like previous fixes in PCAL et al) for HSOLTY - thanks to Jean-Francois Desmurs. Fixed the table copying routine in USUBA to keep going on error as best it can and to close the files no matter what. USUBA needs work but it is too messy to do now. Moved to 15APR99 this date. 10227. April 12, 1999 FITLD Chris Increased the limit in FQIDs to 256 to conform to the rest of AIPS and added a test against this limit in MAPFQI. Moved to 15APR99 on this date. 10228. April 12, 1999 Solaris tape routines Chris Used #ifdef tests to protect reference to MT_ISXY in ZMOUN2, ZTAP2, and ZTPWA2 in APLSOL. This symbol appears to have been removed in Solaris 7. Moved to 15APR99 on this date. 10229. April 12, 1999 GRDFIT Eric UVGRID has a -1 that was not include in GRDFIT so it was possible to disagree by 1 row on whether a gridding would fit without sorting. Added 2 rows for safety to GRDFIT. Moved to 15APR99 this date. 10230. April 12, 1999 CONFI Leonia CONFI has not worked for the array design if the number of elements exceeds 50. The last version of MMA(ALMA) considers 64 element. To meet these requirements the dimmension of the relevant array are increased to allow number of an array elements as large as 100. Moved nowhere 10231. April 15, 1999 CLINV.HLP Leonia The help file for CLINV lists OPTYPE as an adverb, but it has not appeared in the inputs. The HELP file erroniously has not listed OPTYPE in its first part. I fixed it Moved to 15APR99 by Eric 10232. April 18, 1999 LSAPROPO, ABOUT files Eric Ran the update tools and found that LSAPROPO is not updated when a remove of a help file is done. So I putbck a new LSAPROPO with the fixes suggested by SHOPH. Then updated the follow ZZxxxx files for ABOUT lists: ZZADVERB.HLP ZZANALYSIS.HLP ZZCALIBRAT.HLP ZZFITS.HLP ZZGENERAL.HLP ZZHARDCOPY.HLP ZZIMAGE-UT.HLP ZZIMAGING.HLP ZZPLOT.HLP ZZPOPS.HLP ZZTABLE.HLP ZZTAPE.HLP ZZTASK.HLP ZZUV.HLP ZZVERB.HLP ZZVLA.HLP Moved to 15APR99 this date. 10233. April 18, 1999 CookBook Eric Updated $DOCTXT/LaCook13.tex and $AIPSPUBL/COOK13.PS for the new about files. We got lucky and the table of contents did not change although the pagination is less than desirable a couple of places. Moved to 15APR99 this date. 10234. April 19, 1999 PBUTIL Eric Bill pointed out that subroutine PBFSCI also needs to call the good coordinate routine PSNANG. It is not used so far, hence Moved nowhere. 10235. April 19, 1999 XTRAN Eric Fixed to put the angle it finds in the output header. Fixed XTRFIT formats to be more precise. Moved nowhere. 10236. April 21, 1999 FITLD, IMLOD Pat Minor enhancement of an error message, in the situation where the file size is physically shorter than the headers indicate. Some non-compliant FITS writers fail to pad the last block. Moved to 15APR99 this date. 10237. April 25, 1999 UVPRM.FOR Leonia The routine SETANT and CALL SETANT have had different number of parameters: CALL SETANT (50, XANT, XBASE, NANT, NBAS, DESEL) in UVPRM SUBROUTINE SETANT (N, XANT, XBAS, NXANT, NXBAS, IANT, IBAS, DESEL) in $APLSUB As a result the memory confusion occurs. In particular FRQSEL=1 was zeroed. I matched parameters in CALL SETANT with the subroutine in $APLSUB. Moved nowhere 10238. April 27, 1999 $APGOOP/COHER Leonia COHER has given an idea about the coherence property of the baselines and antennas. The new option has been added to create the file with variable coherence time as a function of baselines and time. This file has the format which match the format of input file of the task BLING. So it can be used directely by BLING making the variable solution interval. This option can be usefull for milimeter interferometry, when the phase changes nonlinear and fast. Moved nowhere 10239. April 28, 1999 DATCAL Chris Fixed a typographic error in the application of dispersive delay corrections that corrupted the imaginary part of the visibilities. Moved to 15APR99 by Pat. 10240. May 3, 1999 Number of fields Eric Raised the number of fields to 512, without changing number of fields in the adverbs: PUVD.INC Changed MAXFLD to 512, MAXAFL is still 64. ZREHEX $APLGEN - new routine to decode extended hex AU5C Decode class name with ZREHEX rather than format. SETGDS Changed to not use DFIL.INC common - UNSETG must be called. Changed to use ZEHEX to make class name. UNSETG Simplified to ignore DFIL.INC and just clear status. IMCREA Changed to use ZEHEX to make class name. QCLEAN Removed obsolete test on 64 fields for an equivalence. Fixed TV display which had some odd things and fixed it to do > 64 fields by asking for a field number. MX Limit # fields to MAXAFL, fix adverb declarations, fix class name to EHEX. UVMAP Fix beam classes to xBEM. FLATN Fix class names to use EHEX. WFCLN Fix to limit adverbs and # fields to MAXAFL, name classes with ZEHEX. IMAGR Fix to handle MAXAFL adverbs to MAXFLD fields, to name classes with ZEHEX, and to force 3D when #fields > MAXAFL. Made it be interactive when #fields > 64 and to resume aips from a higher level when not. SCMAP Fixed to handle > 64 fields in TV display. Changed to set RQUICK false. INPUT Changed AV2INT to leave RQUICK at whatever is returned from the lower level routines. EDITR Changed to set RQUICK = .false. EDITA Changed to set RQUICK = .false. SNEDT Changed to set RQUICK = .false. PLAYR Changed to set RQUICK = .false. Moved nowhere. 10241. May 3, 1999 Stopping use of CC file Eric Added FLUX adverb to tasks to specify the minimum (abs) value of a CC component to be used in modeling. LIMFLX was added to DGDS a while back. Now change: SETGDS Changed to set NCOMP to max on NCOMP(i)=0 only if all NCOMP are 0. Otherwise, negative and excessive are set to the max. UVPDIV Upgrade comments UVMDIV Allow gridded method on NONEG UVMSUB Allow gridded method on NONEG GRDSUB Upgrade comments. GRDCRM Test for NONEG and LIMFLX and bail out when encountered. VISDFT Quit when flux level < LIMFLX. VSFDFT Quit when flux level < LIMFLX. VSCDFT Quit when flux level < LIMFLX. QUVUTIL Pick up MODFLUX wherever MODNONEG is picked up. PBUTIL Upgrade comments. Removed sections of unused variables. BLING Changed to handle NCOMP dimension correctly, to honor NONEG, and to add the FLUX adverb. Also help file. BPASS Added FLUX to help and code, adjust MAXFLD to MAXAFL. CALIB Added FLUX to help and code, adjust MAXFLD to MAXAFL. CLPLT Added FLUX to adverb and changed it to do NONEG and LIMFLX cutoffs. Made it more standard in handling NCOMP. CPASS Added FLUX to help and code, adjust MAXFLD to MAXAFL. FRING Added FLUX to help and code, adjust MAXFLD to MAXAFL. GPHAS Serious fixes: this has not worked in 1998 since the BCOMP and NCOMP adverbs were dimensioned 16. Added NMAPS and FLUX adverbs, changed to create multiple field objects for modeling. IBLED Added FLUX to adverb and changed it to do NONEG and LIMFLX cutoffs. Made it more standard in handling NCOMP. LPCAL Fix class naming to ehex, adjust MAXFLD to MAXAFL. KRING Added FLUX to help and code, adjust MAXFLD to MAXAFL. PCAL Added FLUX to help and code, adjust MAXFLD to MAXAFL. UVADC Added FLUX to help and code, adjust MAXFLD to MAXAFL, changed odd code that tried to allow 0 components in a field to the current natural method, changed to allow NONEG. UVCON Added FLUX to help and code, adjust MAXFLD to MAXAFL, changed odd code that tried to allow 0 components in a field to the current natural method, changed to allow NONEG. UVSUB Added FLUX to help and code, adjust MAXFLD to MAXAFL, changed to allow 0 components in a field (if some NCOMP not 0) for all OPCODEs. VPLOT Added FLUX to adverb and changed it to do NONEG and LIMFLX cutoffs. Made it more standard in handling NCOMP. Removed "temporal" statement that eliminated the Z term from non-3D models. Moved nowhere. 10242. May 3, 1999 MAXFLD Eric Changed formats for field number in QCLEAN, GR3DUV, GRDMEM, GRIDUV, IMGMEM, MAKMAP, UVTBGD, UVGRTB, UVGRID. Also format in QCLEAN for interpolated images. Moved nowhere. 10243. May 3, 1999 LWPLA Eric Created $APLGEN/ZTXREN.FOR to rename text files. Changed ZLWIO OPEN function to return the number of pages and max bounding box if the file pre-exists. Changed LWPLA and help file to allow multiple plot files in a single run, making a multi-page PostScript file that is only pseudo encapsulated. If there is one page then it is accepable for true encapsulation and most functions will over-print multiple pages (which gives one ideas...). Moved nowhere. 10244. May 3, 1999 CATN2L in APLSUB Chris Fixed CATN2L to transfer the image sequence number on little- endian machines. Solves problems with TVCPS and perhaps other programs finding erroneously different catalogue headers. Moved to 15APR99 patch area on May 14, 1999. 10245. May 7, 1999 EXTLIST Chris Updated offsets in AU8A which handles the EXTLIST verb and had fallen out of sync with the inputs for CLPLT and VPLOT. Moved nowhere. 10246. May 7, 1999 VPLOT.FOR, .HLP Leonia VPLOT has plotted auto if OPTYPE='auto' and cross if OPTYPE='cros', or ' '. If erroniously optype is selected as something different, VPLOT has not worked. I changed the logic of OPTYPE. Now cross is carried out for any optype but 'auto'. Auto is carried out if OPTYPE='auto'. Moved nowhere. 10247. May 11, 1999 POPSDAT.HLP, ANTUSE.HLP Leonia The new adverb ANTUSE(50) has been created. This adverb should be used when it is required to specify the subgroup of antennas having been selected by ANTENNAS. Moved nowhere. 10248. May 11, 1999 ANTAB.FOR, .HLP Leonia ANTAB has not worked with foreign stations when the time of TSYS measurement appeared between scans. This feature of ANTAB occurs because ANTAB has to know the source ID for the time of the measurement. ANTAB takes this time from NX table and of course failes to do this if the time does not belong to any scan. To circumvent the problem I increase the scan length of both side by the given time interval. The value of the time interval is given by the new adverb OFFSET. Moved nowhere. 10249. May 11, 1999 LDOPTS.SH Eric Added the SAVE parameter in $SYSHP, $SYSHP2, $SYSALPHA, $SYSDEC versions of LDOPTS.SH and changed to true this parameter in the $SYSSUN and $SYSIBM versions. Otherwise links fail if someone is using the load modeule at the time. Moved nowhere. 10250. May 11, 1999 AITOFF Coord Eric I was working with an image making task using AITOFF coordinates and found a problem or two: DIRCOS Allowed L and M to be too large in aitoff ( (L/4)**2 + (M/2)**2 must be < 0.5 ) NEWPOS Required L*L + M*M < 1 which is not needed, a proper test is done later. CTICS Had absurd initial values for parameters designed to bail out coordinate difficulties. Changed them to the reference point value. ITICS Like CTICS. Moved nowhere. 10251. May 12, 1999 GRIDTB (TVFLG) Eric Added messages about no data selected when UVGET returns error 4 with no data read. Moved nowhere. 10252. May 13, 1999 CALIB.FOR .HLP Leonia CALIB has used all antennas selected by ANTENNAS in the calculation of the mean gain modulus (CPARM(2)>0). It can be useful to limit the antennas used for the gain normalization to those with good a priori calibration, especially when using VLBI-style calibration based on system temperature and gains. This prevents the flux scale from being dragged up or down by poorly calibrated antennas including antennas subjected to bad weather. Now this additional selection can be done under contol of new adverb ANTUSE. Moved nowhere. 10253. May 13, 1999 TECOR and APGPS Chris Corrected sign of the dispersive delay correction in both programs. Moved to 15APR99 patch area on May 14, 1999. 10254. May 14, 1999 Solaris 2.x on PCs Chris Added support for Solaris on PCs. The operating system interfaces for Solaris on i86 are identical to those for Solaris on SPARC; the only change needed for AIPS is a new ZDCHI2 that sets BYTFLP to indicate that the host is little-endian. Added APLSOL86 to hold the new ZDCHI2. Added SYSSOL86 for compiler scripts etc. Changes INSTEP1 in INSUNIX to recognize i86 platform for Solaris. Modified the Makefile for XAS to recognize the SOL86 platform. Added a SOL86 section to FDEFAULTS.SH in SYSUNIX. The compilation scripts assume that g77 and gcc are used as the FORTRAN 77 and C compilers (tested with egcs 1.1.2). Moved nowhere. 10255. May 14, 1999 CCNTR Eric Changed help file limits to allow plotting as many CCs as desired. Use this to examine affect of CPARM(8) option newly added to IMAGR. Moved nowhere. 10256. May 17, 1999 IMAGR Eric 1. Added option to do Steer-Dewdney-Ito Clean as well as Clark Clean on images, with it switching at a break point set by the ratio of RESMAX to the level that BGC would load to the AP. If that ratio gets close to 1.0 SDI is done. Also added TV menu option to force SDI and to force BGC Clean at any stage. 2. Added option to remove isolated weak CC components at the end and even during the Clean. This keeps flux from being sucked out of the source by isolated noise bumps. Changed: IMAGR Help file to add options and to allow CPARM in TELL for some and only some of the options. IMAGR Fortran to pick up the options and store them under proper names in the Clean class object. UVUTIL Added subroutine UVRSCR to do a raw uv-data copy from one file to another. The filtering routine requires that we keep a copy of the original data and go back to it after the filter is applied. QCLEAN Added subroutines CLASDI to do SDI Clean, CLFILT to do filtering, CLGRCV to do convolution to a grid for CLFILT. Also changed CLINI, CLNUV, CLNTV to support these options. Added CPARM to TELL options in part. Moved nowhere. 10257. May 19, 1999 GRDFLT Eric Changed it to print messages both on the first call and on any later call in which the convoltion types have changed. It did only the first call before. Moved nowhere. 10258. May 24, 1999 IMAGR Eric Changed to delete the work file unless one of the name adverbs is set in advance (IN2NAME, IN2CLASS, IN2SEQ). Changed help file remarks about this and about OUTSEQ. Moved nowhere. 10259. May 24, 1999 CLIP Eric CLIP assumed that the WEIGHT/SCALE random parameters were the last 2 in the list. This is usually, but not always, true. Fixed it to fetch the right pointer from the header for compressed data. Moved nowhere. 10260. May 24, 1999 FILLR, FILLM Eric Changed help file for FILLR to make the point that the user wants FILLM fairly forcefully. Improved the documentation for FILLM's VLAMODE as well. ' ' => any. Moved nowhere. 10261. May 24, 1999 IMEAN, ACTNOISE Eric Added DOCAT to IMEAN to tell it to save the RMS and mean found by fitting the histogram as ACTNOISE and ACTMEAN. It also picks up these from the header if the adverbs are 0. Changed the AU7A verb ACTNOISE to store ACTMEAN as well inthe header. Moved nowhere. 10262. May 25, 1999 XAS Eric Changed the code controlling character size to be (hopefully) smarter. It wants to make bigger characters if there are a lot of pixels on the screen, but only if those pixels are smaller than our current ones. Moved nowhere. 10263. May 25, 1999 CLIP Eric CLIP assumed that the Stokes axis was axis number 2 when dealing with compressed data. This is not the case for VLBA Correlator output. DTCHK did not fix the address pointers to address VIS(3,*) and so worked only for compressed data. RESEQ did not multiply the pointers by 3 for expanded (previously compressed) data and so failed on compressed data. Moved to the patch area this date. 10264. May 25, 1999 MK3IN Eric A string was continued across two lines, e.g. MSGTXT = 'xxxxxxxxxxxxxxxxxxxxxxxxxx * aaaaaaaaaaaaaaaaaaaaa' which will produce a variable amount of blank space between the x's and the a's, probably longer than MSGTXT. The correct code is MSGTXT = 'xxxxxxxxxxxxxxxxxxxxxxxxxx ' * // 'aaaaaaaaaaaaaaaaaaaaa' Moved nowhere. 10265. May 26, 1999 SNPLT Eric Added DDLY = dispersive delay to the possible things plotted. Moved nowhere. 10266. May 26, 1999 IRING Eric Changed APARM(1,2) to be relative to the reference point rather than relative to the "center" of the subimage selected by the user. This is more general and specific and invariant to changes of window size. Moved nowhere. 10267. May 28, 1999 IMAGR Eric Found problems with the new automatic CCMERG: TABSRT Corrected to copy 0-length tables rather than go through the sort code (and fail). TABMRG Corrected to avoid writing out a data record when there were no good records in. CCMERG Correctd to skip all the sort and merge calls on empty files and copy instead if needed. QCLEAN Added test to see if there are any CCs at the moment before calling the merge and filter routines. Moved nowhere. 10268. May 28, 1999 PCAL Eric Cleaned up the typing a little to attempt to follow the logic. It didn't help much... Moved nowhere. 10269. May 30, 1999 SYSETUP Eric Dropped all the TV files (IC and ID) junk. It produces error messages otherwise. Moved nowhere. 10270. May 30, 1999 Square-root transfer Eric Added Square-root transfer functions to linear and log. Changed: ISCALE To add the function and clean up code Also changed the scaling of the log function, it was to washed out. TVALUE To respond to the function type (when TV only) LWPLA Add sqrt function IMLHS Add sqrt function The following help files were also upgraded to mention the sqrt function: FUNCTYPE BLANK BLSUM LWPLA PLAYR TVHLD COTVLOD TVALL YVCUBE TVLOD TVMOVIE TVROAM Moved nowhere. 10271. June 2, 1999 QCLEAN Eric Made a better test for Clean going wrong (comp abs(flux) going up, sum of flux going down) in OVERLAP 2 mode. Others were handled better. It now keeps track of the min CC flux by field for this case. Moved nowhere. 10272. June 2, 1999 UVHGM Eric Added PIXRANGE to control plotted range when desired. Moved nowhere. 10273. June 3, 1999 OTFUV Eric Changed it to use MAXCHA for max number of channels rather than 2048. Also raised max number of directory entries to 5000. The new correlator needs the former anyway. Moved nowhere. 10274. June 3, 1999 OOP Eric Changed MAXIO from 5 to 6, MAXCLS from 10 to 20, MAXVKW from 50 to 100 in OBJPARM.INC. Changed HBUFF size from 4096 characters to 20480. The class name arrays for 512 fileds of clean get pretty large. Recompile and relink all OOP. Moved nowhere. 10275. June 13, 1999 Memory files Eric Changed INIT and STORES to open POPS Memory file "0" without asking for exclusive use. This allows it to be opened read only if there is a privilege problem with read/write access. Moved nowhere. 10276. June 13, 1999 IMAGR Eric Changed QCLEAN to merge the Clean components only when needed (OVERLAP > 0, SDI possible, CC filtering). Changed the help file to say this and to encourage the merging. Moved nowhere. 10277. June 13, 1999 FITAB Eric Changed coordinate keywords in UV table to match the wcs proposal for tables. Added comments re QUANTIZE <= 0 to help. PRTTP: changed UV table reading to recognize table wcs keywords and to translate them back to normal random groups words so that the standard parser will recognize them. Moved nowhere. 10278. June 14, 1999 $QNOT/XCSHFT.FOR Leonia The logical error has been found in the smoothing part of the routine. The IF (ISMTH.GT.1) THEN appeared instead of IF (ISMTH.EQ.1) THEN I fixed it. $QPGNOT/CVEL and $QPGNOT/BPASS have been comlnked with the new XCSHFT. Moved nowhere. 10279. June 14, 1999 XTRAN Eric Added some explanation of Fortran formats for the kiddies. Moved nowhere. 10280. June 15, 1999 FRING Chris Cleaned up the logic in the main loop of FRNALL and fixed many serious errors including the following. - Failure to set NSUB when DOFIT was used leading to the latter part of the subroutine being skipped (leading to the reference antenna not be set and consequent array out-of-bounds access in least-squares routines). - Many attempts to access index 0 of arrays with a lowest index of 1 (including loop termination conditions). - Failure to correctly add/subtract results for intermediate antennas when rereferencing phase, delay, and rate. The aggregate effect of these was that the least-squares routines would always have been fed incorrect data whenever DOFIT was used or APARM(9) was 1 and delays, rates, and delays were never properly referred to a reference antenna when DOFIT was used. Even with these fixes in place, FRING still fails to find some solutions with APARM(9)=1 that it finds with APARM(9)=0 so some unidentified flaws remain. Moved nowhere. 10281. June 16, 1999 $APGNOT/PCCOR.FOR, HLP Leonia Selection of IFs is added for the case when two bands are observed simultaneously. For example X/S observations typically used in astrometry/geodesy. In this case PCCOR should run twice selecting IFs for the each band. CLCAL should be applied twice to combine the two SN table in one CL table. Multi band delay (MBDELY) of the output CL table will include the sum of the two SN tables MBDELYs. So it does not correspond to the either first IFs or second IFs. If MBDELY for one of the two IFs is wanted, then MBDELY in the relevant SN table should be zeroed. The zeroing of MBDELY can be done under control of the new parameter DELCOR. Moved nowhere. 10282. June 16, 1999 FITLD Chris Found that FITLD went to some effort to make sure that VLBA weights were not proportional to the integration time. Removed this so that weights are now written as w' = w * s * t where w is the tape weight (0 <= w <= 1), s is the VIS_SCALE value and t is the integration time. Incremented version number to 10.5. Moved nowhere. 10283. June 17, 1999 $UPDUNIX/UPDSSHD.PL Pat New file: intended to act as a pseudo-server in conjunction with an appropriate configuration of ssh keys. The idea is that a key is generated on the master machine (done), and a copy of this key is distributed to each MNJ site for placement in the appropriate ~/.ssh/ directory. On the server, a special line is inserted in the ~aipsmgr/.ssh/authorized_keys file that will permit a single command to execute (UPDSSHD.PL) if this key is used. That command in turn retrieves the passed argument to the ssh command on the client via an environment variable, and if this matches an AIPS file, the contents are returned on standard output. The next step is to incorporate this into UPDCOPYSUB, keyed on a switch to be set in UPDCONFIG on a per-site basis. Moved nowhere. 10284. June 17, 1999 TYPUVD Eric Changed TYPUVD in DUVH.INC to be -1 for interferometer data from E-W interferometers using UU-L-NCP. Changed UVPGET to find this occurence and return the -1. Changed DATGET, GETNAN, SDGET, FITTP, SELSD, UVFND, DBCON, UVDIF, UVPLT, GRIDR, SDIMG, MX, and UVMAP to accept this value as equivalent to 0. Changed OOP ONBJMNGR to trap the occurence and return 'UP' for this NCP case. Changed UVUTIL, UVPHAS, IMAGR, and INDXR to accept this value as equivalent to 'UV'. It will be used in a bit. Moved nowhere. 10285. June 17, 1999 -NCP imaging Eric Changed IMAGR To force DO3D false for non -SIN projections. UVUTIL To detect NCP TYPEUVD and set appropriate projection parameters on the image header in U2IDES. Moved nowhere. 10286. June 17, 1999 Large angle imaging Eric The -SIN projection can be used for shifts only up to some reasonable angle [(1 - L*L * M*M) must be positive]. It should be possible to image while keeping all shifts as simple shifts and changing them into -SIN shifts only when required. The previous code changed all shifts into -SIN shifts and then went back to RA, Dec when required (to do 3D for example). This means that user input SHIFT, RASHIFT, DECSHIFT will now be in "simple" [Delta RA * cos (Dec_ref), Delta Dec] or "-GLS" shifts. Changed: BSGRD.HLP To refer to simple shifts QIMAGE In I2IDES changed to simple shifts from -SIN (which may not have been what was being gridded anyway!) SDIMG.HLP To refer to simple shifts with proper info. SDIMG Changed setting of ref pixel and shift header parameters. The ref pixel setting was wrong before in that the shift had already been corrected by dividing by COS(Dec) before it was used to shift the ref pixel. GRIDR.HLP To refer to simple shifts with proper info. GRIDR Changed setting of ref pixel and shift header parameters. The ref pixel setting was wrong before in that the shift had already been corrected by dividing by COS(Dec) before it was used to shift the ref pixel. SDGRD.HLP To refer to simple shifts with proper info. UVUTIL In U2IDES, change handling of SHIFT to simple -GLS type. (I believe SHIFT is only used by SD data.) Fixed ref pixel on ISBEAM (which never seems to be used). UVMOD Minor, perhaps unneeded, change to make the Z shift term 0 for -NCP. W will be zero anyway so... FXSHFT Add comments at start to => -SIN (non 3D) only and now obsolete. PRJMAT Add comments that it is for -SIN projection only. SHFTXY Changed to make simple shifts, not -SIN XYSHFT Changed to use simple shifts, not -SIN SHINCP New: -NCP version of SHIFTXY/SHIFTP pair SHISIN New: -SIN version of SHIFTXY/SHIFTP pair SHFCRP New: figures out the new CRP shift (in arc sec) GRDAT Changed to call SHINCP or SHISIN rather than SHIFTP. SHIFTP Changed comment to point at new routines and => this one is obsolete. IMCREA Changed to use SHFCRP to set ref pixel IMHCOR Changed to use SHFCRP to set ref pixel UVMAP.HLP Upgrade SHIFT comments. UVMAP Changed SHIFT to be simple shift, adding calls to SHISIN/SHINCP and SHFCRP as needed. MX Changed SHIFT to be simple shift, adding calls to SHISIN/SHINCP and SHFCRP as needed. Corrected an equivalence CLBUFF that was 512,512 that should have been 512,64. SHIFT.HLP Changed the nice explain section to do simple shifts. RASHIFT.HLP Fixed definition, mention BOXFILE. DECSHIFT.HLP Fixed definition, mention BOXFILE. HORUS.HLP Fixed SHIFT description. WFCLN.HLP Fixed RA/DEC SHIFT descriptions. IMAGR.HLP Fixed RA/DEC SHIFT descriptions. MX.HLP Fixed RA/DEC SHIFT descriptions. Moved nowhere. 10287. June 17, 1999 BSGRD Eric Changed to manage shifts better and to allow different XTYPEs for the two gridding operations (to Az-El and then to ra-Dec). Moved nowhere. 10288. June 18, 1999 FRING Chris Fixed two sign errors for RFACT in FRNALL. FRING now gives consistent answers with APARM(9)=1 and APARM(9)=0. Results from the coarse search are slightly different using the two modes on some baselines where the target antenna has a lower number than the reference. I suspect that there is some integer truncation in FRNSR2 or QSEARC: this is probably harmless. Moved nowhere. 10289. June 21, 1999 QCLEAN Eric Fixed up a couple of things about IMAGR: 1. CCFILT has to recompute the CC flux in each field. It was not doing so. Added message about it too. 2. The OVERLAP restore step dies if the angle between the 2 fields was too large to compute mutual coordinates. Changed to suppress this "error" (the fields cannot overlap) and continue. Moved nowhere. 10290. June 24, 1999 FRING Chris Fixed REFI/REFJ confusion in referenceing the second antenna of a pair in FRNALL. Moved nowhere. 10291. June 24, 1999 GRBOXS Eric The new box code did not handle the situation of multiple images on a TV plane properly. Changed to draw windows in range of image not of full TV screen. Moved nowhere. 10292. June 24, 1999 IMAGR Eric QUVUTIL would only accept type UV not type UP - the NCP uv data. Moved nowhere. 10293. June 21, 1999 KRING Ketan Fixed a simple typo in the calling sequence for a subroutine. Moved nowhere. 10294. June 21, 1999 CL2HF Ketan Dave Gordon at GSFC has supplied a new version of CL2HF that is reported by R. Gaume at USNO to produce acceptable results on one of the RDV experiments. Another revision to CL2HF is coming soon (although it is only cosmetically different from the version checked in today). Moved nowhere. 10295. June 21, 1999 STFUN Ketan The hard-coded image size limits in STFUN have been removed. Now, the largest image size needed is determined and the necessary amount of memory is allocated. Moved nowhere. 10296. June 24, 1999 KRING Ketan Fixed a simple typo affecting the weights calculation for the doubly indirect baselines when CPARM(3) = 0 or 3. Moved nowhere. 10297. June 24, 1999 IMCREA, IMHCOR Eric These also forced the -SIN string. Changed to test on TYPUVD. So IMAGR did not get right string in header even though shifts done okay. Moved nowhere. 10298. June 28, 1999 GPINI in APLNOT Chris Some table columns were changed from REAL to DOUBLE sometime between the realease of 15OCT99 and 15APR99. Changed them back to REAL so that the GP table access routines work. Moved nowhere. 10299. June 28, 1999 LDGPS Chris LDGPS now reads the KEYIN-style file format agreed with VLA/VLBA operations. Support for CONAN and the Los Alamos summary format has been withdrawn. Moved nowhere. 10300. June 29, 1999 COMB Pat With OPCODE = 'MEAN' this task was not correctly setting the output to the blanked value when both inputs were blanked. Fixed. Moved nowhere. 10301. June 29, 1999 INPUTS GO Eric Remove GPOS from AU1A, AU2 lists of system adverbs. It can be in units other than pixels. Fixed limits in IMFIT and JMFIT help files. Moved nowhere. 10302. June 30, 1999 VISDFT etc Eric These routines dimensioned DONE(field number) as DONE(MAXANT) which used to work (90 > 64) but does not now (90 < 512). Affected VISDFT, VSCDFT, VSFDFT. Moved nowhere. 10303. June 30, 1999 LDGPS Chris Corrected scaling factor for differential biases in nanosec. Moved nowhere. 10304. July 1, 1999 APGPS Chris Forced longitudes read from the GP table header into the range 0 to 360 degrees to avoid problems caused by wrapping around the Earth. Moved nowhere. 10305. July 4, 1999 LWPLA Eric Fixed the handling of ASPMM. It was setting it for the first plot and then using that value if possible for later plots. Moved nowhere. 10306. July 5, 1999 AIPSEXEC Pat Modified this script slightly to permit the command-line DEBUG argument to AIPS itself to specify both the debugger and whether or not to run AIPS under said debugger. Now one can say, e.g. "aips debug=dbx:aips" if you want to run dbx and run AIPS itself under it; or "aips debug=/path/to/oddball/debugger" to use that debugger on tasks (but not on AIPS itself). The old behaviour has not been altered, i.e. "aips debug" will still prompt you for the debugger name and whether or not to run AIPS under it. Moved nowhere. 10307. July 6, 1999 REGRD.FOR Pat Due to a buffer being 2x too small, this program failed in somewhat random ways on different architectures (Solaris for one; Linux seemed unaffected). Resized the appropriate buffer to the correct size. Moved nowhere yet; candidate for a patch? 10308. July 7, 1999 $APGNOT/UVCON.FOR Leonia The range of hour angles has been given for Greenwich. I changed it to the array center (site) for people's convenience. Time attached to a visibility has been rounded to the range 0-1day. The help file has been updated. Moved nowhere. 10309. July 8, 1999 $APGNOT/UVCON.FOR Leonia Having finished the yesterday modification I found inconsistency of different coordinate system in one input file if some antennas are determined in the coordinate system related with Greenwich and other with local meridian. To solve this problem I recalculatecoordinates of all antennas to the RH coordinate system with axis X directed to Greenwich. Then the given hour angle at the local meridian is recalculated to Greenwich. Moved nowhere. 10310. July 9 SD Beam switched stuff Eric Changed: QIMAGE had a bad call sequence to GRDFIX affecting coordinate-scaled convolution functions. BSGRD to allow 2 convolution functions, one for the Az-El image and another for the Az-El to Ra-Dec regrid. Also changed the "rotation" correction to be a shift either done in the initial Az-El gridding or done as a separate interpolation step. BSGRD Help file to explain these things BSTST to remove a bug in a call sequence and to compute and display RMS in non-signal as well as overall. BSTST Help file to allow specification of the non-signal channels. BSMOD New task to create beam-switched data from a model with various errors and defects. Also help file. Moved nowhere. 10311. July 9, 1999 PLOTR Eric New task to plot data points read from a text file. Up to five different sets of data may be plotted with specified symbol (or line) type. Moved nowhere. 10312. July 10, 1999 UVWAIT Eric The frequency scaling array was carried partially in REAL*4 for loading into the array processor and partly in REAL*8 since the numbers need a fair amount of accuracy. Unfortunately, when one or the other is used microscopic differences in round off can occur cauing it to seek wrong rows, getting into all sort of trouble. Why this has not arisen before I do not know. Moved nowhere. 10313. July 10, 1999 VTESS Eric VTESS was blowing up with a zero divide in DOTGRD. Attempting to find the optimum step length it assumed that SCLF was not zero, but at this point near convergence both SCLF and GAIN were zero. EPSSTR = GAIN / (SCLF * RL0) and only RL0 was tested. Changed it to avoid the zero divide and to limit EPSSTR to <= 1 when it would occur. UTESS had the same math. Moved nowhere. 10314. July 11, 1999 POSSM Eric POSSM with SOLINT = x was looping for all possible baselines for each time interval from start to stop by x even though large portions of the data were not at suitable subarrays etc Changed UVGET To make FSTVS3 (a mostly unused DSEL.INC parameter) one less on init DATGET To use FSTVS3 as a counter of the real vis record number. FINDUV New: to find the next time interval start time and the corresponding vis record range. POSSM To use FINDUV to find the time intervals and to require Time-ordered data when SOLINT not 0. POSSM Help - added comment about time order being required. BPASS Used FSTVS3 to recompute NREAD3 and FSTRD3 which do not need recomputing to reinit the I/O. CPASS Used FSTVS3 to recompute NREAD3 and FSTRD3 which do not need recomputing to reinit the I/O. Moved nowhere. 10315. July 12, 1999 $APGNOT/PCCOR.FOR Leonia PCCOR has considered the row of a PC table beeing bad if at least one of the tones is zeroed. Actually PCCOR needs only two tones. The logic has been wrong. For example if one of the three tones, noticed at the table header is zero PCCOR did not worked although it should. I changed the logic allowing PCCOR working if number of measured tones .GE. 2. Moved nowhere. 10316. July 16, 1999 FRING Chris Factored the frequency gridding calculation from CHKAP, FRNSRC, and FRNALL into a single subroutine and fixed a gross error in this calculation that caused FRING miscalculate the AP memory requirement as an absurdly large number and abort when a single channel was selected. Also fixed several tests in CHKAP, FRNSRC, and FRNALL that used the total number of frequencies instead of the number that were actually being search. The behaviour of FRING appears ill-defined if the delay search is turned off and more than one channel is selected (or more than one IF with APARM(5) > 0). This requires some further thought. Moved nowhere. 10317. July 19, 1999 EDITR Eric To assist with the expert mode, changed QEDIUTIL to call TVDXPR rather than an internal routine Also corrected loading of UV data to make sure that current flags are applied when data are loaded; the one-antenna at a time mode was overlooked in this matter. Pick up DOEXPERT from TBEDIT or UVMAST object. Corrected pickup of DOTWO (=COMPARE) in TY editing - it did not use the correct object and so was always the default. TVDEVICE Added routine TVDXPR as interface to TVXPRT. TVXPRT Displays TV menu, highlights single-character codes in the menu, does the expert prompt, read, help, etc previously done in QEDIUTIL. POPSDAT Added adverb EXPERT NEWPARMS.001 Added scalar expert. EDITR Added EXPERT adverb to enter expert mode as task begins. EDITR.HLP Added EXPERT. EDITA Added EXPERT adverb to enter expert mode as task begins. EDITA.HLP Added EXPERT. SNEDT Added EXPERT adverb to enter expert mode as task begins. SNEDT.HLP Added EXPERT. EXPERT.HLP New help file. Moved nowhere. 10318. July 19, 1999 AIPS.FOR Fixed the copyleft statement that is displayed. It said 1998 and should have said 1995-1999. Moved nowhere. 10319. July 20, 1999 CALIB/$APLNOT/GETCTL.FOR Leonia/Eric CALIB has not worked properly when the model is in Lpol, but the UV data have all four polarizations. CALIB has subtructed the model only from Rpol of the data in this case. The problem has been found in the subroutine GETCTL. Some other errors are found at the GETCTL as well. The modified version of GETCTL has been put in the system. The list of the tasks included CALIB has been comlnked with the new GETCTL. Moved nowhere. 10320. July 20, 1999 VGET, VPUT, DEFAULT Eric Created 2 new pseudoverbs to save and get task adverbs in versioned files and a pseudoverb to re-initialize the adverbs of one task only. Changed: POPSDAT Added pseudoverbs VGET, VPUT, DEFAULT, reserved verb numbers for these, verb VGINDEX, adverb VNUMBER. AIPSB Changed TS to TG and added loop to kill all possible versions of TG files. AIPSC Changed TS to TG in loop deleting and in creation of main TG file. Changed CU2 to add VPUT. HELPS Added codes for the 3 Pseduoverb to verb conversions. AU2 Added code to rename old TS files to TG and to name the files by the new schemes including VPUT. AU2A Added code to rename old TS files TG, to name the files by the new scheme, to do VGINDEX and VGET (using the code for TGINDEX and TGET except for the file opening) and to do the new verb DEFAULT. This requires long lists of adverbs with non-0 and non-blank adverb values. We will have problems with maintenance. AU3A Changed to new TG names - the previous version had the wrong names so ALLDEST and TIMDEST did not kill TS files. DISKU Changed to loop over TGET/VGET files. Created new help files for VGET, VPUT, VGINDEX, DEFAULT, and VNUMBER. Moved nowhere. 10321. July 20, 1999 DOCRT Eric Changed things to respond to the width of the terminal window. Changed ZWINC Added number columns to call sequence ZWINC2 Added number columns to call sequence and set them in $APLGEN, $APLSOL, $APLHP, $APLDEC, $APLLINUX (last 4 identical) versions. AU1A Changed call to ZWINC, added code to set DOCRT from Ncols, pick up NACROS from LPOPEN or ZWINC setting. AU7 Changed call to ZWINC, set print width with window. OFMLST Changed call to ZWINC, made DOCRT an in/out variable containing the width in abs value. PRTMSG Changed call to ZWINC, set print width. LPOPEN Changed to return DOCRT = width to use (in abs value) PRTLIN Changed call to ZWINC, return is ignored. PRTALN Changed call to ZWINC, return is ignored. ZDCHIN Changed call to ZWINC, return is ignored. PRINTER Changed to use output of LPOPEN as width and to store that as DOCRT back in the printer object. Changed to use LPOPEN to set the width: CORER LISTR MFPRT PRTAN PRTUV SHOUV UVFND UVHOL UVPRT COHER DTSUM TBDIF OTFIN PRTAB PRTCC SAD UVDIF ISPEC PRTIM PRTTP HFPRT BLSUM TXPL Changed DOCRT write up in HELP files for DOCRT BLSUM BSPRT COHER CORER DTSUM HFPRT ISPEC LISTR MFPRT OFMLIST OTFIN PRTAB PRTAC PRTAN PRTCC PRTHI PRTIM PRTMSG PRTSD PRTTP PRTUV RFI SAD SHOUV TBDIF TXPL UVDIF UVFND UVHOL UVPRT Removed holography from UVPRT help file. Moved nowhere. 10322. July 20, 1999 UVCOP Eric UVCOP determines a mask for the input channels so as to select frequencies and IFs. Sadly, it used the uv-data increments of the output data set to address things relevant to the input data set. Selecting BCHAN > 1 and BIF > 1 would have not gotten the correct results. Moved to patch area July 21. 10323. July 21, 1999 BPASS, CPASS Eric My July 11 change did not work since BPASS and CPASS were effectively using FSTVS3 a common variable as a local variable to count the way through local IO buffers. Changed to local variables, using only NREAD3 and FSTRD3 from common when UV IO is initialized. Moved nowhere. 10324. July 21, 1999 IMAGR, VTESS help files Eric Fixed the adverbs in the help area to be in the same order as in the inputs area for IMAGR.HLP. For VTESS.HLP, added strong remarks about the importance of NOISE to a successful deconvolution. Moved nowhere. 10325. July 22, 1999 GO, TPUT, .. Eric Added message suppresion around the automatic renaming of the old TS files to the new TG files. If users are using aips new and aips tst they can end up with both files and the rename will fail without serious consequence. The message is anoying. Moved nowhere. 10326. July 22, 1999 QCLEAN Eric The routine that found the max/min ignored the field requested and did all fields. Since only the requested field was current, the results were misleading for all other fields and would cause the next field to be the previous field. Moved to patch area this date. 10327. July 22, 1999 $APGNOT/PCCOR.FOR Leonia Direct cal transfer produces some extra lines if the number of tones given at the table header exceeds the actual number of measured tones. Until now these extra lines are filled out by zero. The relevant modification (see #10315) was done to serve in the case of such a table. In the near future the extra lines will start be filling by blanks(nan). I have modified PCCOR to serve both zeroed and blanked extra lines in the PC table. Moved nowhere. 10328. July 26, 1999 FITTP, FITAB Eric A major screw up has been found with FITAB/FITTP etc. If a UV file was written with FITAB, then read back in and modified in some way (UVCOP et al) and then written with FITTP, the output will not be read correctly thereafter. There will be a work around when I figure out how to recover any ffile in this state. For now and quickly - changed FITTP and FITAB to stop writing out any "AIPS" history cards found in the HI file (all others are still written). It is these "AIPS" history cards which are parsed by UVLOD and FITLD and which contain from FITAN information about the UV file structure, information that may no longer be current. FITTAB avoided the problem by writing the current information AFTER the HI file, but FITTP writes the current information before the HI. When I figure out how I will add an option to UVLOD for users to try when UVLOD/FITLD gets an error 4 (unexpected EOF). Moved nowhere will have to be a major patch announcement. 10329. July 27, 1999 Fixes for FITTP/FITAB Eric Added the ERROR adverb to UVLOD, IMLOD, and PRTTP and comments about it in their help files and in the FITLD help file. If ERROR = 2 the AIPS history cards are not parsed. Moved to the patch area this date. 10330. July 27, 1999 DDT Ralph Gaume / Eric The class used by UVMAP is now IBEM again in accord with all other tasks. Changed DDTLOAD.001 to match. Moved nowhere. 10331. July 27, 1999 FLGIT Eric Added an option to flag data close to the V axis (abs(U) < APARM(9) wavelengths. U s computed on a channel-by-channel basis making this more selctive than UVNOU. moved nowhere. 10332. July 27, 1999 IMAGR Eric Changed the OVERLAP=2 loop that decides what field to do next to require the maximum of the field to be > MINFLX, the user's cutoff flux level. The loop includes the average of the field in the test and that may alter the choice and allow it to go below MINFLX which is not desirable. It got into virtually an infinite loop over this. Moved nowhere. 10333. July 27, 1999 ORDER Eric New adverb to specify the order of a polynomial fit - initial value = 1. Changed: POPSDAT Add ORDER, set to 1. NEWPARMS.001 Add ORDER, set to 1. AU2A Add ORDER to list of adverbs with +1 initial value for DEFAULT. IMLIN help file: change NCOUNT to ORDER. XBASL help file: change NCOUNT to ORDER. SDLSF help file: changed NCOUNT to ORDER SDLSF renamed order variable and added it to HI files. UVLSF Add ORDER adverb to allow for 0 order fits as well as first order. Also help. UVLIN Add ORDER adverb to allow for 0 order fits as well as first order. Also help. FLGIT Add ORDER adverb to allow for 0 order fits as well as first order. Also help. ORDER New help file Moved nowhere. 10334. July 28, 1999 QEDIUTIL Eric I failed to fix all of the "FLAG FAST" strings into "FLAG QUICKLY" strings. So the mode was not actually available. Moved nowhere. 10335. July 28, 1999 FILEBOX Eric The handling of the name was not fixed properly for the new extended hex field names. It did the ZREHEX part but tested for validity on the old form of the names. Moved nowhere. 10336. July 28, 1999 FLGIT Eric I forgot to init the LUN and VER for CHNDAT which did not matter for my 1 IF data set but which quickly got some people. Sigh.. This CHNDAT call was added for frequencies needed to do the UVNOU option. Moved nowhere. 10337. July 28, 1999 TV minor bug Eric AU5D, AU5E, AU6, and AU6A all used TV parameters before calling TVOPEN. This works since they are in Common and more or less right xcept at the beginning. TV parameters (NGRAY, NGRAPH, TVLIMG, etc etc) must never be referenced until TVOPEN is called at least now that the TV info is kept in XAS itself. Also found IMLHS called YINIT before TVOPEN - that never worked. Moved nowhere. 10338. July 28, 1999 $APGNOT/PCLOD.FOR Leonia PCLOD has given a wrong result on mashines under LINUX, if COMLNK had been done with optimization. The condition of the difference in time has been given as IF (TTIME .NE. OTIME). If the times are considered equal the previous condition can work differentely in different system. This occured in the machine under LINUX. I change this condition: DDTT = TTIME - OTIME EPS = 1.0E-10 IF ((DDTT .GT. EPS .OR. DDTT .LT. -EPS)... Moved nowhere. 10339. July 29, 1999 ZWINC Eric There was not enough error handling when ZWINC2 returns 0 for the terminal height. Some default - I use -24 - needs to be put in CRTMAX when it starts as 0. Moved nowhere. 10340. August 3, 1999 XTRAN Eric Changed the task to be used only for its "plate-solution" function. OGEOM, LGEOM, OHGEO, et al. can be used for simpole re-gridding. Changed the adverbs needed to run it. Added mode to simply copy the image and alter the output header parameters. Fixed handling of BLC, TRC - it did not do the right thing at all when they were specified and a re-gridding with rotation occurred. It fit for the wrong "center" pixel in that case and then rotated about the center. Changed it to leave the geometry ID alone if there is one and use -TAN otherwise. The task is still a linearized one and so unreliable for large fields. Moved nowhere. 10341. August 3, 1999 FLATN Eric Changed INTERPLATE INTWIN routine to suppress error messages from the position routines and to return IERR=-1 if all coordinates fail and IERR=0 if only some do. Changed IMGHGE to suppress error messages and only use the output of the position routines if the error is zero. This way will allow partial images to work. Chaned FLATN to handle IERR=-1 properly and to make another progress message. Moved nowhere. 10342. August 4, 1999 SETFC Brian Butler / Eric New task: it creates C cards for IMAGR's BOXFILE giving the centers of (a) a fly's eye covering the center region, (b) extra fields covering strong sources taken from the NVSS survey tables, and (c) an extra field on the Sun. There is a help file plus 4 text files in $AIPSTARS call 1000MJY.NVSS, 300MJY.NVSS, 100MJY.NVSS, and 50MJY.NVSS. These files are used by default choosing the one that matches the flux limit given by the user or the user may specify his.her own table. Moved nowhere. 10343. August 5, 1999 ALLTAB Eric Changed some lists of tables copied: SPLIT Added BL table to omission list since it is applied. SPLAT Added BL table to omission list since it is applied. AVSPC Added BP table to omission list since it is meaningless in the output and allow FG tables to be copied if they are not applied. ACFIT If the output is multi-source (which it can be) then allow the copying of the SU table. BLAVG Added a bunch more files to omit. GPHAS Corrected error handling so that it would die on multiple source selection - previously it would emasculate itself and then try to continue. Added a bunch of table types to the omission list. Moved nowhere. 10344. August 5, 1999 XTRAN Eric The interpolation to the new grid neglected to sum the interpolation weights and divide by them. This led to a herringbone pattern in the images output! Added a new array to sum weights and divide by them and added code to blank areas not computed. Also changed it to honor blanked input pixels. Minor updates to help file. Changed XTRFIT subroutine to correct a FORMAT. A 1PDm.n was followed by Fm.n with no 0P and so printed a number 10 times too large for the reference pixel. (The 0P came right after!) Moved nowhere. 10345. August 6, 1999 UVAVG Eric Corrected the call seq to ZEXPND - the FTAB pointer was used where the disk number was needed. This never worked. Moved nowhere. 10346. August 6, 1999 TBOUT Eric Improved string handling with JTRIM to avoid unprintable characters in the output. Moved nowhere. 10347. August 6, 1999 CLIP Eric Added PRTLEV to replace APARM(8) for printing and then added APARM(8) as a parameter to control deletion of whole spectra if some number>= APARM(8) of channels are clipped. We need a CLIPM to do clipping with flag table, multi-source type inputs. Moved nowhere. 10348. August 9, 1998 FITAB Eric Found and fixed 2 exgregious errors having to do with image writing. In the first, quantized floating output did not hvae the data values filled to the buffer, causing garbage to be written. I know I tested this, so I must have added a "simple" logic change that messed everything up afterwards. In the second, the time tests were applied to tables having nothing to do with time and a < instead of <= caused no table data to be written Moved to patch area this date. 10349. August 9, 1998 CC File Eric Changed: PRTCC Type 4 (sphere) was not recognized in PRTCC. SAD SAD was willing to write out deconvolved Gaussians (even if the deconvolution failed) in the CC file iff the output CC version number was 1. Otherwise it wrote convolved Gaussians which VISDFT etc cannot handle. Changed it to write deconvolved or zero width if deconvolution is being done. SAD Help file - clarify the writing to CC file. CCEDT Reordered CPARMS, allow circular flux aperture. Handle 7-column CC tables but require the values in columns 4-7 be constant through the file. Allocate memory to hold the CCs dynamically. Changed to use image increment based tolerances and defaults. CCEDT Help file: Changed CPARM(3,4) to CPARM(1,2) and vice versa to put important parms first. Allow circular flux aperture w CPARM(1) < 0. Moved nowhere. 10350. August 10, 1999 $YPGNOT/CONFI Leonia I added two options, both are important when optimization is carried out with topography constraint: 1. The input file can be now shifted (when it created by previous fitting to the topography) or not shifted under control of APARM(8). 2. Optimization with topography constraint can be done now together with donut or two circle constraints. The help file is updated. Moved nowhere. 10351. August 12, 1999 $QPGNOT/UVCON, $APGNOT/UVSIM Leonia UVSIM has had limit of 50 antennas for history file. When I was writing UVCON I copied the relevant codes from UVSIM. So the limit of 50 antennas was transfered to UVCON. The problem has not been seen in UNIX mashine but was discovered recentely on LINUX. Therefore both UVSIM and UVCON have not worked if number of antennas exceeds 50 (at least on LINUX). I fixed it. Moved nowhere. 10352. August 12, 1999 EXTAB Chris A new task to export AIPS table data to other programs (e.g. spreadsheets and databases) via text files. Also serves as a paraform for translating AIPS tables to text formats with data selection. Moved nowhere. 10353. August 12, 1999 TAPQL Chris Removed TAPQL, which translated AIPS tables to PostQUEL for export to the Postgres database. PostQUEL is no longer used (recent versions of Postgres use SQL) and EXTAB solves the problem of exporting table data to a database in a more general fashion. Moved nowhere. 10354. August 12, 1999 KEYIN subroutine Chris Modified the KEYIN package to treat ASCII carriage-return characters (character 13) as a white space character. This should reduce problems with files that have been prepared under Windows or that have had carriage-return new-line pairs inserted by awkward mail programs. Moved nowhere. 10355. August 12, 1999 CLIPM Eric New task - a CLIP geared to multi-source input files with flag table output. It is based on UVMLN and CLIP. Moved nowhere. 10356. August 16, 1999 CONFI.HLP Leonia A small misprint has been fixed in CONFI.HLP. Moved nowhere. 10357. August 16, 1999 DDT scripts Chris Updated DDTLOAD to set FLUX to zero before running CALIB rather than inheriting the value from the map settings (which causes CALIB to use zero input components and leads to a DDT failure). Moved nowhere. 10358. August 16, 1999 TABLE object interface Chris Added support for bit-field columns to TABDGT and TABDPT where they had silently been ignored before. Bit fields are input and output as arrays of logical values. Moved nowhere. 10359. August 16, 1999 EXTAB Chris Now writes bit-field values to the output file. Moved nowhere. 10360. August 17, 1999 $APGNOT/JMFIT.FOR Leonia JMFIT gives negative errors sometimes because the fitting of the gaussian to histogram gives negative value for its width sometimes. Since the absolute value is right I simple added ABS to the calculation of the width. Besides of that I removed two redandant prints and changed one of the subroutine's output parameter, which was given explicitely equaled zero. Moved nowhere. 10361. August 20, 1999 IMAGR Eric Upgrade to deal with requests: IMAGR Use average BCOMP for fields < 65 as default for fields > 64. Call FIELDF for BCOMP and store it away then. In loop over channels, restore initial BCOMP for each channel and do open CLEAN on each channel (to get BCOMP limits correct for each channels). Store interactive state for CLEAN class use. IMAGR Help file: Add comments about BCOMP setting and interactive use. Upgrade TV EXPLAIN to include menu items for filtering, BGC/SDI selection, prompting for field number and new FORCE A FIELD. More explanation under the OVERLAP 2 paragraph. FIELDF Read B cards for BCOMP. ISUM New routine to sum an integer array RSUM New routine to sum a real array QCLEAN TV option to force a field. Use ISUM to save lots of little loops. Filter CCs on restarts and on Clean failures as well as on request. Correct and improve filtering summaries. Improve estimate of peak residual in SDI clean (it still is pretty lousy). Moved nowhere. 10362. August 24, 1999 VPLOT Leonia VPLOT has declared the absence of SU table in the edit mode (APARM(1) > 0) if the image file as a model is presented. The logical unit number (LUN) for SU table used when the model presented is equal 27. The logical unit number (LUN) for FG (write) was chosen erroniously as 27 also. I fixed the problem changing the logical unit number (LUN) for FG (write) to 30. Moved nowhere. 10363. August 25, 1999 SNPLT Eric Substantially rewritten. It can now plot multiple polarizations and/or IFs and is able to plot them separately or plot both polarizations and/or all IFs on single plots (per antenna). Moved nowhere. 10364. August 25, 1999 UVLOD, FITLD Eric The table reader in UVLOD (and its clone in FITLD) overwrote the error condition in cleaning up and exiting. So a tape error was lost. Moved nowhere. 10365. August 30, 1999 SNPLT Eric Minor problem blocked access to LL-only data. Moved nowhere. 10366. August 30, 1999 KRING Ketan The least-squares stage of KRING has now been completely replaced. This was necessary for reworking the SNR calculation to use the same methods as found in the FFT. For test data, the two stages now produce SNRs that agree to within 1%. Many inputs were removed from KRING including the ability to create/fill an output file for single source data. Since KRING requires an NX table, this was disallowed anyway. Averaging the frequencies within an IF is not allowed anymore either; KRING's more efficient use of memory renders this sort of option unnecessary. The default SNR cutoff threshold has been lowered to 3. The help file has been updated somewhat but likely still needs some user input. The output format has been cleaned up as well to be more readable and the print levels consolidated. Moved nowhere. 10367. August 31, 1999 EDIT Class Eric Fixed all HLPED* (UV, TY, CL, SN) files to say EDIT QUICKLY rather than EDIT FAST. Changed QEDIUTIL - it would not flag the first point in quick mode until you had pointed at and flagged a second point. Moved nowhere. 10368. September 1, 1999 OTF Eric The OTF routines did not work on byte-swapped machines. They once did, but I overhauled the logic and never checked. Changed SDDPTS New call sequence - header not translated from external bit pattern. OTFIN Changed to call SDDPTS before translating full header from external to internal form; fixed setting of swapped. OTFUV Changed to call SDDPTS before translating full header from external to internal form; fixed setting of swapped. OTFBS Changed to call SDDPTS before translating full header from external to internal form; fixed setting of swapped. Moved nowhere. 10369. September 1, 1999 CPASS.HLP Leonia Pre-average interval (CPARM(4)) has not been described clearly. In particular the default has not been named. I copied the full description of CPARM(4) from the explain part to the help part. The default is added to the list of input parameters. Moved nowhere. 10370. September 1, 1999 $APGNOT/QUACK.FOR..QUACK.HLP Leonia QUACK has had two options of flagging data starting the beginning of a scan. 'END ' => flag all but APARM(1) min from end of scan. 'BEG ' => flag first APARM(2) min of each scan. I have added the third option, which allows to flag data near the end of the scan. The old options are renamed: 'ENDG' end of the scan is good (not flagged) 'BEGB' beginning of the scan is bad (flagged) The new(third) option is named: 'ENDB' end of the scan is bad (flagged). The flagging was started 5sec earlier to include the edges of the scan. It was done wrong for the option 'end', because the flagging interval should be expanded to the right to include the right edge of the scan. I fixed it. Additionally the time of the expanding is not fixed now as 5sec. It can vary under control of APARM(3). It is important for the case of very short scans. Moved nowhere. 10371. September 1, 1999 SNCOR Leonia I've added the new option 'NORM' which normalizes amplitudes to one. The help file is updated. Moved nowhere. 10372. September 2, 1999 QUACK Leonia Revising my recent modification of QUACK (10368) I changed back the names of the two old options. So they are named now again as 'END ' and 'BEG '. Moved nowhere. 10373. September 2, 1999 EDIT class Eric The WTSCAL (DOWEIGHT adverb) parameter did not set a default, so DOWEIGHT=0 flagged all the data. Moved nowhere. 10374. September 2, 1999 ZDCHIN Eric Changed NETSP error message to appear only for prgnam = 'AIPS' and added explanation. Moved nowhere. 10375. September 2, 1999 IMAGR Eric Changed UVWAIT and QWTSUM to handle areas which have no data. Like other grid routines, having a large gap between 0 and the lowest abs (u) caused problems. Moved nowhere. 10376. September 3, 1999 CVEL Leonia Added new selection parameter QUAL to select shifted sources by qualifier. The help file is updated. It was written in the help file in source selection: '*' = all. Indeed ' ' means all also. So I changed this statement to '*' or ' ' = all. Moved nowhere. 10377. September 3, 1999 Various MNJ scripts Pat The Secure Shell MNJ has been tested and found to work. There are three necessary components: (1) the "mnj" binary key, which should be put in the ~/.ssh/ directory of the relevant account and set via chmod to mode 0600; contact me for a copy of this; (2) a revised UPDCOPYSUB (already putback; it's backwards compatible and should not break anything; and (3) UpdSSH=YES set in the relevant section of UPDCONFIG. Currently only brador is using this setting, but within a week I expect all NRAO sites to be using the secure shell. Eventually the old rsh way will be phased out completely. This scheme is implemented via a special entry in the authorized_keys file of the aipsmgr account on the server MNJ (kochab): command="/AIPS/15OCT99/SYSTEM/UNIX/UPDATE/UPDSSHD.PL", no-port-forwarding,no-X11-forwarding,no-agent-forwarding, no-pty 1024 37 10939307688004... This restricts the command that can be executed to the UPDSSHD.PL script in $UPDUNIX. This in turn checks the first two arguments to the ssh command; the first is the filename, and the second is an optional string. If the latter is MODE, it returns the file protection mode of the file. If there is no second argument, the contents of the file are returned. If the filename requested is not in one of the allowed hierarchies (/aips1/master/TST, /aips1/master/NEW, /AIPS, or /home/kochab_1/AIPS) then the request for the file or mode info is rejected. NOTE: any $SYSLOCAL copy of UPDCOPYSUB will override the new version. You should not need such a local copy. NOTE 2: anyone wishing to implement this scheme on a "client" MNJ (i.e. where your own machine and not kochab is the "master") needs to generate a ssh key for the relevant account as I did for aipsmgr here, and put the right key in the authorized_keys file on the subservient MNJ machines. Moved to brador this date. 10378. September 3, 1999 Misc Eric Improved the message in ZDCHIN explaining NETSP failure. Changed $SYSSUL and $SYSSOL versions of CCOPTS.SH to compile for 64 bit file indices in NRAOCV. We are all at 2.6 now. Moved nowhere. 10379. September 6, 1999 PRTAC Eric Fixed handling of DOCRT - the LPOPEN call was skipped when DOCRT > 0 so it did not get set to the screen width. Also it did not zero enough of the arrays so some **** appeared in the output. Moved nowhere. 10380. September 9, 1999 $APGNOT/SETJY Leonia I have added the forth name to having existed three names of each of the six calibrators. The new names are taken from ~cwalker/sched/catalogs/sources.vlba. Moved nowhere. 10381. September 7, 1999 UVAVG, SPLAT Eric Changed: LISRAN To clean up typing, to change output meanings, to report unrecognized random parameters, to include WEIGHT and SCALE in the list since they do not have to be last. UVAVG Changed handling of random parameter group to use the output of LISRAN. SPLAT Revised to use same random parameter technique. Changed an RCOPY call to copy only what is needed. The program runs some 20 times faster now. Moved nowhere. 10382. September 10, 1999 SNPLT Eric The attempt to choose the best phase range (-180 to 180 vs 0 to 360) caused it to ignore PIXRANGE under some circumstances. Moved nowhere. 10383. September 10, 1999 EDIT class Eric The data reading loop was too sensitive for linux in dealing with times. Changed it to look sensible for data between half way before to half way after the current time. Moved nowhere. 10384. September 11, 1999 EDITing Eric Changed default SOLINT and DETIME to 0.01 sec and max (2, 5*SOLINT) minutes. Changed EDITR.HLP, EDITA.HLP, SNEDT.HLP and QEDIUTIL.FOR. Also fixed another time test loading 2nd UV data for comparison. Moved nowhere. 10385. September 13, 1999 CLIPM Eric Fixed STOKES handling for cases in which there are fewer than 4 correlators. Moved nowhere. 10386. September 20, 1999 BPASS Chris Removed the declaration of BPVER that had crept into the subroutine BPSIN. BPVER is also declared in DSEL.INC and the double declartion caused compilation to fail under Linux. It should have produced some warnings elsewhere as well. Moved nowhere. 10387. September 23, 1999 POSSM Chris Moved the statement incrementing the number of a plot on a page in MULTPL so that it is only executed if the plot is not blank. This avoids problems setting the coordinates and opening the plot file if the first plot on a page is blank and should also save some paper. Moved from 31DEC99 on this date. 10388. October 18, 1999 VBGLU Chris Fixed an error in retrieving the catalogue number for the output file when updating the history records that caused the header of one of the input files to be corrupted when there were more than 2 input files. Moved from 31DEC99 on this date. 10394. October 18, 1999 FIXWT Chris Revised FIXWT to use the real parts of the visibilities to estimate noise levels rather than the amplitudes. This should give better estimates of the noise at low SNRs. It also seems to reduce the occurrence of isolated points with discrepent weights. Moved from 31DEC99 on this date. 10395. October 21, 1999 FRPLT Chris Removed an assignment to the unused and uninitialized common- block variable UVSCAL that included a division by its uninitialized value. Moved from 31DEC99 on this date. 10398. October 26, 1999 UVMAP Eric Believe it or not, there are still people using UVMAP/APCLN. They discovered that the xshift parameter is in backwards for the output header, messing up self-cals... Moved from 31DEC99 this date. 10399. October 28, 1999 PLOTR Eric Found a bug when plotting symbols. The plot paraform has an internal scaling which was not applied to the symbols coordinates before calling the general symbol plotter. The failure affected only certain ranges in the data. Moved from 31DEC99 this date. 10400. October 29, 1999 ACCOR Leonia ACCOR has not worked if NX table was absent. The routine UVTIME has gotten zero for LREC because UVPGET (which estimates LREC) was called after UVTIME. I moved CALL UVPGET before CALL UVTIME. Moved from 31DEC99 this date. 10401. October 29, 1999 TICINC Leonia/Eric Small plots sometimes end up with only one tick mark which does not say much about the scaling. Fixed a .LT. to .LE. which should require more ticks in general. Moved to 31DEC99 this date. 10402. November 1, 1999 TVBLNK Eric TVBLINK did not work right on XAS with 8-bit pseudo color display. It tried to do split screen which it can do, but it cannot have 2 different LUTs. Fixed code to recognize the less capable situation and to do 1 channel at a time. Moved to 15OCT99 this date. 10403. November 1, 1999 SPLAT Leonia All tables have been copied with subarray=0. I fixed it adding relevant cards in subroutine COPTAB. Moved from 31DEC99 this date.