----------------------------------------------------------------------- ; Copyright (C) 1995, 1997 ; Associated Universities, Inc. Washington DC, USA. ; ; This program is free software; you can redistribute it and/or ; modify it under the terms of the GNU General Public License as ; published by the Free Software Foundation; either version 2 of ; the License, or (at your option) any later version. ; ; This program is distributed in the hope that it will be useful, ; but WITHOUT ANY WARRANTY; without even the implied warranty of ; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ; GNU General Public License for more details. ; ; You should have received a copy of the GNU General Public ; License along with this program; if not, write to the Free ; Software Foundation, Inc., 675 Massachusetts Ave, Cambridge, ; MA 02139, USA. ; ; Correspondence concerning AIPS should be addressed as follows: ; Internet email: aipsmail@nrao.edu. ; Postal address: AIPS Project Office ; National Radio Astronomy Observatory ; 520 Edgemont Road ; Charlottesville, VA 22903-2475 USA ----------------------------------------------------------------------- This file must be used to record any changes made to the the master version of AIPS kept in Charlottesville. An entry is: line 1: Date Program name Your name line 2: What was done (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 15JUL95 *********************************************************************** Changes while 15JUL95 was TST: *********************************************************************** *********************************************************************** 8414. December 23, 1994 TVFLG, SPFLG Eric Changed TVFLG, SPFLG, and GRIDTB to try hard to avoid bad beating between the data times and the Master grid cell times. The programs all integerize the times on a 0.1 of a Master grid cell (DPARM(6)) and look through the first 10000+ times to see what regularity there is in those sample times. Then the Master grid cells are set to be centered on that preferred 0.1 of a cell. Also made fixes that will allow data sampled < 1 sec apart (i.e. OTF) including raising averaging buffers to hold all the data allowed (16384 rows in the master grid). Changed DSMEAR to suppress error messages and clean up its typing. Fixed SPFLG and TVFLG to average across blank rows so long as the time range and source ID are ok. A blank row stopped it before. Also fixed a quirk with the menu and help display. Changed the smooth and scan times displayed to be the integer number of basic intervals rather than try to show seconds. Moved to 15JAN95. 8415. December 23, 1994 OTFUV Eric Changed it to make the beam be 257 (i.e. baseline 1-1). The old scheme confused SPFLG etc. Added code to keep track of data ordering and to put SORT = 'T' in the header iff its true. Added message at the end about it too. Moved to the patch area and 15JAN95. 8416. December 28, 1994 ZLASCL Pat Modified it to shed the ghostview or pageview process instead of waiting for it to finish. It would be nice if there was a way to keep track of a running previewer and feed it new text or graphics... (maybe there is and I haven't found it yet) Moved to 15JAN95 this date. 8417. December 28, 1994 APROPO list Eric Changed HLPA.FOR to put out lines with no trailing blanks, HLPB.FOR to eliminate all duplicates (it got only 1 of multiple occurences). Updated LSAPROPO.HLP in TST and NEW. Otherwise, moved nowhere. 8418. December 27-29, 1994 Plotting Eric (1) Most plotting tasks place their labels too close together, separating by 1.0 characters. Changed this to 1.33 in the vertical direction. (2) Added additional LTYPE values 7--10, which are like 3--6 but omit all labeling except the tick values and X and Y axis labels. (3) LWPLA has been using a character size larger than the actual characters of the font to compensate for the too close spacing. This caused strings to be placed incorrectly, which was partially compensated for by having blank characters expanded some. Changed it to use correct character sizes and to drop the blank expansion. Also added a little to the border. Routines changed: $APLSUB: LABINI, SLBINI, TKSLIN, TKLAB, TKTICS $YSUB: COMLAB, CTICS, CLAB1, CLAB2, ITICS, IAXIS1 $AIPSUB: AU5B, AU5C, AU9B $YPGM: CNTR, CCNTR, PCNTR, GREYS, ISPEC, PROFL, SL2PL $YPGNOT: KNTR, PFPL1, PFPL2, PFPL3, PLROW, PLCUB $HLPFIL: LTYPE, CNTR, CCNTR, PCNTR, KNTR, TVLABEL, TKGUESS, TKMODEL, TKRESID, TKSLICE, GREYS, ISPEC, PROFL, SL2PL, KNTR, PFPL1, PFPL2, PFPL3, PLROW, PLCUB Corrected a bug in TKTICS which caused the tick marks to have 0 length. Shortened the TV tick marks as well. Changed IAT to TIME in CLAB1, CLAB2, TKLAB, PROFL Moved nowhere. 8419. December 29, 1994 many help files Eric Changed the version number limit from 255 to 46655 in a large number of help files. Did the same to 15JAN95 as 15JUL95. Moved nowhere. 8420. December 29, 1994 UVPRT Eric Changed it to trap HOLG mode and make DOCRT negative even when the user left it positive. This makes the task detach. Also removed the qualifier from W in lography mode (mod by 256). Moved to 15JAN95 from 15JUL95. 8421. December 29, 1994 HOLGR Eric Rick Perley has corrected and tested the flat field model routines for both kinds of feeds. Replaced FLATPH and FLATPH2 routines per his instructions. Moved to 15JAN95 from 15JUL95. 8422. December 29, 1994 $APLNOT/DATCAL.FOR Athol Corrected ~10% error in delay smearing correction caused by using an imperfect model for the spectral averaging. Moved to 15JAN95 from 15JUL95. 8423. December 30, 1994 SN tables Phil There was a problem in CLCAL when the user had copied an SN table from a single source file to a multi-source file and then sets the array CALSOU to one or more non-blank values. CLCAL looks in the SN table for the source numbers, but they are zero (because the file from which it originated was single-source) and so decided that there are no valid sources and thus fails to write a new CL table. The most general way to fix this was to add a new keyword, ORIGIN, to the SN table. If ORIGIN is 0 or not there then the CLCAL performs as before and does the correct thing for multi-source files. If ORIGIN is 1 then the SN table came from a single-source file and the CALSOU array is then ignored inside CLCAL. Modified: $QPGNOT/CALIB, $APGNOT/ACFIT & APGNOT/UVCRS to write the ORIGIN keyword. Modified CLCAL to read that keyword only when a single SN table is being applied, otherwise if multiple SN tables are being merged it acts as it did before. Modified SN2CL to act on the ORIGIN keyword in the appropriate manner. Moved nowhere. 8424. December 31, 1994 ACFIT Eric A statement "IF (IRET) THEN" was placed in the code. This is not legal since IRET is not a logical variable. Fixed. Moved nowhere. 8425. December 31, 1994 More plotting Eric Numerous tasks do plots without giving the user any control over the axis labeling. While the pixel and relative types may not be meaningful, types 1, 2, 3, 7, and the negatives are. These tasks used, effectively 3 at all times before. Changed: IMEAN Added LTYPE and added the option to plot the histogram with the counts on the Y axis and flux on X. LWPLA Changed to use more correct character sizes for 2 of the fonts. I'll get more later. IMVIM Added LTYPE. TAPLT Added LTYPE. EXTLIST checked and okay for IMEAN, IMVIM, TAPLT Moved nowhere. 8426. January 2, 1995 UVHGM Eric Changed it to use LTYPE. Also changed it to read compressed data (it did not even check for this!), to do some Stokes conversion, to plot both RR and LL or RL and LR on one plot if requested, to plot more than one channel and/or more than one IF on each plot. Cleaned up handling of parameters too. Changed AU8A (EXTLIST) for changes in help file. Moved nowhere. 8427. January 6, 1995 DISKU Pat Redirected egrep output to the bit bucket; some systems (e.g. linux) have GNU egrep which still outputs stuff with -s. Moved to 15JAN95 this date. 8428. January 6, 1995 Single dish Eric Changed OTFUV to make the output beam be 257 * BIF. This will let it look like antenna BIF's autocorrelation to the data selection routines. Changed PRTSD to check the BEAM random parameter for this patern and to then display the antenna number. Changed DATGET to force DOACOR to true and to skip selection on UVRANGE for single-dish data. Moved to 15JAN95 this date. 8429. January 9, 1995 BLING Chris Fixed a bug in the output routine for 'INDE' mode which meant that the user had to enter 'Q' multiple times to abort the program. Moved to 15JAN95 this date. 8430. January 9, 1995 FRING Leonia The multi-band delay is stored in the SN table as a phase ramp across IFs, there was a bug in the calculation of this phase ramp if the first IF was not the reference frequency. In this case the frequency offset in the FQ table for IF#1 was negative since its sky frequency lay below that of the reference frequency. When this happened FRING calculated the same phase for the first and second IF. The error was caused because the frequency shift for the second IF was equal zero. The relevant change has been implemented in subroutines FRNDR and FRNMS. Moved nowhere. 8431. January 9, 1995 FRING Leonia I have canceled the previous change because it forced the reference IF to #1 that is generaly wrong. It is necessary to evaluate the reference IF using criteria frequency shift equals zero. The change will be implemented after complete investigation of the problem. Moved nowhere. 8432. January 5-6 & 9, 1995 More plotting Eric Numerous tasks do plots without giving the user any control over the axis labeling. While the pixel and relative types may not be meaningful, types 1, 2, 3, 7, and the negatives are. These tasks used, effectively 3 at all times before. Added LTYPE to: SNPLT 3 places labeling is done POSSM Had lots of things stopping single-dish plots. UVPLT also moved string placement some VBPLT, VPLOT, CLPLT, DFTPL, FRPLT, FRMAP saved parameters better Changed AU8A (EXTLIST) to compensate for the changes in the inputs files of VBPLT, VPLOT, SNPLT, POSSM, CLPLT, DFTPL, FRPLT, FRMAP made some long time ago. The last 3 needed new code in AU8A. Also corrected a bad wording in TVINIT's help file and made new plot type numbers (21 - 25) in GINIT's comments. Moved nowhere. 8433. January 10, 1995 FRING Leonia I have implemented changes in subroutenes FRNDR and FRNMS to overcome the problem discribed in 8430. Reference IF is that one for which frequency shift in FQ table is equal zero. Moved nowhere. 8434. January 10, 1995 SDGRD Eric Fixed HI record on weighting to include UVWTFN string. Moved to 15JAN95 this date. 8435. January 10, 1995 SPLIT Leonia SPLIT calculated erroniously header frequency and shift frequency for different IF in the case when the first IF's frequency does not equal zero. It was realized with M. Rupen data in which the first IF frequency is negative. But the task could not work properly with any data if a portion of IFs (starting with not the first IF) was copyed. The relevant change in subroutine SPLCOP has been implemented. Now the first IF's frequency is forced to zero and corresponded correction of the rest frequency is being done. Moved nowhere. 8436. January 12, 1995 WFCLN Eric The routine to read clean boxes only read the 1st 50 cards in the file. Changed that to 5000 and fixed the END and ERR branches to go to a reasonable point. Moved nowhere - should go to 15JAN95. 8437. January 12 & 14, 1995 LABINI Eric Changed a character limit by a little. Moved nowhere. 8438. January 13, 1995 UVFLG Eric Some user has complained about Leonia's English in the UVFLG help file. Actually, it was one of his better efforts and was quite clear, but I reworded it to be more pedantic. Moved nowhere. 8439. January 13, 1995 GREYS, DOGREY Eric Correct GREYS to honor LTYPE in labeling the wedge - it labeld it even when LTYPE = 1 or 2. Create a new adverb DOGREY, making a help file and changing POPSDAT.HLP and NEWPARMS.001. It will be used in KNTR. Moved nowhere. 8440. January 15, 1995 KNTR Eric Changed KNTR to do grey scales and/or contours. Pulled the beam plotting routine, made suitably generic, from the code as a separate routine PLTBEM ($YSUB). Also help file of course plus AU8A for EXTLIST. This means that CNTR and GREYS are obsolete. Changed POPSDAT.HLP and NEWPARMS.001 to add adverb CBPLOT and LABEL. Made new help files for the 2 new adverbs. Moved nowhere. 8441. January 16, 1995 UVFLG Leonia Changed the help to take it more clear (I hope). Changed logic a little bit. Excluded the low limit of weight. Moved nowhere. 8442. January 18, 1995 PCNTR, EXTLIST Eric Changed PCNTR to process I-P-A images as before plus I-Q-U images and I-Q-U cubes. Upgraded EXTLIST (AU8A) to support the improvemnets. Moved nowhere. 8443. January 20, 1995 $APGNOT/SNSMO.FOR Athol SNSMO did not resolve the 2 pi ambiguity before direct smoothing of phases. Corrected this to be consistent with what is done in CLCAL. Moved nowhere. 8444. January 20, 1995 $APGNOT/AVSPC & SPLIT Phil When averaging all channels and all IFs together both routines called AVGIF. AVGIF is flawed and had been replaced with AVGCIF a long time ago. SOmewhow these two tasks were overlooked when AVGCIF was written. Fixed this. Also ensured that SPLIT uses the reference frequency determined by UVGET rather than recomputing it before exit. Although it gets the anwer right I can forsee it coming up with a different answer after all the uvw's have been scaled. So this must be prevented. Moved nowhere. 8445. January 21, 1995 POSSM Eric Fixed it to work when NCOUNT > 0. I had messed it up adding the new labeling, since parameters were not yet known when they were needed. Changed things around to do more sensible scaling and labeling and get things like the velocity label right (and use it only when appropriate). Change frame scaling to put the plots directly next each other with no interior labeling when the plots are fully user scaled. Fixed frame scaling so outer labels did not have a big space away from the plot. Fixed it to put the relevant antennas into XANTEN for EXTLIST. Moved nowhere. 8446. January 23, 1995 $APLSUB/AVGIF.FOR Phil Removed this flawed routine from the system. It had been replaced eons ago by AVGCIF but was still lurking in a couple of routines. Moved nowhere. 8447. January 23, 1995 $YPGNOT/VPLOT.FOR Leonia There was a difference between VBPLT and VPLOT in interaction of adverbs ANTENNAS and BASELINE. VBPLT did not have opportunity to plot the baselines between antennas named in a given list. VPLOT did not have opportunity to plot the baselines between antennas named in a given list and any antenna. I have changed the function REQBAS to have both these options in VPLOT. The help file has been changed to clarify the interaction of adverbs ANTENNAS and BASELINE. Moved nowhere. 8448. January 24, 1995 $APGNOT/LISTR.FOR $HLPFIL/LISTR.HLP Athol Added a new option to OPTYPE='GAIN', namely STOKES='POLD', which, for any CL or SN table, will print the R/L ampl. gain ratio or R-L phase gain difference as selected by DPARM(1) (0=amp; 1=phase). This option is ignored for other values of DPARM(1) or OPTYPE <> 'GAIN'. Moved nowhere. 8449. January 25, 1995 $APGNOT/PCLOD.FOR Phil The LO Freq in the ascii file is used to identify which phase cal tone belongs in which IF. I had assumed that the value given needed to be adjusted by 0.5 * bandwidth. This is not the case. Also if the data had < 4 Stokes parameters PCLOD assumed that there was only one polzn in the tables, this does not deal with the case of RR & LL only. Also provided a little more information on the # of entries read as the task proceeds. Moved nowhere. 8450. January 25, 1995 $YPGNOT/SNPLT Phil When plotting multiple IFs per page SNPLT refused to plot data from PC tables. Fixed it and also ensured the phase-cal tone being plotted was recorded in the accompanying text. Moved nowhere. 8451. January 25, 1995 KNTR Eric Changed PLTBEM to test for beam too large and to give messages. Changed KNTR to set default windows when doing only one image. Moved nowhere. 8452. January 26, 1995 MAXFIT Eric There was a way through AU9 that could generate the FIT FAILS message without setting ERROR. Stopped it - this was when the fit actually worked, but the peak was outside the search area by more than 1 pixel. Moved nowhere. 8453. January 27, 1995 $YPGNOT/POSSM.FOR Athol Change POSSM to append to the text file specified by OUTFILE, rather than rewriting the file. Setting SOLINT will now write a sequence of spectra to the output text file. Also corrected a minor error in POSSPL: pressing button D did not terminate plotting if a text file was written as the GFINIS return code was overwritten by ZTXCLS. Moved nowhere. 8454. January 27, 1995 SETPAR Eric Corrected it - the SP disk keeps the assigned users for only 15 disks, not the 35 that can occur. Changed SETPAR to limit the range of disks to 1 - MIN(15,NVOL), rather than 2 - NVOL. The first disk can be assigned too these days. Moved to the patch area 31 January. 8455. January 30, 1995 $YPGNOT/POSSM.FOR Athol Corrected a logic error in IF range selection when SOLINT <> 0. Also the scan center time and second antenna on each XC baseline were not written correctly to OUTFILE. Moved nowhere. 8456. January 30, 1995 $HLPFIL/UVGLU.HLP; $APGNOT/UVGLU.FOR Athol Modify UVGLU to allow input data files that do not contain exactly the same number of visibility points, perhaps as a result of independent gain calibration. In this case the data are assumed to be in time-baseline sort order. Moved nowhere. 8457. January 30, 1995 MK3IN Chris/Walter Alef Incorporated changes from Walter Alef to allow MK3IN to read experiments with up to 66 baselines. MK3IN now carries a mapping from extent numbers to baseline numbers (LXTNT) since the type 51 extents only store the baseline numbers modulo 64. Also removed the call to MK3ANT from MK3DAT and changed the reference date test to use the value of RDATE in the AN table. Since the AN table keywords are not updated by ANTINI, the calls to MK3ANT from MK3DAT could cause the wrong reference date to be used in the AN table if the data were not written onto tape in sequential order; MK3ANT should now only be called once per run (immediately after the correlator parameters are shown). The revised reference date test should get rid of spurious warnings when appending to an existing file. 8458. January 31, 1995 UVINIT ERic Believe it or not, found a significant bug. If the apparent read size was an integer * 512 bytes then the extra safety margin was set to 0. Then, if the buffer size did not allow this read size, the read size was reset, but not the safety margin. This led to wrong I/O! Changed it to reset the margin. Moved to the patch area this date. 8459. February 2, 1995 WFCLN, UVUTIL Eric Found a bug in UVUTIL (UVFRQS) - it ignored BIF and returned frequencies for IFs 1 through NIF. These were then applied to data of BIF and up leading to wrong scales for images made with BIF > 1. WFCLN was clipping out data outside some UMAX based on the reference frequency. In bandwidth synthesis or spectral-line imaging, this UMAX should be adjusted to the max frequency in each image. Since it was not, it was easily possible for WFCLN to attempt to grid samples well off the grid (and off the address space of the task). Corrected it to (1) scale the U/V in range tests properly to require all channels of a sample to fit inside the used portion of the data grid, (2) apply the usual defaults for GUARD taking into account the taper, and (3) count samples off the grid and the like to report the discarded samples and to suggest better cell sizes. Moved nowhere. 8460. February 2, 1995 FRING Chris Added a check on the AP memory requirement in FRNFIT. This should catch most instances where FRING would fail with SOLINT messages after running for some time. This check should really be moved to a point before the data selection and averaging but it appears that this would require major surgery. Moved nowhere. 8461. February 2, 1995 $APLNOT/NXTFLG.FOR Athol Modified NXTFLG to ignore flag table entries which cover an IF or spectral channel range incompatible with the header structure of the underlying u-v file. Flag entries which do overlap with the IF and channel range in the u-v file are adjusted and retained in the common block flag tables used by DATFLG. This error could cause memory overwrites in DATFLG with unpredictable effects. Moved nowhere. 8462. February 3, 1995 $APGNOT/LISTR.FOR Athol Revise correction #8448 to LISTR. Was not dealing correctly with multi-IF tables in all cases. Moved nowhere. 8463. February 3, 1995 $YPGNOT/POSSM.FOR Phil Trivial correction dealing with an obscure bug when labelling multi plot per page plots. Moved nowhere. 8464. February 3, 1995 UVUTIL, WFCLN Eric Corrected UVUTIL (U2IDES) to take into account the channel averaging which affects the number of frequencies in the output image and the reference and alternate reference pixels. Changed WFCLN to put the number of channels averaged into the UV data object and to take this into account in doing the loop over channels. Also fixed the default GUARD value - forgot the SQRT. Improved the ECHAN description too in the help file. Moved nowhere. 8465. February 6, 1995 MCREAT, UVCREA Eric Changed message format to add a digit for disk number. Moved nowhere. 8466. February 6, 1995 SCMAP Eric Add GUARD to SCMAP inputs. Changed SCMAP to put the Guard-band parameters and UMAX/VMAX into the UV data object for UV2SCR. Changed UVUTIL (UV2SCR) to do the guard-band tests on U and V and to limit the data passed to UMAX and VMAX correctly adjusted for frequency of the highest-frequency channel. Also changed UV2SCR to put the largest U in the UV file header as MAXBLINE. This saves a read of the whole file to get it later in UVPREP. Moved nowhere. 8467. February 7, 1995 UPDAT Pat Martin Shepherd reported this task did not work when it tried to read USERLIST.HLP in $HLPFIL; apparently it used ZTOPEN and ZTREAD to access the file, and ZCLOSE to close it. This caused it to die before it did any conversions. Changed to ZTCLOS. Moved nowhere. 8468. February 8, 1995 $QPGNOT/CVEL Phil Found a stupid screw-up in CVEL. When I did the changes a few months ago I was using an array called AVDELI to track the amount by which the BP entries were shifted so I could flag those channels. What I didn't do was change the original use of that array in $QNOT/DSHIFT. It was used to report the amount by which spectra were frequency shifted. Anyway, as that array was filled with values from DSHIFT more and more channels were flagged as CVEL went through the data until eventually all channels were flagged. The fix was trivial - use different arrays. Also put DSHIFT back into CVEL since it is the only task that calls it. Removed DSHIFT from $QNOT. Will move CVEL to the patch area for 15JAN95 this afternoon. 8469. February 8, 1995 PSEUDO Eric Found bugs when recompiling verbs and pseudoverbs as in RUN NEWPARMS. These bugs caused the wrong info to be put in the wrong place, converting NEWPARMS from a proc to a verb or pseudoverb. WE're lucky that's all it did! Moved to the patch area. 8470. February 13, 1995 $APGNOT/SNSMO Phil When using SMOTYPE='VLMB' the phases for each IF were replaced in the SN table with the wrong sign. The multi-band delay was OK. Thanks to Michael Bietenholz for pointing this out. Moved nowhere. 8471. February 13, 1995 &APGNOT/FILLM Gustaaf FILLM put the wrong reference channel in the header for BW codes 8 and 9. Reason was that FILLM assumed that the sum of the LO frequencies equals the frequency of the beginning of the band. For BW codes 8 and 9 there is a full bandwidth offset between these two frequencies. Must have been there all the time, but not spotted due to relatively rare use of these BW codes. Moved nowhere. 8472. February 14, 1995 $APGNOT/UVFLG Phil A task called TSM generates the Tsys and flagging files for data from the VLBA correlator. These files are in a form that can be read into AIPS with little modification. Previously though, for flag files, a user would have to change the TIMEOFF value, change the day of year in the timerange to 0-relative numbers and change the antenna names to numbers. A new version of TSM is coming out this week that modifies the format of the flag tables slightly. Now the user need make no change to the file UVFLG will read and understand the commands as they are written. UVFLG will also still deal with the old format ascii files. UVFLG.HLP also updated. Moved nowhere. 8473. February 14, 1995 SDGRD Eric SDGRD failed to place the Stokes adverb in the appropriate common for the data reading routines, so it was ignored. Changed the strings for RR and LL too (in help and code). Moved to the patch area. 8474. February 16, 1995 $APGNOT/UVFIX Gustaaf Added the long overdue leapsecond information of 1 July 1995 to UVFIX. Moved nowhere 8475. February 16, 1995 $APGNOT/UVCOP Leonia When UVCOP runs with selection criteria (timerang, sources list antennas/baseline) it used these criteria for selecting data and sources in SU table. CL table has been copyed not paying attention on the selection criteria. The new version of the task copyes the part of the CL table in accordance with the selection criteria. To make it I added some lines in COPYIN to determine the list of selected antennas from ANTENNAS/BASELINE data, attached CLSEL to the text of the task and made germane change in it. I have excluded subroutine COPRST which copied the tables if all IFs need to copy. A subroutine COPTAB intended for a copy of a portion of IFs, can serve in that particular case as well. I have modified HELP file to ilucidate the relation of ANTENNAS/BASELINE in the case of auto correlation. Moved nowhere 8476. February 17, 1995 $APGNOT/VBGLU Phil The frequency scaling of u,v.w was calculated with the expression old freq / new freq instead of the inverse. Fixed. Moved nowhere. 8477. February 17, 1995 $HLPFIL/POPSDAT.HLP $RUNSYS/NEWPARMS.001 Athol Add two new scalar adverbs, TYVER and GCVER, for the system temperature and gain curve tables. Moved nowhere. 8478. February 21, 1995 $YPGNOT/CLPLT.FOR Leonia The task has not worked when averaging the frequency channels. The reading of the first portion of data by UVGET was skipped in subroutine CLOSAV. As a result, the frequency averaging procedure (AVGCIF) fell in confusion. Additionally, the default value of STOKES was 'HALF' instead of correct 'I '. So the task has not worked with the default value of STOKES. Fixed. Moved nowhere. 8479. February 21, 1995 $APGNOT/MK3IN.FOR Athol Increased parameter MXFQID from 3 to 24 to accomodate datasets with a large amount of frequency switching (precipitated by difficulties in loading data from a pre-SVLBI maser survey). Also moved the call to MK3FQ, which writes the FQ table, to MK3TB to deal with the case where the FQ-ID's are not all contained in the first set of header extents. Moved nowhere. 8480. February 21, 1995 $HLPFIL/POPSDAT.HLP, $RUNSYS/NEWPARMS.001 Phil Added a new adverb, DIGICOR, to control the application of the digital corrections in FITLD. Moved nowhere. 8481. February 21, 1995 $APGNOT/UVFLG Phil A user had a VLBA ascii flag file with 3000 entries. One of them had a start time > end time. UVFLG told him this but gave no clue as to where in the file this might lie. Fixed to print out the offending times. Also ensured that in the new format any lower case antenna names are converted to upper case when read in. Moved nowhere. 8482. February 21, 1995 $QPGNOT/BPASS Phil BPASS specifically did not normalize BP tables generated from VLBA data. The reason I did that no longer holds so changed the code to ensure that BPASS does so if requested and prints a message that it is doing so. Moved nowhere. 8483. February 21, 1995 $APGNOT/LISTR.FOR Athol OPTYPE='GAIN' was working correctly for STOKES=' ' but not STOKES='HALF'. Fixed this. Also corrected a long-standing error for OPTYPE='GAIN'; STOKES='L'; this did not work if the gain table contained data for both polarizations. In this case RCP gains were printed in error. Moved nowhere. 8484. February 22, 1995 $RUNSYS/VLAPROCS.001 Dave The default for SNVER was 1 in VLACALIB; I changed it to zero. Moved nowhere. 8485. February 22, 1995 $HLPFIL/DIGICOR.HLP Phil As Eric pointed out I had forgotten the helpfile for the new adverb DIGICOR. It is now written. Moved nowhere. 8486. February 22, 1995 $APGNOT/TBOUT.FOR Dave Michael Rupen pointed out the following: in the header of the disk file created by TBOUT, I*4 columns were quoted as a format of I12; however, I11 was what was actually used for the actual row output and column calculation. Since 11 is used several places in the code, I just changed the string in the header (FORM(4)) to read I11; also ITLEN(4)=11. Moved nowhere. 8487. February 23, 1995 $HLPFIL/TYVER.HLP $HLPFIL/GCVER.HLP Athol Help files for new adverbs TYVER and GCVER (see #8477). Moved nowhere. 8488. February 23, 1995 $APLSUB/TRIM.FOR Dave Another bug from Michael Rupen: TBIN was trashing everything past column 132. It was due to the temporary storage array in TRIM being only 132 characters; I upped it to 1024 and recompiled TBIN. Moved nowhere. 8489. February 23--, 1995 TV Object, revised imaging Eric Changes for TV Object and revised imaging: OLDMAP New: the old version of MAKMAP HORUS Changed to call OLDMAP. UVGET Corrected faulty logic setting EIF defaults. DMPR.INC Changed to add Robust, weight power, weight function. OBJMNGR Changed OBVHKW to check whether a keyword was already defined rather than just automatically appending it to a rather finite (50) list. POPSDAT Help file changed to add ROBUST, UVBXFN, UVSIZE. NEWPARMS Run file changed to add ROBUST, UVBXFN, UVSIZE. AU6 Changed to call TVZOME, TVPSUD, TVFLAM TVZOME New: interactive TV zoom from AU6 TVPSUD New: interactive TV pseudo colors from AU6 TVFLAM New: interactive TV flame colors from AU6 Moved nowhere. 8490. February 23, 1995 PLTBEM, KNTR, PCNTR Eric Changed KNTR to take into account offsets of one image from the other. Changed PLTBEM to offer other modes filling in the beam some. Changed PCNTR and KNTR and their helps to support these modes. Moved nowhere, 8491. February 24, 1995 $APGNOT/SPLIT Phil SPLIT will not now copy the (sometimes rather large) HF table to the single-source output file. 8492. February 24, 1995 $APGNOT/SETJY Phil Ensured that the alternate refernce pixel is set to the true reference pixel as a default when velocities are being set. This was causing some confusion for ATNF data. Moved nowhere. 8493. February 24, 1995 New adverb: WTTHRESH Phil Added another new adverb to POPSDAT.HLP and NEWPARMS.001. Also wrote a helpfile. WTHTHRESH will be used in FITLD to generate an FG table based on the input data weights. Moved nowhere. 8494. February 25, 1995 TV Eric changed IMCHAR Support all 4 angles IMANOT Support all 4 angles GRBOXS To make the number of boxes an output variable up to a given max. AU5C To call the revised version of GRBOXS - simpler than before. TVBOX Changed help file. REBOX Changed help file. Moved nowhere. 8495. February 27, 1995 $APGNOT/FILLM Gustaaf Realtime FILLM on the IBMs failed when NFILES > 0. Strictly speaking this is pilot error, but difficult to track down when it occurs. So I bypassed the file skipping when FILLM is used in its online mode. Strangely enough, NFILES > 0 worked on Suns. 8496. February 27, 1995 BLING Chris Moved call to UVFRQS in SETFRQ to get rid of benign but annoying error messages about BIF/EIF and removed superfluous calls from FIX0, FIX1, FIX2 and FIX3. Moved nowhere. 8947. February 27, 1995 Name conflict Eric A name conflict between OOP and TV has arisen. Changed: COLORF New - was IMFCLR COLORH New - was IMPCLR COLORC New - was IMCCLR the actual conflict COLORL New - was IMLCLR TVMOVI to call COLORC TVFIDL to call COLOR TVFLAM to call COLOR TVPSUD to call COLOL, COLORH, COLORC Moved nowhere. 8948. February 28, 1995 CURVAL, TVBLIN Eric Changed TVBLINK to do all of the blinking operation including the enhancements. Moved it from $AIPSUB to $YSUB. Created $YSUB/TVALUE.FOR to do CURVALUE. Changed AU6A and AU6B to call these subroutines. Now these functions can go in the TV object. Moved nowhere. 8499. February 28, 1995 SETJY Chris Widened the format used to print the alternate reference pixel to cope with channel numbers in the thousands. Moved nowhere. 8500. February 28, 1995 $QPGOOP/SCMAP.FOR Bill C. Fixed bug which occurred when the data contained multiple subarrays. The self calibration was putting the solutions in separate SN tables and then only calibrating the final subarray. Moved nowhere. 8501. February 28, 1995 $APLOOP/UVUTIL.FOR Bill C. SLFCAL, UVFRSC now only force a new SN table if the subarray number is 1. Moved nowhere. 8502. March 1, 1995 $APLCVEX/ZTACT2, ZTKILL Pat Created convex-specific versions of these routines. The problem arises in the decoding of the process id from the lock file, e.g. /tmp/FUBAR1.12345 -> 12345. This is achieved via a scanf with a %d in the format string in the $APLUNIX versions, but pid_t is a short on ConvexOS, and an int just about everywhere else. This resulted in antisocial behaviour such as AIPS deleting the lock file in /tmp before the task was finished. Moved nowhere. 8503. March 1, 1995 $APGNOT/FITLD.FOR Phil A major release of FITLD. I have added the routines written by Leonid Kogan to perform the digital corrections necessary for data from the VLBA correlator. The user has the choice of whether to apply these corrections or not through the use of the adverb DIGICOR. DIGICOR <= 0 will not apply the corrections, DIGICOR = 1 will apply them to all data, DIGICOR = 2 will apply the corrections to all cross-power data and only to the total-power if zero-padding was used in the correlator. 2 other minor problems were fixed: a format statement in BTRTAB was wrong, thanks to Bill C for pointing this out; the LO-offsets can only be set at the antennas to increments of 10kHz, I adjusted the DOPOFF parameter in the IM/CL tables to reflect this. This fixed a bug that results in these numbers when the correlator writes > 256 channels per IF. Also updated the helpfile. Moved nowhere. 8504. March 2, 1995 $APGNOT/SNSMO.FOR Athol Corrected minor bug in SNSMO which disabled clipping of rates in most circumstances as a result of not checking the rate threshold in the correct units. Moved nowhere. 8505. March 2, 1995 New subroutine: $APLNOT/MBDREF Phil Routine to rereference the multi-band delay column of an SN table. This is cosmetic really because the MBD is stored as a phase slope across IFs, but users have noticed that this column is not referefenced when a ref. antenna changes and it worries them. Got to keep the users happy. Moved nowhere. 8506. March 2, 1995 $APLNOT/CALADJ $QPGNOT/FRING Phil Modified to call MBDREF where appropriate. Also stored BCHAN & ECHAN in the history for FRING. Moved nowhere. 8507. March 2, 1995 SPLIT Chris Removed a booby-trap which would prevent the reference frequency being set correctly if (i) channels were to be averaged, (ii) a subrange of channels was selected and (iii) BCHAN and ECHAN did not include the entire range of channel numbers. The culprit was a test for inclusion of all channels after UVGET was called; UVGET modifies the header block to reflect the number of channels it will return after BCHAN/ECHAN selection instead of the original number of channels. Moved nowhere. 8508. March 1, 1995 task help Eric Changed TSKHLP to take a task name argument. This lets all Cleans call it CLEAN and lets there be more than one menu per task. Changed TVFLG, SPFLG, IBLED, TVRGB, TVHUI for call seq change. Moved nowhere. 8509. March 3, 1995 TVBOX, REBOX Eric Changed GRBOXS to do circular as well as rectangular boxes. Added to $YSUB two support subroutines call LMBOXS and DRBOXS to limit the cursor position and to draw the boxes. Changed help files HLPCLEAN, TVBOX, REBOX as well. Moced nowhere. 8510. March 6, 1995 FITLD Phil Added the DIGICOR information to the HI file. Also corrected a mildly amusing typo in one of the messages printed while running. Moved nowhere. 8511. March 6, 1995 CVEL Phil Found the potentially serious bug that was causing the spectra out of CVEL to be subtly different from those out of POSSM. The address of the start of the spectrum in a large buffer was off by 2 words and so the time and antenna number were being passed into the shifting routines as though they were data. This will need a patch. Moved nowhere. 8512. March 7, 1995 CVEL Phil I have been having persistant problems with the sorting of the CL table. Finally sorted them out by ensuring the sort was done early on in the task execution. Moved nowhere. 8513. March 7, 1995 $APLNOT/KEYIN.FOR Athol Mode=1 in KEYIN is meant to echo KEYIN input, which is useful in debugging large input KEYIN files such as are used in VLBI amplitude calibration. Enabled reflection for mode=1, and also introduced a new mode=4 which is simply mode=3 with relection. Only the first 64 characters of each KEYIN line are printed if reflection is activated. Moved nowhere. 8514. March 7, 1995 $APGNOT/ACCOR.FOR Leonia Calibration of a cross correlation spectrum supposes knowing parameters of samplers (digitizers in particular). Unfortunately these parameters are not identical to optimal ones and are not known with enough accuracy. As a result the cross correlation spectrum amplitudes are measured with errors. These errors can be estimated having measured the mean value of auto correlation spectrum of a given interferometer's antennas. The new task ACCOR provides correction using this algorithm. The task estimates the mean values of the auto correlation spectra of all antennas of a given VLBI network and records them in SN table. The following application of the correction is provided by standard procedure of applying the SN table to an CL table (CLCAL) and following application this table with DOCALIB = 1 in relevant tasks. Moved nowhere. 8515. March 8, 1995 $APGNOT/FILLM Gustaaf FILLM greatly exaggerated the number of shadowed visibilities. Reason of this was that the test for shadowing antennas took place *before* any other test of eligibility; consequently many visibilities which wouldn't have made the output data set were reported as shadowed. The fix was simply to count only those shadowed visibilities which met the other criteria as well. 8516. March 8, 1995 $HLPFIL/ACCOR.HLP Leonia I has written ACCOR.HLP with lines longer than 72 characters. So it was difficult legible. I've fixed it. Moved nowhere. 8517. March 8, 1995 $APGOOP/MBDLY.FOR Ed F Made lots of changes to improve MBDLY, especially for use with the Calc/Solve programs. The major changes are: Added TIMERANG adverb to limit MBDLY range of execution. Added BIF and EIF adverbs to pick out only those IF's associated with one frequency. This is needed for dual frequency (S/X) data bases. More careful checking of bad fringe fits various scatter rms's amongst the IF's. Default values improved. Moved nowhere. 8518. March 8, 1995 $HLPFIL/MBDLY.HLP Ed F Improved help file in conjuction with code in previous entry. Added TIMERANG, BIF and EIF adverbs. Improved the help and explain files. Moved nowhere. 8519. March 8, 1995 $APGOOP/CL2HF.FOR Ed F Made lots of changes to improve CL2HF. The major changes are: Fix geometric calculations for converting earth center to station center (Haystack style) rates and delays. Took out INFILE adverb for station names and put selection in code. Include TIMERANG adverb for selection. Fixed several other minor errors in converting from CL to HF format. Source observations below 0.5 degree elevation elliminated. Moved nowhere. 8520. March 8, 1995 $HLPFIL/CL2HF.FOR Ed F Improved help file in conjuction with code in previous entry. Removed INFILE adverb and added TIMERANG adverb. Improved help and explain file. Move nowhere. 8521. March 8, 1995 $HLPFIL/POPSDAT.HLP Athol Add new task-specific adverbs to POPSDAT.HLP, namely TREC(30), TAU0(30) and DOFIT(30), which are used for antenna-specific receiver temperatures, zenith opacities and control of the opacity fitting respectively. Moved nowhere. 8522. March 8, 1995 $APGOOP/HF2SV.FOR Ed F Made lots of changes to improve HF2SV. Added OUTFILE adverb to specify directory of the binary output from HF2SV. Fixed a few small bugs in program. Moved nowhere. 8523. March 8, 1995 $HLPFIL/HF2SV.HLP Ed F Added OUTFILE adverb for location of output binary data. Moved nowhere. 8524. March 8, 1995 $APGOOP/HFPRT.FOR Ed F New task HFPRT prints the contents of the HF extension file. Moved nowhere. 8525. March 8, 1995 $HLPFIL/HFPRT.HLP Ed F Help file for new task HFPRT which prints the contents of the HF table. Moved nowhere. 8526. March 9, 1995 $APLGEN/ZWHOMI.FOR Pat Changed the assignment of graphics device number for the REMOTE category (i.e. where you're at a graphics terminal). Instead of the arbitrary numbers 241-255 (240 + POPS#), use the first 15 devices after the last TV device. Assuming AIPS managers set up their SP files correctly, this should not be a problem. The new rule will be #tek devices = 35 + #tv devices. Also changed $SYSUNIX files TVDEVS.SH, AIPSEXEC and START_AIPS. Moved START_AIPS to /AIPS locally. 8527. March 13, 1995 $HLPFIL/TABED.HLP Leonia It was not clear from the HELP that OUTVERS has to be equal INVERS for header functions (OPTYPE='key'). This statement has been clarified. The erronious statement: "Input file is ignored for OPTYPE = 'key' has been changed for correct: "Output file .." Moved nowhere 8528. March 13, 1995 UVFIT Chris At some time (probably when antenna gain fitting was added) the maximum number of parameters to be fit without was changed from 24 to 52 without changing the sizes of some internal arrays nor the workspace of DVDMIN. This caused anything from infinite loops to nonsensical fits. Changed the internal arrays to be dimensioned by XMXPRM and increased the array sizes in DVDMIN to cope with up to 64 parameters. Also added a frightening error message to DVDMIN in case its limits are exceeded. Moved nowhere. 8529. March 14, 1995 $APGNOT/ACCOR.FOR Leonia The logic of determining whether the disired SN table had existed or not was wrong. As a result there was an annoying warnings in some cases. Additionnaly the task could not work in the very rare case of zero auto correlation spectrum amplitude and non zero corresponded weights. Moved nowhere 8530. March 15, 1995 $YPGNOT/UVPLT Craig/Phil Craig passed along some changes that enable UVPLT to self-scale each axis seperately and automatically when requested. He also tidied up the self-scaling algorithm. Moved nowhere. 8531. March 15, 1995 Gain curve routines Athol Added or updated several routines and include files for the new gain curve (GC) table format. These include: $INC/PGCV.INC - parameter include containing column pointers for GC table. $INC/DGCV.INC - declaration of row variables for reading or writing to the GC table. $APLNOT/GCINI.FOR - initialise a GC table. $APLNOT/TABGC.FOR - read/write a GC table row. $APLNOT/GCKEY.FOR - read/write the GC table keywords. Moved nowhere. 8532. March 15, 1995 $APGNOT/ANTAB.FOR $HLPFIL/ANTAB.HLP Athol New task which, in conjunction with a forthcoming task APCAL, will constitute an alternative to ANCAL. This task (ANTAB) reads an external amplitude calibration text file and updates the TY, GC and BL tables accordingly. Numerous new features have been added to allow greater flexibility in the allowed format of the calibration file. This includes the ability to specify an arbitrary many-to-many mapping of Tsys/Tant column number to IF/polzn, the ability to have IF, polzn. and freq. specific gain curves, the possibility of entering tabulated gain curves and the ability to run the task multiple times with the same tables without difficulty. Moved nowhere. 8533. March 16, 1995 $APGNOT/UVFLG.FOR Leonia/PHIL The new parameter DTIMRANG has been added in INFILE giving opportunity to expand a TIMERANG interval in both side by DTIMRANG seconds. The relevant adding to HELP has been done. An error has been fixed to ensure flagging is applied to all source qualifiers if no qualifier is specified. Moved nowhere. 8534. March 16, 1995 $APLNOT/TABGC.FOR Athol Dec Alpha required that the order of two declaration statements be reversed before it would compile TABGC correctly. Moved nowhere. 8535. March 17, 1995 $YPGNOT/POSSM.FOR Leonia If reverse of X-axis has been desired but labeling was in channels POSSM labeled the axis finishing by zero donot paiing attention on BCHAN and ECHAN. I fixed it. Moved nowhere 8536. March 17, 1995 $APGNOT/SPLIT.FOR Phil When multiple sources are selected and BIF > 1 then after the first source BIF was reset to 1 in subroutine SPLCOP. Fixed. Moved nowhere. 8537. March 20, 1995 $APGNOT/ANCAL.FOR Athol Allow multiple runs of ANCAL if necessary. Also check EIF against the catalog header and set accordingly if not specified. Moved nowhere. 8538. March 20, 1995 CALIB Chris The DOPTMD flag is now set only if IN2NAME and IN2CLASS are both blank. The previous behavious was at variance with the help file and bypassed some consistency checks on the adverbs. Moved nowhere. 8539. March 21, 1995 TABCOP Eric Changed to suppress the copied message if 31990 <= MSGSUP < 32000. Otherwise the message is sent with level 3 to MSGWRT where other MSGSUP rules apply. Moved nowhere. 8540. March 21, 1995 POSSM Chris Subroutine POSSPL was using OPTICL, ALTRFP and REFVEL without setting them. This should have had some observable consequences when plotting against velocity. Modified LABLAX to return these values. Move nowhere. 8541. March 21, 1995 $RUNSYS/NEWPARMS.001 Athol Add TREC(30), TAU0(30) and DOFIT(30) to NEWPARMS. Moved nowhere. 8542. March 22, 1995 TV object Eric Created a TV class and added it to the system. Changed: HLPCLEAN $HLPFIL/ New Clean object TV interactive help file. TVDEVICE $APLOOP/ TV device class does class functions and basic TV functions (by calling $YSUB routines) TVUTIL $APLOOP/ utility functions of simple call sequence to do TV operations GENERIC $APLOOP/ added TV class knowledge QCLEAN $QOOP/ Add CLNTV and CLBCHK routines to do TV. TVMENU $YSUB/ new to display menu and return selection TVTPOS $YSUB/ read TV cursor with prompt & timeout options PLAY $APGOOP/ New task PLAY $HLPFIL/ Help for new task HLPPLAY $HLPFIL/ Interactive help for new task Moved nowhere. 8543. March 22, 1995 IMAGR Eric New task intended to replace all older imaging tasks. It is WFCLN with all of the calibration adverbs exposed. Changed QUVUTIL Set new adverbs into DMPR common. UVUTIL Changed UV2SCR to suppress table copied messages and added Bill's CP2SCR which copies uvdata to a scratch file looping over subarray. QWTGET New: returns a row of weight summing array w robust summing. QWTGRD New: convolves weights to summing grid, counts samples to robust count grid. QWTSUM New: does summing needed by robust option. ROBUST New: Help file for new adverb - uniform weight tempering parameter. UVBOX Help file updated for extended UVBOX meaning. UVBXFN New: Help file for new adverb - box counting function and shape. UVSIZE New: Help file for uniform weighting array size. UVWTFN Help file revised to show the options clearly. UVWAIT New: do uniform and other weightings. WTFUNC New: compute weight counting functions and put them into the AP. GRDFIT New: determine whether the proposed weighting and gridding operations will fit into the AP memory with and without sorting. MAKMAP Changed to call GRIDUV GRIDUV $QNOT new: Does only one pass on data. Moved nowhere. 8544. March 22, 1995 PCAL Chris Incorporated a fix from Martin Shepherd that stops PCAL from overwriting sections of the CATBLK if there is no IF axis present in the input data file. Moved nowhere. 8545. March 23, 1995 IMAGR Eric Changed QCLEAN to display the dirty image (allowing you to set windows at that stage) on the TV. Moved nowhere. 8546. March 23, 1995 KNTR Eric Changed KNTR to allow you to control whether the edge between blanked and good regions is plotted or not. Changed KNTR.HLP, POPSDAT.HLP, NEWPARMS.001 and created DOBLANK.HLP to give this control. DOBLANK might replace other adverbs used for similar purposes some day. Moved nowhere. 8545. March 23, 1995 MBDLY Pat Remove references to external BLAS routines ISAMAX and SMIN as these are not used. Besides, the Dec Alpha compiler for OSF/1 1.2 does not contain the BLAS library. Moved nowhere. 8546. March 23, 1995 POSSM Chris Changed OUTFILE writing to take possible reversals of the AMP and PHASE arrays (caused by APARM(10) = 1) into account. Channel data is now written out in channel/IF order. Previously channel and velocity labels were written in channel/IF order but the amplitudes and phases were reversed. Moved nowhere. 8547. March 24, 1995 $APLNOT/GCTAB.FOR and $APLNOT/GCIO.FOR Athol Removed these two routines which have been replaced by TABGC and GCKEY respectively (see #8531). Moved nowhere. 8548. March 24, 1995 DBCON Leonia Having different header frequencies in concataneted files the task calculated wrong frequencies for a source not existed in the first file. In this case the task took the header frequency from the first file but frequency entries for SU table from the second file. As a result the frequency for such a source was wrong. I added the difference between the header frequencies to the frequency entries for SU table from the second file, if the source exists in the second file but does not in the first one. Moved nowhere. 8549. March 27, 1995 $APLNOT/BPFSEL Phil As BPFSEL copied BP tables it would replace the imaginary column by the real column, obviously due to a typo. Fixed. Moved nowhere. 8550. March 27, 1995 $QPGNOT/CVEL Phil For some unknown reason the copy of CVEL at the AOC was not the same as the copy on baboon. The MNJ must have failed. Fixed up a minor problem with a buffer size and will watch the MNJ carefully tomorrow. Moved nowhere. 8551. March 27, 1995 $APGOOP/HFPRT Pat This task was not compiling under Linux, as there was mixing of integer*2 and integer in calling sequences (the int*2 are necessary for the HF table format, I think). Changed key variables to regular INTEGER, uppercased code, split common into character and numeric commons, shortened variable names to at most 6 characters, and removed a bogus variable from a format statement. Moved nowhere. 8552. March 28, 1995 Gain curve routines Athol Added or updated several new gain curve (GC) table routines used in amplitude calibration tasks such as APCAL. The new routines are: $APLNOT/GETGC.FOR - read all gain curve information for a specific antenna,suba. and fqid. Analogous to GETSOU/GETANT etc. $APLNOT/GCVAL.FOR - compute gain value at a specific sky position after calling GETGC. Analogous to SOUELV. $APLNOT/DGNRAO.FOR - compute Legendre-type polynomials cutomarily used to define the GB 140ft gain curve. $APLNOT/INTERP.FOR - unweighted 1-d linear interpolation routine used for interpolating tabluated gains and elsewhere in APCAL. Updated code: $INC/DGCV.INC - include file used by GETGC and GCVAL corrected for faulty subscripting. Moved nowhere. 8553. March 29, 1995 $APLNOT/GETGC & GCVAL $INC/DGCV.INC Athol Temporarily removed these routines. Large declared buffers to accomodate GC records were causing shared library problems. Moved nowhere. 8554. March 31, 1995 $SYSUNIX/LIBR, $UPDUNIX/UPDREMOVE Pat (Actually most of the work was done by Athol and Martin Shepherd). When I changed the shared library location to $LIBR/.so instead of $LIBR//SUBLIB.so, I updated the code in the LIBR shell script but not the UPDREMOVE one. This caused the .so's at the Midnight Job sites using shared libraries to become out of step with reality. Also, under SunOS, use of the -d qualifier to the ld command is apparently necessary to avoid an unusual failure of ld.so to map large common blocks at runtime. Moved to the AOC (aguila, laphroaig, pilabo, miranda) this date and last night. 8555. March 31, 1995 $QPGNOT/PCAL, $APGNOT/SPLIT Pat Thanks to Colin Lonsdale for finding these bugs. Apparently they both declared a hollerith scalar HSOLT but then proceeded to treat it as an 8-byte quantity. Modified both declarations so that it's HSOLT(2) instead; Colin reports this cured the problem (but not his shaky faith in AIPS :-) Moved nowhere. 8556. March 31, 1995 IMAGR Eric Changed: GETCTL $APLNOT: to put out I polarization.. message iff MSGSUP < 31900, but return the messge in MSGTXT in any case. ALGSUB $QNOT: corrected computation of how many rows of grid are needed for bandwidth synthesis, corrected AP memory usage to allow for more when the # vis limit kicks in, added a message re polarization and channels suppressing the message from GETCTL. This corrected a bug that let channels not be subtracted! UVMSUB $QNOT: commented out the gridded-interpolation model message. ALGSTB $QNOT: reduced the AP core used for vis's to 20% and added a message giving pol and channels, suppressing the one from GETCTL. GRDFIT $QSUB: added tests on size needed for gridded subtraction. IMAGR $QPGOOP: changed call seq to GRDFIT passing the subtraction method. Moved nowhere. 8557. April 3, 1995 $SYSUNIX/TVDEVS.SH Pat Minor modification enabling the user to use the buffered I/O feature of XAS. If you define the environment variable AIPS_TV_BUFFERED to be YES (uppercase), then the TVDEVxx variable will in turn be defined automagically with SSSINB instead of SSSIN, thereby triggering appropriate code in XAS. This should help with slow network links such as a 14.4k modem. Moved nowhere. 8558. April 3, 1995 TV change Eric Giving up only 4 more colors to the graphics planes it is possible to have 8 graphics channels rather than 4. This should make for some interesting further graphics work. Changed GRCH limits in a lot of help files and changed code to think of channel NGRAPH rather than 4 as the background channel. Changed $HLPFIL CNTR DFTPL GAL GAPLT GNPLT GRCHAN GRCLEAR GREYS GROFF GRON IMEAN IMVIM IRING ISPEC KNTR OFFSCROL PCNTR PFPL1 PFPL2 PFPL3 PLCUB PLROW POPSDAT POSSM PROFL SL2PL SNPLT TVSTAR TAPLT TVANOT TVPL TVSCROL UVHGM UVPLT VBPLT FRPLT TVLINE TVDIC GREAD GWRITE TVCPS CLPLT CCNTR CONPL VPLOT PLAY XAS $AIPSUB AU5 AU5B AU5E $YSUB IAXIS1 IMANOT ITICS TVALUE $YPGNOT TVCPS - new coloring scheme $YSERV XAS.SHR Moved nowhere. 8559. April 4, 1995 $QYPGNOT/MX.FOR Athol Warn if AC data are being mapped; these can be de-selected with UVRANG if necessary. Max. U-baseline was scaled twice for the frequency offset, both in MXSEL and the gridding subroutines; corrected this. Also fixed minor bug in checking BMAJ before deciding whether a residual map was required; this would fail for BMAJ in the range -10mas to zero - too coarse for VLBI. Moved nowhere. 8560. April 4, 1995 IMAGR Eric Changed IMAGR to get the correct numerical size of the beam from the BEAM object before deciding on whether or not to sort. Corrected GRDFIT to use the correct ULIM in cells for the image (beam or clean) in its second test rather than in cells for the weighting. Changed IMAGR to insert the zero-spacing in the file copy so that it applies in the weighting as well as the gridding and dropped it from the CLEANO objects so that it does not go into the images. Changed IMAGR help file to explain the new ZEROSP adverb meanings. Changed GRIDUV to check the sort order and if X sort is required to die if the sort is not done. Moved nowhere. 8561. April 4, 1995 $APGNOT/UVCOP.FOR $HLPFIL/UVCOP.HLP Athol Added facility to allow clipping data with weights below a certain threshold. This limit is specified by UVCOPPRM(5) if flagging data with bad weights is enabled. Moved nowhere. 8562. April 5, 1995 $SYSUNIX/EHEX.PL Jeff Perl front end to $SYSPERL/ehex.pl "library" for dec-to-ehex (and ehex-to-dec) conversions. Itty-bitty standalone utility. Moved nowhere. 8563. April 5, 1995 $SYSUNIX/DADEVS.PL Jeff Perl version of $SYSUNIX/DADEVS.SH (much faster than the shell version). Not yet implemented, but will be soon on systems that have Perl versions >= 4.036 available (which will be auto-detected). Moved nowhere. 8564. April 6, 1995 SAD Eric Changed code and help file to: (1) Reject only some components in a multiple distinct component island rather than rejecting them all when any one is deemed inacceptable. If it attempts to fit a complex area with two components and 1 of those is rejected it will revert to the single-componet (poor) fit. (2) Add option to constrain widths to > Clean beam min and to encourage widths >= Clean beam max. Moved nowhere. 8565. April 6, 1995 TVANOT Eric Changed AU5B to handle graphics channel assignment better. GRCH = 8 means letter in 8, not letter and background in 8. Moved nowhere. 8566. April 6, 1995 FITLD Phil I added the code necessary to deal with FREQIDs that are actually the same but that get renumbered at correlator job boundaries. FITLD will now tell users about FREQIDs as it reads the data in. Removed subroutine CA2CL which is not used. Moved nowhere. 8567. April 7, 1995 $YPGNOT/FRMAP Leonia I installed the default value for TIMERANG, which is equal to the time range of the data itself. The help file is updated and improved. Moved nowhere. 8568. April 7, 1995 $HLPFIL/TVCPS.HLP Phil The upper limit for the value of GRCHAN in the helpfile was too low following Eric's changes to the # of graphics channels. Moved nowhere. 8569. April 9, 1995 $SYSLOCAL/NOSHARE.LIS Athol Add FRING to NOSHARE.LIS. Moved nowhere. 8570. April 10, 1995 $QPGNOT/CVEL.FOR Leonia A floating point error occured in ALPHA (CfA). It happened because upper limit of one of the loop on antennas MAXANT is much bigger than number of antennas in data. The calculation for the unexisting antennas leaded to the floating point error. I change the upper limit from MAXANT to NSTNS-number of antennas in data which is calculated in GETANT. Moved nowhere. 8571. April 10, 1995 TVCPS Chris Replaced faulty code for colouring the graphics planes with code to XOR the graphics planes together. I assume, in the absence of any comments explaining the intent of the code, that this is what was intended. The previous code had the end effect of setting the colours of all the graphics planes to black and potentially accessed an array beyond its declared bounds. THIS IS WRONG - it does not work that way anymore - I WILL DO SOMETHING ELSE. (ERIC) Moved nowhere. 8572. April 10, 1995 SAD Eric Changed it to scale the width constraint by DPARM(9) rather than 0.1 which is probably too small. 0 => 0.3 and the limits are 0.01 to 1.0. Moved nowhere. 8573. April 11, 1995 TVCPS et al Eric Changed TVCPS by fixing my version. The colors of the graphics channels are Xor'd for channels 1-4 only. Channels 5,6,7,8 appear only if no lower numbered one is on. Also made Phil's fix to all affected help files: GRCLEAR, GROFF, GRON, OFFSCROL, TVANOT, TVDIC, TVLINE, and TVSCROL. Moved nowhere. 8574. April 11, 1995 IMAGR Eric Changed $QNOT UVSORT to respond to errors from SRTIN. Changed SRTIN to use non-excluive ZOPENs if the two files being opened are the same. Apparently, the file locking used at the AOC (flock) does not allow the same process exclusive use of the same file twice while the network-wise fcntl has other superior wisdoms and does allow this operation. Moved nowhere. 8575. April 11, 1995 START_AIPS, AIPSPATH, RUN, etc Pat Inserted support for DADEVS.PL (the perl equivalent of DADEVS.SH) in the various shell scripts. First, AIPSPATH.SH/CSH now check the path specified to perl in the first line of $SYSUNIX/DADEVS.PL and see if it's really there (default /usr/local/bin/perl; I'll have INSTEP1 set this). If so, this sets an environment variable that START_AIPS, RUN, and START_TPSERVERS all know about. Each of these in turn will check if that environment variable is set, and if so, run DADEVS.PL (and then "dot" and delete the temporary output file it produces); otherwise they just "dot" the DADEVS.SH script. The improvement in speed of DADEVS.PL over DADEVS.SH cannot be understated; it's impressive (I didn't do DADEVS.PL, Jeff did). Moved relevant scripts to $AIPS_ROOT in CV today. 8576. April 11, 1995 TV Eric Changed IENHNS to use graphics channel NGRAPH-1 rather than 1. Changed TVFLG and SPFLG to know this and take less defensive action than before (usually). Also removed unused variables. Changed TVDEVICE for IENHNS and to support up to 8 channels and graphics planes. Changed TVUTIL to handle 8 also. Changed QCLEAN to turn off all graphics channels initially. Moved nowhere. 8577. April 11, 1995 FITLD Phil My fix #8566 did not deal with all of parameter space. I fixed the immediate problem I was dealing with but then broke most others since I did not update the FQ table with the new FQIDs (if necessary of course) at job boundaries. This resulted in disconnected data within the file. This is now fixed and should now work for all data - famous last words. Moved nowhere. 8578. April 11, 1995 *TESS Eric Changed DTCIO.INC to increase buffers to 16K words. Recompiled a shit load of subroutines in $APLNOT and $QNOT. Changed UTESS, LTESS, STESS, and IMERG to set the buffer size correctly - most were setting it to 4K words or 8K words, not XBUFSZ words. Relinked these and VTESS which did it correctly. Moved nowhere. 8579. April 11, 1995 $APGNOT/ANTAB.FOR Athol A few added features: allow DPFU = RCP,LCP as in ANCAL; print bad times in HH:MM.MM format to be compatible with calibration file format; improve consolidation of TY records for the case of Ta/Tsys entries. Moved nowhere. 8580. April 12, 1995 $SYSUNIX/DADEVS.PL Jeff Now catches signals and removes /tmp/DADEVS.$$ file (forcing DADEVS.SH script to run on some OS's). Returns exit status of 1 if signal caught. Calling scripts should probably now check its exit status and act accordingly, if the desired effect of a signal is to abort the startup under all OS's. Moved nowhere. 8581. April 12, 1995 $SYSUNIX/*.PL $SYSPERL/*.pl Jeff Minor fixups to work under Perl version 5. Added library function to be called by scripts to check for a "tested" version of Perl (currently only 4.036 and 5.001 are tested), with a non-fatal warning message displayed if an untested version of Perl is being used. All scripts modified to call this. Moved nowhere. 8582. April 12, 1995 FITLD Phil A rounding error in comparing the frequency from the catalogue header and the freq. keyword in the FR table was causing spurious freqids to be generated, especially with high frequncy observations. I used the knowledge that all frequencies to be encountered from the VLBA correlator are modulo 10kHz and so overcame this nasty little problem. Moved nowhere. 8583. April 13, 1995 $YPGNOT/FRMAP.FOR Leonia I have discovered several small bugs in the programm having tested it with the data having big value of preavrage interval. The rounding of time worked wrong when time's step is named equal to the preavrage interval of the data. The program does allow to have namber of solution interval more than 10. There was not a warning about this limit. The problem are fixed. Moved nowhere. 8584. April 13, 1995 $YPGNOT/APCAL.FOR $HLPFIL/APCAL.HLP Athol New task for amplitude calibration, primarily of VLB data. Takes as input the system temperature (TY) table and gain curve (GC) table produced by ANTAB and writes a SN table containing amplitude gain corrections. Can solve for atmopsheric opacity and plot the resulting solutions or use a pre-defined receiver temperature in making the opacity corrections. Opacity algorithms still under evaluation but placed in TST for wider access. Use plot type 1 for now; update AU8A.FOR shortly. Moved nowhere. 8585. April 13, 1995 FITTP et al Eric FITTP has been writing an extra header-like record in place of table data when there is no table data. Changed EXTREQ to skip records looking for the next table (or TABNAME old tables or SIMPLE a new data file) and count those skipped. Changed MLREOF to take the count as an input/output variable. Changed IMLOD, UVLOD, GSCAT and FITLD to handle this count and report it. Changed PRTTP to skip records, count, and report also. Moved nowhere. 8586. April 13, 1994 IMAGR, MX, WFCLN Eric All Clean tasks have the idea that, when the clean is within 10% of the max number of iterations, a major cycle ought to keep going. While, with small jobs this is reasonable, with CasA and 1000000 iterations this means that the last major cycle is about 100000 iterations. But, with CasA, all other major cycles are about 3000 iterations or less. This ain't right. Changed QCLEAN and MX to make the limit 0.5 of the last major cycle or of CLNLIM, whichever is smaller. This means that a cycle is extended no more than half of its expected length. Moved nowhere. 8587. April 13, 1995 VPLOT Leonia The task did not work with negative SOLINT. Inspite of the fact that only positive or zero values were recomended in HELP for SOLINT, the failure of the task when SOLINT is negative bothered users. Now both SOLINT=0 and SOLINT<0 correspond to abcense of time averaging. Help file is updated. Moved nowhere. 8588. April 17, 1995 OOP imaging Eric Preparing for SCMAP, corrected bugs and changed: UVUTIL (UV2SCR) to do coordinate rotation and correct the sort order for it. GRIDUV Changed to require the AP for no more than 2048 vis recs rather than simply using a quarter of the AP. This changed a limit, for example of 239 rows to 314 rows - enough to fit a 512x1024 in this case. ALGSTB Was using too much AP by far for data - causing us to skip the sort and then do multiple passes in ALGSTB. ALGSUB Was also using too much AP for data. GRDFIT Changed call seq and code to do specific tests or all of them depending on a TYPE argument. Correct computation for GRIDUV, ALGSUB, ALGSTB requirements. GRDFLT Added message control levels. QUVUTIL Changed OUNFWT and OUVIMG to decide about sorting and sort if needed for the weighting or gridding resp. QCLEAN Added CLNSRT and changed CLNUV to call it to make the same decisions re gridded component subtraction. IMAGR Dropped IMSORT, leaving the work to the above. Added CPARM(10) option to control beam image size. IMAGR Help - changed for CPARM(10). Moved nowhere. 8589. April 17, 1995 SCMAP Eric/Bill Suprisingly enough it seems to work. Added all the new imaging adverbs - UVSIZE, ROBUST, UVBXFN, ROTATE, MAXPIXEL, FLAGVER, DOTV - plus BPARM and CPARM to allow sub-array weighting and different reference antennas for each subarray. It really zips along for VLB data sets. Moved nowhere. 8590. April 17, 1995 $SYSUNIX/REVENV.C Pat Small, not very smart program to do (fast) reverse lookup of an environment variable. In other words, you give it the value, it gives you the first name to match that value. Like its sibling PRINTENV.C it chdir()'s to "/" to avoid the usual PWD ambiguity. This will shortly be used in the various shell scripts that currently do ugly things with PRINTENV and "grep -v". Moved nowhere, but have to allow for it in MNJ's and INSTEP1... 8591. April 17, 1995 $APGNOT/ACCOR Leonia I added a source ID, FREQID, SUBARRAY in the SN table. The first task had forced these parameters to constant. Moved nowhere. 8592. April 18, 1995 $APGOOP/HFPRT, MBDLY Pat Corrected typos in several format statements after reports from the Bonn HP-UX based midnight job that they failed. AIPS Programmers should note that the HP Fortran compiler is EXTREMELY fussy about correct syntax in FORMAT statements. If you leave out ONE comma (e.g. FORMAT (1X,'Hi there' I4) no comma between the text and the I4), the compile will fail. Moved nowhere. 8593. April 18, 1995 SAD Eric Changed a bad write statement in the history routine. Moved nowhere. 8594. April 18, 1995 IMAGR Eric Changed TVDEVICE to implement read/write graphics colors and read/write (!) the TV window. Changed QCLEAN (CLNTV) to test the window size and increase it if needed. Moved nowhere. 8595. April 18, 1995 SCMAP Eric Changed program and help to start implementing a TV optiion to be doen at the end of each full Clean. So far all it lets you do is look and stop the self-cal. It should be able to do more such as looking at the restored image and editing vis data! Moved nowhere. 8596. April 18, 1995 $APGNOT/SNSMO.FOR Athol Fixed some IF statements in subroutine REFFAZ which relied on short-circuit Boolean evaluation and thus failed on the DEC Alphas. Also check whether the preceding point is blanked before extrapolating phase to resolve 2pi ambiguity. Moved nowhere. 8597. April 19, 1995 $APLULTRX/ZTAP2.C Pat This has been removed (though I have a copy stashed away). Apparently on Ultrix systems with DAT drives it didn't work, in the same way an old version of the Alpha equivalent didn't. A site in Italy had a problem so I recommended they try the Alpha version (in $APLDEC) which worked perfectly for them. Nothing to move. 8598. April 19, 1995 IMAGR et al Eric Fixed QCLEAN (CLNSRT) to put the correct dimension on the sort order requested. Sorting was failing if attempted at this point. Also changed it to do the sort only if there will be components to subtract and to know if the dirty map (or 1st residual) is both the first and the last image and so not do all the TV interaction, just the TV display. Moved nowhere. 8599. April 19, 1995 Alpha found Eric POPSGN $AIPPGM: used POTERR without setting it APCLN $QYPGM: used R1 without setting it, replaced with 1.0. BBUILD $AIPSUB: used PRGNAM with the needed DATA. PRINTER $APLOOP: mis-spelled INDXF in PRTIMH. UVDPAD $APLNOT: continued a string constant across a line which probably included a large number of blanks UVMSUB $QNOT: also continued the same string across lines. UINIT $AIPSUB: had all the POPS errror management code and never branched to it. MFILL $APLNOT: never branched to one error type. IMREF1 $APLNOT: never branched to two error types. ZTTOPN $APLDEC: skipped lots of code - removed it. ZSSOPN $APLGEN: never branched to one error condition, removed an unused variable too. IMGMEM $QNOT: never branched to the MSGWRT call. APUTIL $QOOP: removed all 995 CALL MSGWRT statements - none were used QCLEAN $QOOP: removed some of the extra MSGWRT calls at the end since some were not called. ARRAY $APLOOP: removed 12 of 990 CALL MSGWRT not used. CLEAN $APLOOP: removed 1 excess MSGWRT call. IMAGE $APLOOP: removed 1 excess MSGWRT call. IMAGEUTIL $APLOOP: removed 1 excess MSGWRT call. UVDATA $APLOOP: removed 1 excess MSGWRT call. UVUTIL $APLOOP: removed 3 excess MSGWRT calls. Moved nowhere. 8600. April 19, 1995 COMLNK, FC Pat Modified COMLNK so that it now marks the COMLNK.UPD "midnight job" log file ONLY if the file compiled/linked successfully. Also modified COMLNK and FC to allow compiling and linking to be done in a single command. Moved nowhere. 8601. April 20, 1995 AP in OOP Eric Nowhere in AIPS OOP do we defend ourselves from one object using the AP while another object thinks it has it all to itself. Therefore: APDEVICE $APLOOP: new class allows one AP open at a time. GENERIC $APLOOP: recognize AP device calls. APUTIL $QOOP: add APOBJ to do usual AP open/close (NOTE: this is not a QINIT or QRLSE which are done separately wherever. The AP object name has the subroutine name in it, allowing for unique ID and traceback of failures (if any). QARRAY $QOOP: add calls to QINIT, QRLSE and APOBJ to ARRFFT. QCLEAN $QOOP: add calls to APOBJ in CLNCYC, CLRGRD, CLGRID QINTER $QOOP: add calls to APOBJ in IMGINT, IMGHGE, IMGIAE QUVUTIL $QOOP: add calls to APOBJ in ODFT, OUNFWT, OUMSUB, OUMDIV, OUPDIV, OUCDFT, OUFDFT, OUVIMG, OUVSRT IMAGR $QPGOOP: add calls to APOBJ in IMTAV, IMACPY WFCLN $QPGOOP: add calls to APOBJ in WFTAV, WFCOPY CXCLN $QPGOOP: add calls to APOBJ in CXLEAN. Moved nowhere. 8602. April 21, 1995 $APLSGI/ZTPWAD Pat/Jeff Pedelty Under SG-Irix 5.3 (unlike 5.2 and earlier versions) the error return for an access to a record beyond EOF in a direct access FORTRAN file is not 27 or -1, but 168. Modified this routine to check for all three. Moved nowhere (tested in 15JAN95 at GSFC by Jeff). 8603. April 21, 1995 $SYSALPHA/LINK Pat I created this at the start of the week in an attempt to let the linker (ld) know the optimization level. Due to the increibly baroque manner in which the compile/linking is done in these (adjectives fail me) scripts, and the inadequacies of the sun4 version of awk, I was unable to achieve this short of gutting the scripts altogether and starting over (in perl). Removed this version of LINK. Moved nowhere. 8604. April 24, 1995 $APGNOT/BLAVG.FOR & BLAVG.HLP K.Leppanen/Athol New task created by Kari Leppanen to average cross-polarized data over baseline and time, allowing a more robust estimation of R-L delay and phase differences. Moved nowhere. 8605. April 25, 1995 $RUNSYS/FXAVG.001 $HLPFIL/FXAVG.HLP Athol New procedure to set the keyword SPEC_AVG in the AN table so enabling or disabling the delay de-correlation amplitude corrections required for the VLBA. This is basically a wrapper for TABED, a task which can present problems for the uninitiated. Moved nowhere. 8606. April 26, 1995 $APGOOP/SPECR.FOR Dave SPECR was not properly dealing with single-source files; it would fail when no SU table was found. I added a call to MULSDB and put an IF in front of the call to SUUPDT. Moved nowhere. 8607. April 26, 1995 $SYSUNIX/*.PL $SYSPERL/check_version.pl Jeff Changed calling convention for Perl version-checking to use process name vice a passed parameter. Simpler and cleaner. Moved nowhere. 8608. April 27, 1995 $APGNOT/SNSMO.FOR Athol Correct IF statements relying on short-circuit Boolean evaluation. These brought to my attention by Martin Shepherd at Caltech. Moved nowhere. 8609. April 27, 1995 $HLPFIL/BPCOR.HLP Athol Correct wrap-around of a line in BPCOR.HLP. Moved nowhere. 8610. April 27, 1995 $APGNOT/ANCAL.FOR & ANCAL.HLP Athol A default b-factor of 2.5 was assumed if keyword BDEFAULT was not set in the calibration file. This is an unsafe practice given that ANCAL now processes data from a variety of correlators. Changed the default to 1.0 and emphasised in the HLP file that this keyword should be set explicitly in the calibration file. Moved nowhere. 8611. April 27, 1995 SWPOL Chris The GEODLY array in SWPCAL was not large enough to hold the six polynomial terms used for VLBA data. This resulted in array overruns and possible crashes when processing VLBA data. Increased the size of the array to 8 elements and doubled the size of TABUFF to agree with the dimension required by TABCAL. Moved to the patch area on this date. 8612. April 28, 1995 $APLGEN/ZWHOMI Pat (Actually April 27; I forgot to document what I did). The remote graphics terminal option ("aips remote" on command line) was not working. The startup scripts were setting 35 environment variables TKDEVxx=/dev/tty where xx starts at the number of TV's (ehex) plus 1, and then setting TKDEV to "REMOTE". While ZWHOMI picked up on the "REMOTE" string, it did nothing to the TKDEV environment variable. Thus, any attempt to use the graphics "device" met with "no such device" errors. Modified the code so that it changes the TKDEV environment variable to TKDEVxx as appropriate, based on NTKDEV, the assigned graphics device (=#tvs plus POPS number). The "aips remote" functionality is now restored with the following caveat: the information in the image catalog for the tek devices is *site-wide*, whereas the assignment of remote graphics terminals is only host-specific. In other words, two people could get the same graphics device number on two different hosts in a single site, and they would then be using the same slot in the IC file. This will affect things like cursor lookup. One possible solution is to make the ICv000000; file host-specific instead of site-wide (in SYSETUP, INSTEP1). Moved nowhere. 8613. May 4, 1995 $APGNOT/ANTAB $YPGNOT/APCAL Athol Neglected to check for valid JLOCIF in determing the IF axis dimension. Moved nowhere. 8614. May 7, 1995 IMAGR Eric Changes: QCLEAN to limit the TV blow up factor of 256x256 images to 2 so as to leave some of the screen for other things. TVDEVICE to use a variable rather than a constant for the type parameter to IENHNS - it is an in/out variable. This caused Solaris machines to abort. UVWAIT to avoid 0 divides without changing low but real sums of weights to 1. This led to "noise increases" < 1. Also added code to write the WTNOISE to the image header as a keyword. IMGMEM to read the WTNOISE from the UV file and put it in the image file header. MAKMAP to read the WTNOISE from the UV file and put it in the image file header. Moved nowhere. 8615. May 9, 1995 OOP and IMAGR Eric Changed: IMAGR to get the paralactic angle for multi-source files and to not claim to die if that trivia fails. TVDEVICE to use the full buffer size - not half - IN TVLOAD. QUVUTIL to use the full buffer sizes - not half - all over QARRAY to use full buffer sizes in calls to PASS1 and 2. IMAGR help file to limit ti 8192 not 4096 - may not work that big however! ARRAY Maxrow -> 8192 CXIMAGE Maxrow -> 8192 CXUTIL ROW -> 16384 QCFFT DWORK -> 16400 QRFFT TABLR, TABLI -> 2048 QXFOUR IBITRV -> 8192, C, S -> 2048 QCLEAN SIZROW -> 16384 QIMAGE MIN (-> 8192. Moved to the IBMs to try it out today. 8616. May 10, 1995 Compiler scripts Chris Modified the AOC-specific compiler definitions to override the normal compiler pathnames only if the OS version is 5.3 since the compilers have revert to their normal location with the Solaris 2.4 upgrade at the AOC. Moved nowhere. 8617. May 10, 1995 buffer sizes Eric Increased buffer sizes in: $INC: DBUF (WaWa IO) $APLSUB IOSET $APLPGM: AHIST, COMB, MCUBE, PBCOR, PRTIM, SQASH, SUBIM TAFFY, XBASL, XGAUS, XMOM, XPLOT, XSMTH, SAD, SUMIM (from 2048!) $YSUB CONDRW $YPGM: CCNTR, CNTR, GREYS, IMEAN, IMVIM, PCNTR, PROFL $YPGNOT: KNTR, PFPL1, PFPL2, PFPL3, PLCUB, PLROW Found that a lot oof the OOP code was passing buffer sizes in words to routines that expected bytes. Thus, we were using only 1/2 of the buffers. Fixed ARRAY.FOR May 11. Moved nowhere. 8618. May 14, 1995 TVMOVIE Eric Fixed AU5D to apply defaults to TXINC, TYINC, and TZINC. Divides by zero and odd things when TXINC = -2 (et al.) used to occur. Moved nowhere. 8619. May 14, 1995 imaging, IMAGR Eric Corrected ALGSTB - the gridded model subtraction routine for data with rotations or non-XY sorts. It truncated a max U computation before frequency computation wher it should have rounded after increasing by the max frequency. This led to very bad vis's in the residual file! This could have affected many tasks - it was found in IMAGR. Changed HLPCLEAN.HLP to put in the TURN OFF DOTV option. Moved nowhere. 8620. May 14, 1994 SCMAP Eric Corrected: (1) Units for UV-range display (2) Handling of TVFIELD parameter so that the TV display at the end of each self-cal cycle could know what to do with windows even when the Clean TV display is off. (3) Display of final Clean image even if the Clean cycle TV display is off. (4) Increase of TV size to display the image adjusted. (5) Created HLPSCMAP.HLP for on-line helps. Moved nowhere. 8621. May 15, 1995 CONPL Eric Had the wrong scaling for the convolution function - resulting in too narrow Gaussians. Corrected to 4 / BPARM(9)**2 rather than 2 / DPARM(9). Moved nowhere. 8622. May 15, 1995 OOP, SCMAP Eric Changed: INPUT to close onjects on error. We run out of buffers if we continue without doing this. SCMAP to correct the lists of adverbs copied on TELL, modify where the TELL calls are placed, etc. to get the TELL to work. Moved nowhere. 8623. May 16, 1995 IMAGR Eric Added a large explain file section, partly stolen (and edited) from MX. Moved nowhere. 8624. May 17-18, 1995 imaging Eric All CC model subtraction and division programs EXCEPT MX go through GRDAT to initialize para,eters. This routine made a simple linear computation hard and incorrect by doing a non-linear one instead. With some A-array data at declination 86 degrees this made a 0.2 cell error which would give the gridded subtraction subroutine a bad time. GRDCRM reports problems > 0.05 cells but probably should complain on even smaller errors. FXSHFT has been complaining about shift and reference pixels not agreeing. Changed the code to do an exact forward computation first to see if the invers transcendental equation even has to be solved. I found that the forward and invers computations agree and in fact the header IS WRONG. I do not know why for sure - perhaps because IMCREA used REAL*4's for some of the variables. Changed that too and we shall see. This theory may be no good - there are 2 IMCREA's and the one in APLOOP probably won in IMAGR - so long as MAKMAP never had to make a map file. Changed $APLOOP/IMAGR to rename IMCREA in there to IMCRET. Typo fixed in GRDAT on the 18th - after DDT fell apart. Moved nowhere. 8625. May 18, 1995 $APGNOT/VBGLU.FOR Athol/Chris Constants were being passed to TABKEY in certain calls in subroutine GLUTAB where variables were required. On SUNs this precipitated curious I/O behaviour and generated an I/O error on a valid internal write further downstream (perhaps due to an overwrite of something crucial in I/O). IBM's survived this error without failure or apparent corruption of the tables. Moved nowhere. 8626. May 18, 1995 Misc Eric Changed SAD to drop the width constraint option which did not work right and raise the number of digits in the display of RMS. POPSDAT to fix TVALL to clear the background graphics channel of the labeling (now 8 not 4). KNTR fixed format of panel labeling to take needed accuracy into account. Moved nowhere. 8627. May 18, 1995 TVWEDGE Eric Moved the wedge writing to a new subroutine called IWEDGE and took it out of AU5C which now calls IWEDGE. All in preparation for adding functionality to the TV class. Moved nowhere. 8628. May 19, 1995 $APGNOT/ANTAB.FOR Athol Corrected a minor bug in parsing the INDEX keyword which caused the program to fail complaining that it did not recognize 'X', a valid INDEX argument. Moved nowhere. 8629. May 22, 1995 $QPGNOT/CONVL.FOR Athol Warn the user when the deconvolution calculation fails, and, in all cases, print the CLEAN beam size and convolving beam size used by CONVL. This information previously only recorded in the HI records. Moved nowhere. 8630. May 22, 1995 $APLNOT/AFRFI.FOR Leonia The new subroutine AFRFI estimates the values of frequency, initial phase and amplitude of a complex exponent given (with noise) in the data set. A three step process is used, first an FFT stage with zero padding. The errors of the parameters determined in this stage are less than 0.5/N for the frequency; 45 degrees for the phase and 10% for the amplitude, where N is the number of points in the data. In the second stage a search of the frequency and phase inside of the ranges mentioned above is performed. The theoretically known dependence of the phase change on the frequency change is used here. The final solution of the amplitude, frequency and phase is found by least square method. The output of the frequency is dimentionless. It has to be devided by period of data samples to obtain the correct value. Moved nowhere. 8631. May 23, 1995 $APGNOT/CLCOR.FOR Athol As a result of a user request, enhance validation of input parameters for opcode 'ANTP', which corrects for antenna and source position offsets. These are: i) Abort with an error message if ANTENNAS=0 is used for this opcode; ii) check the array name in the AN table header in deciding whether VLA data are being corrected, rather than relying exclusively on CLCORPRM(7). Moved nowhere. 8632. May 23, 1995 TV IO errors ERic When the TV connection is broken during a TV operation, the attempt to clean up - off the cursor, off any display hold, close, ... - usually causes an interupt for broken PIPE and an abort. Changed: DTVD.INC Added parameter (TVBROK) to tell ZSSSXF not to do any more IO. ZSSSXF Test TVBROK to skip the IO, set TVBROK on error ZSSSOP Reset TVBROK before open ZSSSCL Reset TVBROK after close ZSSSX2 Set IO codes to 333 and 444 to be safer since ierr is also used for the XAS error code (usually -1). Removed unused DTVD.INC from: AU5C, AU5E, AU6C, IENHNS, IMANOT, TVALUE, TVFIDL, TVLOAD, IWEDGE, OFMGAM, OFMCON Moved nowhere. 8633. May 24, 1995 TV Class Eric Changed TVDEVICE Add functions TVDWED, TVDLAB, TVDOFM TVUTIL Add functions to use these. AU5E Change calls to OFMCON, OFMGAM, OFMMOD. OFMCON Move to $YSUB, change call seq OFMGAM Move to $YSUB, change call seq OFMMOD Move to $YSUB, change call seq Moved nowhere 8634. May 24, 1995 TV changes Eric Changed: ITICS to support labels to right and above and to specify the graphics channel IAXIS1 to specify graphics chan to use to support labels to right and above and to pass the chan on to ITICS IWEDGE to specify which side of the image the wedge is on in the call seq and to bury that info away in the header for IAXIS1 to use. IMANOT to not specify channel 4 as the background AU5B to call new IAXIS1 AU5C to call the new IWEDGE PLAYR to demonstrate all the new options of the TV class. PLAYR help from PLAY.HLP to have new options. Dropped the older task PLAY in favor of a new name and added options called PLAYR. Moved nowhere. 8635. May 25, 1995 $YPGNOT/FRMAP Phil The two arrays ANT1 & ANT2 were declared to be of size MXBAS (local variable) rather than MXBASE. Fixed up and re-arranged order of includes slightly. Moved nowhere. 8536. May 25, 1995 SAD Eric Corrected formats of RMS displays. I had messed up one and not fixed the other. Moved nowhere. 8537. May 25, 1995 CONPL Eric Added a text file output option. Moved nowhere. 8538. May 25, 1995 SDGRD Eric Corrected an error in which the upper level routine asked the low level one to grid one extra sample each time (except the last). This sample must have been 0 most of the time, but something on the IBM caused it to give trouble in the form of a 5x5 square of constant largish amplitude (added to any real data at that point). The amplitude was variable and 0 in some of the channels: it had values of 2, 4, 6, and 8 in my tests in other channels. Why this should occur I have no idea - I even put in traps for large weights - since weight maps had large values at this squares - but nothing was triggered. Sigh... Moved nowhere should go to patch. 8539. May 26, 1995 SDGRD Eric Changed the computation of the max number of channels that we can grid at a time. It left little room for data. Yesterday's bug is better understood: it was getting data from the uniform weighting in the AP and got a lot of weight by gridding it many times rather than by an intrinsic large weight. There was also a chance due to another bug of overwriting one of the convolution functions with data. Moved to patch area. 8540. May 26, 1995 REMAG Eric Threw away the old version entirely and modified TAFFY to do what we wanted. The old one used WaWA IO so poorly that a lot of its power was lost and the output image was correct only in the first plane, the rest were zero. Moved nowhere. 8541. May 26, 1995 IMEAN Eric The reversed plot (flux on x axis) had the X axis label going vertical. Moved nowhere. 8639. May 29, 1995 SDGRD Eric Changed the default weighting to natural - uniform really does not make a lot of sense when doing interpolation images. Added a mode to compute an image which estimates 1/sigma**2 for the convolved image. This is te lead up to making a weighted image summing task. Meant revising QGRIDA to have mode 5 and making QGRD8 for this and QGRD7 for direct calling. Changed SDGRD and QGRD5, QGRD6, QGRD7, QGRD8 to do the multiplication of the uniform weighting factor in the gridding, and to find the maximum gridded/convolved weight to use to scale REWEIGHT(2) and the pure convolved and weight output images. Moved nowhere. 8640. May 29-30, 1995 GET4NAME, CLR4NAME Eric Changed POPSDAT.HLP, NEWPARMS.001, and AU8.FOR to make new verbs to get and clear the 4th file name. New help files too. Moved nowhere. 8641. May 30, 1995 WTSUM Eric New task to do a weighted sum. The input weight images may be planes of cubes and may be of the rms or of 1/rms**2. Had to create HENCO4 to store some of the history. Moved nowhere. 8642. May 30, 1995 UVFLG Eric It was changed by someone to demand an antennas file. For single dish, this means that it does not work. Changed it to test for AN files and to die if they really are needed (I hope). Also changed the handling of BDROP and EDROP to test for TABLE true and die since vis numbers do not go in the FG table and to apply BDROP and EDROP sensibly to single dish, rather than using "sample" numbers. Moved nowhere. 8643. June 1, 1995 GNU preparations Eric Changed: $SYSUNIX/PP.FOR to omit GNU copyleft statements found in include files from the output .f file $APLSUB/OFMDIR to ignore lines beginning with a semicolon on read and to write the gnu copy left on write. Moved nowhere. 8644. June 1, 1995 Misc Eric CALIB Will not do point solutions if GRID or IMAG is specified. The help does not say this nor did the task. Changed the latter. UVFIX Switched the pointer INFILE for CNO but not for disk. Changed to switch for both. Thanks to Michael Bietenholz. Moved nowhere. 8645. June 2, 1995 SCMAP Eric Changed NXTAVG in UVUTIL to read the uv data set to guess the data interval if the averaging time is zero on OPEN. Changed the TV load in SCMAP to avoid excessive blowups of the image. Changed other default handling as well. Moved nowhere. 8646. June 2, 1995 IMAGR Eric Corrected QCLEAN to put the data product in the output image headers. Changed warning msg format in IMAGR. Moved nowhere. 8647. June 2, 1995 XAS Eric Changed it to avoid using the lowest color values in the color table, therby leaving your first-created colors alone when it makes its non-default color map. Then, your xterms should remain readable even when the cursor is in the TV window. Moved nowhere. 8648. June 3 & 4, 1995 Misc Eric GETJY Added HI writing of input file name LISTR Changed handling of ANTENNAS and BASELINE in LIST mode. It should work now without leaving things out or looping Also fixed an abort gotcha when there was no data in MATX. Something has caused my TVGAMA to be reset to 1.5 more than once. I cannot figure out what causes that (say a TVOPEN, then a YTVCIN call and then a TVCLOS), but I am removing some possibilities. YTVCIN should not be called in normal tasks - removed it from AIPS, APCLN, SDCLN, TVCPS, UVMAP, UTESS, VTESS, and TVDIC. YTVOPN, YTVCLS also should not normally be called - changed them to TVOPEN and TVCLOS in TVCPS and TVDIC. Found a way that TVCPS could reset TVGAMA (most likely to 1.0) and fixed that. Moved nowhere. 8649. June 5, 1995 FITLD Phil My memory is failing me, thought I'd done this. As from 3 weeks ago the correlator labels the weight function as HANN instead of HANNING. Of course nobody told us and FITLD looks for HANNING. As a result the amplitude scaling for some projects is wrong (by a constant factor). Fixed FITLD to recognize both strings. Moved nowhere. 8650. June 6, 1995 FITLD Phil An obscure one. If a user read files off the tape which were written into separate files on disk (say C, S & X band) then stopped and deleted one of the bands. Then carried on with FITLD loading more files and appending to the existing files then, when FITLD had to create a new file after having appended to an existing one, it failed complaining about a lack of an FQ table. This happened because the logical DOFMAP was not reset correctly. Moved nowhere. 8651. June 6-7, 1995 CookBook Eric Anticipating changing chapters 5 and 6 into 1 chapter, I have revised most of the chapters of the CookBook ($AIPSPUBL/COOK*). Chapter 2 and Appendices A, B, G, and Y were not changed. In Chapters 3 (basic), 6 (display) , 8 (line), and 9 (VLB), references to MX were changed to discuss IMAGR instead. GET4NAME was added to chapter 3 and PLAYR to Chapter 6. Chapter 13 (the lists of help files available via ABOUT) was updated by running SHOPH, changing all relevant ABOUT help files, and then making the same changes to COOK13.TEX. The HYB.HLP file gave bad information (it called itself a task for starters) which was corrected and SHOPH re-run. Help files changed were: ZZVLBI ZZTASK ZZRUN ZZIMAGING ZZCALIBRAT ZZAP ZZVERB ZZUV ZZTV-APPL ZZTV ZZTABLE ZZSPECTRAL ZZSINGLEDI ZZPROCEDUR ZZPLOT ZZOOP ZZMODELING ZZINTERACT ZZCATALOG ZZASTROMET ZZANALYSIS ZZADVERB Moved nowhere - the WWW will be updated later. 8652. June 7, 1995 ZFREE Eric Now that user 1 is automatically given access to reserved disks, ZFREE need to say "resrved" rather than "not you" to user 1. Moved nowhere. 8653. June 7, 1995 UVPLT Eric Changed the help file to explain "Craig's" scaling changes - he made 0.0 always mean self-scale when one might want it for a limit. Added the option to weight the data in binning. Moved nowhere. 8654. June 7, 1995 MCUBE Gustaaf When creating a new output cube, in which the first input map is not located at the first pixel in the output map, the location of the alternative reference pixel was wrong. Fixed this by changing the pixel location, not the value. Moved nowhere. 8655. June 8, 1995 $APGOOP/DTSUM Phil DTSUM only dealt with single freqids. Fixed it so that it now prints information for all freqids in the file. Moved nowhere. 8656. June 9, 1995 500 boxes ! FILEBOX Eric Changed TVUTIL, GRBOXS, AU5C, PLAYR, IMAGR, WFCLN, SCMAP to use the new include file DCLN.INC which gives the numebr of clean boxes from adverb CLBOX (50) and max allowed (now 500). Wrote new verb FILEBOX to do REBOX with output to a text file (for IMAGR and WFCLN). Changed POPSDAT.HLP, NEWPARMS.001, AU5C.FOR, and (new) FILEBOX.HLP for this. Also COOK6.*. Moved nowhere. 8657. June 11, 1995 FILEBOX Eric The way I designed this was all wrong. Changed AU5C and FILEBOX.HLP to read in the file if it pre-exists, modify those boxes in the currently chosen field which fit on the TV image, and write the lot back out. Created new adverb BOXFILE to use for this: changed IMAGR and WFCLN helps to use this adverb, changed POPSDAT.HLP and NEWPARMS.001 to create it, and wrote a help file. Also COOK6.TEX. To make it work I needed to use a scratch file. So changed ZTXOPN, ZTXOP2, and ZTXZAP to accept any length file name and to call ZFULLN only if there is a colon in the file name. Thus, one can either use it with a "standard" aips name or with a fully expanded name. AU5C then calls ZFULLN to construct a scratch name and uses ZTXOPN to open it and ZTXZAP to delete it. Fixed a bug in IMAGR's reading of the BOXFILE. It did not read round boxes correctly. Moved nowhere. 8658. June 11, 1995 UVCOP.FOR Leonia UVCOP has not modified the FREQID entries in IM table when splitting off a specific FREQID. Since FITLD loads all frequencies into one file this made some programs (INDXR for example, which looks at the IM table to make the new CL table) very confused. I added several selection critera including FREQID, ANTENNAS, IFs, TIMERANGE and SOURCES. IM table was excluded (in subroutine COPYHI) from the tables list copied by subroutine ALLTAB.FOR. The new subroutine IMSEL.FOR has been written for copying IM table with the selection criteria. FREQID is modified now to 1 if a specific FREQID is given in input adverbs and is copied without modification if FREQID=0 in input adverbs. Moved nowhere. 8659. June 12, 1995 $APGNOT/ANTAB.FOR Athol Tidy up the reporting of bad time stamps found in reading tabulated TSYS or TANT data. Check time validity before searching NX table. Moved nowhere. 8660. June 12, 1995 $APGNOT/APCAL.FOR Athol Improve clarity of error message concerning missing source flux densities in the SU table. Moved nowhere. 8661. June 12, 1995 $QPGNOT/PCAL.FOR Athol Increase internal buffer size to double the number of allowed visibilities after averaging from 10,000 to 20,000. Also add code forwarded by Eric to add the full filename description to HI file. Moved nowhere. 8662. June 13, 1995 New CQ table routines Athol Add new routines for correlator parameter frequency (CQ) table. Needed for VLBA FX correlator delay decorrelation corrections within the calibration system. $APLNOT/CQINI.FOR - Creates/initalizes CQ table $APLNOT/TABCQ.FOR - I/O to CQ table $INC/PCQV.INC - Parameter include for CQ table $APLNOT/CQMAKE.FOR - Utility subroutine to build a CQ table if missing. Moved nowhere. 8663. June 13, 1995 $APGNOT/FXVLB.FOR $HLPFIL/FXVLB.HLP Athol New task to create a CQ table using information contained in the FQ and MC tables and the catalog header. Will replace $RUNSYS/FXAVG.001 shortly, and is provided for backward compatibility with files for which FITLD has not written a CQ table. Moved nowhere. 8664. June 12, 1995 Clean up old mail Eric Changed: HOLGR $HLPFIL - Rick's rewording suggestions. UVPRT $HLPFIL - changed the holography added info to the modern version. UVPLT $YPGNOT - corrected the handling of UVGET calls for the multiple FQ id mode. An empty/missing FQ could kill it because of an useless UVGET INIT call ahead of the GINIT call or if the data were not self-scaled. Moved nowhere. 8665. June 17, 1995 ROBUST Eric Changed IMAGR, QUVUTIL, and UVWAIT to use 10**ROBUST rather than 100**ROBUST. Changed IMAGR and ROBUST help file for this and ran ispell on IMAGR's help file and otherwise cleaned up some wrong things. Moved nowhere. 8666. June 17, 1995 FILEBOX Eric Changed meaning of NFIELD = 0. Now, on this it tries to get the field number from the CLASS of the loaded image. It takes 1 as a last resort. Changed AU5C and the help file. Moved nowhere. 8667. June 19, 1995 UVCOP Leonia The new subroutine IMSEL added to UVCOP (see #8658) has been detached from UVCOP.FOR and put in the $APLNOT. Moved nowhere. 8668. June 20, 1995 CookBook Eric Put back the new draft chapter on imaging - COOK5 - and revised the table of contents and a number of other chapters (in their references to 5). Also updated some ZZ*.HLP and the FXVLB.HLP to add VLBI to its categories. Moved to the WWW/ftp area. 8669. June 21, 1995 CQ routines Athol Modified $INC/PCQV.INC, $APLNOT/CQINI.FOR and $APLNOT/TABCQ.FOR to add an extra field for FFT overlap. Corrected channel bandwidth filled by $APLNOT/CQMAKE.FOR. No impact as CQ tables not yet activated in calibration system. Moved nowhere. *********************************************************************** Changes while 15JUL95 was NEW: *********************************************************************** 8670. July 3, 1995 $SYSUNIX/LINK Pat Fixed a subtle but nasty bug in LINK. If the LINK environment variable set in the $SYSLOCAL/LDOPTS.SH file contained a qualifier that started with -n (e.g. -native on SunOS), the bourne shell command "LINK=`echo $LINK | sed -e ...`" to change it to something else will not work as expected. "echo" will see the "-n" and apparentyly ignore the rest of the string, thus resetting the variable to empty. In the LINK script, this was happening to the LDOPTS variable when it tried to "squeeze out extra blanks" via a sed command like the above. Removed the offending command (we can live with a few extra spaces). Moved to 15JAN96 this date. 8671. July 4, 1995 ZACTV8, FILEBOX Eric A couple of fixes I could not do during the freeze: AU5C Fix the default naming for the 1st field ZACTV8 Be self-defensive on the returned path name Moved to 15JUL95 this date. 8672. July 11, 1994 Help files Eric Fixed the one-liner for WTSUM and re-ran SHOPH. This stated that changes were needed to 8 ABOUT files: ANALYSIS, ASTROMET, IMAGING, SINGLEDI, SPECTRAL, TABLE, TASK, and VLBI. Updated these help files and changed COOK13 and COOK0 and RECIPE.TEX to adjust for it all. Moved to 15JUL95 as needed. 8673. July 12, 1995 FILEBOX Eric Redisigned FILEBOX in AU5C to write a complete backup file in the user's area with a related (but scratch) name and then zap and re-write the output file. The backup is kept if there is any error and its name reported to the user. The $FITS area is used only if the user does not include a logical area in the input file name. Moved to 15JUL95 this date. 8674. July 12, 1995 GNU Eric Created a GNUGPL.HLP from the short text in the help and the full GPL text in the Explain sections. Changed AIPS.FOR to put out some lines witb info. Moved to 15JUL95 this date. 8675. July 12, 1995 ZACTV9 in APLUNIX Chris Stopped ZACTV9 changing the process name of the debugger to that of the task. Revision 3.0 and later of the SPARCworks debuggers use the process name to search for icon files and to obtain license-manager tokens and will not run if the process name is messed with. Moved from 15JAN96 on this date. 8676. July 13, 1995 PP Eric Removed "new version" message - it's not that new anymore. Corrected variable name that caused it to include the gnu copyleft stuff at first level includes. Moved to 15JUL95 this date. 8677. July 13, 1995 SCMAP Eric Corrected saving of WINdows - it used the old 50 boxes dimension rather than the parameterized one (now 500). Later routines were correct and OOP defended itself. Moved to 15JUL95 this date. 8678. July 13, 1995 PLAYR Eric Corrected minor errors in help file. Typed up a new HLPPLAYR.HLP file since the old one was lost in a putbck screwup. Corrected the code to allow for HLPPLAYR and to see that graphics planes are off when empty. Moved to 15JUL95 this date. 8679. July 14, 1995 OTFUV Eric Changed OTFUV to add time smoothing - an averaging interval and an output increment - to the task. This should save disk space without comromising the imaging. It has suprisingly large effects on SDGRD's output, but after baseline removal an averaging of 2 or 3 has only a little effect. Moved to 15JUL95 this date. 8680. July 18, 1995 $APGNOT/PCLOD.FOR $HLPFIL/PCLOD.HLP Athol Upgrade PCLOD, which reads VLBA pulse-cal information from an ASCII text file, to correctly handle data with multiple freq. ID's. In addition, now reads the NX table to assign source ID's to the SN table entries and is more robust in handling mixed state count/pulse cal. modes. Moved to 15JAN96 this date. 8681. July 18, 1995 EGETNAME Eric Fixed AU8 to do EGETNAME again. It got lost in the code clean-up accompanying the new GET4NAME... Moved to 15JUL95 this date. 8682. July 18, 1995 IMAGR, SCMAP Eric Fixed both tasks to force the IMSIZE to be a power of two. Changed SCMAP to set NITER to 300 if the user sets it <= 0. Moved to 15JUL95 this date. 8683. July 18, 1995 AIPS Eric Tried centering the GNU message to see if that's more readable. Moved to 15JUL95 this date. 8684. July 20, 1995 $YSERV/XAS.SHR (Makefile*) Pat Made some changes to the Makefile that avoid the need to edit the file on most systems. Ultrix (4.3) needs its own; its version of make is more primitive. Otherwise it works without editing on all our other architectures. It now includes the -R option for Solaris 2.x systems to burn in the library locations. Moved from 15JAN96 this date. 8685. July 20, 1995 ZMSSRV Jeff/Eric Corrected the length argument to gethostbyaddr. Giving it too much was okay with everybody until Solaris 2.4. Moved to 15JUL95 this date. 8686. July 20, 1995 HISTORY, SCMAP ERic Added more adverbs to history from SCMAP and corrected HISTORY which had too small a buffer for possible adverbs to go to history. Raised it from 50 to 500 - CLBOX requires 200. Moved to 15JUL95 this date. 8687. July 21, 1995 $QPGNOT/LPCAL.FOR & LPCAL.HLP Kari Leppanen via Athol New polarization calibration task, contributed by Kari Leppanen. Algorithm is described in his thesis, also Leppanen, Zensus and Diamond, 1995, ApJ (in press). Moved to 15JAN96 this date. 8688. July 21, 1995 $APGOOP/CCEDT.FOR CCEDT.HLP K.Leppanen/ Athol New features in CCEDT which are required by new polarization calibration task LPCAL. These include the ability to write multiple CC output tables and an option to automatically select CC boxes satisfying a flux constraint. Moved to 15JAN96 this date 8689. July 26, 1995 $YPGNOT/VPLOT.FOR Leonia The task plots different baselines for ANTENNAS=0, BASELINE=list and BASELINE=0, ANTENNAS=list. Inspite of exhaustive comment in HELP complains occurs somtimes. I have added a couple of warnings. Moved to 15JAN96 July 25 8690. July 26, 1995 $APGNOT/ACCOR Leonia The task has put unit in the SN table in the case of data absence for a given time interval. It provided some errors in result of application of output SN table near gaps of the data. I have changed the unit on blank. Moved to 15JAN96 July 25 8691. July 26, 1995 $APGNOT/PCCOR.FOR Leonia The new task PCCOR corrects instrumental phase and delay offsets using pulse-calibration information in the PC table. Pulse-calibration tones are injected into the VLBA receivers and, in conjunction with the cable calibration system, can be used to measure and track instrumental variations in delay and phase between separate baseband converters. The measured phases of the tones allow the connection of the phase across all the IFs. The phase difference between two tones in the same IF allows the estimation of the instrumental single-band delay for that IF. The measurement of single band instrumental delay is however subject to ambiguity due to the 2pi*N ambiguity in measured phase. To resolve this ambiguity, the task uses observed visibility data taken on a calibrator. The ambiguity factor is determined by the condition of maximum coherence. The ambiguity factor is determined in this manner for each polarization, IF and antenna relative to the reference antenna. The adverbs, TIMERANG, CALSOUR and FREQID select both UV and PC table calibrator data. Currently PCCOR assumed only two tones are measured for each IF. Each FREQID and SUBARRAY must be processed separately. Moved to 15JAN96 July 25 8692. July 26, 1995 $APGNOT/CLCOR.FOR Leonia Right handed coordinate system is used in VLA. Left handed coordinate system is used in VLBA. I have changed the codes to have an agreements with HELP and make the task provide a correct correction with default value of CLCORPRM(4)=0 for VLBA and VLA. The relevant change of HELP has been done as well. Moved to 15JAN96 this date 8693. July 26, 1995 $APGNOT/UVCRS.FOR Leonia The task has not worked if a start day is more than 1. I have fixed it Moved to 15JAN96 this date 8694. July 28, 1995 New adverb: DELCORR Phil Added a new adverb to control the generation of the CQ table in FITLD. The default is to create the CQ table. Once generated, and when delay corrections are applied to VLBA data at a later point, the 2nd order corrections described in Athol's AIPS Memo 90 will be applied to the data. Moved from 15JAN96 today. 8695. July 28, 1995 FITLD Phil Several changes: We have known for a few weeks now that AIPS has problems in some areas, especially fringe-fitting, if the reference frequency is not in IF 1. FITLD would always sort the incoming VLBA data into increasing freq. order but would not adjust the reference IF from that labelled at the correlator. Now it performs the adjustemnt so that IF 1 always contains the ref. freq. Also increased the size of a couple of buffers to deal with the large number of channels being routinely generated by the VLBA & ATNF. When FITLD encountered a file with a large number of FQIDs the output FQ table did not correctly describe the data due to an indexing bug in labelling the rows of the FQ table as it was being written. Fixed, also increased the maximum number of FQIDs to 100 from 10. Moved nowhere yet. Added the DELCORR adverb and ensured that CQMAKE is called if necessary. Moved from 15JAN96 today. 8696. July 30, 1995 Update HLP files Athol Update some help files in preparation for the new release: FXVLB.HLP: Add EXPLAIN section ANTAB.HLP: Document DPFU=RCP,LCP mode more clearly PCLOD.HLP: Change contact person for PCLOD text files. Moved from 15JAN96 today. 8697. July 31, 1995 OTFUV Eric Change output header to put the SD 12m beam width in the Clean beam major and minor axis. Moved to 15JUL95 this date. xxxx. July 30, 1995 Convolving functions Eric Add circular functions and a type 6 = BESSJ1(x)/x * exp. Changes: Moved nowhere. 8698. July 31, 1995 SDGRD Eric Change CONVFN to add type 6 and circular storage, GRDFLT to add type 6 and circular; clean up code and create BESSJ1 function. Changed QGRD6 and QGRD7 to do circular as well as separable convolving functions. Changed SDGRD by appending all these (so that they will not affect other tasks) and to allow for circular functions, Bessel functions, and functions specified in arc seconds rather than cells. Moved by appending method from 15JAN96. 8699. Aug 1, 1995 FITLD Phil Very minor change to avoid trying to create a new CQ table when one already exists. Moved from 15JAN96. 8700. Aug 1, 1995 VLBA delay decorrelation Athol Perform delay decorrelation corrections for VLBA data using information in the CQ table. These include spectral averaging and segmentation loss corrections, which are further described in AIPS Memo. 90. This method replaces the previous use of the SPEC_AVG keyword in the AN table so that these corrections are implemented in a more robust form that is less sensitive to changes in the frequency structure of the uv-data during calibration. Several options which are possible in the VLBA correlator, but not yet implemented, have been anticipated in the new format. For further information, consult the AIPS Memo. mentioned above or type EXPLAIN FXVLB. Changes include: $INC/DSEL.INC - Add new arrays to contain the information required for the segmentation corrections. $APLNOT/SELINI.FOR - Reset new DSEL.INC variables. $APLNOT/GAININ.FOR - New call format to DSMEAR $APLNOT/DSMEAR.FOR - Re-written to use CQ information $APLNOT/FXSEG.FOR - New routine to calculate segmentation losses. $APLNOT/DATGET.FOR - pass subarray number to DATCAL $APLNOT/DATCAL.FOR - add segmentation corrections, also warning if VLBA data with no CQ information but delay solutions are being applied. Move from 15JAN96 Aug 1, 1995. 8701. Aug 1, 1995 $APGNOT/SPLIT.FOR Athol Disable copy of the GC and CQ tables. Also add a history record noting whether the CQ table in present and hence VLBA delay decorrelation corrections are activated. Move from 15JAN96 on Aug, 1, 1995. 8702. Aug 1, 1995 $APGNOT/VBGLU.FOR Athol Disable copy of CQ table. Also increase MAXBAS to 1000000 as recently modified in TST. Moved from 15JAN96 on Aug 1, 1995. 8703.