; ----------------------------------------------------------------------- ; Copyright (C) 1996, 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). MAX 72 COLUMNS. line n-1: Where has this change been moved (e.g. 15APR97, nowhere) line n: Blank PUT A TAB IN THIS FILE, AND YOU WILL BE WHIPPED WITH A WET BANANA PEEL! 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. Likewise, the n'th line should be COMPLETELY blank (no spaces). Changes to 15APR97 *********************************************************************** Changes while 15APR97 was TST: *********************************************************************** *********************************************************************** 9167. October 23, 1996 GPSDL Chris Added a test to protect against the cosine of azimuth going out of range near culmination due to rounding errors. Also fixed a formatting error in the .HLP file. Moved to 15OCT96 on this date. 9168. October 23, 1996 LDGPS Chris Now requires the user to enter the time offset of the GPS receiver (which is variable) rather than using a look-up table. Also corrected the scaling of TEC data and modified the Los Alamos table reader to deal with variable amounts of spacing between columns. Moved to 15OCT96 on this date. 9169. October 23, 1996 CLCAL Pat Replaced old SORT routine with HSORT, a fast, public-domain heap sort. Thanks to Fred Schwab for donating it. Moved to 15OCT96 on this date. 9170. October 23, 1996 This file Pat Almost forgot to trim it for the new version. Note the new admonitions above. Moved nowhere. 9171. October 24, 1996 CLCOR Leonia Some times ago I unified the concept of left (right) hand coordinate system for VLBA and VLA (OPCODE='ANTP'). The relevant change was done at the help file. But comments at fortran codes of the subroutine ANTPOS stayed old and wrong. I brought to conformity the comments with the actual reality. Moved to 15OCT96 on this date. 9172. October 24, 1996 $YPGNOT/ELINT.FOR Leonia When there is only one antenna observing the given source, there is no opportunity to fit a polinom in the data and to determine a correction of the source flux. The task has died without an explanation meeting such data. I have fixed the problem excluding the flux correction for the case of the only antenna observing the given source. Moved to 15OCT96 on this date. 9173. October 25, 1996 $APLPGM/PRTTP.FOR Gustaaf At some time during the eighties, VLBI data sets had a total of 0 antennas in their headers. PRTTP checks for the type of tape; part of the test for an archive tape is whether the number of antennas is between 1 and 27. Finding 0, PRTTP assumed it to be NOT an archive tape, and the listing would go horribly wrong. I now modified PRTTP to allow a total of 0 antennas as well. Moved to 15OCT96. 9174. October 28, 1996 FILLM.HLP Gustaaf Added some detailed explanation on the conventions of time stamps in FILLM and AIPS uv data in general. Moved nowhere. 9175. October 28, 1996 $YPGNOT/ELINT.FOR Leonia The task excluded undefined points in the input SN table during fitting but not during plotting of the data. It happened because the data were multiplied by the gain factors before plotting and as a result the points were no longer equal to FBLANK. Corrected this by not scaling FBLANK values. Moved nowhere. 9176. October 29, 1996 $APGNOT/BLOAT.FOR Athol Copy multi-Stokes data, as requested by NRAL. Leave the infrastructure for now, which makes some assumptions and needs work. Moved to 15OCT96 on this date. 9177. October 30, 1996 $QPGOOP/OMFIT.FOR Ketan Added dynamic memory allocation to improve execution times. Also, finally, added capability of processing polarization data as well as fitting multi-channel components. Still lots to do - but not today! Moved to 15OCT96 on November 1, 1996 9178. October 30, 1996 $HLPFIL/OMFIT.HLP Ketan Much revised help file - thanks to Michael Rupen's efforts. Documented new model components. Made some features simpler to understand. Moved to 15OCT96 on November 1, 1996. 9179. October 30, 1996 $APGNOT/MK3IN.FOR & .HLP Athol Implement new fractional bit shift correction as derived and coded by Kari Leppanen (JIVE) in the course of high dynamic range imaging tests using the EVN. The FBS correction now includes amplitude corrections if desired. In addition the band centre for the existing phase correction was found to be in error by one half of a channel, and has been corrected. Moved to 15OCT96 on this date. 9180. October 30, 1996 $APGNOT/FITLD Phil Made a minor but vital fix needed for the new correlator code. The amplitude scaling change I added on Oct 3 did not work if files were skipped on tape since a crucial parameter was reset. Also added a much more minor change in correcting the reference pixels when IF and channel selection were used. Moved to 15OCT96 on this date. 9181. October 30, 1996 VLB DDT Athol Minor change to DTCHK to compare angles in the correct quadrant. Also add new file $RUNSYS/DTSIM.FR_1 for the VLB DDT. Moved to 15OCT96 on this date. 9182. November 1, 1996 $DOCTXT/USERNO.LIS Ketan Wrote PERL script to make updating the AIPS user numbers list easier. In the process, reformatted the file and made it a bit easier on the eye. Now, to add new user numbers, only the first two columns need be filled in. The script automatically figures out what belongs in all the rest of the columns. Moved to 15OCT96 this date. 9183. November 1, 1996 $APGNOT/CLCAL.FOR Ketan Replaced matrix inversion code inside CLCAL with LAPACK equivalent code. Also, cleaned up the subroutine FITSC quite a bit to make it more efficient. Moved from 15OCT96 this date. 9184. November 1, 1996 BLING Chris Added protection against overflow problems in the SNR calculation. Also removed a lurking PRINT statement left over from debugging. Moved to 15OCT96 on this date. 9185. November 2, 1996 $APGNOT/CLCOR.FOR Athol Correct the amplitude factor used for the opacity correction for OPTYPE = 'OPAC'. The amplitude factor was the square of what it should have been. Moved to 15OCT96 this date. 9186. November 3, 1996 $SYSPERL/check_version.pl Jeff Added Perl version 5.003 to "known-good" list of Perl versions, eliminating the pesky warning at the AOC. Moved nowehere, though should probably be put in 15OCT96. 9187. November 4, 1996 PCCOR.FOR, PCCOR.HLP Leonia The new adverb CUTOFF have been added. The adverb allows to swich off the cable delay. It is useful for geodesy people. Additionnally the task has been simplified excluding the referencing to the reference antenna at the stage of ambiguaty resolving of the target sources data. Moved to 15OCT96 this date. 9188. November 5, 1996 $INSUNIX/INSTEP1 Pat Fixed some typos (missing quotes) and changed several things. Added a new logging facility (what happens now goes in INSTEP1.LOG for the most part), always build PRINTENV from the C program, use gmake for XAS if we can, remove problematic LD_LIBRARY_PATH and LD_RUN_PATH setting, replace with warnings if these variables are set for SOL, use UNSHR to unpack XAS always, modified the tape device (TPDEVS.LIST) part and added a section to take care of setting up the TPHOSTS file. Moved to 15OCT96 this date, and to the tape-cut area. 9189. November 11, 1996 $APGNOT/FITLD Phil I finally added weight-based flagging due to pleading by some VLBA users. The implementation method is not totally general in that I do not write an FG table (is complicated if wish to do so efficiently). The data are permanently flagged since they are not copied from tape to disc if the visibility weight in any IF drops below the specified threshold. This was deemed adequate by several VLBA users. Also updated FITLD.HLP. Moved nowhere. 9190. November 11, 1996 SSSIN, AIPSMTn Pat Using network service names in upper case letters exposes bugs in operating systems. In particular, a NIS/YP server running SunOS 5.5 was found to reject connects for service names (in getservbyname) in uppercase. Changed $APLUNIX routines ZSSSO2.C, ZVTPO2.C, and ZVTPO3.C, $YSERV archive XAS.SHR (though SSS.SHR and XVSS.SHR should also be fixed if anyone uses them anymore), $APLGEN routines ZMOUNT.FOR, ZTPOPR.FOR ZVTPRO.FOR and ZVTPO2.FOR, and $SYSUNIX (and AIPS_ROOT) routine TVDEVS.SH to name these services in lower case. The tekserver and message server already used lowercase. Moved from CVX this date, and moved to SOL MNJ site in CV. Should also be moved to a patch area for 15OCT96 once it exists. 9191. November 11, 1996 BLAPP Chris Inserted code to ensure that the BS table is sorted in time order before entering the gridding loop (which assumes that times are non-descreasing). Moved nowhere. 9192. November 12, 1996 $SYSPERL/hlpitgen.pl Jeff/Pat This was in the aipssrc area which is not distributed with AIPS, and anyone adding tasks to AIPS will need to generate the HLPIT.LIS file in $HLPFIL after adding the new help file. Currently needs perl 5 though this can be changed with a little work. Moved nowhere. 9193. November 13, 1996 $YPGNOT/IRING Ketan IRING only plotted the incremental intensity distribution if requested. I added the option to plot the cumulative intensity distribution as well - by user request. Also, added an old modification to IRING that I had lying around. This modification allows the user to select a range of azimuthal angles over which to compute the intensity distributions. Moved nowhere. 9194. November 21, 1996 PRTAC Pat Changed it to regard real times of 0.0 with cpu times <= 1 as "good" rather than aborts (by setting real = 1.1*cpu). The really fast things these days get reals of 0 a lot due to the integer seconds of ZTIME. Moved from CVX this date. 9195. November 21, 1996 MAKE.MNJ Pat/Shoshana This script was not doing anything with the UPDCONFIG file, and it should at least make sure it's in $SYSLOCAL and copy it if not. It should not put it in $TST/$ARCH/UPDATE/ as it was doing. Moved nowhere. 9196. November 25, 1996 $YPGNOT/FRMAP Leonia The maximum number of baselines (100) was not enough to use the task for the VLA data when all 27 antennas provide the data. The relevant parameter was increased to 400. The format of print a baseline number was changed from I2 to I3 to print baseline numbers more than 99. Moved nowhere. 9197. November 26, 1996 $INSUNIX/INSTEP1 Pat Allow zero disks per host to be legitimate; some sites use the .dadevs mechanism exclusively so zero is the right answer for these hosts. Also comment out the host/IP verification code as it was not working (regexps in case statements are extremely lobotomized). Moved nowhere yet, maybe to the patch area if I get complaints.. 9198. November 28, 1996 VLB DDT data files Athol Add several new VLB DDT data files to $RUNSYS, namely DTSIM.BPASS_1, DTSIM.BPASS_2, DTSIM.CALIB_1 and DTSIM.CALIB_2. Moved nowhere. 9199. December 3, 1996 DOPOL Athol Implement new polarization correction modes DOPOL=2 and DOPOL=3, to allow second order polarization corrections when using the linear D-term approximation within AIPS. The mode DOPOL=3 applies more rigorous flagging, removing visibility points with any missing polarization correlation pairs. The mode DOPOL=2 applies the same second-order correction but makes several approximations for any missing polarization correlation pairs when computing the correction matrix. The mode DOPOL=1 makes the same first-order correction as before. The new modes are currently implemented only for 'VLBI' polarization calibration within AIPS, and have no effect on other solution types (eg. VLA or orientation-ellipticity). The second-order corrections have recently been shown to be important when correcting EVN data which may have individual D-term amplitudes as high as ~20% in isolated cases (as reported by K. Leppanen/JIVE, who submitted private code used to correct this effect). Note that 2nd order corrections have always been applied for the orientation-ellipticity VLBI polarization model in AIPS. Made the following changes: $INC/DSEL.INC - changed DOPOL from logical to integer. Change DOPOL handling from logical to integer in: $APLNOT: DATGET, DGHEAD, FNDPOL, SELINI and UVGET. $APLOOP: CALEDIT. Further changes in $APLNOT: DATPOL - add check for missing cross-hand correlations; use approximations where necessary; If DOPOL=3 flag all data with any missing polzn. pairs; polzn. correction matrix was only computed every 0.1 min or when the parallactic angle changed by more than 2 deg - this led to a 2 deg phase sawtooth in the RL,LR data at times; change to update the matrices for any change in time or parallactic angle - the computational overhead seems reasonable. VLBPOL - implement the new 2nd order elements of the polzn. correction matrix for DOPOL > 1. Do not assume low linear source polarization. Change DOPOL handling from logical to integer in: $APGNOT: DECOR, LISTR, SHOUV, SPLIT, UV2TB, UVFIT, UVPRM, UVPRT and SPCAL. $APGOOP: MAPBM. $QPGNOT: BPASS, HORUS and CPASS. $QPGOOP: IMAGR and OMFIT. $YPGNOT: POSSM, FRPLT, UVPLT, VBPLT, CLPLT, FRMAP and VPLOT. Recompiled/linked all affected tasks/libraries. Moved nowhere. 9200. December 4, 1996 DOPOL Athol Update .HLP files affected by #9199. These include: DOPOL, DECOR, LISTR, SHOUV, SPLIT, UV2TB, UVFIT, UVPRM, UVPRT, SPCAL, MAPBM, BPASS, HORUS, CPASS, IMAGR, OMFIT, POSSM, FRPLT, UVPLT, VBPLT, CLPLT, FRMAP and VPLOT. Moved nowhere. 9201. December 5, 1996 $YPGNOT/VPLOT.FOR, CLPLT.FOR Leonia Only four symbols have been allowed to use in IN2CLASS, because the other two are reserved for the images number. It has not arisen any problem in standard cases, because the CLASS of the output images of a CLEAN procedure uses usially only four of six symbols (ICLN, RCLN ...). But sometimes people desire to rename the CLASS of the image, using all six symbols. In this case the last two symbols are substituted by an image number and as a result the image can not be found by subroutine MAPOPN and plot of the model is not being plotted. To overcome the problem the tasks provide now the substitution of the last two symbols only if more than one image is used for the model. Moved nowhere. 9202. December 9, 1996 VLB DDT Athol Revise $APGNOT/DTSIM.FOR and $APGNOT/DTCHK.FOR to support several new features: i) instrumental polarization errors; ii) multiplicative gain errors; iii) rest frequencies and LSR velocities in the SU table and catalog header; and iv) baseline-dependent integration times. Also updated $HLPFIL/DTSIM.HLP and added new DDT data files DTSIM.DOPOL_1 and DTSIM.DOPOL_2 to $RUNSYS. Changes to $APLNOT/DTPARM.FOR, $APLNOT/DTCALS.FOR and $INC/DTPM.INC as part of this update. Moved nowhere. 9203. December 10, 1996 $APLOOP/UVUTIL Ketan Fixed IF/channel selection in UV2SCR. UV2SCR invokes UVGET and later, OUVSCR which then invokes OUVSLT. Both UVGET and OUVSLT attempt to check channel/IF selection. Before calling OUVSCR, the channel/IF selection parameters are reinitialized in the input UV object - they are restored after return from OUVSCR. This seems to fix the problem of channel/IF selection being applied twice. Moved nowhere. 9204. December 10, 1996 $QPGOOP/GPHAS Ketan Added new options for REFANT selection, fixed model division. Previously the REFANT was created as a new VIRTUAL antenna. Now, the REFANT can be specified explicitly, or taken to be the most commonly occuring REFANT in the CL table. Model division would fail if a source name was not specified, even if only one source existed in the file. Now, a source name is required for CC model division unless the data file is a single source data file. Moved nowhere. 9205. December 10, 1996 $APGNOT/FRING Ketan FRING now implements the ORIGIN keyword for single source data files. When CALIB and BPASS run on single source files, they add the ORIGIN keyword to the output SN table. Fring has now been modified to do this as well. Moved nowhere. 9206. December 10, 1996 $RUNSYS/VLBDDT.001 Athol Preliminary script to check some VLB tasks using simulated data. Also added .HLP file. 9207. December 12, 1996 $APLNOT/GETDEL.FOR Athol Change GETDEL to deal with IM table access that may not be in time sequence. Will backspace IM table records in this case. Also, do not compare subarray numbers when searching for IM table entries; match source id., freq. id., antenna no. and time only. Moved nowhere. 9208. December 12, 1996 $APGNOT/INDXR.FOR Athol Subroutine WRITCL was not calculating the number of CL entries correctly. An extra entry was written, which for short scans, may have fallen outside the scan boundary. Moved nowhere. 9209. December 17, 1996 $QPGNOT/CVEL.FOR Phil Before all the buffer changes were made earlier in the year the call to TABSRT for the CL table used KBUFSZ to set the size of the scratch buffer. During the changes this was modified to call UVBFSZ but no check was made that UVBFSZ was set under all circumstances. If there is no FG table (and hence no FG sorting) then the call to set UVBFSZ was not made and so the CL sorting failed. Fixed so that UVBFSZ is always set, whether an FG table is present or not. Moved nowhere. 9210. December 18, 1996 $APLNOT/DTPARM.FOR Ketan DTPARM did not properly initialize the arrays containing the source model information. This resulted in faulty data set generation when using multiple source components in DTSIM. Moved nowhere. 9211. December 18, 1996 $APGNOT/DTSIM.FOR Ketan The noise calculation in DTSIM was faulty. The 'uniform Gaussian deviate' generated in the DTSIM subroutine ADNOIZ was being generated with unit amplitude. Thus only phase noise was being generated, no amplitude noise at all. Moved nowhere. 9212. December 18, 1996 $APGNOT/UVCOP.FOR Ketan UVCOP was reporting that it had 'written' the Nth visibility when in fact it had 'read' the Nth visibility. 1-liner. Moved nowhere. 9213. December 18, 1996 $YPGNOT/UVPLT.FOR Ketan UVPLT can now be told to generate UVW coverage plots with uniform limits in the U,V,W directions. If BPARM = x,y,2,0 is selected, where x and y are chosen from 6,7,8 [U,V,W respectively], then the plot is generated with identical limits along the X and Y axes. Moved nowhere. 9214. December 18, 1996 $HLPFIL/UVPLT.HLP Ketan Documented change #9213. Also, the axis codes for UVPLT are now shown when HELP UVPLT is invoked [A minor inconvenience.] Moved nowhere. 9215. December 19, 1996 $YPGNOT/ELINT.FOR, $HLPFIL/ELINT.HLP Leonia Several modification have been made: We have implemented an OPTYPE command to additionally fit for the opacity (following a law of the form exp(-tau/cos z)) as well as the gain polynomial. This option can be useful in situations of high optical depth (such as at the upper end of Q-band, or a soaking wet day at K-band. This option is experimental and we plan to continue testing it. OPCODE = 'SECZ' has been discontinued. The task can also now deal with elevation ranges exceeding 90 degrees, when antennas observe "over-the-top". The fitted polynomial can now take two forms: i) three coefficients of the polynomial; ii) three coefficients which determine position and value of minimum correction (maximum gain). The print output format has been improved and the help file has been revised. Moved nowhere. 9216. December 19, 1996 $APLNOT/BPCOEF.FOR Athol Allow autocorrelation data to be corrected by cross-power bandpasses for polynomial bandpass solutions. Would previously trap this case and halt with an appropriate error message. Moved nowhere. 9217. December 20, 1996 $YPGNOT/FRMAP.FOR Leonia For some reason threshold, detecting the fringe rate maxima, was an integer variable. It has provided some inconvinience when the threshold is required to be fractional. So the type of the variable THRES is changed from INTEGER to REAL. Moved nowhere. 9218. December 24, 1996 $APGNOT/DTCHK.FOR Athol Clean up an erroneous ATAN2(0,0) that was caught by the Dec Alpha. Moved nowhere. 9219. December 30, 1996 $APLNOT/GETANT.FOR Athol GETANT reasonably assumes that the station numbers in the AN records are positive. If the station number is zero the array subscripting errors can overwrite file control information in some cases, and result in misleading error messages concerning unrelated system files. For this reason, modify GETANT to check for valid station numbers. The ultimate cause of the problem (likely DBCON) is being pursued. Moved nowhere. 9220. January 5, 1997 $APLNOT/TABOTT.FOR Gustaaf The over-the-top routine TABOTT was based on TABNDX. A relic from TABNDX, testing for an older version of NX tables, was removed. It created problems in TABOTT. Moved nowhere 9221. January 6, 1997 $APGNOT/LWPLA.FOR Gustaaf LWPLA's default behavior for the font size was not well documented. As before, fontsizes =<0 default to 13 point. Font sizes exceeding the maximum (17pt) will now default to the maximum font, and the user will be informed of this. Moved nowhere. 9222. January 7, 1997 GPSDL Chris Fixed a transcription error in GETSIP where the sine of the antenna latitude was used in place of its cosine in the formula for the azimuth of the source at the ionosphere (which lead to errors in the magnetic field projection). Moved nowhere. 9223. January 8, 1997 LIBR.DAT in many areas Pat Changed $SYSUNIX, $SYSSOL, $SYSSUN, $SYSIBM, $SYSLINUX, $SYSALPHA LIBR.DATs to not use directory logical names in comments (the OOP tasks did). Need to do $SYSSOL and $SYSSUN versions of LIBR.DAT.SHARED if anyone still uses shared libraries. I think I got them right, may need to watch the next few COMLNKS and COMRPLS... Moved from CVX this date. 9224. January 10, 1997 $YPGNOT/ELINT.FOR Leonia Several minor changes have been made by request of Rick Perley: some formats were changed to allow print of the output at the larger range; the headers of the plots are changed; and so on. The codes connected with OPCODE='PWRN' have been simplified. Moved nowhere. 9225. January 15, 1997 $YSERV/XAS.SHR Pat Removed the peculiar CV-specific (and now obsolete) location for the C compiler from the Makefile. Also inserted a few minor comments so that it's obvious what to do on Suns if you don't want to use openwin. Moved from CVX this date. 9226. January 17, 1997 OVLBI simulator Leonia/Athol Add support for new perturbed orbit delay and phase errors. New routines written to compute the orbit perturbations are $APLNOT/TAUORB.FOR and $APLNOT/ORBERR.FOR. New utility routine $APLNOT/DTMATC.FOR used by DTCHK and DTSIM was added to match indices in the simulated data common blocks. The calling sequence of $APLNOT/DTCALS.FOR was changed with supporting modifications in $APGNOT/DTSIM.FOR and $APGNOT/DTCALS.FOR. Moved nowhere. 9227. January 17, 1997 $APGNOT/PCLOD.FOR & . HLP Athol Revise calculation of reference frequency to correctly take into account the frequency reference pixel. Task would fail without being able to match the pulse-cal frequency set to the IF frequencies in the uv-file. Add new adverb FQTOL to set the tolerance used in the frequency match. Moved nowhere. 9228. January 20, 1997 BLING Chris Reset IRET after calling TABEXI to test for the existence of an index table. The real problem is that ISTAB (which is wrapped by TABEXI) sets IRET to an non-zero value if the table does not exist although this is a condition that we call ISTAB to check; ie. ISTAB doesn't distinguish between a table not exising and some error condition preventing us from finding out whether a given table exists. This should really be fixed in ISTAB but I'm not sure that sorting this out wouldn't have adverse consequences for other tasks that use ISTAB. Moved nowhere. 9229. January 20, 1997 $APGNOT/VLOG.FOR & .HLP Athol New task to segment and re-format the external calibration file produced by TSM for each VLBA project. The resulting output files can be used directly by ANTAB, APCAL, UVFLG and PCLOD and should help to automate a priori calibration for a broad range of VLBA projects. This task is VLBA-specific and makes some narrow-minded assumptions about the format of the input TSM and gain curve files. Nonetheless should be useful in a significant number of cases. Trial version placed in the system to allow broader testing. Moved nowhere. 9230. January 21, 1997 $APGNOT/FILLM.FOR Gustaaf The over-the-top (OT) table that FILLM wrote on request did not contain sufficient information for Leonia's task ELINT. I have redesigned the OT table format. This is not as serious as it seems, since - apart from ELINT - until now only the local task BASFT (which does the occasional baseline fitting) used this table. 9231. January 23, 1997 $APGNOT/JMFIT.FOR Athol JMFIT used a fixed format to print fitted baseline parameters. This yielded zero output for sufficiently small parameter values and led to some confusion. Corrected the appropriate format statements. Moved nowhere. 9232. January 29, 1997 $SYSUNIX/READLINE.SHR Pat There were some (4) lines in the Makefile.in that state "comment-me-out-in-distribution"; these were not commented out and were causing problems with Linux builds under Red Hat 4.x distributions. Field reports indicate that commenting them out fixes the problem. Duly commented out. Moved nowhere, perhaps should move to patch area? 9233. January 29, 1997 $YGEN/YCHRW.FOR Pat/Craig Changed the multiplier for font size adjustment so that it does not override a small font size for 1600 x 1200 monitors. The threshold is now 1800 pixels in X whereas it was 1400. Moved nowhere. 9234. January 29, 1997 $AIPSPUBL/AIPSMEMO94.TEX & .PS Athol/Chris New AIPS memo. containing results of the recent SGI benchmark tests. Moved nowhere. 9235. January 30, 1997 PCLOD Chris Made PCLOD friendlier for multifrequency observations. PCLOD no longer aborts if it encounters a frequency group that does not match an FQ-ID in the uv file but prints a warning message and skips pulse-cal groups until it finds a frequency group that does correspond to an FQ-ID in the data file. 9236. January 31, 1997 OPTIMIZE.LIS Jeff Turned off optimization for FILL.FOR under Linux. (Who keeps turning this back on?!?) Moved nowhere. 9237. February 3, 1997 $APLLINUX/ZMOUN2.C Jeff Now checks the block size on SCSI tape drives (Exabytes and DAT's) when they are mounted. If the block size != 0 then I explicitly set it to zero, and then restore the old block size upon dismounting the tape. As I recall, we've had similar problems on other systems (e.g. AIX) with tapes when the block size is non-zero. Also, ZMOUN2 now reports a bit more about the type of tape that is mounted; in addition to reporting it as an "ANSI SCSI 1" or "ANSI SCSI 2" device, it will report it as one of the following: "DDS", "EXB-8200", "EXB-8500", "DDS-2", "EXB-8505 compressed", or "EXB-8205 compressed". Moved nowhere, though this should probably be released as a patch. 9238. February 3, 1997 $INC/AIPSPROTO.H Jeff Added #define of MAXEHEX (value: 36). Currently only used by Linux ZMOUN2 for tracking "original" block sizes on multiple tape drives. Moved nowhere. 9239. February 3, 1997 $APLLINUX/ZMOUN2.C Jeff Corrected an oversight (thanks for the question Pat!): I need to keep the original block sizes in a static array since multiple tapes can be mounted and unmounted, in arbitrary order, and thus each needs to know its original block size separately. Also, if MT_Debug is #define'd as non-zero (i.e. debugging is in progress) then the tape will only rewind (*not* rewind + eject) at dismount. This eliminates a really time-consuming annoyance. Moved to 15OCT96 patch area this date. 9240. February 5, 1997 PRTAB, $APLNOT/TABOTT.FOR Gustaaf Modified PRTAB to accept OT tables as suitable for the DOHMS=1 option. Also fixed a type mismatch in TABOTT. Moved nowhere. 9241. February 5, 1997 $APLNOT/CHEBYD.FOR Athol New subroutine to calculate the first and second derivatives of the Chebyshev polynomials. Used in VSOP fringe-fit tests. (see $APLNOT/CHEBY.FOR). Moved nowhere. 9242. February 5, 1997 $APGNOT/DTCALS.FOR Athol Revise this subroutine to calculate rate and acceleration calibration errors using CHEBYD. Moved nowhere. 9243. February 5, 1997 $APGNOT/DTSIM.FOR Athol Revise to use new calling sequence for DTCALS; also compute true inverse variance weights if noise specified. Moved nowhere. 9244. February 5, 1997 $APGNOT/DTCHK.FOR Athol Use new DTCALS calling sequence. Also modify to check BS tables produced by BLING, and for rate and acceleration errors where possible in BS and SN tables. Moved nowhere. 9245. February 5, 1997 $QPGNOT/BPASS.FOR Athol Subroutine BPSEL passed a real array to CHNCOP where a double precision array was required. This causes alignment problems if double load/store is enabled (eg. default under IRIX). Declared local DP array to avoid this. Moved nowhere. 9246. February 5, 1997 SEARCH Athol Add new adverb SEARCH(10) used by FRING to specify alternate rerference antennas in initial FFT search. Modified POPSDAT.HLP and NEWPARMS.001 accordingly. Moved nowhere. 9247. February 6, 1997 $APLOOP/UVUTIL.FOR Ketan Added the capability to form scratch files using UV2SCR in the OOP system with baseline selection. This should help make scratch files smaller for some tasks. Although some AIPS tasks allow baseline selection through BASELINE and ANTENNAS, the underlying AIPS code only allows all selection in the sense of all baselines to a list of specified antennas. Now, via ABSELECT, ASELECT, and BSELECT, an OOP call to UV2SCR can form a scratch data file with only those baselines the user desires. This frees the calling task from having to discard unwanted baselines that made it into the scratch file. Moved nowhere. 9248. February 6, 1997 $QPGOOP/GPHAS.FOR Ketan Added flexibility regarding the choosing of the reference antenna. APARM(1)=1 works as before creating a new fictitous new antenna. APARM(1)=0 chooses the most commonly occuring reference antenna in the chosen CL table as the new reference antenna. APARM(1)=2 lets the user choose via APARM(2) the reference antenna number. APARM(3) also allows the user the option to normalize the visibilities before stacking them together. [This may be the only right thing to do.] Moved nowhere. 9249. February 9, 1997 $APGNOT/DTSIM.FOR Athol Fix bug introduced in #9243 which caused an error in the polarization solutions copied to the AN table. Moved nowhere. 9250. February 10, 1997 $APLOOP/UVUTIL.FOR Ketan Baseline selection as implemented in #9247 gave lots of error messages. Thanks to some friendly advice, these error messages have now been suppressed. Moved nowhere. 9251. February 12, 1997 SVANDT in APLNOT Chris New subroutine to solve the transposed Vandermonde system of equations used in polynomial interpolation (I'm surprised that we didn't have this in a subroutine already). Moved nowhere. 9252. February 12, 1997 FRSRCI in QNOT Chris New subroutine to perform a gridded fringe search with interpolation. This is similar to the QSEARC routine but extracts the delay and rate indices from the combined 1-D index and estimates the fringe position to a fraction of a cell by performing simple, quadratic interpolation around the peak in each dimension. Moved nowhere. 9253. February 12, 1997 BLING Chris Now interpolates using quadratic interpolation in delay, rate and acceleration. This has allowed the default padding to be greatly reduced so that BLING should be much faster than before. The user still has the option of adding extra padding using negative values of APARM(5) and APARM(6) but preliminary tests indicate that this gives little improvement in precision at a large cost in run-time. Errors quoted in the BS table are now wildly pessimistic (by a factor of 10 to 100) in most cases. Also fixed a sign error in writing out the acceleration term. Moved nowhere. 9254. February 13, 1997 FRSRCI in QNOT Chris Fixed faulty logic used to interpolate over the discontinuity between cells -1 and 0. Moved nowhere. 9255. February 18, 1997 $APGNOT/VLOG.FOR Athol Corrected a bug affecting INDEX record generation for uv-datasets with multiple FQ_ID's. Also added a full history section, which was neglected as an oversight in the first release. Moved nowhere. 9256. February 18, 1997 Orbit tables Athol Introduce table access routines for the new spacecraft orbit (OB) tables required for OVLB data: $INC/POBV.INC - OB table parameters $APLNOT/OBINI.FOR - initialize/create an OB table $APLNOT/TABOB.FOR - read/write records in an OB table $APLNOT/OBREFM.FOR - re-format OB table Moved nowhere. 9257. February 19, 1997 BLING Chris Revised the logic for calculating the time spacing in a solution interval to reject information from baselines that do not contribute to the current solution. This should reduce the chances of the rate window being set too wide and including aliased fringes. Moved nowhere. 9258. February 19, 1997 $QPGNOT/FRING Ketan Added two new features, exhaustive fringe searching and subset solve, to FRING. The FFT stage in FRING used to give up too easily when searching for fringes. Now, by setting APARM(9)>0, FRING will exhaustively search all baselines, beginning with those baselines to the reference antenna specified by REFANT. Optionally a new control adverb called SEARCH can be filled in to specify the order in which to use the remaining antennas to search for fringes to antennas when REFANT was inadequate. See the help file. The Least Squares Stage now allows you to specify a subset of antennas to solve for. I.e. you could specify ANTENNAS =1,2,3,4 to include data on all baselines between antennas 1-4 but also use DOFIT=1,2 to specify that solutions only be computed for antennas 1 and 2 [effectively, baselines 1-2, 1-3, 1-4, 2-3, 2-4 are used to find solutions for antennas 1 and 2]. DOFIT=0 specifies that all antennas are solved for, as before. Moved nowhere. 9259. February 19, 1997 $APLNOT/GCALC2 Ketan New routine to be invoked by CALIB instead of GCALC when subset solve feature is requested. [ see #9258 and #9261 ] Moved nowhere. 9260. February 19, 1997 $APLNOT/CLBPA Ketan Added to the call sequence so that GCALC2 is called instead of GCALC when the user requests subset solve in CALIB [ see #9261 ] Moved nowhere. 9261. February 19, 1997 $QPGNOT/CALIB Ketan Added the subset solve feature described in #9258 to CALIB. As in #9258, the adverb DOFIT can be used to specify a subset of antennas to solve for. DOFIT=0 specifies that all antennas are to be solved for, as before. Moved nowhere. 9262. February 19, 1997 $RUNSYS/*.001 Ketan Updated the run files in $RUNSYS to also initialize DOFIT=0 for CALIB and FRING and SEARCH=0 for FRING. The affected files are COMAP.001, CROSSPOL.001, CRSFRING.001, DDTHLOAD.001, DDTLOAD.001 HYB.001, MAPIT.001, VLALLOAD.001, VLACLOAD.001, VLAPROCS.001, VLBA.001, MYVLA.001 and TESTVLA.001. [ see #9258 and #9261 ] N.B.VLBDDT.001 is also affected but is currently check'ed out. Moved nowhere. 9263. February 19, 1997 $APGNOT/FILLM Gustaaf PRTAB aborted on encountering NaN values in the TSYS table as written by FILLM. In previous AIPS versions PRTAB would print these values anyway, but now PRTAB chokes on them. There was already a check for NaNs in FILLM for solar data only; this was now extended to all data. Moved nowhere. 9264. February 20, 1997 $APLNOT/ORBIT.FOR Leonia The subroutine had a PI error calculating the angle between vector of satalite position and its velocity. The error had a place only for low half of elliptical trajectory. The error did not caused a mistake in the position but caused the mistake of velocity for low half of elliptical trajectory. Moved nowhere. 9265. February 20, 1997 XTRAN Chris In 1992 the solution code from XTRAN was moved into an APLSUB subroutine (XTRFIT) without making any provision to pass back the solution vector for the non-linear case. This resulted in XTRAN failing with NPOINTS=6. Changed XTRFIT to pass back the solution vectors as output arguments. XTRAN could probably be made a lot simpler than it is. Moved nowhere. 9266. February 25, 1997 Argument counts in APLGEN Chris Fixed the number of subroutine arguments for ZV20MC, ZVTPGC and ZVTPO3 in APLGEN. The argument lists for ZVTPGC and ZTVPO3 now agree with the UNIX versions of these routines and ZV20MC has a dummy FTAB argument to agree with the single call made to it (in ZV20XF). Moved nowhere. 9267. February 26, 1997 EDITA Eric/Chris New TV task to edit uv data based on calibration tables (imprted from the CVX version of AIPS). The following files are new from CVX EDITA.FOR in QPGOOP EDITA.HLP FCINI.FOR in APLSUB TABFC.FOR in APLSUB HLPEDITY.HLP The following have been modified TABLEUTIL.FOR in APLOOP (added OFCINI and OTABFC) QEDIUTIL in QOOP (aligned with CVX) SNEDT.FOR in QPGOOP (aligned with CVX) SNEDT.HLP (aligned with CVX) HLPEDISN.HLP (aligned with CVX) Moved nowhere. 9268. February 26, 1997 BLING Chris Fixed faulty stacking logic in DOSOLI that caused the intermediate baseline buffer (IBBUFF) not to be cleared when baseline stacking was turned off; this lead to phantom results for scans without data. Moved nowhere. 9269. February 27, 1997 COMLNK Eric/Pat It makes no sense to COMLNK in debug mode only to have the sources disappear on you due to PURGE=TRUE being the default. Added NOPURGE to the options when DEBUG is found on the command line. Moved to CVX this date. 9270. February 28, 1997 UV_DESC object Eric/Chris Added a TYPEUVD keyword to the UV_DESC object which indicates whether the data is true uv data or single-dish data. Merged changes from CVX into OBJMNGR.FOR in APLOOP UVDESC.FOR in APLOOP UVPHAS.FOR in APLOOP (checks for TYPEUVD = 'UV') QUVUTIL.FOR in QOOP (checks for TYPEUVD = 'UV') Moved nowhere. 9271. February 28, 1997 EDITR Eric/Chris New task (from CVX) that allows interactive editing using a display of up to 10 baselines to a single antenna and that allows editing based on a secondary dataset (eg. residual visibilities). Moved nowhere. 9272. February 28, 1997 SCMAP Eric/Chris Enabled the "Edit Data" menu command (from CVX) and added APARM(9) to give the averaging time for data editing. Moved nowhere. 9273. February 28, 1997 FRING Ketan Change #9258 to FRING left behind two typos in the subroutine FRNMSP. These typos affected only mode APARM(5)=2. In this mode, after finding multi-band delays, the typos made FRING less likely to be able to find valid refinements to the single-band delays. Modes using APARM(5)<2 were not affected by these typos. Moved nowhere. 9274. March 3, 1997 BLAPP Chris Added checks for COLMAP(I) = 0 (indicating no solution expected) in code sections that extract results from the least-squares solution routines. This eliminates the phantom solutions that were being written to the SN table for missing antennae. Moved nowhere. 9275. March 5, 1997 $APLNOT/TFILTR.FOR Leonia/Athol New subroutine to calculate rate smearing corrections for different types of filtering during time averaging. Includes the OVLBI filters implemented in the VLBA correlator. Moved nowhere. 9276. March 5, 1997 $APLNOT/TABHDR.FOR Athol Modify to recognize the SPACECRAFT_ORBIT table (OB) now produced by the VLBA correlator, and containing the orbital ephemeris as calculated by SPICE. Moved nowhere. 9277. March 6, 1997 Correlation id. Athol The VLBA correlator can change correlation modes with great flexibility, leading to time-variable rate and delay decorrelation corrections which depend on the type of frequency and time filtering performed in the correlator. The most general solution to this problem is to implement a new correlation_id random parameter for VLBA datasets which points to the recorded correlation modes stored in the existing CQ table. This change required for full VSOP support. Changed: $INC/DUVH.INC: add variable ILOCID. $APLSUB/UVPGET.FOR: add support for random param. CORR-ID. Moved nowhere. 9278. March 6, 1997 CQ table update Athol Add a TABREV keyword to the CQ table and include a check for a mismatch between the AIPS version and the table version. This does not constitute a format revision. Changed: $INC/PCQV.INC: add revision level and new keyw. parameters $APLNOT/CQINI.FOR: add support as indicated above. Moved nowhere. 9279. March 6, 1997 Decorrelation corrections Athol Implement new form of decorrelation corrections, using the new correlation_id random parameter if it exists. This change is backwards compatible with older AIPS versions and existing CQ tables. New system allows time variable correlation mode changes in general, although only time variable OVLB filtering is activated at present. This is required for full VSOP support. Changed: $INC/DSEL.INC: add new decorrelation arrays ITFVBA, TVGVBA and ICQVBA. Change the decorrelation array indexing from (subarray,IF) to (corr_id,IF). $INC/PUVD.INC: add new global parameter MAXCID, equal to the maximum expected number of correlation_id.'s within a given file. $APLNOT/DATGET.FOR: extract the corr_id if it exists and pass it to DATCAL. $APLNOT/DATCAL.FOR: use new (corr_id,IF) indexing in the decorrelation parameter arrays; call new subr. TFILTR for rate smearing correction which supports the new OVLB filters. $APLNOT/SELINI.FOR: initialize ICQVBA; use new DSEL.INC. $APLNOT/DSMEAR.FOR: load CQ table using new decorrelation parameter arrays. $APLNOT/GAININ.FOR: change DSMEAR call to reflect new input parameters. Moved nowhere. 9280. March 6, 1997 $APGNOT/FITLD.FOR Athol New FITLD revision 9.0 to transfer the new spacecraft orbit (OB) tables and write output data with the new VLBA correlation_id random parameter. The correlation id.'s are built up as the data are read, starting from an existing CQ table if concatenation is enabled. The orbit table is reformatted to comply with AIPS requirements through subroutines OBREFM and TABREF. FITLD will not concatenate to earlier output files lacking a correlation_id random parameter. Also, the multiple integration time flag will be set if the use of an OVLB filter is detected, and data will be accepted out of time order, generalizing the existing condition. Moved nowhere. 9281. March 6, 1997 $APGNOT/DTCHK.FOR & .HLP Athol Add adverb TIMERANG to allow time selection in deriving the test compliance statistics. Moved nowhere. 9282. March 6, 1997 $APGNOT/DTSIM.FOR Athol Fix a bug in subroutine DTANTB which prevented the orbital elements from being properly written to the AN table. Moved nowhere. 9283. March 7, 1997 $YPGM/TAPLT.FOR Athol TAPLT has not worked since #9003 due to an insufficient buffer size for TABUF. Increased from 256 to 512. Also add new feature to exclude blanked table values. Moved nowhere. 9284. March 7, 1997 $APGNOT/OBTAB.FOR(HLP) Leonia A correlator creates the orbiting table at the case of an orbiting antenna at an orbit (in particular VSOP). This table includes three coordinates of the orbiting antenna and its velocity. The new task fills out the table by additional important information such as angle distance between the Sun and observing source, condition of the Sun eclipse of the antenna by the Earth and position angle of the antenna feed. The task reads the line of the specified input OB table, calculates the new data and records the modified line at the specified output OB table. The task provides (optionaly) the estimation of the six orbit parameters having used the information about the position and velocity of the satalite at the orbit. The found six orbit parameters are used to fill out the relevant column at the AN table. Moved nowhere. 9285. March 7, 1997 $APGNOT/MSORT Ketan Introduced new task which reproduces the functionality of UVSRT. MSORT does an in-memory sort of a UV-data file. It reproduces all the functionality of UVSRT. At best, MSORT should be about 3 times faster than UVSRT [ in case when the data is only weakly mis-sorted - as is the case for normal VLA and VLBA data]. At worst MSORT should take no more than 50% longer. In either case, MSORT requires no ancillary disk space - making this the preferred sorting method for large data files. MSORT does not currently meet AIPS coding standards - this will be remedied in the next few days. Moved nowhere. 9286. March 10, 1997 $APGNOT/MSORT Ketan Cleaned up the coding of MSORT. Removed the BADDISK and ROTATE adverbs. BADDISK is not relevant as MSORT does not use scratch files. ROTATE [see UVSRT.HLP] is no longer supported. MSORT is primiarily needed for VSOP support and proper ROTATE implementation may take more time than is currently available. [One day _real soon_ ROTATE support will be re-introduced]. Moved nowhere. 9287. March 10, 1997 $APGNOT/OBTAB Leonia A mistake has been discovered at the case of several orbiting antennas. It has been fixed. Moved nowhere. 9288. March 11, 1997 $SYSUNIX/TVDEVS.SH, START_AIPS Pat These scripts, and their possible counterparts in the AIPS_ROOT area, were expanding the AIPS_ROOT variable before performing any remote shell command. This is not a good idea in a hetero- geneous environment with possibly different AIPS_ROOT values. CVX has had this problem corrected for some time; imported this here finally today. NOTE: changes in START_AIPS need to be merged with midnight job sites' versions in AIPS_ROOT!!!! Moved from CVX and to lemur's /export/laips/AIPS/ today. 9289. March 11, 1997 $YPGM/BLANK.FOR Leonia Michael Bietenholz noticed that the task blanks with 0.0 when opcode 'TVCU' is used; even though it claims to blank with DPARM(6) if DPARM(5) >0. Originally there was no parameter DPARM(6) at all. So the task blanked with 0.0 if DPARM(5) >0. Having introduced DPARM(6) the relevant changes were made in subroutine DO1BLK, leaving subroutine DO1TVB without change. So DPARM(6) has been used in all opcodes except opcode = 'TVCU'. I have fixed it correcting several comments as well. Moved nowhere. 9290. March 11, 1997 $APGNOT/FITLD.FOR Athol Fix bug introduced in #9280. FITLD produced unintelligible data for an older revision of the correlator output format. Now corrected. Moved nowhere. 9291. March 12, 1997 $YPGNOT/VPLOT.FOR Leonia The selection criteria UVRANGE was introduced for SOLINT > 0. It was erroniously missed early. Moved nowhere. 9292. March 13, 1997 $RUNSYS/VLBDDT.001 Athol Add support for new adverbs recently introduced for FRING, CALIB and DTCHK. Moved nowhere. 9293. March 14, 1997 $APGOOP/DTSUM.FOR Pat/Martin Shepherd DTSUM prints a table of visibility counts versus baseline. When this table becomes too wide to show on the printer in one go, the table is split into two parts. The bug caused the second part of the table to show repeat values of the first part of the table and to label antenna number columns as all zero. This was actually due to two bugs. The first was that variable II was being used for two exclusive purposes within the same loop, and the second was that an antenna index was getting reset to zero on the second pass instead of taking on the value for the first antenna of that pass. Moved from Caltech today (finally! Sorry for the delay - Pat) 9294. March 14, 1997 $APGNOT/SPLIT.FOR Leonia SPLIT has sent zeros for RESTFQ and LSRVEL to catalog in the case of single source file. It was happenning because these values were taken from SU table which is absent for the single source file data. At the same time any single source file is a result of SPLIT aplication to an original multi source file. At this stage the RESTFQ's and LSRVEL's values are recorded at the catalog. At the single source file zeros were recorded to the catalog. So I left this recording only for multi source data saving the original values for single source files. Moved nowhere. 9295. March 14, 1997 FRSRCI in QNOT Chris Added simple phase interpolation to this fringe search routine (used by BLING). Moved nowhere. 9296. March 14, 1997 BLING Chris Stopped BLING from writing garbage to the BS table for scans where data is missing. Moved nowhere. 9297. March 17, 1997 $APGNOT/DBCON.FOR Athol Update DBCON to mirror the new common block structure in $INC/DUVH.INC resulting from the introduction of the CORR-ID random parameter in #9277 above. Modify /CATHDR/, which expects to be a copy of /UVHDR/; also update the number of words copied in FRMAT used to enforce the equivalence. Moved nowhere. 9298. March 19, 1997 $APGNOT/MSORT Ketan Noticed that in #9286 I had missed some references to UVSRT, the progenitor to MSORT. Removed them. Moved nowhere. 9299. March 19, 1997 $APGNOT/RESEQ Ketan SVLBI will require the ability to alias stations together at some points in the data reduction stream. This is the first step for that requirement. RESEQ is a new task which will, via the ANTENNAS input, renumber antennas in a UV file. RESEQ will renumber all antennas specified in the ANTENNAS adverb to ANTENNAS(1). also placed RESEQ.HLP in $APGNOT. Moved nowhere. 9300. March 20, 1997 $APGNOT/TBOUT.FOR Athol TBOUT failed with a floating point exception on DEC Alpha systems due to an invalid floating point comparison in subroutine PUTPAS. Other architectures are not affected. Use H2CHR to avoid the FPE. Moved nowhere. 9301. March 20, 1997 $APGNOT/VBGLU.FOR Athol VBGLU was failing on DEC Alpha systems for certain direct Hollerith comparisons in VBGLIN, where the comparison was made as floating point values. Replace with CHCOMP calls to avoid FP exceptions at times. Moved nowhere. 9302. March 27, 1997 $APGNOT/RESEQ Ketan RESEQ only allowed remapping of multiple antennas to one specific antenna. Now, via an optional INFILE, RESEQ can remap antennas numbers to some desired order. Also fixed up the access routines to UVDISK and UVINIT. Also updated RESEQ.HLP to reflect the new options. Moved nowhere. 9303. March 27, 1997 $APGNOT/MSORT Ketan The computation of buffer size for disk access via UVINIT and UVDISK was screwed up so that for very large record length data files, this task died just after startup. Fixed up the routines accessing UVINIT and UVDISK. Moved nowhere. 9304. March 29, 1997 $APGNOT/DTSIM Ketan A typo in the subroutine DTFILL caused file access problems via ZEXPND when DTSIM is used to generate _very_ small files. Moved nowhere. 9305. March 30, 1997 $APLNOT/AVERT.FOR Leonia/Athol New subroutine to find the integration times in a uv dataset if the random parameter ILOCIT is defined. Moved nowhere. 9306. March 30, 1997 $QPGNOT/FRING.FOR Leonia/Athol First pass at supporting baseline stacking for multiple integration times, as required by VSOP. Sparse baselines are padded with duplicate values before the initial FFT step. Weights and the SNR calculation are adjusted accordingly. Will search the data using AVERT if DPARM(4)=0. Will be refined in the coming weeks. Updated .HLP accordingly. Moved nowhere. 9307. March 30, 1997 $APLSUB/DSORT.FOR Athol New DP sort routine, cloned from RSORT. Moved nowhere. 9308. March 30, 1997 $APGNOT/DTCHK.FOR Athol Add antenna selection; also non-parametric pseuodo-variance estimation to be less vulnerable to single outliers than using only RMS. Needed for VSOP fringe-fit tests. Modify DTCHK.HLP accordingly. Moved nowhere. 9309. March 31, 1997 $HLPFIL/FRING.HLP Leonia/Athol DPARM(4) description has been changed in accordance of last change of FRING.FOR (see 9306). Moved nowhere. 9310. April 2, 1997 $APLNOT/BACOOR.FOR Leonia The misprint at the sign of one formula describing rotation of coordinates has been discovered. This misprint affected model plot of orbiting VLBI plotted at the VPLOT. It fixed. Moved nowhere. 9311. April 3, 1997 $YPGNOT/VPLOT.FOR Leonia Compile VPLOT with the new BACOOR (see 9310) Moved nowhere. 9312. April 3, 1997 $YPGNOT/CLPLT.FOR Leonia Compile CLPLT with the new BACOOR (see 9310) Moved nowhere. 9313. April 6, 1997 $QPGNOT/FRING.FOR Ketan There was a variable in FRNSRC [IS1] that long ago had been discovered to be mis-declared. Unfortunately, the subroutine FRNALL, which implements the new SEARCH mode, was based upon an older version of FRNSRC. Unfortunately, FRNALL still had the old value of 2048 - this has now been updated. Moved nowhere. 9314. April 7, 1997 $YPGNOT/VPLOT.FOR Leonia I have added new information at the header of plots, following requests of people. VPLOT has provided information about the CC table version of the image ising in the modeling and has not informed the image file name itself. I have added the image file name at the third line of the header. I have added also information about CL table number (if. DOCALIB) and FG table number at the second line of the header. The information about CL and FG tables is not given at the case AMP. and PHASE plot (BPARM(2) < 0) because lack of room at the line at this case. Moved nowhere. 9315. April 7, 1997 PUVD.INC Gustaaf CALIB failed on data taken with a 84-element Heliograph. This was caused by a too low maximum baseline table size XBTBSZ as defined in PUVD.INC. XBTBSZ was increased to 16384, causing a massive recompile and link of lots of AIPS routines. 9316. April 7, 1997 START_AIPS Pat The secure shell (ssh/slogin) and its daemon (sshd) complicate the login sequence somewhat when the X tunneling feature is implemented (e.g. you're on foo, slogin bar, DISPLAY will be set to bar:1.0 and "tunnelled" in an encrypted form through the login back to foo's server and not via the usual X port). Fortunately there's a way to detect this (SSH_CLIENT is set in the environment). The part of this script that attempts to figure out "where are you" -- which is absolutely NOT guaranteed; it's at best an educated guess -- should now work correctly in such an environment. The part that will start the TV on the remote host is still a bit flaky, and will only work if AIPS_ROOT is defined in a remote shell (rsh or ssh) environment. But that's another problem... Moved nowhere yet. 9317. April 8, 1997 $APLPGM/PRTAB.FOR Ketan PRTAB did not print out the keyword values to full precision even though there is plenty of space since each keyword/value pair occurs on a separate line. motivated by SVLBI considerations. Moved nowhere. 9318. April 8, 1997 $QPGNOT/FRING.FOR Ketan There was a typo in FRNSRC [and so also in FRNALL] relating to the amount by which the FFT padding size was decimated when the AP size limititation became an important consideration. Moved nowhere. 9319. April 15, 1997 $APGNOT/CLCAL.FOR Ketan $APLNOT/SN2CL.FOR CLCAL now allows pass-through calibration, as well as 'proper' FQID and SOURCEID selection. By choosing 'CALP', CLCAL can now be made to pass CL table records for which no SN table record is found. This is dangerous but necessary if SN tables are constructed piecemeal as may be necessary for VSOP data. SN tables generated by single source data files and by OOP tasks may contain FQID=-1 and/or SOURCEID=0. Such entries should match ALL CL table records. This is perfectly obvious for FQIDs but a bit ambiguous for SOURCEIDs. Moved nowhere. 9320. April 15, 1997 $QPGNOT/FRING.FOR Ketan The format statements in FRING have been revised. This was instigated by the need for more digits to be printed during HALCA fringe searches. Moved nowhere. 9321. April 17, 1997 $APLNOT/ORBIT.FOR Leonia The subroutine returns back the input parameters of the orbit. Because the input parameters are changed inside the subroutine it could not be used in cycles. I renamed the input parameters inside the subroutine. Moved nowhere. 9322. April 18, 1997 $APGNOT/DTSIM.FOR, $APLNOT/BACOOR, ORBERR Leonia I have comlnked (comrpled) $APGNOT/DTSIM.FOR, $APLNOT/BACOOR.FOR, $APLNOT/ORBERR.FOR due to change of ORBIT.FOR. (look #9321) Moved nowhere. 9323. April 18, 1997 dbgwrap utility Chris Added the source for a small program that allows an AIPS user to switch debuggers on the fly to SYSUNIX. If dbgwrap is specified as a debugger the user will be prompted for a debugger to use every time he types GO with debugging enabled. This was motivated by a need to run AIPS tasks under the SGI performance monitoring tools which require a number of command- line arguments that can not be supplied using the standard AIPS debugging interface. New files are DBGWRAP.C in SYSUNIX (program source) DBGWRAP.DOC in SYSUNIX (documentation) Note that DBGWRAP.C is C++ not C (the Standard Template Library reduces the amount of code needed for this utility by a significant amount). This utility is, of course, not necessary for running AIPS. Moved nowhere. 9324. April 20, 1997 $QPGOOP/CL2HF.FOR Ketan $QPGOOP/HF2SV.FOR Dave Gordon of USNO has made extensive[?] modifications to CL2HF and HF2SV which was originally written by Ed Fomalont. Ed has blessed the upgrade of these two tasks by Dave Gordon's replacements. I have pushed Dave's modifications through the SOL compiler and have removed most non-standard FORTRAN stuff. Compilation still gives some warnings about I*2 which Ed says is necessary to use. CL2HF.HLP and HF2SV.HLP will be modified in the _near_ future to recommend the user to Dave Gordon's guide for Geodesy/Astrometry data reduction in AIPS. Moved Nowhere. 9325. April 20, 1997 $APGNOT/JMFIT.FOR Ketan Although JMFIT is properly driven by using small regions around the image features being fitted, sometimes these regions can have lots of pixels across them. I have encapsulated the declarations of the max. number of pixels that JMFIT can deal with and increased it from 2025 to 10000. Moved Nowhere. 9326. April 21, 1997 BLING Chris Inserted a test on AEXIST(A) to the baseline stacking loop in DOSOLI so that BLING does not examine baselines for which they can be no data. Not only did the lack of this test introduce unneccessary computations into this loop but the memory hits on unused areas of the intermediate data buffer, IBDATA, lead to cache-line contention on the Origin 200 (and probably other machines) making baseline stacking a severe performance bottleneck. Run-time has improved from 107 sec to 17 sec on the Origin for the VLBDDT test case: the L2 data cache hit rate has gone from 29.8% to 98.3%. Moved nowhere. 9327. April 16, 1997 PRINTER Pat Added the ability to change your AIPS printer on the fly from the POPS command line. This was achieved via a new verb PRINTER and some revision of the shell scripts. New files added : $AIPSUB/PRINTR.FOR: show current printer, show a list, or change to a different printer. $HLPFIL/PRINTER.HLP: describe how to use it: PRINTER 0 or PRINTER to show current printer; PRINTER n (n <= max printer) to change printer; and PRINTER n (n > max, e.g. 999) to list available printers. Changed were: $AIPSUB/AU1.FOR: code for new verb added. $RUNSYS/NEWPARMS.001: add new verb $HLPFIL/POPSDAT.HLP: add new verb $SYSUNIX/START_AIPS: new method for defining printer environment variables, plus faster loop (in awk) to display the menu; this significantly speeds up the startup. NOTE: $AIPS_ROOT version of this needs to be merged with these changes. $SYSUNIX/ZLASCL: use new variables, remove obsolete code $SYSUNIX/ZLPCL2: same thing. With this new way of addressing printers, the PRDEVS.SH file is no longer needed. INSTEP1 needs to be modified to not bother with it, and perhaps $SYSUNIX/PRDEVS.SH itself needs to be deleted from the system. Note: It may be necessary to remove the older versions of ZLPCL2 and ZLASCL and replace with symlinks to the newer versions for sites that maintain more than one version of AIPS. Moved from CVX (#9399), with minor enhancements (null printer). 9328. April 24 1997 Printers under UNIX Chris The previous change introduced an environment variable named LPTYPE into the AIPS environment: this variable is also interpreted by the SVR4 print service resulting in a complete failure of AIPS printer support under Solaris 2.x. Changed LPTYPE to PRTYPE in START_AIPS, ZLPCL2, ZLASCL (all in SYSUNIX), and PRINTR.FOR (in AIPSUB). Moved nowhere. 9329. April 30 1997 $APGNOT/OBEDT.FOR Leonia The orbiting VLBI provides some additional requirements which do not have a place at the case of the classic ground based VLBI network. For example the Sun can be eclipsed for the orbiting antenna and as a result the battery will not be charged properly and the whole orbiting hardware will not work properly. So the user can want to flag such data in advance. He can want to flag data using another parameter (distance, speed ..) as a criterium. The new task OBEDT allow to flag the orbiting antennas data under control of OPTYPE (five options) and the lower and upper limit of the selected parameter. The data are flagged if the selected parameter is between these lower and upper limits. In principle the value of the selected parameter could be taken from the OB table. But possible big gaps in time at the OB table could provide errors at the estimation of the edge times. To circumwent this problem the task uses the AN table reading the six parameters of the orbits. If the AN table is not filled out by the orbits parameters, the task reconstructs the orbits using OB table data. If OB table does not exist, an user can use the orbit determined by selected six parameters. In the all three cases the value of the selected (by OPTYPE) parameter is calculated with the selected time step. Moved nowhere. 9330. April 30 1997 ZMOUN2, ZTAP2 in APLSGI Chris ZMOUN2.C and ZTAP2.C in APLSGI have been revised for IRIX 6.x. Non-rewind, variable-blocksize devices should be used (eg. an Exabyte 8500 as device 5 on SCSI adaptor 1 should be listed in TPDEVS.LIST as /dev/rmt/tps1d5nrv.8500) and mediad should be configured to ignore any tape devices used by AIPS (mediad makes it impossible to use variable blocksizes). Exabytes, DATs and 9-track tapes are nominally supported. Exabyte compression modes are not supported at this time, however. I have verified that FITS files can be exchanged between IRIX and Solaris systems on Exabytes at both high and low density: other file formats and device types have not yet been tested. The old ZMOUN2.C and ZTAP2.C, which do not compile under IRIX 6 have been renamed ZMOUN2.CI5 and ZTAP2.CI5. Moved nowhere. 9331. April 30, 1997 TABINI Chris Modified TABINI to use manifest logical constants rather than variables initialized to .TRUE. and .FALSE.. On the SGI F was being initialized to 00000000 hex if TABINI was not in a library and ffffff hex (interpreted as true) if it was in a library (at least when called from FITTP). This is clearly an IRIX bug but the code clean-up is probably desirable anyway. Similar problems may be lurking elsewhere. Moved nowhere. 9332. May 2, 1997 VBGLU Chris Added some missing terms to the tests that determine which file a given IF comes from in the 3-file and 4-file cases. The output FQ table should new be correct for 3 and 4 file. Moved nowhere. 9333. May 5, 1997 $QPGNOT/CPASS.FOR Athol Correct a word alignment error in BPSEL, which passed a real array to CHNCOP where a double precision array was required (see #9245 above). This failed on the SGI system which has double load/store as a default, but had no effect on other architectures. Also added check in XPSCAL to avoid divide by zero for null spectra, and changed erroneous reference to CPASSPRM to the correct BPASSPRM. Moved nowhere. 9334. May 6, 1997 $YSUB/PLINIT.FOR Ketan $YSUB/PLAXES.FOR $YSUB/PLPNT.FOR $YSUB/PLPGO.FOR $YSUB/PLLINE.FOR $YSUB/PLDRW.FOR $YSUB/PLCLOS.FOR $INC/DPLD.INC Athol has implemented a _small_ subset of 'PGPLOT-like' routines within APCAL. These routines have now been used within OBPLT as well. Therefore, these routines have been extracted from within APCAL.FOR and are now going into $YSUB and $INC Moved nowhere. 9335. May 6, 1997 $YPGNOT/OBPLT.FOR Ketan $HLPFIL/OBPLT.HLP OBPLT is a new task which allows various columns of the OB table to be plotted against each other. By design, the inputs are minimal. Moved nowhere. 9336. May 7, 1997 $YPGNOT/OBPLT.FOR Ketan I tried to pull one over on the compilers, the LINUX compiler found me out and complained about calling sequences not being aligned. Moved nowhere. 9337. May 8, 1997 FNDEXT Athol Space VLBI requires support for more than 20 tables, as currently accomodated in the catalog header. This limit is already exceeded for some VLBA projects. Start changes to increase this limit by encapsulating access to this part of the header, and then re-use the existing WCS pointers, which are defined but not used, to allow more tables. This defers a format change for a few years, and is backwards compatible. Adopt an incremental approach to this update; as a first step replace direct catalog header access with FNDEXT calls in the following $APGNOT tasks: CANPL CLCAL CLSMO DBCON GETJY SNSMO LWPLA SDCAL SDTUV SLCOL SPLIT SWPOL TABED TXPL SPCAL Moved nowhere. 9338. May 8, 1997 $RUNSYS/NEWPARMS.001 Ketan $HLPFIL/POPSDAT.HLP $APLSUB/ALIASA.FOR $APLSUB/ALIAS.FOR $YPGNOT/VPLOT.FOR $HLPFIL/VPLOT.HLP A new adverb called ALIAS has been introduced for use in SVLBI applications. Together with this adverb, two subroutines were added. The subroutine ALIASA should be used to rearrange the ALIAS adverb list into a more suitable form. The subroutine ALIAS should be used to determine if any antenna is supposed to be aliased into another antenna. The ALIAS keyword has been implemented in VPLOT. Search VPLOT for instance of ALIAS for an example of the implementation of the ALIAS keyword. Moved nowhere. 9339. May 8th, 1997 $YPGNOT/APCAL.FOR Ketan The plotting routines in #9334 were based upon routines in APCAL and in this revised version of APCAL, those routines no longer appear internally. Moved nowhere. 9340. May 13, 1997 FILLM Gustaaf Decided to put back the version of FILLM with the Van Vleck correction as a - hidden - option. I am pretty sure it works, but so far we have been unable to find data that would benefit and I could test it on. Main reason is a saturation effect in the online system that clips high correlations - which are the ones that benefit from Van Vleck correction in the first place. We need to wait for D-array to properly test this; until then the option will remain hidden, but the code is all there. Moved nowhere. 9341. May 13 1997 GPSDL Chris Added a fiddle-factor to adjust the predicted Faraday rotation to agree with that measured in the April test run. Added an option to provide diagnostic printout. Changed the TECMOD subroutine to use the LAPACK routine SGELSS and to buffer GP table data. Moved nowhere. 9342. May 14, 1997 FILLM Gustaaf In # 9340 I left some test code in which I thought was innocuous. It wasn't. Took it out. Moved nowhere. 9343. May 15, 1997 FITLD Phil Several minor problems had accumulated. Saved the fixes until now: (1) Users were complaining that FITLD was finding subarrays in their data when there weren't any. In fact there are, since when changing sources rapidly the log based nature of the correlator results in fuzzy scan boundaries since some telescopes get to the new source more quickly than others. The VLBA on-line group are looking at this problem but in order to demonstrate that subarrays exist I print the offending information with the subarray message. (2) Some old data have the occasional source number of 0 in the distribution FITS files. Fixed FITLD to skip such entries. (3) The correlator now writes CALCODES. The FITLD code to check whether a source in an incoming file matches an existing source in the file on disk was changed to compare only the first character of the CALCODE string. Previously the comparsion was failing and the SU table was bloated with the result that SPLIT would run several times on the same source. (4) Ensured that the IM table is correctly sorted after each file is loaded. This was causing problems for rapid source switching observations. (5) Ensured that the CATBLK is updated after the CQ table is created. Moved nowhere. 9344. May 15, 1997 $YPGNOT/FRMAP.FOR, FRMAP.HLP Leonia The FRMAP.HLP file provided wrong information about used polarisation. The label of the plots at the case STOKES = ' ' was wrong. The FRMAP can handle two polarization combination; RR or LL. If STOKES = ' ' RR is handled. The relevant small change at FRMAP.FOR and FRMAP.HLP is provided. Moved nowhere. 9345. May 16, 1997 FILLM Gustaaf The data format on the Modcomp tape involves a scale factor and its exponent. Due to saturation of very strong sources the VLA online code was recently modified to allow this exponent to be of the opposite sign. This caused a logical AND in the FILLM code to fail. This AND (which was there for very old formats only and until now worked fine on new format data too) has now been removed for data formats after 1988. Moved nowhere. 9346. May 18, 1997 FITLD PHIL In change 9343 I forgot to update FITLD's welcoming message. Now done. Moved nowhere. 9347. May 20, 1997 BLING Chris Added summary reports giving delay, rate and acceleration for each solution interval and polarization for which fringes were found. Moved nowhere. 9348. May 20, 1997 $QPGNOT/EDITA, EDITR, SNEDT Leonia The task EDITA has not worked when RATE editing. The problem was units. The units of delays are sec and of rates are sec/sec. The normal values for these are then around 1.e-9 and 1.e-12. In $QOOP/QEDIUTIL.FOR, subroutine EDMXMN there was a 1.E-8 test on reasonable which was two big for rate. It was changed to 1.e-14. The problem was figured out by Eric Greisen. EDITR and SNEDT use $QOOP/QEDIUTIL.FOR also so they were comlnked as well. Moved nowhere. 9349. May 20, 1997 FNDEXT Athol Continuation of #9337. Add FNDEXT calls in: $APLPGM: PRTPL, QMSPL, TACOP, TKPL. $QPGNOT: BLCAL, BPASS, CALIB. Moved nowhere. 9350. May 21, 1997 PCLOD Leonia The default value for FQTOL has confused people. It was exact match. For some reason exact match does not have a place in many cases. At the same times in major cases there is no requirements to the frequency match. That's why I changed the default for FQTOL from exact match to ignore match. It was done adding additional condition in subroutine MATFID. Moved nowhere. 9351. May 23, 1997 POSSM Phil Trivial change to tidy up a format statement. Moved nowhere. 9352. May 23 1997 PARALLEL verb Chris Added a PARALLEL verb to AIPS which controls the number of processors that will be used for AIPS tasks that have been compiled with multiprocessor support. Changed AU2.FOR in AIPSUB and POPSDAT.HLP in HLPFIL. Added ZPARAL.FOR in APLGEN, APLSOL and APLSGI and PARALLEL.HLP in HELPFIL. Moved nowhere. 9353. May 23, 1997 $APLNOT/APLPBI Pat/Martin S. Implemented a bug fix at Martin Shepherd's request: APLPBI was failing to apply the primary beam cutoff to non-VLA (i.e. gaussian) primary beam corrections. Moved nowhere. 9354. May 27 1997 Astrometry Tasks Chris for David Gordon Updated CL2HF, HF2SV and their corresponding help file. Added a HELP ASTROMETRY file giving an overview of astrometry/geodesy processing in AIPS. Moved nowhere. 9355. May 30 1997 PRTAB Eric/Chris Merged in changes to allows column and array-range selection from CVX. Also added BS and GP tables to the list of tables recognized as having TIME columns and fixed the long-lived character transposition in `ASCII'. Moved nowhere. 9356. June 4, 1997 g77-inspired fixes Eric/Jeff $AIPSUB/GETFLD.FOR: Fixed call to GETSTR to pass length argument of 80 vice 132. This fixes the problem with POPSGN et al. crashing under g77 (and sometimes f2c/gcc). $AIPSUB/KWICK.FOR: Increased size of STRING from 80 to 132 to handle max string size now allowed in AIPS. $APLSUB/GETSTR.FOR: Set full out string blank w/o length parameter, use LMAX = min (NMAX, len (outstr)) instead of NMAX. (This is where the stack-stomping was taking place under g77, thanks to the incorrect argument passed in from GETFLD.) Moved from CVX this date. 9357. June 5, 1997 $SYSUNIX/CAD Pat Created a perl script to Check an AIPS Data area (CAD). This will, given a usernumber and a directory name, check for minimal consistency in the files that represent the user's catalog on disk. All slots represented by the existing files are checked to make sure they have a CB file, and either a MA, UV, or SC file as well. Zero-length files and missing files cause it to scream. For each catalog slot in which something is wrong, it lists exactly what it found. When programs get killed unexpectedly (e.g. power outage), this can be invaluable in seeing if one has to clean up a user's catalog with RUN RECAT. Moved to CVX this date (change #9446). 9358. June 10, 1997 $AIPSUB/AU3A, $APLSUB/MDESTR Pat Added sanity checking to these routines so that they only check for realistic extension numbers and types. Logic moved from CVX's FXHDEX today. 9359. June 10, 1997 $YPGNOT/IBLED and .HLP Athol Modify IBLED to allow an overlay plot of model amplitude as derived from an external CLEAN component table. This plot can be selected or de-selected using a new "PLOT MODEL" option of the left-hand menu. The model plotting supports multi-field data, and is useful for editing gravitational lens datasets in particular. Moved nowhere. 9360. June 12, 1997 $APGNOT/SNSMO.FOR and .HLP Athol Miscellaneous fixes and changes accumulated in local versions: i) allow smoothing of tables attached to single source files; ii) add new smoothing option 'DELA' for smoothing of delays only; iii) add feature to control whether existing flagged solution entries are interpolated; activated using BPARM(6); iv) make messages more informative; no message regarding clipping unless selected; message regarding re-referencing now includes percentage of entries re-referenced; v) replace SNVER adverb with INVERS and OUTVERS to avoid preliminary TACOP step; vi) fix what appears to be a bug in SNREF; phase always re-referenced but delay and rate were only re-referenced for VLBI and VLMB smoothing modes. This allows the possibility of partial re-referencing; corrected this. Moved nowhere. 9361. June 17, 1997 $APGNOT/SNCOR.FOR Leonia When SNCORRPRM(1) > EIF, the parameter SNCORRPRM(1) has been forced to one for several opcode where IF's were involved. In particular if OPCODE = 'cpsn', the task has coppied the first IF to the all determined range of IFs, if SNCORRPRM(1) > EIF. If EIF=0 then the condition looks reasonable because EIF is recalculated to the maximum available IF at the data. But if EIF is fixed and EIF is less than maximum available IF at the data the condition provides a mistake copping the first IF instead of desiring one which is more than EIF. To fix this bug I have change the condition (in subroutine SNCCOR) SNCORRPRM(1) > EIF on SNCORRPRM(1) > maximum available IF at the data. Moved nowhere. 9362. June 17, 1997 $APLNOT/POLINO.FOR Leonia The new routine POLINO.FOR fits a polinom up to 19 degree by least square method. Moved nowhere. 9363. June 17, 1997 $APGNOT/ACFIT.FOR Leonia Lincoln Greenhill comlained that algorithm of fitting polinoms to find a baseline of spectra (in ACFIT) gives erronios result on his data. The algorithm used by ACFIT is very cloudly. I decided to write a separate general purpose routine for fiting a polinom to data (see #9362) and in particular to use it at ACFIT. Moved nowhere. 9364. June 20, 1997 $APGNOT/FITLD Phil Some changes for SVLBI and two bug fixes. 1) The VLBA correlator group would like to ensure that FITS tables are no longer than 200 Mbytes and so they split a uv output table into several inside a single FITS file. FITLD could not handle this due to shortcuts taken in the file handling when it was first written. Corrected this. 2) Ensured that the CQ table is not generated for data from correlators other than the VLBA. 3) Penticton is writing the data weights as a spectrum - a legal thing to do. When the code was added for correlation-ids the handling of the weights was disrupted slightly. Fixed. 4) A bug in the concatanation of files with source selection was fixed. The symptom was that source selection was ignored. I'm surprised this one hadn't bitten before, the circumstances in which it was seen were not bizarre. 5) Ensured that the antennas are renumbered from file to file when necessary in the MC table. This is done for all other tables but was missed in this case. Moved nowhere. 9365. June 23, 1997 $YPGNOT/VPLOT Leonia Long time actually since the beginning VPLOT has printed at the header CC table version even if the model is not going to be plotted. It is annoyed people. Now CC table version together with the related image file name is printed only if the model is supposed to plot. Moved nowhere. 9366. June 23, 1997 $APGNOT/CLCAL.FOR & .HLP Athol Merge some CLCAL changes accumulated in local versions into TST. These include: i) add new adverb, CUTOFF, to specify a maximum time interval over which interpolation is to be performed; ii) do not pass blanked CL entries to the output CL table. Modified: $APLNOT/SN2CL.FOR - implement (i) and (ii). $APGOOP/BLAPP.FOR - modified call sequence to SN2CL. Several files in $RUNSYS to include new CLCAL adverb: MYVLA.001 TESTVLA.001 VLACLOAD.001 VLALLOAD.001 CROSSPOL.001 VLBA.001 VLAPROCS.001 Moved nowhere. 9367. June 24, 1997 FNDEXT Athol Continuation of #9349. Update table access using FNDEXT in: $QPGNOT: FRING.FOR, PCAL.FOR and LPCAL.FOR. $QYPGNOT: MX.FOR $YPGM: GREYS.FOR and TVPL.FOR. Moved nowhere. 9368. June 24, 1997 $APLSUB/FILEXT.FOR Athol Add new subroutine to initialize the extension file data in the catalog header, as part of the continuing effort to encapsulate such information so that the table limit can be increased to meet the requirements of Space VLBI. Modify the following tasks to use FILEXT for now: $APGNOT: IMFIT, IMLOD, JMFIT, MULTI and XTRAN. Further updates to follow. Moved nowhere. 9369. June 25, 1997 ZPRMPT.C Jeff The ongoing g77 port turned up a null-termination related bug in GNU readline based prompting; ZPRMPT.C versions fixed under: $APLLINUX, $APLHP, $APLSOL, $APLSGI, $APLSUN, $APLDEC, $APLIBM. Moved nowhere. 9370. June 26, 1997 $APLNOT/IMSORT Phil Extracted a subroutine from FITLD that sorts IM tables into time-antenna order. This is now needed in INDXR also. Moved nowhere. 9371. June 26, 1997 $APLNOT/GETDEL Tony Beasley/Phil Tony rewrote GETDEL to take advantage of the fact that IM tables will now be sorted before entering GETDEL. In addition he improved its extrapolation behaviour to avoid the GETDEL error messages we see on occasion due to a small number of VLBA records being > 2 minutes from an IM entry at the end of correlator jobs. Moved nowhere. 9372. June 26, 1997 $APGNOT/INDXR Phil Now calls IMSORT before calling GETDEL. Moved nowhere. 9373. June 26, 1997 $APGNOT/FITLD Phil Subr. IMSORT extracted and placed in $APLNOT. Moved nowhere. *********************************************************************** Changes while 15APR97 was NEW: *********************************************************************** *********************************************************************** 9374. July 9, 1997 $YPGM/PCNTR Phil I got fed up never knowing what p.a. rotation had been applied to a particular image so I added the ROTATE value used to the information in the plot file. Moved from 15OCT97 today. 9375. July 9, 1997 $YPGNOT/KNTR Phil While in the mood I decided to remove the irritating message concerning the block being plotted in KNTR. This is not really necessary. Moved from 15OCT97 today. 9376. July 9 1997 INDXR Chris Increased the size of the buffer used to sort the IM table and shuffled LUN assignments to avoid clashes with TABSRT. Moved from 15OCT97 on this date. 9377. July 9 1997 GPSDL Chris Disabled delay corrections (which are not reliable), leaving only the Faraday rotation correction, and made some changes to the diagnostic output. Moved from 15OCT97 on this date. 9378. July 9 1997 $APGNOT/UVMOD Ketan UVMOD now will process compressed data. Moved from 15OCT97 on this date. 9379. July 10 1997 LPSEC in APLNOT Chris Split LPSEC from UVFIX to make it easier to find out where leap seconds should be inserted. Also made it more compact added the leap second for July 1 1997. Moved from 15OCT97 on this date. 9380. July 11, 1997 $QNOT/GRIDUV Rick/Eric/Pat There was no lower bound check on the chunk of the UV data read into the AP in cases where it would not all fit, and the next visibility was so far down in the second chunk that the conv. fn. did not overlap into the first chunk. This caused negative offsets into an array in such cases with disastrous results. Spotted by Rick, fixed and tested by Eric, implemented here by yours truly. Moved from CVX and 15OCT97 this date. 9381. July 16, 1997 $APLNOT/CALREF Ketan CALREF accepted the FREQOF parameter as a REAL when it is actually sent in as a DOUBLE PRECISION. This effectively rendered some parts of the code in CALREF non-functional. The declaration of FREQOF was changed, and the relevant code was commented out. Moved from 15OCT97 this date. 9382. July 16, 1997 $QPGOOP/IMAGR Ketan $QPGNOT/CALIB When operating on a multi-source file, if a single source name was not specified, IMAGR and CALIB did baddd things. Now they warn that they will do baddd things. Moved from 15OCT97 this date. 9383. July 16, 1997 $APGNOT/MSORT Ketan In the case when there are no records to sort [the file _is_ in the proper sort order] MSORT incorrectly left an empty file behind. This has been fixed. Moved from 15OCT97 this date. 9386. July 18 1997 CL2HF Chris for David Gordon Merged in changes allowing CL2HF to read items from the interferometer model (IM) table for VLBA data. Moved from 15OCT97 this date. 9389. July 25 1997 POSSM Chris Corrected a manifest real constant in a call to BPCOEFF where a double-precision value was expected. Moved from 15OCT97 on this date. 9390. July 25, 1997 $APGNOT/SNSMO.FOR & .HLP Athol Corrected a re-referencing bug affecting multi-IF calibration tables for the second secondary reference antenna and above. Subroutine REFFAZ was modifying the table column pointers directly (RF1KOL etc) thus causing problems when REFFAZ was called for secondary reference antennas other than the very first one encountered. Use local variables in REFFAZ to avoid this. Also corrected a typo. in the ref. antenna selection in REFRAT, and the update condition in RFFXFG. Added new subroutine RFFIX to clean up after re-referencing, but this should not be required in normal processing. All the above are long-standing problems. Only the REFFAZ error is likely to have been significant, causing outliers in the smoothed output table. The effect of the other bugs is limited. Also changed BPARM(6), introduced recently to DOBLANK. Moved from 15OCT97 on this date. 9391. July 25, 1997 UVSUB Chris Rearranged a common block to avoid data alignment problems. Moved From 15OCT97 on this date. 9392. July 29, 1997 New task: CLFIX Chris for David Gordon Added CLFIX to the APLCONTR (contributed) area and updated ASTROMET.HLP to refer to it. CLFIX adds geometric and atmospheric delay information and clock offsets from IM and MC tables to CL table 1. This is only needed as a stop-gap until INDXR is able to read information from VLBA MC tables. Moved from 15OCT97 on this date. 9393. August 4, 1997 $QPGOOP/IMAGR Ketan Change #9382 was improperly implemented in that IMAGR printed dire warnings even if sources had been selected. The warnings are now only printed IF more than one source makes it past the OOP data selection system. Moved from 15OCT97 on this date. 9394. August 6, 1997 $QPGOOP/OMFIT Ketan OMFIT has been rewritten in many ways. Many new models have been included and subjected to user testing. The error-analysis has been considerably improved. This is still development code and problems should be pointed directly to me. Moved from 15OCT97 on this date. 9395. August 6, 1997 $QPGOOP/IMAGR Ketan $QPGNOT/CALIB Changes 9382 and 9393 [which merely printed out warnings] are to be back'ed out of NEW pending further testing of the claimed effects. Ive used the RCS command co to revert the NEW versions of NEW and TST to the pre-#9382 state. Moved nowhere. 9396. August 7, 1997 $QPGOOP/OMFIT.FOR Ketan Fixed incorrect statement involving complex variables spotted by the DEC and the IBM [via Pat]. Moved from 15OCT97 on this date. 9397. August 7, 1997 $HLPFIL/OMFIT.HLP Ketan Last minute changes contributed by Michael Rupen. Moved from 15OCT97 on this date. 9398. August 8 1997 FILLM Chris Fixed a bad test condition that would turn off solar mode. Moved from 15OCT97 on this date. 9399. August 7, 1997 $SYSUNIX/OPTIMIZE.LIS Ketan Turned down the optimization for OMFIT on the IBMs to zero. This makes OMFIT run like a dog, it is recommended that anyone wishing to run OMFIT on an IBM compile the code in pieces. Contact me directly for more info. Moved from 15OCT97 on this date.