----------------------------------------------------------------------- ; Copyright (C) 1995 ; Associated Universities, Inc. Washington DC, USA. ; ; This program is free software; you can redistribute it and/or ; modify it under the terms of the GNU General Public License as ; published by the Free Software Foundation; either version 2 of ; the License, or (at your option) any later version. ; ; This program is distributed in the hope that it will be useful, ; but WITHOUT ANY WARRANTY; without even the implied warranty of ; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ; GNU General Public License for more details. ; ; You should have received a copy of the GNU General Public ; License along with this program; if not, write to the Free ; Software Foundation, Inc., 675 Massachusetts Ave, Cambridge, ; MA 02139, USA. ; ; Correspondence concerning AIPS should be addressed as follows: ; Internet email: aipsmail@nrao.edu. ; Postal address: AIPS Project Office ; National Radio Astronomy Observatory ; 520 Edgemont Road ; Charlottesville, VA 22903-2475 USA ----------------------------------------------------------------------- This file must be used to record any changes made to the the master version of AIPS kept in Charlottesville. An entry is: line 1: Date Program name Your name line 2: What was done line 3: Where has this change been moved (e.g. 15OCT90) line 4: Blank The details of line 1 are important since programs are used to find specified strings and do conversions on them. It MUST be typed as follows: 1 blank, 4 digits, a period, two blanks, then the Month, day, comma, year, some space, a title, some more space, your name. Changes to 15JUL94 *********************************************************************** Changes while 15UL94 was TST: *********************************************************************** *********************************************************************** 7974. January 28, 1994 $HIST/CHANGE.DOC (this file) Pat Initialized this file as part of the quarterly update. Moved nowhere, but also edited NEW (15JAN94) version. 7975. January 28, 1993 BLING Chris Fixed an inappropriate use of IDAMAX that caused gridding problems if any IFs had negative offsets. This should only have affected VLBA correlator data in multi-band solution mode. Moved nowhere. 7976. January 30, 1994 AIPS, VERBS Eric Moved the VERBS* subroutines inside the AIPS* program files. VERBSC referred to some routines which are in AIPSC only, causing unresloved external references. The VERBS routines are only used by the single programs anyway. Moved to 15JAN94 this date. 7977. January 30, 1994 ZDCHIN Eric Added the PAPC include and tests on KAP2WD. This parameter was being used before the call to QINIT and hence the software did not have a chance to defend itself from bad settings in the SP file. Now one should set the SP file to the largest for any of the 3 versions and the others should be able to take cae of themselves. Moved to 15JAN94 this date. 7978. January 31, 1994 LIBR Eric Changed most of the LIBR.DAT files - they specified $AIPSUB for $AIPGUNIX which is incorrect. The $SYSUNIX one and a couple of others had already been corrected. Moved to 15JAN94 yesterday. 7979. January 31, 1994 PRTLIN, PRTALN Eric/Dave Changed PRTLIN and PRTALN to call ITRIM rather than JTRIM for the page header line. JTRIM was dropping the line due to the special character (form-feed) we wanted in the line. Created WORD to get the first word in a character buffer. This lets PRTLIN and PRTALN give the first word of the system name without any of the "junk" that follows (i.e. give "rhesus " rather than "rhesus ("). Moved to 15JAN94 this date. 7980. January 31, 1994 MX, ZEROSP Eric Dropped the option of allowing ZEROSP(2-4) to give the widths of a finite gaussian as the zerospacing. It takes too many cells to represent one for the gridding to handle and messes up the beam. We need to meet this need in other ways - placing extra vis samples in the uv data set. Also shortened long line in MX output. Moved to 15JAN94 this date. 7981. February 1, 1994 FILLM.HLP, BAND.HLP Gustaaf Added the Q-band information to the HLP files. FILLM already handled Q-band for some time, but there was no way for the user to know. Though not in the 'urgent' category, moved it to 15JAN94 nonetheless. 7982. February 1, 1994 TBSUB, TABLE Eric Corrected TABLE.FOR subroutine TABOPN to set EXISTS = STATUS.EQ.'READ'. Otherwise, it set NCOL et al. to zero when trying to create a table (with a default OUTVERS). It then failed of course. Fixed two character declarations in TBSUB, one of which caused the first character of each column to appear as the column label of the first output column, and the other output labels were blank. Moved from 15JUL94 this date. 7983. February 1, 1994 UVCOP Phil UVCOP forced the first IF-offset in the FQ table to be 0.0 even when no IF selection or freq. scaling was being done. Added an IF statement that forces this only when necessary. Moved to 15JAN94 this date. 7984. February 1, 1994 FRING Phil FRING selects/calibrates/flag etc the data to a scratch file and does the fitting from that scratch file. However it sets up its solution intervals based on the NX table from the original input file. When any data selection was done the solution intervals could be wrong because there was no check made to see if the NX record read was a deselected one. Added those checks and we now get the correct solution intervals on short-scan VLBI data. Moved to 15JAN94 this date. 7985. February 2, 1994 $APLOOP/UVUTIL.FOR Bill C. UVCLIP wasn't handling the case of doing Ipol when one of the two parallel hands was already flagged. Now Ipol is computed from either of the two parallel correlators if only one is good. Moved to 15JAN94 this date. 7986. February 2, 1994 $APGNOT/MAPBM Bill C. New task. Uses a VLA raster made in holography mode (with no reference antennas) and makes images of the BEAM in I, Q, U and V with RMS uncertainties in each. The input is a calibrated multisource uv data set and the total and polarized intensities of the source. In Q, U and V the resultant images are fractional polarization and the Q and U images can be used by VLABP to correct widefield snapshot VLA images for off axis instrumental polarization. Also help file. Moved to 15JAN94 this date 7987. February 2, 1994 $QPGNOT/VLABP Bill C. New task. Task to apply widefield instrumental polarization corrections to a snapshot VLA polarization image. The imput images can be either a set of I, Q and U images or a polarization cube containing the three and a set of instrumental polarization images. This task interpolates and rotates the instrumental polarization image to align with the observed image, rotates between Q and U to counter the effects of the on axis calibration of the off axis instrumental response, multiplies times the Ipol image and subtacts from the (Q + iU) image. The parallactic angle of the observation can be passed either as a header keyword or an adverb. This will ONLY work for snapshot images. Also help file. Moved to 15JAN94 this date 7988. February 2, 1994 $QPGNOT/FRING Phil The FRING fix I put in yesterday did not check for the code running off the end of the NX table. Now does. Moved to 15JAN94 this date. 7989. February 2, 1994 FITLD Phil 3 small changes. 1) allow FITLD to look at more than one disc when deciding on concatanation; 2) Benson has moved the correlator history records to the base header (where they should have been all along) - added code to ensure that the string FITLD is appended to these records; 3) when an antenna number is changed during concatantion and as a result a baseline direction has to be changed it was not doing it, fixed now. Moved to 15JAN94 this date. 7990. February 3, 1994 $SYSSOL/LIBR.DAT Pat The $AIPGUNIX area only references two libraries: APLSUB and APLSOL (in this case). It needs to repeat these for the loader to resolve all references, at least when shared libs are not used. Moved to 15JAN94 this date. 7991. February 3, 1994 $SYSSUN/COLOR,PRINT Pat These shell scripts were never part of AIPS formally. Modified them to use the AIPS printing scheme, automatically selecting the first appropriate printer if none is specified, and allowing the PR=n command line printer selection. Moved to 15JAN94 this date. 7992. February 3, 1994 AJAX.FOR Pat As part of the clean-up (sorry, couldn't resist!) I do with the quarterly update, I noticed that this task failed to link on Solaris. It has a call to DESCR which is in AIPSUB, and AIPNOT tasks are not supposed to call AIPSUB routines. So I moved it to AIPPGM. This cures the problem. Moved to 15JAN94 this date. 7993. February 4, 1994 WFCLN.FOR Bill C. Fixed serious bug in WFTAV which averaged the times in the uv data set to determine the average observing geometry. This was being done in solar time which caused serious problems if there were observations over several days. Now it averages in sidereal time modulo 1 day. Moved to 15JAN94 this date 7994. February 4, 1994 ZLASCL, ZLPCL2 Pat Made yet another exception to the feeble attempt at distinguishing SYSV systems from Berkeley (the TZ variable). Linux may or may not have it but it always has lpr/lpq. Moved to 15JAN94 this date. 7995. February 7, 1994 BLAPP Chris Increased the singular value threshold from machine epsilon to NDATA * epsilon. Also added a warning message if the solutions for any antennae are unconstrained. Moved nowhere. 7996. February 7, 1994 ANCAL.HLP Joan/ Phil Removed the erroneous section on calibrating the VLA for VLBI and replaced it with a reference to the Tsys file's documentation. Moved nowhere. 7997. February 8, 1994 FITTP, UVLOD, FITLD ERIC Changed FITTP to leave the projection type on the U, V, and W random parameter. It does this as 'UU -SIN' for DONEW false and as 'UU---SIN' for DONEW true. Changed UVLOD to recognize either 'UU ' or 'UU--'. FITLD will be changed by Phil. Moved all 3 to 15JAN94 tomorrow. 7998. February 9, 1994 $APLNOT/SOUELV.FOR Leonid Changed coefficient before STNLON (1.002738D0 => 1.0) at the line ANTLST... Moved nowhere. 7999. February 10, 1994 FITLD Phil Put in Eric's change (#7997) to deal with projection types. Also fixed a minor bug that caused an antenna to be left out of the CL table. Moved nowhere, Eric did NEW. 8000. February 10, 1994 TCOPY Eric Added the trick, already in FITTP, of doing a BAKF before an output tape close in order to prevent IBM from appending an extra EOF after the one we just wrote. This acts as an EOI (double EOF) for most tape opertations. But an append will go after the 2 EOFs, BAKF after 2 EOFs goes over both, and other confusions. Moved nowhere yet - should go to NEW. 8001. February 11, 1994 $SYSUNIX/TVDEVS.SH Pat In Multi-site configurations (i.e. where there is more than one SITE specified in HOSTS.LIST; this is useful when you want opposite endian architectures to share the same $AIPS_ROOT) the selection of HOSTS, TVSERVS and XTERMS was not discriminating on the SITE name. Fixed via a simple "grep". Will move to NEW momentarily. 8002. February 14, 1994 $APGNOT/AFILE.FOR Phil In the sorting section I was not sorting the last scan, fixed to now do this. Moved nowhere. 8003. February 15, 1994 $APGNOT/ACFIT.FOR Leonid There was a confusion at the SN table at the multy IF case. The task could not find a solution if a specified timerange was not coincided with the time interval of the template file. Moved nowhere. 8004. February 22, 1994 $APGNOT/ACFIT.FOR Leonid The different Tsys for different IF are available now as an array DPARM(10). Moved nowhere. 8005. February 23, 1994 $APGOOP/RFI Bill C. Fixed a bug in applying the amplitude proportional rms cutoff level and modified to look at the rmses of the real and imaginary parts of the visibilities. This catches interference which has a constant amplitude but a variable phase. Also RFI.HLP. Moved nowhere. 8006. February 24, 1994 $APGNOT/CLCOR Leonid Added a correction of source position into opcode 'ANTP'. Changed local hour angle on hour angle of Greenwich in equations for delay, phase and fringe rate. The using of local hour angle instead of greenwich one had to give uncorrect result at correction of delay, phase and fringe rate for the errors of antennas position. Moved nowhere. 8007. February 24, 1994 $APGNOT/CLCOR Leonid Changed the units of RA correction from sec of time to sec of arc. Moved nowhere. 8008. February 24, 1994 $APLLINUX/ZDCHI2.FOR Pat Modified TTYCAR to say that Fortran carriage control is not needed. Moved nowhere. 8009. February 27, 1994 $APGNOT/SPLIT.FOR Phil 2 changes to SPLIT curiosuly both in the same area. (1) When IF and/or channel selection is performed the ref freq is updated and the uvw's are scaled on output. However the frequency in the AN table was not changed. This could cause problems further downstream if position offsets were inserted using UVFIX, because UVFIX used the An table freq during the shifting process. Fixed so that output AN freq and ref freq are the same for AN table 1. Only do this to subarray 1 because do not really know how other subarrays are effected. (2) When DOPOL = 1 in SPLIT the polzn corrections are applied, however the feed corrections are written to the output AN table during the SPLITing process, thus if DOPOL is left as true the feed corrections can be applied twice. So now reset the polzn corrections in the output AN table. Moved nowhere - do we need a patch for this? 8010. February 28, 1994 FILLM Eric Changed FILLM tape read error handling to accept error 6 (end of information) as morally equivalent to a double end of file. The SUN driver on 9-track tapes returns eoi on double eof. Moved nowhere - should go to patch. 8011. March 2, 1994 $APGNOT/LISTR Leonid Add listing of amplitude of auto correlation at OPTYPE = 'MATX'. Fix a bug in polarisation. The blank value of STOKES had not worked when a data had only one polarization. Moved nowere 8012. March 3, 1994 $YPGNOT/CLPLT Phil Added the ability to plot closure phase/models against GST. Moved nowhere. 8013. March 4, 1994 $APGNOT/SLICE.FOR Dave Added the option to specify an OUTFILE to print out the slice information. Moved nowhere. 8014. March 7, 1994 $APLSUB/DCOPY Phil New routine to copy dp variables from one array to another. Odd we were missing this one. Moved nowhere. 8015. March 7, 1994 FRING Chris Reduced the sizes of the accumulation arrays VIMAG, VREAL, TIME and WTT after noting that they were dimensioned on the assumption that MAXTIM.LE.XASIZE when the stronger condition MAXTIM.LE.(XASIZE/2) holds. This reduces the size of the BSS segment by 2 million words and lessens the chances of blowing virtual memory by running FRING. Moved nowhere. 8016. March 8, 1994 IM* routines Phil For some reason the IM tables being written by the correlator did not match those expected in AIPS, in fact neither matched the description in the VLBA FITS definition. That description was deemed correct and so Benson changed what he wrote, labelling the new format tables as revision 2. DIMV.INC now reflects the correct version also. Fixed up IMINI and IMTAB to read the correct version, also fixed up IMINI to set the correct # columns when creating a new table. Moved nowhere. 8017. March 8, 1994 $APLNOT/IMREF1.FOR Phil New routine to translate IM revision 1 to IM revision 2. Moved nowhere. 8018. March 9, 1994 $YPGNOT/POSSM.FOR Phil Due to an unintialized variable, when the user selected IFs > 1 to be plotted and APARM(9) = 1 (i.e. all IF's on the same plot) then POSSM plotted the correct number of IFs but started at IF #1. Fixed. Moved nowhere. 8019. March 9, 1994 FRING Chris FRING now issues a warning message if any least-squares fit terminates due to non-convergence. Previously, it would silently move on to the next IF leaving the user to wonder what happened to his solutions. Moved nowhere. 8020. March 10, 1994 PRTTP Eric Changed TAPIO to check for end-of-file on read completions when doing ADVF and BAKF. If an EOF occurred, then do no ADVF or 2 BAKFs, rather than always doing 1 ADVF or BAKF. Changed PRTTP help to offer 2 new adverbs: BCOUNT to let the user specify the file number when NFILES > 10000 (i.e. start where the tape now is) and DOEOT to specify how many consecutive null files are allowed before the program quits. Changed PRTTP to use these options, to force all error messages to the printer file, to try to continue running on most errors (such as premature EOF), to do something useful with unrecognized formats (i.e. list the record lengths), to apply more tests before assuming the format is an "IBM map", and to avoid an abort on an empty table. Moved nowhere except to offer it in the patch area. 8021. March 10, 1994 FRING Chris Corrected and moved a test which checked the number of frequencies and IFs in FRNFIT. The original test used the number of selected frequency channels multiplied by the number of selected IFs where the number of selected channels would have been appropriate (and therefore caused FRING to abort when it need not) and was applied after the frequency look-up tables were filled in (so that array bounds could be exceeded before the test was applied. Moved nowhere. 8022. March 10, 1994 FITLD Phil Yet another series of changes. One day this thing will settle down, if only users would stop asking for more features. (1) Gave the user control over the frequency tolerance for deciding if files should be concatanated. Default is a 2MHz difference in the reference frequencies, maximum is 500 MHz (i.e. a front-end bandwidth). (2) Changed the maximum number of freqid's from 100 -> 10. (3) If user selects source(s) and they are not present in the SO table for a given file then that file will be skipped rather than each record read and chacked. (4) If the file on disc has no visibilities after FITLD then it will be deleted on exit, especially useful when doconcat = -1 and many files are created. (5) The SU table will be tidied up aftre FITLD has run. This means that sources which do not occur in the uv-data will be deleted from the SU table. (6) The IM tables are now edited on the fly as they are read in, i.e. the source and antenna renumbering are performed, and also the user specified data selection is applied. This will be done to more tables on the next round of editing. (7) the master VLBA table reading routine was tidied up. (8) If IM revision 1 tables are read in they will be reformatted to revision 2. (9) Avoided the annoying little message from UVPGET about the freq being zero when the uv-data file is created. (10) User now gets a message about which source is currently being read, what time the source was encountered and whether the source number was changed due to remapping. (11) The IFs are now reordered so that the output IFs are in increasing frequency order. The cross-bar switching routine is a little on the hairy side. All IF dependent varaibels on the tables are also reordered. All this is done efficently by setting up an indexing array that indexes the output position to the input position. Moved nowhere - too much for a patch also, but those who need it should get it via MNJ. 8023. March 11, 1994 $APLUNIX/ZLASC2.C,ZLPCL2.C Pat These routines use popen()/pwrite()/pclose() to perform a shell command (sleeping for n seconds, then delete the file already spooled to the printer). On AIX, the return from pclose() can be -1 and the errno set to ECHILD if whatever piped to pwrite() caused a wait() or waitpid() and I guess a sleep from the shell does that. This was causing the temporary files in /tmp to not be deleted which in turn can cause grief if space becomes short. Now the routines ignore ECHILD but report other pclose() errors. Moved nowhere but probably will be in the patch area soon. 8024. March 11, 1994 $APGNOT/FITLD Phil When reading a series of files, if an image file was read then FITLD would check the # visibility points, decide it was zero (because it was an image) and then delete a random uv-file on disk. Fixed to not do this. Also increased the max # files that can be read to 9999. Moved nowhere. 8025. March 14, 1994 TVCPS, LWPLA Eric Changed Fortran and help files to have the output of both tasks always be encapsulated postscript. This meant adding a few words to LWPLA.FOR and dropping the non-encapsulated PS option from TVCPS.FOR. Changed both help files to reflect these and added the Adobe reference and copyright to LWPLA.HLP. Moved nowhere. 8026. March 14, 1994 INPUTS, GO, TV verbs, FITTP Eric Changed AU1A and AU2 to trap BLC, TRC, TBLC, and TTRC and force them to have an upper limit of 8192. Previously, they had either 4096 or infinite and I did not want to change 95 help files! Changed buffer sizes in AU5A (TVLOD), AU5D (TVMOVIE), AU6D (IMSTAT), FITTP, and IMLOD to allow 8192 size images. Moved nowhere. 8027. March 14, 1995 FITLD Phil Stupid error. The version of FITLD I put back before had an older version of subroutine DETORD that caused the data to come out in decreasing freq order rather than increasing freq. order. Fixed. Moved nwohere. 8028. March 15, 1994 FRING Chris Increased the dimensions of VREAL and VIMAGE incorrectly decreased on March 7th. Moved nowhere. 8029. March 15, 1994 $APLSUB/DPCOPY Phil Discovered that there is an APLNOT routine called DCOPY so changed the APLSUB version to DPCOPY. They do different things. Moved nowhere. 8030. March 15, 1994 $APLNOT/PSHIFT & GETDEL Phil 2 new routines. GETDEL obtains the group delay from an IM table and shifts it with apolynomial shift (evaluated by PSHIFT) to the time grid in a CL table. It then updates the geometric delay in the CL table with the shifted value. Follows the same algorithm as used by the VLBA correlator. Moved nowhere. 8031. March 15, 1994 $APGNOT/INDXR Tony Beasley/Phil Added the ability to INDXR to use the information in the IM table to recalculate the geometric delays in the CL table if CPARM(4)=1. This will be useful if the user decided that the CL interval they used in FITLD was wrong and they need to regenerate the CL table with a different interval. Also updated the help file. Moved nowhere. 8032. March 15, 1994 FITLD Phil Changed to remove PSHIFT, GETDEL & DPCOPY. They were internal subroutines that now live in APLSUB & APLNOT. First change in a while that made FITLD smaller. Moved nowhere. 8033. March 16, 1994 ZCREA2.C Pat Four flavors of this routine ($APLBERK, $APLDEC, $APLULTRX, $APLSGI) try to figure out if there's enough space for the file creation requested (the others just die if they can't create it). These were all calculating the number of bytes available on the filesystem and using long ints. For 32-bit systems this will overflow if the free space exceeds 2 Gigabytes. Modified the calculation so that blocks are used instead. For systems where a block is 512 bytes or more, this ups the limit on filesystem size AIPS can handle to 1 Terabyte. File size is still limited to 2 Gigabytes though, for 32-bit versions of AIPS. Moved to the patch area this date. 8034. March 16, 1994 COMB Bill C. Added call to KEYCOP to copy header keywords from the first input image to the output image. Moved nowhere. 8035. March 16, 1994 $APGNOT/INDXR Phil Ensured IM table only opened when geometric delay being recalculated. Moved nowhere. 8036. March 16, 1994 $APLNOT/TABHDR Phil Added 2 new tables to the lookup table that decided the 2 letter code. The 'CALC' table will become a CT table, the 'MODEL_COMPS' will become an MC table. Also renamed the antenna charactersistcs table (AC) to AT in order to avoid confusion with the accounting file. Moved nowhere. 8037. March 16, 1994 AC -> AT Phil Routines ACINI, ACIO & ACTAB renamed to ATINI, ATIO & ATTAB to reflect the change of the AC table to the AT table. Also changed DACV.INC to DATV.INC. Removed the now unused AC... files. Moved nowhere. 8038. March 16, 1994 FITLD Phil Modified to reflect the AC -> AT change. Moved nowhere. 8039. March 16, 1994 $APGNOT/DBCON.FOR Tony Beasley/Phil When the first database in the pair has its reference freq and pixel changed this was not reflected in the reference frequency in the AN table. Modified it also update the AN frequency. Moved nowhere. 8040. March 17, 1994 $APLPGM/PRTTP.FOR Gustaaf The maximum number of entries PRTTP can handle was increased from 100 to 2000. Reason for this is that on export tapes sometimes there is a large number of small dummy files. This happens when during observing the observe file is deleted (this happens!). 8041. March 18, 1994 $APLOOP/BEAM.FOR Bill C. Added member function to copy the beam from one object to another. Moved nowhere 8042. March 18, 1994 $QPGOOP/VLABP.FOR Bill C. Copy Beam object from input to output. Moved nowhere. 8043. March 21, 1994 Various tasks, scripts Pat Received an extensive installation report from Issei Yamamura and cleaned up the various little bugs, etc he found. All of the fortran problems were due to missing commas in format statements [hey guys, watch out for them!] and the LDOPTS.SH and FDEFAULT.SH scripts needed some extra options. Plus the install docs had a goof. Moved nowhere. 8044. March 21, 1994 $YPGNOT/POSSM Phil Increased an array size in POSSUV, the 512 channel, 4 IF data now available from the VLBA correlator cause problems otherwise. Moved nowhere. 8045. March 22, 1994 $APGNOT/POSSM.FOR Phil In change 8009 I did not make the polzn cal value arrays large enough and as a result initializing them in some cases caused an overflow that caused TABIO to fall on its face. Fixed. Moved nowhere, will put in the patch area. 8046. March 22, 1994 $APGNOT/FITLD Phil The handling of IM tables and their reformatting was very clumsy. I tidied this up extensively with the following list of changes. (1) Parameterized the current revision of the IM table for easy changes in the future. (2) IF the IM table needs reformatting, the old table is now read to the scratch file that contains the HI information, is reformatted in place there and then attached to any existing IM tables already in the main file. This avoids the need to do a lot of copying of tables and general pissing around. (3) The new version of $APLNOT/GETDEL removes the need to sort the CL and IM tables into antenna-time order. Again cuts down on the book-keeping and table shuffling. Moved nowhere. 8047. March 22, 1994 TVCPS, LWPLA Eric Make the scaling parameters more flexible and add an informative message to LWPLA. Add adverb DOINVERS to TVCPS to reverse the coloring from white on black background to black on white background. Some printers do not make a uniform black background well. Also altered both help files. Moved nowhere. 8048. March 22, 1994 $QYPGNOT/VTESS.FOR Dave VTESS was not succesfully writing the history file of the convolved image (the .VTC file); PRTHI was saying that there were zero records. It turned out the VTC HI table was never being closed. Added a call to HICLOS which seemed to do the trick. Moved nowhere. 8049. March 22, 1994 $APLNOT/KEYIN.FOR Phil Have received many requests for ANCAL to be able to read Tsys data for 14 IF's. The 80 character limit in KEYIN prevented this. Changed it to 200 characters and all now works well. Moved nowhere. 8050. March 23, 1994 $YPGNOT/SNPLT.FOR Phil In the IFDF mode SNPLT was plotting values for which the value in the 2nd IF was indefinite. This slipped through with no problem because the variable VALUE was not reset before the IFDF test. Moved nowhere. 8051. March 24, 1994 OHGEO.FOR Bill C. The program was not copying header keywords or the convolving beam from the input. Moved nowhere. 8052. March 25, 1994 SNPLT.HLP Dave A. Clarified the OPTYPE options in the inputs, per a gripe suggestion. Moved nowhere. 8053. March 27, 1994 TVRGB, TVHUI Eric Corrected both tasks to do enhancements properly. Previously, the formulae used went the wrong direction from the standard and with very little sensitivity to cursor position. Also improved the coincidence checking to say more when the images appear not to be coincident and to be more careful about image dimensions when checking axes 3 through 7. Changed TVRGB to display pixels at which one or two of the images are blanked in the remaining colors. Moved to the patch area for 15JAN94. 8054. March 27, 1994 TVCPS Eric Changed program and help to offer the user the ability to put a string at the bottom of the picture (appended to the user number string if that is displayed). Changed both to have the program read the image from disk, with TBLC, TTRC, TXINC, and TYINC applied, using the image name, transfer function, LUTs, and OFMs currently displayed on the TV. Corrected the handling of graphics overlay colors to match the scheme actually used in AIPS TVs. Moved nowhere. 8055. March 28, 1994 UVPRT Pat A couple of format statements had I5 specifiers for a real variable (internal writes). This causes problems at run-time on the Dec Alpha (surely elsewhere too?) and was caught by JingPing Ge at Brandeis. I replaced the USERID variable in these writes with IUSER. Moved nowhere. 8056. March 28, 1994 GTTELL Eric This subroutine prohibited batch aips numbers from using TELL. Since one is now supposed to be able to TELL things to batch, changed it by dropping the prohibition. Moved nowhere. 8056. March 28, 1994 FRING Leonia The program stopped occasionally complaining of too large a solution interval informing the user at the same time that this interval is 0.00. The problem (rounding error) was resolved by moving two pairs of lines to their correct places. This caused the end of scan to be adjusted by 2sec as originally intended. Moved nowhere - Phil will place it in the patch area. 8057. March 29, 1994 $APGNOT/UVCOP.FOR Phil Added a subroutine to tidy up the SU table after source selection has been performed. A count is kept of which sources are written into the output file and then at the end of the run the SU table is edited so that only those sources that remain in the output file are actually kept in the SU table. Moved nowhere. 8058. March 29, 1994 KNTR Eric Fixed the number of parameters - it was ignoring GRCHAN and TVCORN. Corrected the handling of labeling parameters and LOCNUM. Now it uses LOCNUM = 2 for plotting star positions independent of LTYPE. It uses LOCNUM = 1 to plot axes affected by LTYPE, but saves and locally handles the axis TYP and PREFix strings to control tick labeling. Moved to the patch area. 8059. March 29, 1994 $APGNOT/FITLD Phil Removed the ability to read VLBA archive tapes, it was only in as a temporary measure. Distribution tapes are the principal mode of data export. I also tidied up the death agonies FITLD goes through when encountering bad data - it tries very hard to leave the file in a state that can be appended to once FITLD is restarted. Moved nowhere. 8060. March 30, 1994 VLABP.FOR Bill C. The interpolated antenna polarization scratch file was not being having it's geometry properly defined. It now calculated the correct reference pixel. If OUTSEQ=0 and there already existed files with the same name and class then the corrected Q image was written to one of these if the input was a polarization cube. This is now fixed. Moved nowhere. 8061. March 30, 1994 OHGEO.FOR Bill C. It was computing the reference pixels of the output image and storing them BEFORE coping the image info (including the reference pixel) from the second input image. The order was switched. Apparently correct behavior in the distant past is mystifying. Also now keeps original observing position. Moved nowhere. 8062. March 31, 1994 POSITION.FOR Bill C. Added member functions PSNPIX, PSNVAL to find the pixel corresponding to a coordinate value or the coordinate value corresponding to a given pixel in an image. Moved nowhere. 8063. March 31, 1993 VLABP.FOR Bill C. Improve method of determining pixel of original pointing center using PSNPIX. Moved nowhere. 8064. March 31, 1994 FFT Eric Corrected call to DSKFFT. The NX and NY parameters were switched, messing up rectangular images. Moved nowhere. 8065. March 31, 1994 FITLD Phil Oops. When removing the OPCODE parameter the other day I mis-edited the # input parms expected and so it no longer read the user supplied freq. tolerance. Fixed. 8066. April 1, 1994 FITLD.HLP Bill C. Dispite previous claims to the contrary the limit on the number of images was 999 rather than 9999. Fixed. Moved nowhere. 8067. April 4, 1994 UVUTIL.FOR Bill C. Self cal phase rereferencing routine SLFREF had a logic error for single polarization, multi IF data; the SN table pointers passed to REFFAZ for IFs higher than the first indicated the presence of a nonexistant second polarization. Mostly affects SCMAP. Moved nowhere. 8068. April 5, 1994 $APLPGM/SQASH.FOR Dave A gripe reported that SQASH was not reporting the correct BLC,TRC in the history file. Changed the array printed in the history file from ISWIN (which is used as a counter) to IDWIN (which remains unchanged). Moved nowhere. 8069. April 5, 1994 $APGNOT/QUACK.FOR Dave A gripe had pointed out that QUACK was not reporting the proper APARM to the history file. Changed APARM(1) to the proper APARM(2). Moved nowhere. 8070. April 6, 1994 $APGNOT/UVFIX.FOR Gustaaf UVFIX will now automatically take frequency dependence into account when applying shifts to multi IF, multi channel data. In cases where speed is deemed to be more important than precision, this new method may be switched off using UVFIXPRM(9). Moved nowhere. 8071. April 6, 1994 $YPGNOT/CLPLT.FOR Leonia The program plotted only one point if DOCALIB=TRUE and frequency channels have been averaged. It occured because NUMPOL was corrupted in subroutine CLOSAV when it was used as input parameter for subroutine AVGCIF. NUMPOL has been changed for NUMTEM at this place Moved nowhere. 8072. April 7, 1994 $SYSUNIX/TVALT Pat The GNU version of "grep" has different qualifiers to the regular grep on Linux, Dec systems and no doubt others too. Removed all references to "egrep -s" or "egrep -q" and just do a simple "grep" with stdout and stderr redirected to the bit bucket. Moved nowhere. 8073. April 7, 1994 $APGNOT/FILLM.FOR Gustaaf In realtime filling, the default subarray was not clearly specified. An extra warning was added, and the HLP file was updated. Moved nowhere. 8074. April 8, 1994 $APGNOT/FILLM.FOR Gustaaf Oops. I had added some zero length strings and the Dec alpha compiler did not like that. Moved nowhere. 8075. April 11, 1994 $APLDEC ZTAP2, ZTPWA2 Eric Created an $APLDEC version of ZTPWA2 with full mt status fetching and error handling. Changed $APLDEC version of ZTAP2 to accept errno = 28 (no space on device) as End-of-information. Moved to patch area. 8076. April 12, 1994 $QPGNOT/IMERG.FOR Dave It appears that IMERG had not been working for quite some time; certainly as far back as 15JUL93. Changed the local include since most of the info is now in DTCIO. Took out some unused variables. Changed some buffer array sizes. Substantially cleaned up the assignment of output CLASS. At some point it might be useful to go through this code and clean up the style to make it consistent with other code. Moved nowhere. 8077. April 12, 1994 ANCAL Chris Parameterized the number of Tant entries allowed per antenna and increased it from 200 to 2000; this makes it easier to handle YCAL output from the VLA, which contains large numbers of Tant entries. Also added some notes on the use of ALLCHAN and INDEX to the explain file; ALLCHAN applies only to the polarization specified in the control group (removed a comment from example 2 that suggested otherwise) and INDEX does not override the number of IFs given in the control group. Moved nowhere. 8078. April 13, 1994 FARAD Chris Corrected a scaling error that caused the magnetic field to be calculated at the incorrect altitude. Moved nowhere. 8079. April 13, 1994 TECB.91 Chris Corrected the labelling of data for May and July which were dated for 1992 (this gave FARAD fits). Moved nowhere. 8080. April 13, 1994 FARAD Chris Normalized the East longitude of the TEC monitoring station, removing a one-day error in calculating LMT. Moved nowhere. 8081. April 14, 1994 FITLD Phil If the first file read from tape returned error 12/13 from GET3DT (i.e. too many times out of order, too many bad day numbers - both indicative of correlator screw-ups) but some good data had been read, FITLD still crashed and burned and detroyed the file. Fixed. Moved nowhere. 8082. April 14, 1994 COOKZ.TEX, .PS Pat Finished a first pass at updating it. I tried to address the general issues as well as the CV-specific aspects. Someone else will have to do the AOC parts. Moved nowhere; should it go to the ftp/web area? 8083. April 15, 1994 New task: $APGNOT/VBGLU Phil New task. VBGLU is designed to glue together data correlated in multiple passes on the VLBA correlator. It can take up to 4 input files (MkIII Mode a - 28 IFs) and will glue them together to form a file with a large number of IFs. It will also bond the tables with IF dependencies. Also VBGLU.HLP. Moved nowhere. 8084. April 15, 1994 PRTSD Eric Corrected a format, handling of titling, handling of null sort strings, and, importantly, changed the setting of BCHAN to 1. It was offsetting the first selected channel, not once but twice, so that wrong data were displayed. Moved to patch area. 8085. April 18, 1994 ANCAL Chris Fixed an array overrun bug: ISTATS was initialized in a loop with upper bound MAXSRC instead on MAXANT. Moved nowhere. 8086. April 18, 1994 $QOOP/QCLEAN.FOR Bill C. The divergence test in CLACLN was getting into trouble when a CLEAN was restarted. Local variable PCCMIN was only properly initialized when a CLEAN started from no components. PCCMIN is now carried in common and initialized in CLNINI. Moved nowhere. 8087. April 18, 1994 VLBA FL table Phil Benson has trouble writing out a bit array for the polzn flags in the FL table. SO he will write the flag array as integers. Modified the FL table routines to deal with this. Moved nowhere. 8088. April 18, 1994 FITLD Phil Minor change to deal with new FL table format. Moved nowhere. 8089. April 19, 1994 CHKNAME Eric Changed AU8 to avoid using a 0 subscript, to give a better error message on disk out of range, and to do the new verb CHKNAME. Changed POPSDAT.HLP and NEWPARMS.001 to define CHKNAME and wrote a new CHKNAME.HLP. Cleaned up the help files for GETNAME, GET2NAME, GET3NAME, and GETONAME to show the output adverbs. CHKNAME counts the number of catalog entries that match a specific set of name adverbs (with defaults, wildcarding, ...) and returns ERROR = 1 - count. Thus ERROR is true if there are no matches. This should be useful in procedures. Also changed SUBS, MASSGN, KWICK, and AU9C for possible references to stacks at 0 subscript. Moved nowhere. 8090. April 20, 1994 $APGOOP/SPECR Phil SPECR lost the keywords in the SU table as it modified it. Fixed. Moved nowhere. 8091. April 20, 1994 $APGNOT/UVFLG.FOR Phil Increased the number of flag lines in the input text file from 1000 to 3000. Also modified an error message that was giving trouble. Moved nowhere. 8092. April 21, 1994 MX Eric Got the DO loop limit wrong for multiple plane images. Moved to the patch area. 8093. April 22, 1994 AIPSASSN.* Pat Modified the SYSUNIX version of these so that it sets the environment variable PSALLOC to "early". This is important on IBM RS/6k AIX v3.2.5 and later; see /usr/lpp/bos/README.PSALLOC for the gory details. Briefly, it forces programs that use malloc() to "get the memory NOW" rather than the controversial behaviour of AIX where it didn't bother getting the memory until it was accessed (see, e.g. the trade mag "Advanced Systems", February 1994, page 52, "AIX's late allocation _malloc_ revisited"). SYSADMINS PLEASE NOTE: doing this may cause swap space requirements to increase, especially if you have large PSAPs in use. Moved to $AIPS_ROOT in CV (won't be effective otherwise) today. 8094. April 22, 1994 POSSM Phil When plotting bandpass tables some info from the SU table is needed, but not that tied to a specific source, so I had to call to GETSOU for source 1. But there may not alwys be a source #1 in the SU table and if not then POSSM died. Fixed. Moved nowhere. 8095. April 25, 1994 New task: $APGNOT/MFPRT Phil A quick and dirty new task. MFPRT is a heavily edited version of PRTAB that will list all MF tables (Model File tables produced by SAD) in the format required by Athol Kemball's external dynamical/geometric modelling software. Will edit the output based on flux and/or component size.. Moved nowhere. 8096. April 25, 1994 $YPGNOT/FRPLT Leonia The fringe rate spectum was shifted erroneously at two points. The program did not work if APARM(1) (pre-average time) was specified as 0. Now APARM(1)=0 correspond to pre-average time = 2 sec. Relevant title is appeared on the screen for both fringe rate plotting and visibilities plotting. Two comments are added for TIMERANG and APARM(1) in FRPLT.HLP Moved nowhere. 8097. April 25, 1994 UVFLG Eric Changed the code to default the second time of a time range to the forst time, to expand equal times by ).5 seconds rather than 2, and to leave fully specified times alone. Changed the help file to say this and to encourage users to do their own expansion as needed. Times were not being entered accurately enough for new fast sampling modes. Moved nowhere. 8098. April 25, 1994 Disk access Eric Changed $APLGEN/ZFREE.FOR and IAMOK to treat disk 1 in the same way as other disks. Since users can now control which disks they get, they can ""accidentally" get themselves a disk 1 to which they do not have rights. Before, when only Managers controlled such things, we had sympathy with the user; no longer. Moved nowhere. 8099. April 25, 1994 QCLEAN, WFCLN Eric If the user sets NBOXES = 0, but leaves BOX set to stuff, then WFCLN (actually all OOP cleans through QCLEAN's INI routine) would reset NBOXES to 1 and use the 1st BOX of the users input! This is not what they expect (or get from MX). Changed to default the BOX when NBOXES <= 0. Moved nowhere. 8100. April 26, 1994 START_AIPS, RUN, START_TPSERVERS, DADEVS.SH Pat First, modified DADEVS.SH so that it gives an informative message about which file it's using for disk definitions (site-wide DADEVS.LIST, host-specific, or personal .dadevs). Then realized that the other three scripts never check to see if the number of chosen disks is zero. Added a check on this and don't start anything if there are no disks (nothing would work anyway; it's just cleaner this way). Moved to AIPS_ROOT today. 8101. April 26, 1994 CLCOR Chris CLCOR now uses the zenith angle in degrees rather than radians when applying a gain curve correction. This brings the code into agreement with the help file and the comments. Moved nowhere. 8102. April 26, 1994 DDCH.INC Eric Added a parameter to the DDCH.INC commons to set the delete delay time for print jobs. Also added a dummy array so that one will be able to add more parameters without recompiling everything. Changed ZDCHIN to set and pick up the parameter. Changed SETSP, SETPAR, FILAI2, and FILAIP to do this as well. Recompiled some 600 subroutines and every program. Changed ZLPCLS and ZLASCL to use the new parameters, passing it to the lower level routines and ZLPCL2 and ZLASC2 ($APLGEN dummies and $APLUNIX real ones) to receive the delete code as a delay time. Corrected ZLASC2 on 29 April for a bug that caused no /tmp files to be deleted. Moved nowhere, moved the 4/29 correction to the AOC. 8103. April 27, 1994 PRTLIN, PRTALN Eric Were not using the cleaned up host name for CRT printing, just line printer/file printing. Moved nowhere. 8104. April 28, 1994 FITLD Phil Gustaaf discovered that when reading normal (non-VLBA) FITS files, if there was a disk with no CA file then FITLD would die. This is because FITLD attempts to build up a list of UV files to which it might wish to add VLBA FITS data, it doesn't know it advance of course whether or not there are any VLBA FITS files on the tape. Added a check to the return codes from NXTMAP to avoid FITLD crashing and burning. Moved nowhere. 8105. April 29, 1994 UVPRT Eric Changed the formats, doing away with the visibility number which means next to nothing within the calibration package. Dropped all of the old holography modes and added HOLO for printing and HOLG for writing files for HOLGR. These will do one antenna pair at a time, but will loop to use up all antennas in BASELINE. ANTENNAS(1) is the reference antenna and is the only one used from that array (in these modes). The phases are reversed when the reference antenna number is greater than the other antenna number, to correct for reversals done by FILLM. Adverbs CLINT, BDROP, and EDROP were added for these modes to define the basic time interval used for each point in the raster and how much of each interval should be deleted. Moved nowhere. 8106. April 29, 1994 TVCPS Eric I forgot to put my new parameters into the TVCPS help file except just enough to test the code. Sigh... Moved nowhere. 8107. April 29, 1994 TIMDEST, DISKU Eric Both TIMDEST and DISKU ran into difficulties with user files on assigned disk 1 for non-assigned users (over which they need to loop). DISKU showed all the error messages while TIMDE hid them better. However, both of these should go after ANY user on the disks, not just those that are supposed to be there. So changed both to zero DASSGN before doing their thing. TIMDEST, being a verb, has to restore DASSGN when it is done. That one is in AU3A. Moved nowhere. 8108. April 29, 1994 $QPGNOT/FRING Phil Fixed a long-standing bug. When users asked for the default search windows (Nyquist) quite often the values of residual delay and rate produced by FRING were obviously incorrect. What appeared to be happening was that the FFT solution was picking up an alias and the lsq routines zeroed in on that alias. At the start of FRNSRC, the routine that does the FFT solution, there is some code to determine the number of delay and rate channels to search. This code then ensures that the specified number of channels is always odd. However, immediately after, this number can be adjusted to ensure that the AP size is big enough for the specified search. In this adjustment the number of delay and/or rate channels can become even again, this caused the aliasing later on. Fixed by ensuring that only odd numbers of channels are used in QSEARC. Moved to patch area. 8109. May 2, 1994 BLING Chris Added a trap for files with one frequency channel; BLING will now abort if only one channel is present instead of continuing with insufficient information. Moved nowhere. 8110. May 2, 1994 POSSM Phil Added the option to reverse the direction of the spectrum when plotting. Enables users to generate soectra that conform to the old standards concerning velocity direction. Rick pointed out that I was one channel off in the labelling when users asked or frequency labelling. Is now correct. When plotting VLA cross-power spectra in the multi-plots per page mode the baseline label in the upper LHC of the amplitude plot chopped off the VLA pad numbers, now fixed. Moved nowhere. 8111. May 3, 1994 FRING Phil Bugs always travel in pairs. The CIT group sent some data that showed up a subtle bug in the LSQ routines. When narrow windows were set the penalty term dominated the Chi-squared and resulted in the solutions from the lsq section being biased towards zero. After consultation with Bill and Chris we decided that the correct solution was to set the penalty terms to zero. Moved nowhere - maybe should place in the patch area. 8112. May 3, 1994 ZTAP2 Eric Changed the DEC version of ZTAP2 to make it test more carefully for errors on the various backfiles. It was assuming that all errors on bakf => BOT in one place and taking them seriously in another. I noticed that other versions of ZTAP2 as well as this one, effectively did 2 bakf's at the start of the tape on the BEGW operation. So I changed the HP, SOL, SUN, and IBM versions of ZTAP2 as well. The SOL change was more extensive since it had not been updated to avoid writing the EOFs at the start of the tape on a BEGW. Moved nowhere. 8113. May 4, 1994 $APGNOT/UVCOP Leonia The criteria of selected sources (additionally to have existed criteria of timerange and frequevcy) has been installed to minimize the size of output file. Moved nowhere. 8114. May 5, 1994 POSSM Phil When plotting IF > 1 and multiple plots per page the normalising weight array was never reinitialized properly and so the weight array became bigger and bigger and the spectra had lower and lower flux. Fixed. Also when reversing the plotted spectra I did not switch the direction of the channel number labels, although I did do frequency and velocity labels. The channel labels are now also consistent. Moved nowhere. 8115. May 5, 1994 PUVD.INC Eric Raised MAXANT from 45 to 90 to accomodate the new Japanese solar telescope (and Tim Bastian). Raised MAXCHA from 2050 to 4100 to accomodate the AT, one hopes. Recompiled and relinked a whole bunch, hopefully enough. Moved nowhere. 8116. May 9, 1994 $QPGNOT/FRING.FOR Phil Under some conditions the hessian array passed to the lsq routines could be singular. When that happens the standard linpak routine to solve for the Newton correction can fail with an arithmetic exception, instead routine GM (Greenstadt's modifcation) should be used. Changed FRING to force that. Moved to patch area. 8117. May 9, 1994 $APGNOT/VBGLU.FOR Phil In the table glueing routine there was an error caused when I used a constant in a call sequence rather than a variable. Changed to always use a variable. Moved nowhere. 8118. May 9, 1994 ANCAL Chris Fixed an index bug that caused the INDEX keyword to be completely ignored. Also fixed the code handling the INDEX keyword to loop over the number of columns expected rather than the number of IF/polarization combinations expected and trapped out-of-bounds array accesses arising from 'X' (ignore columns) codes. Moved to patch area on this date. 8119. May 9, 1994 help files Eric Changed old general list help files to point to ABOUT: ADVERB ANALYSIS APTASKS BATCHJOB CATINFO CUBE CURSOR DELETE DOTASK GENERAL INDEX MAPETC PL2D PSEUDO SL1D TAPE TASKS TVCOLOR TVGEN TVINTER UVPR VLBI Changed new ABOUT files, adding precursor remarks: ZZADVERB ZZVERB ZZPSEUDOVE ZZTASK ZZBATCH ZZCALIBRAT ZZSPECTRAL Removed totally obsolete and inaccessible: ROAM AIPSHELP TV3DIM VERBS ISCOMP Changed others as: SECONDRY added ONED category SECONDARY added ONED category MSGSERVER new - alternate spelling TEKSRV new - preferred spelling ABOUT added ONED category CALIBRAT removed lists, left all the rest HELP changed descriptions of alternatives MSGSRV changed category NOADVERB reword to allow more general use of RUN NEWPARMS and to mention picking up new verbs and pseudoverbs TVROAM Add explain section from old ROAM.HLP USERLIST Improved wording Added ONED and other categories to: SLICE SL2PL TKSLICE TKASLICE SLFIT TKSET TK1SET TKMODEL TKAMODEL TKRESID TKARESID PLCUB PLROW XPLOT XGAUS PFPL2 Moved nowhere. 8120. May 9, 1994 SHOPA Eric Changed it so that it will get rid of all directory names preceding the file name if the input file has that format. Moved nowhere. 8121. May 9, 1994 ABOUT files Eric Upgraded the ABOUT files since SHOPH said to do so: ZZADVERB ZZANALYSIS ZZAP ZZBATCH ZZCALIBRAT ZZCATALOG ZZGENERAL ZZIMAGE ZZIMAGING ZZINFORMAT ZZINTERACT ZZMODELING ZZOBSOLETE ZZOOP ZZPLOT ZZPOLARIZA ZZPOPS ZZSPECTRAL ZZTABLE ZZTAPE ZZTASK ZZTV ZZUV ZZVERB ZZVLA ZZVLBI ZZONED Moved nowhere. 8122. May 10, 1994 $APGNOT/SPLIT Phil Somehow, sometime, the portion of SPLIT that would update the FQ table when channel averaging occurred was broken. The effect was that the channel separation and total bandwidth after SPLIT with APARM(1)=1 were incorrect. Fixed. 8123. May 10, 1994 $APGNOT/CLCAL Phil An unforseen consequence of the increase of MAXANT in PUVD.INC. CLCAL calls subroutine ANTFND - a GREP shows that it is the only task to do so - which assumes that array ANTENS is dimensioned by MAXANT. It fills up ANTENS and, in the course of doing so, overwrites some other, crucial, variables. The result was that CLCAL stopped writing CL tables. What was disturbing about this was that it took 4 days for me to find out - nobody said anything! Moved nowhere, but I bet there are more. 8124. May 10, 1994 DEC tapes Eric Pat found a new way to get more status info on OSF/1 (DEC alpha) systems. It involves devio in addition to mtio. So changed $APLDEC versions of ZTPWA2, ZTAP2 and ZMOUN2 to detect more about the system's status, type, and the like. Had to create $APLULTRX versions of ZTAP2 and ZMOUN2 since not all symbols (e.g. MTSEOD to go to eod and Exabyte 8500's) are known to the old ULTRIX systems. Changed the ZTAP2 to use skip to EOD on SCSI tapes. Let's hope it works - it does on the DAT we have. Moved to patch area. 8125. May 11, 1994 $SYSIBM/CCOPTS.SH Pat Added code (and explanatory comments) so that PSALLOC gets changed from "early" to "" (blank) before the C compiler is called. Apparently IBM has seen fit to create a compiler that is incapable of compiling "hello world" on a machine with 128 megabytes of memory and 256 megabytes of swap space; the error one gets is "not enough space" and there is abundant disk space. Moved nowhere. 8126. May 11, 1994 $SYSUNIX/XASERVERS Pat Modified this script to remove the explicit "xterm" from the "ps" commands that tried to detect already running TEKSRV and MSGSRV servers. Also generalized the configurable interface by providing two new user- or installer-settable variables: AIPS_TEK_NAME_FLAG and AIPS_MSG_NAME_FLAG to isolate the xterm- or other emulator specific command flags to set icon names, iconic behaviour, etc. Finally, shed teksrv and msgsrv as grandchildren (I think) the same way XAS is. One does not always want these to die if AIPS itself does (though on the Dec Alpha, a ^Z still apparently kills them and XAS...) Moved nowhere. 8127. May 12, 1994 $APGNOT/FITLD.FOR Phil Added the code to supply the variables needed for the digital corrections. The corrections themselves will be applied soon. Moved nowhere. 8128. May 12, 1994 $APGNOT/FITLD.FOR & UVLOD.FOR Phil Data written be sde is often in the form of Stokes I & V, FITLD/UVLOD were unable to read such data due to some logic in subr UVFHDR that assumed positive Stokes CRVAL & NAXIS < 4 was a non-valid combination. Added a line of code so that if CRVALn = 1, NAXISn = 2 and CDELTn = 3 the data would be passed through labelled correctly. Moved nowhere. 8129. May 12, 1994 F2PS.C Eric Stand-alone program to read print files (either with or without Fortran formatting commands) and convert them to PostScript. The font used is Courier and can be in -S (7.4 pt), -R (8), -M (9), and -L (10) sizes with corresponding loss in max number of lines. It will take a -n argument, with n being the max lines/page used in the print task, and convert to the nearest size to support that max. It takes a -d option to => duplex and then shifts even pages oppositely to odd from the centered position. It will also take a -A4 option for non-U.S. sites. Moved nowhere. 8130. May 13, 1994 HOLGR, UVPRT Eric Corrected the data reading routine to use proper parsing of the input fields. It was assuming a format and returning with no error when that format failed. Cleaned up the code to look more like AIPS code. Added to UVPRT more automatic holography data flagging: when the W's are 40256 the antennas are in a "turn-around" mode and the data are useless. Also when W = 40257, the sample is on a CLINT interval of time, and a 40256 was within CLINT, the antenna is anticipating being in a good mode, but isn't yet. Moved nowhere. 8131. May 13, 1994 ZFORMF Eric Dropped the cute code that had it return blanks on the first call and real form-feeds thereafter. It is used only when writing to text files and this cuteness meant that there was no form feed at the start of appended print jobs to separate them. The new F2PS will not produce a blank page if it is used to print a file with a form feed at the beginning. Moved nowhere. 8132. May 13, 1994 $QPGNOT/CALIB - $APLNOT/GCALC1 Leonia/Phil Calib failed for all anntennas for a given IF when that IF was missing (visibilities = 0) on one antenna. It only occurred if SOLTYPE = 'L1'. The reason was in subroutine GCALC1. Baselines including the 'bad' antenna were not excluded, because the weight of baselines is determined from all IFs. In the case of SOLTYPE = ' ' (standard least square, subroutine GCALC), the weight of baselines including the 'bad' antennas (vis.=0) is forced to zero and such baselines were excluded. Thats why the problem did not occur in the case of standard least square. The relevant change has been made in subroutine GCALC1. Additionally the logic of the choice of model: CLEAN component, SMODEL or point source from SU table has been clarified (subroutine CLBMOD). Explanation of this logic as well as comment about L1 algotithm has been added in HELP file. Moved nowhere. 8133. May 13, 1994 $APGNOT/UVFLG Leonia If you erroneously input BIF bigger than number of IF the program flagged the last (maximum) IF. The same thing occurs for channels. This unfriendly behavior has been changed by terminating the task with a warning about the error to the user. Moved nowhere. 8134. May 13, $QPGNOT/CVEL Phil I'm sure I fixed this before. In CVEL I use LUNs 37 & 38 in a couple of subroutines, because of the large number of tape drives accessible by kiowa these LUNs are reserved for tape drives and so ZMIO refuses to use them. Changed to higher numbers. Moved nowhere. 8135. May 16, 1994 $INC/DSEL.INC Phil Had a second case of the new MAXANT parameter value causing problems. This time in what is the application of the bandpasses and was related to the size of the ANTENS array. ANTENS was declared to be of size 50 in DSEL.INC but in some place it was in a loop over MAXANT. This was causing problems so I changed ANTENS to be of order MAXANT in DSEL.INC. Means a recompile/link of the calibration system again. Moved nowhere. 8136. May 17, 1994 $APGNOT/UVFIX.FOR Gustaaf The default value for UVFIXPRM(12) (IAT-UTC) meant that UTC was not corrected for accumulated leap seconds. For observations from 1994 this difference is 28 seconds, enough to cause distortions in some cases (A array, large shifts). Now, for VLA data, a value of 0 defaults to the correct value, which is returned by a subroutine I added to UVFIX. When new leap seconds are announced, this subroutine will have to be adapted accordingly. When the default is used, the user is informed about this, and also about the last update to this subroutine. Moved nowhere. 8137. May 17, 1994 $QPGNOT/FRING.FOR Phil Added a minor change to deal with the spectral line case of fitting for residual rates only. The LSQ solution would report non-convergence because it was set up for full delay fitting. So if not fitting for delay changed NA3 to be NA2. Moved nowhere. 8138. May 18, 1994 $APGNOT/AVSPC.FOR Phil When using AVSPC with option SUBS the output channel separation was incorrect. Fixed. Moved nowhere. 8139. May 18, 1994 $QPGNOT/BPASS.FOR Phil In most cases the array declerations inside BPASS were set by parameter statements and the bookkeeping for allocating freqs/IF/baselines etc inside the arrays were handled down in the subroutines. In one case (BPINTP) this was not done. I fixed it up so that it now is - this avoids potential problems with the new values of MAXCHA and MAXANT. Moved nowhere. 8140. May 18, 1994 HOLGR Eric Changed the fortran to handle negative V convolution types the same way it handles negative U ones (set UNIWT true and take the absolute value of - now - both the U and V parm(1)). Added more remarks to the help file about what is - and is not - defaultable. Changed both to make uniform weighting the default and natural weighting selected by a convolution type < 0. Moved nowhere. 8141. May 18, 1994 F2PS Eric GNU C has a non-standard C language word "inline", while my code had a variable called inline. Despite rather clear context, this blew GNU's cookies. Respelled the variable xxline. Moved nowhere. 8142. May 18, 1994 SPFLG, TVFLG Eric We originally pulled TVFOAD and SPFOAD from TVFLG and SPFLG so that they could be fully vectorized on Convexes. Unfortunately, Glen did not figure this out and OPTIMISE.LIST has been explicitly doing TVFOAD at 2 (the general default) and ignoring SPFOAD. TVFWED and SPFWED were moved into $YSUB and SPFBLT was moved into $APLSUB. This was done because TVFOAD and SPFOAD referred to them and Solaris required that all references be locally resolved. Moved these routines back into TVFLG and SPFLG with their includes. Then deleted $INC/DTVF.INC, $INC/DSPF.INC, $APLSUB/SPFBLT.FOR, $YSUB/TVFOAD.FOR, and $YSUB/SPFOAD.FOR. Moved nowhere. 8143. May 19, 1994 ZTAP2 Eric Found a bug of a serious nature in the SUN version of ZTAP2. It (effectively) positioned itself before the EOF that it was supposed to be after on the beginning-of-write EOFs. So if you used DOEOT -1 in FITTP between May 3 and today you have a problem - no EOFs between files. Moved to VLA today by hand. 8144. May 19, 1994 UVDIF Eric Corrected it to work with compressed data. When I fixed it to handle swapping order of data, I forgot that LREC is not always the actual length of the data after expansion. Sigh... Also fixed it to write out RATMAX anyway and complain when no data samples were even compared. Moved nowhere. 8145. May 20, 1994 $APGNOT/AVSPC.FOR Phil There were 2 problems with the SUBS option in AVSPC. First the bandwidth written into the FQ table was incorrect. Second the reference pixel of the output file was incorrect since I had assumed that the left-edge of the pixel was the relevant value for my calculation rather than the centre of the pixel. Both fixed. Also tidied up some of the code, added an ATCA option for Neil, and made the help file slightly clearer. Moved nowhere. 8146. May 24, 1994 $SYSUNIX/F2TEXT.C Pat Copied Eric's F2PS program, stripped out all the postscript stuff and modified it so that it's a simple fortran text to plain text filter. Handles "0", "1" or space in column 1 but not '+' for overprinting. Some Linux installations (i.e. ours) don't have fpr or asa. Moved nowhere. 8147. May 24, 1994 SAD Eric (1) Corrected (at Becker's suggestion) the assignment of the clean beam position angle to the initial guess of the component position angle. The former is in degrees, but the latter must be radians, so added a scaling. This occurred several places. (2) Added a column for cube plane number in the MF file and made all column numbers parameters with an include at the beginning to define them. Changed it to put the current plane number in its column and to subtract previous components only if the plane numbers match. Moved nowhere. 8148. May 24, 1994 TCOPY Eric Changed to detect the bug in ZTAP2 which caused FITS files to be run together without separating EOFs and to write the extra EOFs in the output tape. Also added error tests for writing data to disk (only FITS data works). Moved nowhere. 8149. May 24, 1994 PRTTP Eric Changed it to put something in the output file of PRTLEV=-3 for formats other than basic FITS. It was not puting anything there for VLBA FITS tables, VLA archive, VLA Export, "IBM" map, or miscellaneous formats. Corrected it in its handling of null files - it was forgetting the error condition due to a call to a print routine. Changed the reading of unknown records following FITS data. It used to read up to 20000 and then do nothing but report. Now, if it hits a 100000 limit, it reports it as >= 100000 and advances to the next file. We were getting infinite loops when there were > 20000 otherwise. Moved nowhere. 8150. May 24, 1994 ZDM2DL, ZRM2RL Eric Changed the $APLUNIX versions to swap 16-bit words on the way in to the loop and to swap 32-bit words (in the double prec one) on the way out. This is of course only for byte-swap machines. Moved nowhere. 8151. May 26, 1994 IMLIN, XBASL Eric Changed IMLIN to detect when all points in a row to be fit are blanked and to take appropriate measures. Added test to make sure some baselines were actually fit and to destroy the output if not. Dropped a test designed to save work if the baseline region hadn't changed. It could not know about the previous line's blanking and hence could not know this. Cleaned up the handling of units - it should recognize your axis now. Made XBASL also check for no allowed points and take defensive measures. Moved IMLIN to the patch area. 8152. May 26, 1994 Linux tapes Eric Tapes can actually be made to work directly on Linux-system PCs. Changed $APLLINUX versions of ZTAP2.C, ZTPWA2.C and ZMOUN2.C to handle mag tapes as well as can be done with the current driver (see below). Changed FITTP and PRTTP to remove unused variables (not seen by SUNs since they did appear in DATA statements) and to clean up the calling sequence data types. The problems with the Linux driver are summarized as: (1) Tape opened for write, but the last operation was not a write: others just close (they DO write an EOF iff the last op was a simple write) linux writes an EOF anyway (which could destroy the tape contents, create null files, ...) (2) Write data at beginning of tape: others do it linux returns errno 28 (3) Write data in the middle of the tape: others sometimes require that we write an eof to make it the end of information before actually writing. We do this by positioning just before the last EOF to be kept, writing a new EOF replacing it, and then writing our data. linux returns errno 28 when we try to write our own data (it allows the EOF writing part) (4) Write data at the end of tape in multiple files: others do it linux allows us to write the first data file starting at the end of information, to write the EOF after the data, but then returns errno 28 on the next attempt to write. (5) The tape statuses are not set: good systems set mt_gstat (or the equivalent) so that GM_EOF, GM_BOT, GM_EOT, GM_EOD, (or the equivalent) actually work. On SUNs, the fileno, and blkno are even meaningful. linux never sets the status, the fileno is usually off, and one detects BOT by an error on back-space-file, EOT by an non-EOF error on read, and other unsafe, kludgy means. The only error codes seen are 5 and 28 in errno, -1 in the ioctl return value. We need more help to find out what is going on with the tape. (Note - AIX is no better at present although they are making promises.) Problems 2, 3, and 4 can be managed by writing an EOF mark, backing up over it, and then spacing forward past it (for 3 and 4 only). Writes are then allowed. We can live with the assumptions about tape status, but they are shaky. Our only tape writing tasks have the tape at its end so the extra EOF is not a disaster (problem 1). However, if users set DOEOT 1 or do not manually reposition their tapes, then they will have lots of null files on their tapes and some systems will assume that these are the end of information and ignore the following data. Moved nowhere. 8153. May 31, 1994 IMLIN, ABOUT, HOLGR Eric Minor correcttions: (1) IMLIN, when DOUTPUT was false, had not set one parameter and so, two parameters deeper, failed to realize that good solutions had been obtained. (2) AU1A (code for ABOUT) requires that each category appear in a DATA statement, unlike HELP which just checks the directory contents. I forgot this when I created the new category ONED. (3) HOLGR applied the uniform weighting weight to the observations, not to the weights used for the beam pattern or the output weight map. This made those wrong, although the main output images used to find dish errors should not be affected. Moved nowhere. 8154. May 31, 1994 $QPGNOT/CVEL Phil CVEL could only handle up to 60 source numbers. Was OK until I tried to push a VLBI OH survey through. It can now handle 300. Moved nowhere. 8155. May 31, 1994 $QPGNOT/FRING Phil With the switching off of the penalty functions I had neglected to deal with the case of no search in delay (common with line observations). Added one line of code to fix this. Also updated the .HLP file to reflect the change. Moved nowhere. 8156. June 1, 1994 New task: $APGNOT/FXPOL.FOR Phil Task to correct the header for dual polzn VLBA data. If there are no cross-hand data then the correlator software labelled all polzns as RR and just gave them different IF numbers. This was causing problems. FXPOL will siwtch the FREQ & STokes axes, change the number of Stokes pixels from 1 to 2 and halve the number of IF pixels. It will also correct the FQ table and delete the IM, CL, SN, BP, BL, FG and NX tables. The user will have to regenerate what he/she can with INDXR etc. Moved nowhere. 8157. June 1, 1994 VLBPOL in APLNOT Chris Corrected an inconsistancy in the sign of the parallactic angle that led to 'VLBI' polarization corrections being applied incorrectly. Moved nowhere. 8158. June 2, 1994 $APGNOT/ACFIT Phil ACFIT had always been able to deal with dual polarization data internally but did not have the input hooks to allow users to specify different Tsys and/or gains for the two polzns independently. I changed ACFIT.HLP to allow that and also ensured that all was done correctly in the code. Moved nowhere. 8159. June 5, 1994 HOLGR Eric Changed the gridding routine to check for samples outside the grid and to quit when none are inside. It used to depend on everything being okay! Moved nowhere. 8160. June 5, 1994 EXTDEST Eric Changed AU8 so that it will not delete CL and HI versions 1 in batch and will ask for confirmation before doing so interactively. Moved nowhere. 8161. June 6, 1994 F2PS Eric Changed the parameters like print size and the like. It was only getting about 124 columns on the page since my test print file (before) was too narrow! Now we can get 97 lines to the page for young guys anyway. Moved nowhere - must be rebuilt in SYSLOCALs. 8162. June 7, 1994 $APGNOT/FILLM.FOR Gustaaf FILLM never loaded 2BD CH0 data correctly. Culprit was a wrong pointer in subroutine MCINI. Checked the VLA database, and noticed 2BD only had been used by Juan Uson before (who had spotted the error back in November 1993). Moved nowhere 8163. June 8, 1994 UV2MS Bill C. Added adverb REFDATE to allow specifying the reference day for the output uv data set. This is only used when the output data set is created. The default is to use the reference day of the input data set. Moved nowhere. 8264. June 9, 1994 $APGNOT/ACFIT Phil Added a check to ensure that the # channels in the template spectrum is the same as that in the main database. This one caught a user unawares. Moved nowhere. 8265. June 9, 1994 ZMOUNT, TPMON Eric Changed ZMOUNT to maintain the error return. It was forgetting the error code on mounts, causing TPMON to think things were good when they were not. Changed TPMON mount section to return the alternate error messages such as tape already mounted as well as the ones generated by ZMOUNT. Moved nowhere. 8266. June 9, 1994 PCAL Chris Swapped the sign of the parallactic angle in the RPCALC subroutine. The 'RAPR' option was (perhaps not surprisingly) afflicted by the same problem found in VLBPOL about a week earlier. This may explain why the linear approximation hasn't worked well for VLBI data in the past; corrections derived using 'RAPR' and 'ORI-' now appear to be of similar quality for the test data set (transatlantic Mk3). 8267. June 13, 1994 F2PS Eric Changed F2PS. I forgot that the close parenthesis was a special character that would have to be "escaped" in trying to display it (along with open parenthesis and the escape character itself). Fixed it. Moved to the IBMs locally. 8268. June 13, 1994 HORUS Eric Changed the default OPTYPE to be LINE if the frequency axis has more than 1 point on it, but SUM otherwise. Added a warning message when EIF forced to BIF for LINE mode (and not = BIF in advance). Changed help file for this. Moved nowhere. 8269. June 13, 1994 SOLCL Eric Changed it to test for the max CL table. If it is 1, changed to copy version 1 to 2 and then to use 2. If the user-specified CL does not exist, die with reasonable error message. Moved nowhere 8270. June 14, 1994 $APGNOT/FXPOL Gustaaf In Phil's absence, put back his last minute fix. Moved nowhere 8271. June 16, 1994 CALCOP, MAKMAP Eric CALCOP ends with a compress of its output file. Then when the file is reused on the next call to CALCOP, it can be too small. Changed it to test for output file size and make a new SC file if needed. HORUS had a problem when, say channels 9-16 were flagged. It would compress the SC file extra for those channels and then blow up on channels 17-24 which had the normal amount of data. It turns out now that, for HORUS, it replaces the SC file on each set of channels since, for multi-source input files, there is a possibility of many more samples than one actually uses. I still left in the compression since CALIB and others also use this routine. Also corrected message level for a happy message in MAKMAP from 6 which means error to 5. Moved nowhere (except UFL). 8272. June 17, 1994 HOLGR Eric Changed the beam computation to declare full size arrays in BEAM, but do the real work in BEAM2 where the arrays can be passed as the dimension needed (now 4 x image rather than 4 X maximum image). Moved to VLA test area for Rick. 8273. June 17, 1994 UWRITE Eric Changed it to handle an Export-format file starting with records that are not either RUN or SOUR. It forgot to init a type code. Moved nowhere. 8274. June 17, 1994 VBPLT Leonia Some plots (really the middle one or the only one if we plot one plot) did not have tick labels at vertical axis if the weights were ploted. The bug has been fixed in SOBROUTINE PLTVB. Moved nowhere. 8275. June 18, 1994 PRTTP Eric Changed the Export-format print routine to print a line on RUN records giving the observation date. Cleaned up the format a little. Moved nowhere. 8276. June 19, 1994 Calibration package de-selection Eric I discovered that the antenna numbers were being made all positive and the leading minus signs on source names were being removed by low-level routines, especially SOUFIL which is called on UVGET in INIT mode and other places. Thus, what was de-selection of antennas or sources on the first call becomes selection of them on later calls. This led, for example, to CALIB determining the calibration for all but antenna 4, followed by applying it only to antenna 4! Found numerous minor logic problems in history sections too. Changed subroutines: ANTFND stop forcing values to positive, use abs value in tests CALSEL use absolute value on antenna number list DATGET to test with abs of ANTENS, rather than straight value FNDSOU stop removing the leading minus, add tests for that in name recognition and clean up logic SLCTD to make the test on absolute value when INCLD is false SOUFIL stop removing minus signs from source and calibrator names, add tests for minuses later, stop making antenna numbers positive, add abs function later SOUFND stop removing the leading minus, add tests for that in name recognition Changed programs: ACFIT remove leading minus signs in history records, changed it to put antenna numbers of the users' sign into common - deselection was being turned into selection even with the 1st call to UVGET - and fixed the setting of the antenna list used by ACFIT to do deselection or selection as needed. BLCAL remove leading minus signs in history records BPASS remove leading minus signs in history records CALIB remove leading minus signs in history records CLCAL remove leading minus signs in history records FRING remove leading minus signs in history records FRPLT treat de-selection of source as error, test for de-selection of antennas in labeling GETJY stop changing sign of antenna numbers, use abs value as needed, remove leading minus signs in history records, stop dropping minus signs which caused HI to have sources and calibrators always shown as included LISTR stop changing antenna numbers to positive, use abs in all testing PCAL remove leading minus signs in history records POSSM remove leading minus signs on labeling SPFLG remove leading minus signs in history records TVFLG remove leading minus signs in history records UVIMG remove leading minus signs in history records Moved nowhere. 8277. June 23, 1994 WFCLN Eric ********* WFCLN.HLP: ********* (1) Changed default and explanation of NPOINTS. ********* WFCLN.FOR: ********* (1) Changed the defaulting on IF and CHAN - to get the IF defaults into the history and to treat as errors bad IF and channel specifications by the user. (2) Changed the default number of channels to average to 1. It was set (echan-bchan+1)/chinc a large number for chinc of, say 2. Then, later it would reset chinc to nchavg! (3) Changed to allow - in LINE mode - CHINC < NCHAVG. (4) Changed test that chose whether it is in LINE or SUM modes. It was comparing NCHAVG input channels to (echan-bchan+1)/chinc which are output channels. (5) Added test - it was making an arbitrary size cube as output and then trying to write, potentially, > 255 CC files. It now is limited to 255 spectral channels at a time. (6) Changed it to write only the channels it needs into the WFCLN file. It was writing BCHAN to ECHAN channels on each loop even if the current channel was almost ECHAN. Thus it read and wrote lots of trash. (7) Changed it to test that NCHAV + ICHAN - 1 was a legal range of channels (no more than ECHAN). ********* QUVUTIL.FOR: ********* (1) Changed default NCHAV to be all channels, not the bizarre (echan-bchan)/chinc + 1 it had before. (2) Changed OUVIMG to make only 1 (NFIELDS actually) image per call. The only call to this is in QCLEAN (CLNINI, CLNUV) and both want one image with channel averaging. Before, it was either doing one image with NCHAN averaging in CHTYPE SUM, or NCHAN 1-channel images. The latter was not desired. (3) Changed OUVDFT to do a loop over channels and ODFT to do one channel with averaging (new call sequence). ********* QCLEAN.FOR ********* (1) Changed default NCHAV to 1 from all channels after UVCHAN less 1 (so it would loop and grab only 1 the next time!). ********* GRDMEM.FOR ********* (1) Added message about frequencies used (like in UVGRTB et al.) (2) Added test for zero divide with frequencies. If the number of channels to average exceeds the number in the data, it divided by 0. Moved nowhere. 8278. June 24, 1994 TVSTAR Eric New verb to plot star files atop TV images. Changed: POPSDAT.HLP, NEWPARMS.001, AU5B.FOR, and created TVSTAR.HLP. Moved nowhere. 8279. June 28, 1994 $APLSOL/ZMOUN2.C Pat The density setting was changing the wrong character in the device name (xlated[xlnb-1] when xlated[xlnb] was the terminating null), assuming device names of the form /dev/rmt/0ln for example. Changed to xlnb-2. Harry Payne at STScI noticed this first. Moved nowhere. 8280. June 27, 1994 SAD Eric Changed: (1) When trying 2 components in messy regions, SAD used to set the width of the 2nd component (the "halo") to double the first. That's fine, but only when widths are being fit. Added test. (2) The routine that uses moments to guess the widths of the component before fitting was capable of very wrong answers. Added clip limits (0.8 and 4 times the clean beam) to the guesses. (3) The errors for the position angle were incorrect. Changed the computation to give 90 degrees when the component is virtually circular and to do a correct ATAN otherwise. This may over-estimate the errors a bit, but better that than the reverse. (4) Moved the scaling cutoff down some to stop it from switching scales for one weak component or 2. (5) More later. Moved nowhere. 8281. June 29, 1994 UVPRM.HLP Glen Updated UVPRM.HLP to reflect reality. Removed all mention of plotting. Moved nowhere. ***********************************************************************