----------------------------------------------------------------------- ; Copyright (C) 1998-1999 ; 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 15APR99 *********************************************************************** *********************************************************************** Changes while 15APR99 is TST *********************************************************************** *********************************************************************** 10026. September 20, 1998 $QPGOOP/OMFIT Ketan Indexing error caused SN table time-tags to be off by one solution interval when self-calibrating data. Moved to NEW tomorrow. 10027. September 21, 1998 $QPGOOP/OMFIT and HLP Ketan Indexing error caused covariances between self-calibration parameters to be ignored - slowing convergence when self-calibrating data. Also loosened convergence criteria until self-calibration error-bars are properly sorted out. The weights reported in the SN table may not yet be correct. Moved to NEW this date. 10028. September 21, 1998 SPLIT Eric Changed INDXIN To return -1 when there is no error except that no matching index record is found UVGET To handle the -1 return from INDXIN SPLIT To handle -1 returns from UVGET better (keep going if looping over subarray, else go on to next source). Also to not send subarray 0 into UVGET at all. Changed default for NPOINTS to all channels. Fixed count of channels averaged to apply to FQ table values. SPLIT Fixed help for new NPOINTS default. DATCAL Changed logic counting samples so that those either previously flagged or flagged by gain would be known as dead. Moved to 15OCT98 this date. 10029. September 21, 1998 TVFLG Eric There were several arrays of 5000 meant to handle #subarrays * #baselines. Changed these to 15000 for Frazer. Moved to AOC today in SUL and to 15OCT98 this date. 10030. September 22, 1998 $QPGOOP/OMFIT Ketan Strange typo introduced yesterday is now fixed. Also, found an array out-of-bounds condition that was only triggered for large self-calibration problems. Moved to NEW this date. 10031. September 22, 1998 FITLD Eric Added comments in DOCONCAT section about setting OUTSEQ=0 for concatenating. Moved to 15OCT98 this date. 10032. September 22, 1998 UVCOP ERic Selecting by source in the CL copying can cause problems with CL tables that are not set up well wrt the sources being copied. Changed it to call CLSEL with 0 for the number of selected source (which means no selection applied). Moved to 15OCT98 this date. 10033. September 23, 1998 $QPGNOT/CVEL.FOR Leonia Several bugs have been found in the routine ACSHFT. The hanning smoothing-convolution has been added to the routine XCSHFT, which shift cross spectra. Originnally the smoothing of the shifted spectrum was provided only for VLBA(?) and only for auto spectra. Now the hanning smoothing-convolution is carried out under control of APARM(9). The hanning smoothing- convolution help to exclude(minimize) the faken phase jumps, which appear in the case of the strong narrow feature in the spectrum. The HELP file is updated. Moved nowhere 10034. September 23, 1998 $APLNOT/ACSHFT.FOR Leonia I fixed several bugs in this routine which have brought errors in the shifted spectra. Moved nowhere 10035. September 23, 1998 $QNOT/XCSHFT.FOR Leonia I added the smoothing option to the process of the cross spectrum shift. So the new input parameter ISMTH is added. Moved nowhere 10036. September 23, 1998 $QPGNOT/SDVEL.FOR Leonia It is comlnked with the new version of ACSHFT.FOR Moved nowhere 10037. September 23, 1998 $QPGNOT/BPASS.FOR Leonia It is comlnked with the new version of XCSHFT.FOR. The new input parameter ISMTH put zero. Moved nowhere 10038. September 28, 1998 $APGNOT/UVFLG.FOR Leonia If APARM(6)>0 but APARM(5)<1 the task flagged the data following input parameters but not intervals near small fringe rate as it should be. But it is reasonable in this case not flagging at all but provide the user information about the locations of the small fringe rate in accordance of APARM(6)>0. I changed the logic to do it. Flagging of intervals near zero fringe rate can not be carried out for single source data with FLAGVER=-1. The relevant warning is added. Moved nowhere 10038. September 30, 1998 $QPGNOT/KRING.FOR Ketan Fixed a typo that prevented multi-band delay solutions from being transferred to all IFs in the solution table. This only affected the 'SMBD' and 'MBD' OPTYPES. Moved to 15OCT98 this date. 10039. September 30, 1998 $HLPFIL/SAD.HLP Ketan At user request, added missing description of DPARM(6) to inputs section of help file. Moved to 15OCT98 this date. 10040. September 30, 1998 $QPGOOP/OMFIT.FOR Ketan Squashed some bugs, added some new ones. Fixed SN table writing routine. I was writing out each solution twice with different times. Also, fixed the writing of GAINS model info to the SN table so that GAINS model solutions can be used with other AIPS tasks via the SN table. At user request, I added a new model, the BLOB model. This model allows for a Gaussian component with variable position. Currently, the only linear variability with time is implemented. This could easily be generalized to non-linear variations with time, frequency, channel, IF, etc. [any requests?] Moved to 15OCT98 this date. 10041. October 8, 1998 $APLAXLIN/various Pat Placed the APLLINUX versions of ZGRLHI, ZGRLTC, and ZPRMPT in APLAXLIN to test them with the new readline v2.2 library. Readline now works under this architecture too. Moved to 15OCT98 this date. 10042. October 8, 1998 $AIPSPUBL/AIPSMEM101.TEX Ketan $AIPSFIG/AM101[A,B].EPS Fringe-fitting memo is now checked in. 10043. October 8, 1998 $QPGNOT/KRING.FOR Ketan $HLPFIL/KRING.HLP New SNR calculation has been moved into place in KRING. The HLP file now reflects the new calculation, proper references to the Schwab and Cotton paper, AIPS MEMO 101, and fringe-fitting now in explain section of HLP file. Moved to 15OCT98 this date. 10044. October 15, 1998 $APGNOT/PCCOR.FOR Leonia PCCOR has not worked in the case of LSB/USB, if the Pcal tones are not located symmetrically near the edges of IFs. Previously the tones were located on equal distance from the left and right edges of IFs. After recentely change in SCHED it is not right. PCCOR calculated phase difference of the given IF relatively the first IF using distance between the IFs beginning but it should be done for distance between the first tones. I fixed it. The minor change has been done with more correct using of the reference pixel. Moved nowhere. 10045. October 15, 1998 FITLD Chris Added code to add the appropriate number of days to the TIMERANG entries in FITS interferometric data interchange FLAG tables. Moved nowhere yet. Needs to be added to the 15OCT98 patch area when it becomes available. 10046. October 16, 1998 $QPGNOT/CVEL.FOR Leonia CVEL has not worked in 15OCT98 and 15APR99 if SUBARRAY=0 is used. SUBARRAY has not been recalculated to 1 as it is written in HELP. I added this recalculation. Moved nowhere. 10047. October 19, 1998 $QPGNOT/CALIB.FOR Leonia The new option has been added to CALIB. If CPARM(2)>0, CALIB provided normalization of mean gain to one over all antennas, polarizations, and IFs. Rick Perley requested to print the value of the global normalized factor. It has been done. Additionnaly the gain, averaged over all antennas for each IF and polarization is printed also. Moved nowhere. 10048. October 21, 1998 $APGNOT/SPLAT.FOR Leonia SPLAT has been died if ASSEMBL option aplied to single source file. I changed it. Now SPLAT transforms the single source file to another single source file giving a relevant warning. The HELP file is changed a little bit to cut short the title into one line. Moved nowhere. 10049. October 21, 1998 GETSIP in APLNOT Chris New subroutine to calculate the geometry associated with a line of sight through the ionosphere; taken from GPSDL and generalized to allow the altitude of the ionosphere as a parameter. Moved nowhere. 10050. October 21, 1998 TECOR Chris New task to perform ionospheric Faraday rotation and dispersive delay corrections using maps of the free electron content of the ionosphere in standard IONEX format. Moved nowhere. 10051. October 22, 1998 Calibration Eric CSLGET had a nasty feature added to it in 1992 when gain normalization was added. If a solution was flagged in the SN or CL table, then the previous solution was left in the arrays used for the interpolation although the time was updated. Changed it to put the new solution into the interpolation arrays even if it is flagged --- which was intended but one wanted to avoid multiplying a magic blank by something which would make it not a magic blank anymore. This error/feature caused data that should have been flagged to be calibrated with good solutions from (potentially much) earlier times. Moved nowhere - this needs testing before going to the patch area. 10052. October 26, 1998 CSLGET in APLNOT Chris Fixed an error in initializing the IFR table in which the ionospheric Faraday rotation was read into the first entry in the table for each antenna while the second entry was never initialized. This would lead to the Faraday rotation being interpolated incorrectly in CGASET. CSLGET, CGASET and DATPOL all assume that the IFR is valid at all times; the assumption is invalid and more work is required to deal with this. Moved nowhere. 10053. October 26, 1998 DATCAL in APLNOT Chris The logical return value DROP was set from ALLBAD which was never assigned a value. It is now set from ALLDED which indicates that all channels were either previously flagged or have been flagged due to bad gain solutions. Moved nowhere. 10054. October 27, 1998 Calibration Eric CSLGET/CGASET had a "feature" that was designed to reduce excess computation but which, beginning August 12, caused problems when a CL/SN table had large gaps for some of the antennas but not all. The calibration for the good antennas was not updated often enough because the update interval was made large by the gaps in the "bad" antennas. Changed the meaning of CALTIM(1,2) to be the max lower time and the min upper time over all antennas rather than the average IN CSLGET. Changed CGASET to do the computation more often even with this new interpretation. Needs to go to NEW! 10055. October 28, 1998 $APGNOT/CLCOR.FOR Leonia When correcting the source position RA correction was given in CLCORPRM(5) in arcsec as the RA correction itself without recalculation by COS(DECL) to the picture plane. The people who took the the desired correction in RA from an image shoud devide it by COS(DECL) to convert into actual RA correction. It has been found not comfortable. So I implemented this division in the codes. The relevant changes are made to the HELP file. So now CLCORPRM(5) means correction in the RA direction in the picture plane. Moved nowhere. 10056. October 30, 1998 $APGNOT/SPLAT.FOR Leonia For some reason (I do not remember it) SPLAT has not copied SU table if there is only one source in the table. I removed this limitation. Moved nowhere. 10057. November 3, 1998 $YPGNOT/CONFI.FOR, HLP Leonia The minimum spacing requirement has not worked in the case of fixed elements at the outer circle. I fixed it and add relevant explanation in HELP file. Moved nowhere. 10058. November 8, 1998 $QPGNOT/FRING Phil Occasionally FRING will fail to converge in a fit. This is not a problem unless the user is performing a rate-only fit (usually spectral line observers). In this case FRING aborts with a floating point overflow because various values of the Hessian have not been filled correctly. Modified the code slightly so that the matrices are accessed in different ways for different types of fitting. Moved nowhere. 10059. November 9, 1998 LISTR Chris Added code to deal with missing antenna table on the assumption that the data comes from the VLA to GANSET (cloned from COLSET). Revised the message issued under these circumstances so that it doesn't wrap over more than one line. Moved nowhere. 10060. November 9, 1998 XMOM Eric Instead of destroying the first axis, XMOM now moves it to the last axis with 1 point and appropriately changed reference pixel and increment. This allows coordinate pairs and center frequencies to be retained. Moved nowhere. 10061. November 9, 1998 GREYS et al Eric Changed GINITG to return an error for grey scale plots that are too big for the TV. Previously it complained but did not quit. Moved nowhere. 10062. November 9, 1998 UVSUB Chris Inserted a trap for multisource input files in UVSUB. Moved nowhere. 10063. November 9, 1998 MERGECAL procedure Chris Added a procedure (available trhough run file MERGECAL) to merge redundant calibration information that has been created by concatenating calibration transfer data in FITLD. Changed the FITLD help file to refer to this procedure. Moved to 15APR98 patch area on March 9, 1999. 10064. November 10, 1998 FLGIT Eric Changed: 1. APARM(4 and 5) now control the multiplication factor of the rms for flagging. They default to 6 and 5, the previous fixed values. The previous use for these APARMs seemed to generate no additional flagging. 2. Added display of the fraction of channels flagged on each baseline and improved the computation of when a full spectrum is flagged. The display of the fraction of fully flagged is now done only when there are any. 3. Changed meaning of FLAGVER = 0 from none to highest which is the standard default in cal application tasks. Moved nowhere. 10065. November 10, 1998 Tucson references Eric Tucson no longer stores its compilers in non0standard places. Changed $SYSUNIX/FDEFAULT,sh, $SYSSUL/CCOPTS.SH, and $SYSSUL/LDOPTS.SH. Moved nowhere. 10066. November 10, 1998 DFTPL Michael Bietenholz / Eric The weight was applied wrongly in computing the rms. Also the sum of the weights was used to adjust the rms to the expected error rather than the count-1 of samples in the average. Moved nowhere. 10067. November 10, 1998 SPLIT Eric Changed the number of channels written when frequency smoothing to write as many as possible even if the last one does not get as many averaged into it. If DOUVC=-1, as it should with channel-dependent weighting, then the weights will take care of this partial averaging. Changed AVGSMO to watch for the upper limit. Previously it averaged one too many output channels, reading and writing data potentially outside of arrays. Moved nowhere. 10068. November 11, 1998 BLCAL Eric Added CMETHOD option to fortran and help file. Moved nowhere. 10069. November 11, 1998 ABOUT files Eric Updated the following about files before putting all in LaCook13.tex. ZZADVERB.HLP ZZANALYSIS.HLP ZZCALIBRAT.HLP ZZEDITING.HLP ZZINTERACT.HLP ZZOOP.HLP ZZPLOT.HLP ZZPOPS.HLP ZZPROCEDUR.HLP ZZSPECTRAL.HLP ZZTABLE.HLP ZZTASK.HLP ZZTV-APPL.HLP ZZUV.HLP ZZVERB.HLP ZZVLA.HLP ZZVLBI.HLP Moved nowhere. 10070. November 11, 1998 Help files Eric Deleted some of the obsolete (and now empty) list files from many years ago. Fixed up precursor remarks and categories for LPCAL, HISTART, HIEND, FITTP, IMLOD, UVLOD, DOPOL. Updated the affected ZZ files again. Moved nowhere. 10071. November 13, 1998 ZWINC2 Eric It turns out that we have working ZWINC2.* routines only for DEC, SOL, and LINUX architectures. For IBM and HP, the $APLGEN version was called and that returned 0 lines. Changed it to return 24, which is a workable number. Moved nowhere. 10072. November 18, 1998 $APGOOP/DTSUM.FOR Leonia DTSUM has not payd attention on subarray. So in the case of the overlapping scans with different subarrays DTSUM has included numbers tothe both scans without selection of the different subarrays. I fixed it. Additionaly I found DTSIM has not worked properly if the source IDs do not look like 1, 2,3.. in NX table. I fixed it also. Moved nowhere. 10073. November 19, 1998 $APGNOT/LISTR.FOR Leonia In the "SCAN" option, LISTR has labeled the number of visibilities for each scan in the wrong form if the scans have different subarrays and were overlapped. I changed it to print the start and the end visibility number instead of their difference. Moved nowhere. 10074. November 23, 1998 UVINIT Eric The computation of record and word offsets to reach vis = VISOFF did not take the > 2 Gigabyte issue into account. Changed these to DOUBLE PRECISION since I do not see a way to do them in blocks. Moved nowhere. 10075. November 23, 1998 QCLEAN Eric Restored the testing on clean limit (NITER) to allow the DO3D with OVERLAP 0 or 1 to exceed NITER letting each field get a full major cycle. It was not closing things right when it did some but not all of the fields in the last round and this is better than finding that bug. Moved nowhere 10076. November 23, 1998 $APGNOT/PRTUV.FOR(HLP) Leonia PRTUV carries out additional pass of the data for scaling to find the relevant format of print. If the data are large, the time for this additional pass can be rather big. (sometimes minutes). I implemented the new parameters DPARM which allows excluding the automatic scaling forcing the ranges of U, V, W, AMP, and WEIGHT. The default is the automatic scaling. On the request of C. Walker I added printing subbarray and fqid. Moved nowhere 10077. November 23, 1998 Source selection Eric It is not a good idea to select on sources when copying the CL table. Changed FLGIT and FLGIT to stop doing this. I guess we should leave it in CLSEL in case there is ever a real need for it. Moved nowhere. 10078. November 23, 1998 $INC/DTPM.INC Ketan Increased MAXFQD, the number of FQIDs allowed in a simulation from 5 to 8. Also, increased the implicit number of allowed antennas from 10 to the value of MAXANT in $INC/PUVD.INC. Need to COMRPL DTBAND, DTCALS, DTINDX, DTMATC, DTRC, DTPARM and to COMLNK DTSIM and DTCHK. Moved nowhere. 10079. November 23, 1998 $APLNOT/DTPARM Ketan $HLPFIL/DTSIM.HLP Added easier methods of specifying antenna information in the INFILE. These include methods for easily specifying information for individual VLBA and VLA antennas as well as specifying the entire VLBA and various VLA arrays at a time using a single line in the ANTENNA section of the INFILE. For example, STATION='ALL_VLBA' now specifies all 10 VLBA antennas. See the DTSIM.HLP file for more details. Moved nowhere. 10080. November 23, 1998 $QPGOOP/OMFIT Ketan Compilation on a DEC alpha revealed several gotchas. Some floating point constants introduced to prevent division by zero were too small and have been adjusted upwards. An undeclared variable was found that affected only the HAL modeltype. Another undeclared variable was found that probably slowed convergence for the limb-darkening coefficient in the DISK model. Finally, the interpretation of the a priori thermal noise [APARM(1)] was fixed up to match the description in the HLP file. Moved nowhere. 10081. November 23, 1998 $QPGNOT/KRING Ketan Finally, a simulation to test the multi-band modes of KRING has been built using DTSIM. The numbers entered in the SN table were correct but the results reported to the user were not properly labelled for SOLMOD='SMBD' [corresponding to APARM(5)=2]; in this case, KRING was printing the multi-band delay followed by the (single band delay - multi band delay). Now, KRING prints the single band delay followed by the multi band delay. The simulation tested the ability of KRING to recover the correct solution using a standard frequency setup from RDV09, a geodesy experiment spanning frequencies from 2.22099 to 8895.99 MHz. Tests of KRING vs FRING confirm previous results: KRING runs faster in many cases, uses less memory, is able to handle larger problems then FRING. This simulation was also used to test MBDLY to confirm the correctness of the multi-band solutions. KRING got the right answer in 14 cases while MBDLY got the right answer in 2 cases and either failed outright or produced incorrect answers in the other 12 instances. The results of the simulation are available by email to kdesai@nrao.edu. [An AIPS memo may follow at a later date.] Moved nowhere. 10082. November 23, 1998 Calibration routines Chris Introduced support for dispersive delay corrections into the calibration routines. Dispersive delay corrections are automatically enabled when the values in the DISP 1 column of the CL table are not zero. Also revised code handling Faraday rotation to cope with blanked entries in the IFR column of the CL table and fused three loops at the end of DATCAL to improve array access. Note that DSEL.INC now carries wavelegths for each channel rather than squared wavelengths as before (squaring a number is more efficient than taking its square root). The following files were changed. DSEL.INC in INC CGASET.FOR in APLNOT CSLGET.FOR in APLNOT DATCAL.FOR in APLNOT DATPOL.FOR in APLNOT GAININ.FOR in APLNOT ORIPOL.FOR in APLNOT VLBPOL.FOR in APLNOT PCAL.FOR in QPGNOT FRMAP.FOR in YPGNOT (to fix a name clash) Note that LPCAL and SPCAL ignore Faraday rotation (and probably should not). Moved nowhere. 10083. November 23, 1998 PRTUV Eric The change listed above had several errors - the headers for types that did not display subarray and FQid indicated that they did, there was no test for the existence of an FQid, and the lines were too long. Cleaned up the format handling and these matters. Moved nowhere. 10084. November 24, 1998 R*8 keywords in tables Eric Change all R*4 keywords to R*8 keywords in table headers. Fixed: GAINI On write a new table had all the interpolation nodes writing messed up. Probably this is used only on read. MCINI On write a new table had many keyword values in the wrong place. Probably this is used only on read. Changed to force R*8 internally even if not externally: TABKEY ANTINI BPINI CALINI GPINI SNINI AGIO ATIO BAIO BCIO CAIO CMIO FRIO IMIO SOIO CLCAL IMLOD UVLOD UNCAL FITLD Moved nowhere. 10085. November 24, 1998 APGPS Chris Renamed GPSDL to APGPS (APply GPS). Added antenna selection adverbs and an option to calculate dispersive delay. Moved nowhere. 10086. November 25, 1998 SOUFIL, SOUFND Eric Both of these had the logic that QUAL and CALCODE modify the list of sources selected or deselected. Changed this so that QUAL and CALCODE select a subset of sources and then these are selected or deselected on source name. This means that if source names are deselected (SOURCE = '-xxxx'), that a list of sources INCLUDED is made of CALCOD or QUAL are non null and a list of sources EXCLUDED is made if they are both null. The list is 2500 long which should be adequate. Moved nowhere. 10087. November 27, 1998 FITLD Eric Messed up TABKEY when it re-wrote a keyword the type code was re-used as a loop counter and then stuck into the data record. Got is a new loop counter. FITLD did not die on this error (or numerous other ones), so I have added a bunch of IF (IERR.NE.0) thingees. Moved nowhere. 10088. November 30, 1998 BPASS Chris Set DOWTCL true after the call to SELINI so that weights will be calibrated when calibration is requested. This should reduce the influence of weaker antennae in the least-squares fit and improve the quality of bandpasses derived from cross-correlation spectra for arrays with large ranges of sensitivity (eg. VLBA plus HALCA). Moved nowhere. 10089. November 30, 1998 SNCOR Chris Added an operation to clip by weights (suggested by George Moellenbrock). Moved nowhere. 10090. December 1, 1998 ACFIT Eric This task did not put the correct REFANTs (1 per pol per IF) in the SN table. Fixed it also to use dynamic memory so that it can support whatever size averaging problem is thrown at it. Previously it blithely ignored the possibilities of having more than 4 IFs and 1024 channels. It also messed up the number of samples when copying to the output buffer. Moved nowhere. 10091. December 1, 1998 FITLD Eric In attempting to have FITLD actually pay attention to error returns I messed up the sign of the comparison and so failed to reopen the AN file for writing. This then caused AT2AN to fail and the job to die. Moved nowhere. 10092. December 1, 1998 POSSM Eric Changed error message on IERR=-1 from UVGET (INIT...) Moved nowhere. 10093. December 2, 1998 CVEL Chris Removed an unused leap-second function (IAT) from CVEL. Moved nowhere. 10094. December 4, 1998 BP cal Eric removed the subtraction of 1.0 for parallel hand autocorrelations in the bP application code. This assumes that the AC data are around 1 (usually the case but not exactly) and this causes apparent non-circularities in the various cal apoplications. Moved nowhere. 10095. December 4, 1998 Map size parameterize Eric Created PMAD.INC to parameterize the maximum image size and the correspnding buffer sizes. MAXIMG is set to 16384, MABFSS is set to this also and MABFSL to 4 times this. Changed: DBUF.INC Include PMAD, set buffer accordingly, drop RMAX and RMIN (never used) MAPIO Drop RMAX, RMIN never otherwise referenced DTCIO.INC Include PMAD and use to set buffers. MAKMAP Drop equivalence to polarization array, raise buffer size to 4*MABFSS and use a 2nd buffer size indicator to get the most out of our buffers. OLDMAP Drop equivalence and raise buffer size. The routines used by OLDMAP all require 3 buffers so the last trick above does not work. Changed to use PMAD to set limits & buffers: AU1A AU2 AU5A AU5D AU6D AU7 IOSET WRBLNK WRPLAN ARRAY CXIMAGE CXUTIL INTERPLATE QCLEAN QIMAGE Changed formats AU5 Moved nowhere. 10096. December 4, 1998 size bugs Eric In researching the image size business, fixed: AFILE Used 8192 for a 256-word buffer DECOR Used 8192 for 3*MAXCIF uv data buffer. LISTR Used 3,8192 for 3,MAXCIF uv data buffer. MSORT Used 8192 for 512 word buffer OTFUV Used 8192 rather than MAXCHA to set default upper limit. TBIN Used 8192 buffer where only dummy needed; used MAXCHA where MAXCIF would be more correct. TBOUT Used 8192 buffer where 2 words would have been enough and used XBPRSZ where MAXCIF would do. TFILE Used 8192 buffer where 256 words would do. UV2TB Had unused JBUFSZ, used 8192 where 3*MAXCIF was needed, and used 8192 where 3+MAXCIF was needed. UVFLG Used 8192 for max source where XSTBSZ would do. UVLOD Used 8192 in testing for data too large although the buffers will now hold UVFBSS and added test of NCOR < MAXCIF. Moved nowhere. 10097. December 4, 1998 PMAD.INC Eric Changed to use PMAD to set buffer size DPLT.INC CCNTR CNTR GREYS IMEAN IMVIM ISPEC PCNTR PROFL SERCH TVHUI TVRGB IRING KNTR PFPL1 PFPL2 PFPL3 PLCUB PLROW Changed image I/O and: BLANK Rearranged declaration of buffers as well. BLSUM Rearranged declaration of buffers as well. GAL Raised buffer from 1024 for row length. IMLHS Raised buffer from 512 for displayed row length. No image I/O change, but changed: IBLED Removed MINIT from a comment. SPFLG Removed some 16384's from comments TVFLG Removed some 16384's from comments Moved nowhere. 10098. December 4, 1998 FRING Eric FRING set up a priority array for antennas in the initialization routine. Then it used it in the fitting routine to set baseline pointers and then re-arranged the list for another purpose. But, the main routine can loop over subarray, so this rearrangement cost dearly. Added a save and restore for this priority array so that the looping will work. Also added controls so that a subarray in the loop with no data will not cause it to die abruptly. Moved nowhere. 10099. December 7, 1998 More PMAD.INC Eric Added PMAD.INC to control buffer and image sizes in AHIST COMB CPYRT HISEQ PBCOR PRTIM REMAG SAD SDMOD SQASH STFND SUBIM SUMIM TAFFY WTSUM XSUM APGS APVC STEER Made further changes as well in: MCUBE Routine PLNWRT for very small output images could not have worked since it treated inputs as if they were integer images (!) TRANS Made things all explicitly floats and parameterized the very large buffers in this task. XBASL Made all buffers floating, dropped equivalence XGAUS Ditto XPLOT Ditto XSMTH Corrected limit to 2048 hidden in code and removed silly equivalences. APCLN Changed to use POWER2 and parameter MAXIMG to test for legal powers of 2. RSTOR As APCLN. MX Replaced non-standard definition of UVBFSS and made change to POWER2 as well. SDCLN As APCLN. UVMAP As APCLN. IMAGR Changed only the upper limit to be MAXIMG. WFCLN Changed only the upper limit to be MAXIMG. Made only cosmetic changes: UTESS VTESS Moved nowhere. 10100. December 7, 1998 AIPS Eric Changed AU1A and AU2 so that more adverbs will get tested with system limits rather than with the limits in the inputs files. Things like IMSIZE, CLBOX, ECHAN, etc were added to the lists. Moved nowhere. 10101. December 7, 1998 COMB, SUBIM Eric Changed them to save fully blanked images as well as other constant ones. Moved nowhere. 10102. December 8, 1998 FLATN Eric Added the adverb EDGSKP to have FLATN skip pixels around the edge of each field (BLC/TRC do not work since the fields can be different sizes). Also parameterize image size in FLATN. Changed POPSDAT.HLP and DDTLOAD.001 to make EDGSKP a regular adverb rather than a special DDT one. Upgraded the EDGSKP help file. Moved nowhere. 10103. December 8, 1998 Image size param Eric Inserted PMAD.INC into: IMTXT PASTE CONVL DSDG.INC SDIMG IM2UV DCONV GLENS IMLOD IMMOD Inserted PMAD and cleaned up buffering: BLWUP CANDY FETCH IMFIT IMFLT IMLIN JMFIT Minor typing changes: BSMAP FITLD CONVL Fixed buffer size and 2 limits to 4096. FFT Fixed buffer sizes and changed NX, NY tests to use POWER2 and parameters GRIDR Clean up buffer management which was a mess of small buffers and equivalences. HORUS Fixed image size tests to user POWER2, fixed buffers from 4096, clarified buffer assignment. LPCAL Remove unused references to MINIT and MDISK. Moved nowhere. 10104. December 8, 1998 FITLD Eric I missed the critical place where various parameters were from the keywords BEFORE they are put into the aips tables. It did not show up in grep's. This caused scaling, refpix, etc to be off. Moved to the AOC and MPI today. 10105. December 9, 1998 FITLD Eric I did not get yesterday's fix correct. I think it is right now. Moved to aoc and mpi today. 10106. December 9, 1998 POSSM Eric Found a hard-coded bug in computing the bandpass from CPASS coefficients causing it to write at addresses just before the bandpass array. This then caused the start channel to go nuts etc. Moved nowhere. 10107. December 9, 1998 FFT Q routines Eric Fixed array declarations in QCFFT, QRFFT, QXFOUR to be large enough for the new image max size. Moved nowhere. 10108. December 9, 1998 TAPES Pat New verb that shows the printer definitions. It queries the value of NAIPST to see how many tapes there are, then parses fields in the TPDEVS.LIST file and produces a list similar to what you see on AIPS startup. It will be possible to enhance this verb to accept the REMHOST adverb in the future, and try to list the tapes for a remote host (assuming said remote host is listed in TPDEVS.LIST). Moved nowhere. 10109. December 10, 1998 PMAD more Eric Simply added pMAD.INC to set row and buffer sizes to: MANDL MATHS PADIM POLCO REGRD RGBMP RM SHADW SKYVE SLICE SMOTH SUMSQ UVIMG WARP Also changed: MODVF Parameterize not only the buffer, but also the output image size and the number of rings it can handle. Aded correct tests for overflowing these. MOMNT Parameterize row length out and allow up to 1024 channels (it was limited to 256 with no meaningful message to say why it quit). Rearrange buffers and equivalences as well as parameterize: MWFLT NINER NNLSQ PATGN Moved nowhere. 10110. December 11, 1998 PBUTIL Eric The component subtraction routine squirreled away all the image names for safekeeping and should have done so for all fields. Unfortunately, it did it for only 1 field in the 1 at a time mode and so caused nasty things to happen. Moved nowhere. 10111. December 11, 1998 UVGET Eric Changed it to check for a FG table before getting mad about DOFLAG with a non-time sort order. Moved nowhere. 10112. December 14, 1998 AXREF limits Eric Fixed the limits on AXREF for ALTDEF, AXDEFINE, and MCUBE. The 1st two have no absolute limits, the last has 16384 or so. Moved nowhere. 10113. December 14, 1998 TABED Eric Converted the KEYVALUE to integer without examining the data type. Then, when the the value was 10**10 (a frequency) it overflowed computing a value it did not need. Added tests. Moved nowhere. 10114. December 14, 1998 AIPSC Eric The CU2 version of AU2 was way behind in the list of system adverbs and had a bug causing GO tests to fail if there was > 1 GO in the job. Moved nowhere. 10115. December 15, 1998 EPOSWTCH Eric The pointing position was not done, only the reference position. Added code to fix this oversight. Moved nowhere. 10116. December 15, 1998 FLATN Eric Changed INTERPLATE to add a method INTWIN to determine the window in the output image influence by the input (smaller) image. Changed call seq of IMGHGE in INTERPLATE to give the x-axis limits for computing since the whole row must be written (and blank filled). Changed FLATN to init the output and weight files and then to use INTWIN in the loop of fields. It also checks for full 0 or blanked images and skips them. This makes things faster. Changed OHGEO for change to IMGHGE. Moved nowhere. 10117. December 16, 1998 PADIM Eric It ignored the presence of > 1 plane! Changed it to loop over planes. Moved nowhere. 10118. December 16, 1998 TRANS Eric A logic error in setting the depth in the x-y transpose routine was exposed (after > 10 years I suppose) by a cube that started as an N x M x 1 x 1 x K x 1 going to K x N x M x 1... It worked fine in the usual cases even though it was in principal wrong. Moved nowhere. 10119. December 16, 1998 UVCON.FOR, .HLP Leonia This new task generates a u-v database for an interferometric array whose configuration is specified by the user. Visibilities corresponding to a specified model, and gaussian noise appropriate for the specified antenna characteristics are calculated for each visibility. The output is a standard AIPS u-v data file. This task replaces the old procedure which required use of the AIPS tasks UVSIM, UVSUM, UVMOD and verb PUTHEAD. The array geometry can be specified in three different coordinate systems: equatorial, local horizon, and geodetic. There is an option of using set of different frequencies to improve UV coverage. Moved nowhere. 10120. December 17, 1998 LIBR script in SYSUNIX Chris Prevented LIBR from calling lorder on SGI machines, where it does not exist. This appears to have lead to some unreported problems with the midnight job. Moved nowhere. 10121. December 17, 1998 BPINI Chris The KEYVAL/KEYVAR array was declared too small for its contents causing the keyword types array to be overwritten on SGIs. Moved nowhere. 10122. December 18, 1998 OHGEO Eric I made changes for the changed call seq to IMGHGE in INTERPLATE. Unfortunately, I messed up the order of these simple changes making it abort on take off. Moved nowhere. 10123. December 18, 1998 FITLD Eric Corrected the task to check the requested BIF, EIF, BCHAN, and ECHAN and to use them only if they are reasonable. It happily used BIF 5; EIF 12 on a 2-IF input data set finding its data, table values, etc somewhere! Also fixed it to adjust the REFPIX for BCHAN and to put the new NOCHAN and NOBAND and REFPIX in the tables. It went to some real effort to put the old values in the tables while some really clever code managed to build the tables correctly for the selected IFs. Moved nowhere. 10124. December 21, 1998 $APGOOP/MBDLY Ketan As part of my exhaustive check-out of KRING in its various modes, I discovered that MBDLY was hard-wired to only use FREQID=1 and SUBARRAY=1. I have now introduced FREQID and SUBARRAY to MBDLY. Moved nowhere. 10125. December 22, 1998 $QOOP/QEDIUTIL Ketan In the expert mode of EDITR, introduced a default action to be taken when the user enters return without pressing any other keys. Now, the default action is to repeat the last action - on initial entry to the expert mode, the default action is to exit immediately. 10126. December 22, 1998 $QPGOOP/OMFIT Ketan Fixed many minor bugs revealed during recent testing and usage. The processing of multiple source models was falling down when one source was explicitly chosen [revealed by OMFIT test suite]. A bug in the processing of the GAIN adverb, introduced in the last change, has been discovered and is now fixed [revealed by the OMFIT test suite]. The map noise estimate formula was coded incorrectly [missing square root] - it still seems a bit fishy though...]. Data from the ATCA revealed deficiencies with the processing of FQ-less files. APARM(8) was broken and was just sending the input visibilities to the output file. A new component, a 'CORE' model-type has been introduced; this is a point source with a flux spectral index. This model-type was introduced so that results from 'mfclean' in miriad could be tested in AIPS. 'YCORE' is implemented as well. Both seem to have worked when used with real data from the ATCA. Moved nowhere. 10127. December 22, 1998 INCS.SH Eric The SYSSUL version of INCS.SH referenced INCSOL rather than INCSUL as it should. Moved nowhere. 10128. December 22, 1998 APCLN Eric There was an equivalence at buffer(4097) that escaped greps. This was fine so log as buffer was small enough, but when I made it bigger the RA's overran the Decs in component restoration or removal. Also did SDCLN and RSTOR. Moved nowhere. 10129. December 22, 1998 UVPLT Eric Fixed an unfortunate choice of LUN. Using any LUN 31-39 in now illegal on some machines. Moved nowhere. 10130. December 22, 1998 INTEGER*2 Eric Compilers are messing up again on I*2. Changed the declaration in $APLGEN/ZADDR.FOR and $APLUNIX/ZADDR.C to drop the *2 from the INTEGER and call them char *'s in the C. Changed: ZADDR $APLGEN - alter declaration to INTEGER ZADDR $APLUNIX - alter declaration from short int to char ZBFLIP $APLGEN - new routine to call ZBFLI2 with the flip parameter from the Fortran common ZBFLI2 $APLUNIX - new C routine does all 4 values of BYTFLP from a simple copy to all the word swaps and does it for buffers of 2-, 4-, and 8-byte values ZR32RL $APLGEN - drop I*2s, call ZBFLIP. ZI32IL $APLGEN - drop I*2s, call ZBFLIP. ZR64RL $APLGEN - drop I*2s, call ZBFLIP. ZRLR32 $APLGEN - drop I*2s, call ZBFLIP. ZILI32 $APLGEN - drop I*2s, call ZBFLIP. ZRLR64 $APLGEN - drop I*2s, call ZBFLIP. Moved nowhere. 10131. December 28, 1998 GRBOXS Eric Split slice setting into a separate subroutine GRSLIC, dropping it from GRBOXS and DRBOXS. Changed DRBOXS to plot boxes that fit only partially on the TV image and to return IERR=-1 for these and IERR=-2 for ones that don't fit at all. It uses INTEDG to compute all this. Changed GRBOXS to allow resetting only of the boxes that are fully on the displayed image while keeping the other 2 categories fixed. Changed AU5C to use GRSLIC for TVSLICE and to send in all boxes on FILEBOX. This keeps them in the same numerical order which may be useful for memory and record keeping. Fixed limits and comments in the help files for TVBOX, REBOX and FILEBOX. Moved nowhere. 10132. December 28, 1998 IMAGR Eric Changed: INTEDG to fix bug when all 4 coordinates are outside the BLC-TRC area. TVDEVICE 1. to call GRSLIC where appropriate rather than GRBOXS and to issue better messages after it finishes. 2. to reset the status when the window size is made to or is found to have changed QCLEAN 1. to allow OVERLAP 2 without requiring DO3DIM 2. to compute the average absolute value in the windows and to have OVERLAP=2 use 0.9*max+0.1*avg to select the next field to clean. This prefers larger structures to single high points. IMAGR $HLPFIL: to allow OVERLAP 2 without requiring the 3D imaging option. Moved nowhere. 10133. December 28, 1998 SNPLT Eric Changed the meaning of CUTOFF. Weights <= CUTOFF are not plotted when CUTOFF >= 0. Thus CUTOFF=0 plotsd only good data. Plotting bad data with Xs rather than +s was not good enough apparently. Moved nowhere. 10134. December 28, 1998 ACFIT Eric Found a hard-coded 0 divide if the template file has no data for the specified REFANT. Added test and message. Moved nowhere. 10135. December 30, 1998 VPLOT.FOR Leonia The reference frequency has been used for U,V,W in the case of positive NMAPS (U,V,W are taken from the data). It has brought an error in the model calculation which is visible in the case of multifield model with large distance between field. The scaling of the U,V,W being foreseen has not worked actually. The scaling is done for the mean (of selected) channel and the first selected IF. The relevant warnings are added. Moved nowhere. 10136. December 29, 1998 ZIPAC2 Eric I thought I had tested this extensively! But it failed on in place ZI16IL's, blowing the TV code out of the water. Changed it to use a temp variable. Moved nowhere. 10137. December 29, 1998 YCREAD Eric This $YSS routine did not set IERR if it worked! So in loops where it failed to find anything on the 1st plane, it failed to report its success on the 2nd plane! Moved nowhere. 10138. December 30, 1998 XAS Eric Found a bug that was bothering me: in shared 8-bit mode the image memories were not fully cleared by init and clear operations. This turns out to be from bytes_per_line not being SCREEN_WIDTH in this case. Changed the init code to use the same method of finding bytes_per_line as the image access code. Moved nowhere. 10139. January 1, 1999 $YPGNOT/IRING.FOR, .HLP Leonia The collumn of sigma is added to the printing table on request Rick Perley. The default for the ellipces center is changed from the map bottom left corner to the map center. Moved nowhere. 10140. January 1, 1999 UVCOP Eric It was ignoring channel/IF selection when testing data weights and flagging. If an IF that was not to be copied was good, then the fully flagged IFs were copied. Also made it test for flagging or high enough weights in compressed data. Moved nowhere. 10141. January 2, 1999 CLCAL Leonia The setting of CUTOFF has been added to the History file. Moved nowhere. 10142. January 2, 1999 $APGNOT/OTFUV Leonia The was a misprint in printing BCOUNT, ECOUNT in the History file. It's fixed. Moved nowhere. 10143. January 3, 1999 IMAGR Eric Fixed: QCLEAN The recent fix caused it to ignore the request to re-make all images on OVERLAP 2. Moved nowhere. 10144. January 3, 1999 SPLIT Eric Added code to keep track of the time order and to mark the header as unsoretd only if it is. Before, any appending of subarrays caused it to be marked as unsorted. 10145. January 3, 1999 SETGDS Eric SETGDS marks the Clean images as being in the READ status and no one comes along to undo that! In programs that loop over IF, subarray, or whatever with a significant number of fields, the 128 limit on NCFILE is easily blown. Changed: SETGDS Fix message and make it test NCFILE properly. UVGET Fix message (referred to SDGET) UNSETG New routine to unset the read status and reduce NCFILE if possible. Changed to call UNSETG: CALIB PCAL LPCAL KRING CPASS BPASS UVSUB UVADC BLCAL FRING UVCON IBLED VPLOT CLPLT. Moved nowhere. 10146. January 3, 1999 Gridded subtraction Eric The Q routine that does the interpolation and subtraction expects the W term of the filed position to be scaled to the first channel. It even says this in its precursor comments. Unfortunately, it appears that ALGSUB and ALGSTB have never honored this request. Therefore, the W term was done incorrectly if the first channel was not at the reference frequency. If the first channel was at the reference frequency, then multi-channel subtractions were done correctly. Fortunately, this is at some level a second order affect since it is in the W correction and in the (delta freq) / freq0, but it was enough to make my test case diverge. That case had a model object 3450 arcsec from the field center with cells of 2.9 arc sec and a beam around 10-13 arc sec which was based on a real case presented by Bill. Moved nowhere. This must be a patch and an announcement. 10147. January 4, 1999 PBUTIL Eric Changed POSITION to add a method to convert an arc-sec offset to an angle from the pointing position. The method is public and call PSNANG. Changed PBUTIL to use this method - previously it was doing a more linear computation that ignored the fact that the reference pixel and the pointing position may not be the same (esp in DO3D imaging). Moved nowhere. 10148. January 6, 1999 K2TEXT Eric Fixed it to not put close quotes on lines. With close quotes, lower case POPS strings became all upper case on COMPRESS. Movedd nowhere. 10149. January 6-7, 1999 IMAGR, SCMAP Eric Changed: AU6 Pull scroll from AU6 to TVSCRL. TVSCRL New - does interactive TV scroll TVDEVICE Add OFFS and SCRO operations to TVDFUN to do off scroll and interactive scroll. TVUTIL Make utility interfaces OTVOFS and OTVSCR to the TVDFUN operations. QCLEAN Call OTVOFS to turn off scroll. Add code to write OBOXFILE if REBOX or TVBOX and to allow TELL to change the OBOXFILE. SCMAP Call OTVOFS to turn off scroll. Changed TV loading code in self-cal TV to allow an image window not including all Clean boxes. Added code to write out the WINDOW if OBOXFILE is set and REBOX or TVBOX are used. Added the adverb to help and task inputs and its CLEAN object. IMAGR Added OBOXFILE adverb. POPSDAT Add adverb OBOXFILE. NEWPARMS Add adverb OBOXFILE. OBOXFILE New help file WINDF New subroutine pulled from IMAGR FIELDF New subroutine pulled freom IMAGR Moved nowhere. 10150. January 9, 1999 ZIPAC2 Eric Still trying to get this right. Changed all "char" to "signed char" since IBMs treat char as unsigned. All IBM tasks should be relinked but I am not going to relink everything for the mnj. Moved nowhere. 10151. January 10, 1999 RUN files Eric Changed $APLGEN versions of ZTOPEN and ZTXMAT to allow RUN files with user number 1 as well as with the user's number in area VERSON when VERSON is not one of the standard ones (it uses RUNSYS when VERSON is a standard one). Also modified the quaint help file for RUN although I did not try to change the detailed description of SOS in editing a RUN file! Moved nowhere. 10152. January 11, 1999 FITLD Chris Addressed a number of issues that were found while examining FITS-IDI data from DRAO. - The correct SOURCE_ID column label is now recognized in the UV_DATA table. Previously FITLD expected the incorrect heading SOURCE ID that was produced by the VLBA correlator. This old label is still recognized to support legacy data. - Antenna characteristics (ANTENNA) tables are no longer required in FITS-IDI file. Antenna polarizations will not be labelled if these tables are missing and a warning message will be produced. Memo 102 does not require antenna characteristics tables to be present and they contain little useful data. - Data is no longer discarded if it is out of time sequence in a UV_DATA table (DRAO data was in BT order). Data out of time sequence is now noted and treated in the same way that multiple integrations were before. Out of order day numbers will still cause problems. Memo 102 does not require T* ordering. Also fixed some miscellaneous problems. - Removed 10 kHz rounding of antenna table reference frequency in AG2AN. - Ensured ANNCAL is set to zero before it is used to set the number of polarization calibration parameters for a new table in AG2AN (it could have been used without being initialized). - Removed code that used an uninitialized variable to set the FREQID keyword in an AN table from AT2AN. Moved nowhere. Should go to 15OCT98 patch area after further tests. 10153. January 14, 1999 VPLOT Leonia Plot the model at each data sample (NMAPS>0) has worked only for SOLINT=0. I modified VPLOT to exclude this limitation. If SOLINT>0 U,V,W are calculated as average value of the samples of the average interval. Moved nowhere. 10154. January 19, 1999 CGASET in APLNOT Chris CGASET does not refresh the interpolated calibration values if the time for which calibration is required differs from that for which the last interpolated values were calculated is smaller than a given fraction of the time span between the values to be interpolated. The logic for this test ignored the possibility that the values to be interpolated between might have changed since the last interpolated value was calculated. This could lead to stale calibration values being used at the ends of scans (if the gap between scans was sufficiently large the difference in time between the last data point in a scan and the previous one could be a small fraction of the gap even if it was a similar size to the spacing between calibration values within the scan. Modified CGASET to force new interpolated calibration values to be calculated whenever the values to be interpolated between change. Moved nowhere. 10155. January 20, 1999 LDGPS Chris Expanded longitude range from 0 to 360 to -180 to 360 degrees. Added some missing calls to MSGWRT. Moved nowhere. 10156. January 20, 1999 FLGIT Eric Corrected the call sequences to IMSEL, MCSEL, and PCSEL which were changed for UVCOP in August, but forgotten in FLGIT. Moved to patch area. 10157. January 21, 1999 UV table tapes Eric Wrote FITAB. It does two new and dangerous things. 1. For images it can apply the integerization based on a least bit flux provided by the user 2. For uv data, it writes tables of uv data rather than random groups. There is the option to write compressed uv data (like on disk) and to break the operation into multiple output files, each covering a specified time interval. Other changes so far: NPIECE New adverb help file. QUANTIZE New adverb help file. POPSDAT Add NPIECE, QUANTIZE NEWPARMS Add NPIECE, QUANTIZE DFUV.INC Added Table keywords as well as random parameter keywords and piecemeal uv parameters. VFUV.INC Added Table keywords as well as random parameter keywords and piecemeal uv parameters. FPARSE Fixed to handle the extra words properly. The data loading tasks will need these fixes too. Changed call sequence. CHPACK Like REFRMT except that it removes all blanks from a string. PRTTP Added the option to read disk files and to do ADVF rather than read long tape files on PRTLEV=-4 (otherwise like -3). Fixed it to recognize the new UV format and make an intelligent display. ZGDATE New dummy APLGEN and real APLUNIX routines to return the current date at Greenwich. MSGHDR Fixed the pixel-length descriptions to be more readable, making the same change in PRTTP. ZTPOPN $APLGEN - removed a bunch of blank lines ZTPOPD $APLGEN - had it check MSGSUP and honor 32000 to suppress the naming of the file it is trying to open. AU4 Add attempt to open INFILE with a "1" appended if INFILE not found. TPHEAD Add info the HELP file re this INFILE handling. FITTP Changed to call ZGDATE to meet standards and fixed the addresses of the weights for DOSTOKES true UPOL output. Also changed it to stop putting out magic values when the Stokes conversion fails. It should put out 0 for the vis as well as the weight. UVLOD Fixed regular data reading section to change magic blanks into 0 with 0 weight. Added adverb NPIECE and quiet attempt to open INFILE and INFILE with a 1. Add looping for uv tables and read of uv tables. RUVTAB New - reads uv table data to uv file. TABKEY Fixed to set IERR to 0 at start and to exit quickly if NUMKEY <= 0. Moved nowhere. 10158. January 21, 1999 FITLD Chris Prevented FITLD from overwriting the saved values of NOBAND and NOCHAN that are written to table headers. The unmodified values must be written to tables in order that calibration transfer and concatenation work with IF selection in effect. Moved nowhere. 10159. January 21, 1999 CCBOX Chris Removed CCBOX (which is given in arcsec and has no limits) from the system adverbs lists in AU1A and AU2. Corrected the description of CCBOX in CCBOX.HLP. Moved nowhere. 10160. January 21, 1999 GETSIP in APLNOT Chris Added protection against the sine of the azimuth falling out of range near +/- 90 degrees due to finite precision arithmetic. Moved nowhere. 10161. January 22, 1999 FITLD Eric Added the UVLOD uv-table reading changes to FITLD. The resetting of name parameters (when OUTNAME was given) has been changed so that only OUTSEQ is set to 0 and the others are reset back to the input adverbs. Also changed: IMLOD To handle the changed error returns from MAKTAB and to close tables on error TBIN To handle the changed error return from MAKTAB. Moved nowhere. 10162. January 22, 1999 UVPOL Eric Changed it to omit all samples which do not have at least one good cross-hand observation. The doubly flagged times were affecting the imaging through uniform weighting and sum of weights. Moved to Japan. 10163. January 27, 1999 SPLAT Eric SPLAT also was not updated for the changed call sequence of IMSEL and MCSEL. Added calls to CQSEL and PCSEL and changed it to not copy SU tables to output single-source files. Moved nowhere. 10164. January 27, 1999 UVAVG Eric Fixed it to expand the file if it writes more than expected rather than dying. Moved nowhere. 10165. January 27, 1999 PCLOD Chris Phase cal amplitudes in VLBA log files are percentages. Changed PCLOD to divide the amplitudes by 100 to be consistent with direct cal transfer. Moved nowhere. 10166. January 27, 1999 FITLD Chris Protected calls to MAPCQI and CORID with tests on DELCOR. Without these tests FITLD would die of unnatural causes when appending to an existing file with delay decorrelation corrections turned off. FITLD was trying to read a CQ table in circumstances where one does not exist. Moved nowhere. 10167. January 27, 1999 USUBA Eric Changed OPTIMIZE.LIS to avoid optimizing USUBA on HPs. It worked more or less but could not tell that time ranges overlap... Moved nowhere. 10168. January 29, 1999 CVEL Eric Changed the smoothing option in CVEL to require APARM(9) > 1 for smoothing cross correlation spectra, > 0 for autocorrelations. Modified XCSHFT to avoid excess multiplies when not smoothing. Corrected ACSHFT to use the average autocorrelation value rather than 1 for frequency NFREQ+1. Use of 0 for spectra that do not average 0 introduces ringing from the high frequency end of the spectrum. Moved nowhere. 10169. February 2, 1999 FITLD Chris Modified AT2AN to reset the AT table row number when looping over subarrays. Moved nowhere. 10170. February 2, 1999 FITLD Leonia Changed the action of WTTHRESH (C. Walker request). The old scheme was that, if any IF has a weight below the threshold, the whole record is flagged. Now if weights for all IFs are =< WTTHRESH the visibility record is not recorded. If a weight of some IFs (not all) =< WTTHRESH such a visibility is recorded but the weight of these IFs put to zero. Moved nowhere. 10171. February 2, 1999 FITLD, UVLOD Eric Reading an old INTEGER*2 UV-data FITS tape (!!) a bug in the size of the integer buffer appeared. It is not 720 but 1440 to hold all the words after ZI16IL. Added more tests on MAKTAB outputs as well to FITLD's image part. IMLOD was already okay. Moved nowhere. 10172. February 2, 1999 PRTTP Eric Made 2 minor changes affecting PRTLEV=-4: Fixed it to set DOCRT=-1 so that a progress message will appear and fixed it to avoid duplicating entries in the output text file when it does skip a file. Moved nowhere. 10173. February 3, 1999 INSTEP1 Pat Added a section to allow a user to configure a local version of the XHELP cgi perl script. In general doing this from within INSTEP1 requires write access to the cgi area, something that is probably unlikely. However, the section should at least alert people that the capability is there. This shell segment needs tested (but I always get to test it anyway at each release). Moved nowhere. 10174. February 4, 1999 FITLD Pat Discovered that when a FITS disk file is too short (not padded out to a multiple of 2880 bytes), the message from FITLD is rather obtuse. Added detection of IERR=4 (end-of-file) and a more understandable error message. Moved nowhere. 10175. February 5, 1999 ZMOUN2 (Linux) Pat If an error occured during the MOUNT process, this routine did not close the tape device, resulting in a lockout until the process (AIPS) terminated. Put in code to close the device on all relevant error branches. Moved to the AOC this date. 10176. February 10, 1999 TAPES Eric Changed: AU4 To pick up REMHOST and send it to TAPSHO TAPSHO To remove Fortran errors and infinite loops, to have it search the local tape list for REMHOST (or the local machine by default) and then to call ZRMTAP if none found. The parsing and listing follows this. ZRMTAP APLGEN: new - opens the remote TPMON as if for a disk file and expects back 2880 bytes of tape info. It sorts this into an output set of strings and closes the remote. TAPES Added REMHOST. TPMON Trap the file name on ZTPOPN and call ATPSHO rather than ATPOPN and added ATPSHO which is similar to TAPSHO without the remote host and the actual display. Moved nowhere. 10177. February 17, 1999 FITLD Chris The subroutine GET3DT assumed that the ARRAY column in UVDATA tables contained a filter number (presumably a historical oddity in VLBA data at some point) with the result that subarray numbers were discarded if they were explicitly assigned in the incoming data. FITLD now assumes that the ARRAY column in a UVDATA table is a real subarray number for arrays other than the VLBA and turns off automatic subarray allocation if the array number is not zero. Moved nowhere. 10178. February 17, 1999 INDXR Chris Removed a test on the maximum antenna number since INDXR can cope with arbitrarily large antenna numbers as long as there are no more than MAXANT antennas in any subarray. Also increased the maximum number of subarrays to 20 to cope with DRAO galactic plane data (which can have at least 13) and to match the limit set in FITLD. Moved nowhere. 10179. February 18, 1999 FITLD Chris FITLD no longer treats array numbers as filter IDs in VLBA data. This oddity was inserted to deal with some early test data from the VLBA and should not be required to deal with any data in the VLBA archive. Moved nowhere. 10180. February 24, 1999 $APGNOT/CLCOR Leonia The correction of delay and rate (option ANTP) have been calculated calculated using only linear terms in the series of the corrections. In the extreme case the neglecting of the higher terms can lead to the visible errors. To eliminate the problem of the high term at all the new version of CLCOR calculates the correction of the delay and rate as the difference of the delay and rate in the shifted and original position. The hour angle is calculated by SOUELV as REAL. I atached the local version of SOUELV(SOUEL) to have the hour angle with double precision accuracy. CLCOR has died at 'ANTP' if ANTENNA = 0. I modifyed it. Now it dies only if antenna position is corrected but continues working if only source position is corrected. Help file is updated. Moved nowhere. 10181. February 25, 1999 $SYSUNIX/LINK, LIBR Pat These two sometimes conspired to trick COMLNK into thinking that everything worked, when it did not. Tried to close a few holes where the status was not necessarily caught correctly. Moved nowhere. 10182. February 25, 1999 $HLPFIL/UVFIX.HLP Leonia Sometimes ago I modified UVFIX including orbiting antennas, the effect of the differential abberation (Lorentz contraction) and recalculation of the U,V,W to the J2000 in all cases. These modifications were not shown in the help file. I added four lines to emphasize these modifications. Moved nowhere. 10183. March 1, 1999 FITLD Chris SO2UV now calculates approximate apparent source positions from the mean positions if they appear to be the same in the incoming table and if the mean equinox is 2000.0 or later (so that there is a good chance that Julian precession is appropriate). This calculation does not include GR terms, aberration, or polar motion. This is a crude fix for programs that trust the apparent coordinates in source tables and are flummoxed by data from the VLBA correlator (which writes mean positions in these fields). Moved nowhere. 10184. March 2, 1999 OBOXFILE Eric Fixed WRIBOX to initialize the scratch file name. Otherwise, it got into trouble when creating a new file in that it failed to delete the non-existant scratch file. Moved nowhere. 10185. March 3, 1999 FITLD Chris FITLD now issues a warning on finding sources with the same name and qualifier but different positions. It would be better to generate a new source qualifier in these circumstances but this would require revising the source-handling logic to keep track of the qualifiers that are in use: this will wait for a more thorough overhaul of the source table handling in FITLD. Moved nowhere. 10186. March 6, 1999 QEDIUTIL Eric Several places required there to be more than one sample at a time before they would examine that time. This did not work when only one antenna was included for sure and could have gotten in trouble occasionally at other times. Affected SNEDT, EDITA and EDITR. Moved nowhere. 10187. March 7, 1999 COMB Eric COMB assumed that the FREQ axis was the same axis in both images which causes problems on occasion with SPIX amd ROTM. Fixed it to be more careful. Note - > 1 point axes must still be the same type by axis number. Moved nowhere. 10188. March 8, 1999 WHATSNEW Eric Added info re 15APR99 and 15OCT98 to this file. Moved to my www page as well. 10189. March 9, 1999 Calibration transfer documentation Chris Cleaned up references to calibration transfer in the help files for ANTAB, APCAL, MERGECAL, PCCOR, and PCLOD. Moved to 15OCT98 patch area on this date. 10190. March 9, 1999 PASTE Eric Used CRPIX for default PIXXY on non-existing axes, but it is 0 there and => bad BLC, TRC. Pick up number real axes and set defaults 1 for non-real axes. Moved nowhere. 10191. March 10, 1999 Imaging Eric Ketan has a lovely data set (small) that brought out the worst in things. When making an 8192x8192 image, it has only small swaths in the AP at any 1 time and found itself running out of data with the swath from -4 to +10 or so yet to do. So it happily told the weight summing and phase rotating routines to include those non-existant rows. Fixed for this: GR3DUV $QNOT: Failed to know it was out of data and did a bad UVINIT call APSUM $QNOT: Would work on non-existant rows APFNSH $QNOT: Would work on non-existant rows There is also a problem about fiting in a small AP: QUVUTIL $QOOP: Did not prepare NCHAVG before call to GRDFIT so first test did not check for bandwidth synthesis GRDFIT $QSUB: which was a good thing since it would have failed the test. Corrected test to see if bandwidth range will fit around max U rather than full image and weight grid size which should be 2-4 times larger and is not needed. Fixed error message as well. Moved nowhere. 10192. March 10, 1999 UVLOD, FITLD Eric Fixed RUVTAB. It did not copy the right number of correlators for multi-IF and multi-channel FITAB files. Moved nowhere. 10193. March 11, 1999 Imaging Eric Changed UVWAIT that demanded too much memory for weighting and changed GRDFIT to match better and to use FMAX to scale the Max U for the does everything fit test. Moved nowhere. 10194. March 11, 1999 SCMAP Eric Changed SCMAP - its reporting of rms levels was messed up causing it to leave out one message and to have the others be one iteration behind. Removed a debug message from APSUM as well. Moved nowhere. 10195. March 11, 1999 CSLGET Eric When filling in the earlier non-existant time's data from the later time, CSLGET neglected to copy the IFRTAB and DDTAB values along with the time and gain values. This left FBLANK around when it did not need to be exposing us to problems which have been addressed earlier, but lurk. Moved nowhere. 10196. March 12, 1999 SOURNU Eric The routine returned #source=0 in two cases: all are selected and none of the requested ones were found. Changed the latter to return also IRET=-1. Changed calls to SOURNU: PCCOR Changed to test for error! QUACK Changed to test for JERR<0, no call SOURNU when NOSOUR UVCOP Changed to report error and quit rather than copy all. UVFLG Changed to treat a failed match as an error and quit. CLPLT Changed to test error and quit SNPLT Changed to report sources not found - continues an all SU table erors incl not found VPLOT Changed to test error and quit Moved nowhere. 10197. March 12, 1999 UVCOP Eric FGSEL and SUSEL copied records from the old table to the new ignoring the fact that they were flagged in the old table. All the other *SEL ones deleted such records, so fixed these to do that too. Moved nowhere. 10198. March 17, 1999 Linux BUG Eric Changed $APLUNIX versions of ZSSSL2.C and ZSSSO2.C to remember which TV they opened and thereby avoid certain translations plus gethostbyname and getservbyname. This last leaves a file descriptor open under Linux when YP is needed to find the service and this can lead to "too many open files" and program death. Moved nowhere. 10199. March 18, 1999 ZSSSO2 Eric I messed up Unix sockets with not moving all the arguments to the new nomenclature. Sigh... Moved nowhere (e-mail Ralph Gaime). 10200. March 18, 1999 BATQ Eric When submitting a job with a delay time, the stored times were too large by 1900 years. Moved nowhere. 10201. March 22, 1999 UVPRT Eric Added OFFSET to offset time intervals for holography auto-editing. Moved nowhere. *********************************************************************** *********************************************************************** Changes while 15APR99 is NEW *********************************************************************** *********************************************************************** 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 from 15OCT99 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 March 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 March April 8. 10205. March 27, 1999 FLGIT Eric Added option to flag on VPOL, fixing up displays and help file. Moved to 15APR99 March 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 March 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 March April 8. 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 from 15OCT99 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 from 15OCT99 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. 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 April 8. 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 to 15APR99 April 8.. 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 from 15OCT99 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 from 15OCT99 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 from 15OCT99 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 from 15OCT99 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 from 15OCT99 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 from 15OCT99 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 from 15OCT99 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. 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 April 16. 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. 10230. 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 from 15OCT99 this date. 10231. 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 from 15OCT99 on April 29 by Pat. 10232. 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 from 15OCT99 on April 29 by Pat. **********************************************************************