----------------------------------------------------------------------- ; 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) line n-1: Where has this change been moved (e.g. 15JUL94, nowhere) line n: Blank The details of line 1 are important since programs are used to find specified strings and do conversions on them. It MUST be typed as follows: 1 blank, 4 digits, a period, two blanks, then the Month, day, comma, year, some space, a title, some more space, your name. Changes to 15JAN95 *********************************************************************** Changes while 15JAN95 was TST: *********************************************************************** *********************************************************************** 8182. July 7, 1994 $SYSUNIX/FDEFAULT.SH Pat Changed all -fpe3 options for Dec (Ultrix, Alpha) to -fpe0. Apparently the former setting made underflow cause a f.p. exception which makes tasks crash & burn. The latter setting allows the task to continue. Moved to 15JUL94 today. 8183. July 7, 1994 SAD Eric More changes: (1) Changed flux scaling to use RMS when appropriate to set scaling. (2) Removed bug in subtracting previous components: a limit on X was used in testing the max Y. (3) Changed to keep track of the max flux in each island and then to fit them in descending order of max flux. Previously they were fit in the order found, roughly bottom to top in Y. The flux ordering should remove the possibility of overlapped fitting windows allowing the fit to wander off extrapolating the edges of a strong component. (4) Changed help and code to add adverbs GAIN and DPARM. The retry and warning levels will now use sqrt (level*level + (gain*peak)**2) rather than just level. The DPARM's are used to set min flux, max rms, max width, max departure from the fitting area and max departure from the image before a fit is rejected. (5) Changed SUBFIT to return if there are no gaussians. (6) Changed it to recompute the rms, sum, and extrema around each component after all components are subtracted. The numbers then apply to each component, not to a full fit area, and apply after nearby sources have also been subtracted. Normally, this lowers the "errors", but, if objects remain just outside the search area, then the "errors" may increase. (7) Added a bit to the help file about the output, including the H, L, S codes for questionable residuals. (8) Added to code and help PRTLEV to display the areas that are rejected for selected reasons. (9) Raised max number of sources to 4000. (10) The P.A. in pixels is actually measured counterclockwise from the X axis! This means that the conversion of the clean beam to pixels and the fit components to arc seconds had to be changed. This conversion was entirely wrong - it assumed that the pixel spacing varies across the image but that the direction to North does not. This is backwards. The coordinates rotate as one moves around a projection, but the real increments change only very slightly. This is fortunate - CLEAN would be wrong otherwise. (11) The deconvolution area needs work badly - but I will do that later. Moved to 15JUL94 this date. 8184. July 7, 1994 FITLD Eric Changed: (1) $APLDEC/ZTPWAD.FOR - added another error code to mean end of file. This code arises when the last record is not a full 2880 bytes. (2) $APLGEN/ZR64RL.FOR - changed to pick up the correct word when looking for underflows, NANs, etc. Dropped the $APLDEC version which commented out this section rather than figuring out why it did not work. (3) FITLD.FOR - in subroutine LOADUP, changed the code to refer to tape data only when needed. If one multiplies a valid number by a random place in a buffer, then one can get underflow, overflow, NaNs, and other fp abort conditions! One should always use only data that are initializaed. In this case, it used the uninitialized data and then did an IF to replace it with good stuff. An IF-THEN-ELSE structure is required. (4) FITLD.FOR - changed a bunch of message levels. >5 is for errors, 2 is for progress messages. (5) FITLD.FOR - changed it to do the listing of possible output files only when the format really is VLBA and DOCONCAT is true. This listing was updating the last access date, potentially on whole catalogs. Moved to 15JUL94 this date. 8185. July 7, 1994 SUBIM Eric Initialized the max and min with the initial header min and max, resp. If the header had garbage, this could leave garbage in the max/min. Fixed the initialization. Moved to 15JUL94 this date. 8186. July 7, 1994 PRTTP Eric The call to print VLB information used IERR but tested IRET. Thus it took several q's to quit. Fixed it to test the error returns more carefully. Moved to 15JUL94 this date. 8187. July 7, 1994 IMEAN Eric Changed the program to compute a histogram at all times. DOHIST now controls only whether it is plotted. Changed the OUTFIL option, so that the histogram is printed into the output file. Raised the maximum number of boxes to 1024 and the default to 101. Moved to 15JUL94 this date. 8188. July 7, 1994 CPYRT Eric Task to replace HI file with contents of a file and to add an NRAO copyright notice to the top of image files. Moved to 15JUL94 this date. 8189. July 7, 1994 WFCLN Eric Turn off the option to restart using an old WFCLN file - it is purported not to work and it never seems to do a quick restart anyway. Changed only the HLP file to do this. Moved to 15JUL94 this date. 8190. July 8, 1994 AIPSPATH.CSH,SH Pat/Ruth/Rick The logic of these scripts in using TPATH and PATH was flawed. TPATH is set to the old contents of PATH when the script is called (e.g. through LOGIN.SH or LOGIN.CSH), but the old scripts only did this if TPATH wasn't set already. If a user changes the PATH after calling AIPSPATH and then inadvertently calls it again, the changes made to the PATH since the first call are lost. Removed the test on contents of TPATH. Moved to 15JUL94 and /AIPS this date (needs done at AOC). 8191. July 8, 1994 $APLSOL/ZMOUN2 Pat/Ruth/Gustaaf Made it use "m" instead of "h" to indicate high density on exabyte 8500 devices. Apparently "h" means use compression AND high density. Moved to 15JUL94 this date and sent manually to the AOC (mail). 8192. July 8, 1994 $APLUNIX/ZIGNAL.C Pat/Bob Garwood Bob noticed a typo in this routine where there was a - that should have been a _. Linux apparently detected this too, as it was embedded in a #ifndef __linux, and the correct version (with the underscore) was used for Linux. SunOS and other systems had the relevant symbol #define'd so that the section was totally skipped. Moved nowhere as yet, probably should move to 15JUL94? 8193. July 9, 1994 HLPB, LSAPROPO Eric Why did I think this worked? HLPB swaps the list of words back and forth between two arrays as it sorts them, removes duplicates and drops excluded words. The output loop then used WORDS(i) while the last swap put the good stuff in CATS(i). Also changed HLPB to trim all the output lines. Also fixed LSAPROPO.HLP which was messed up by this. Moved to 15JUL94 this date. 8194. July 10, 1994 HOLGR Eric Changed help and code to (1) raise the size of the maximum image to 512, (2) drop the degree of interpolation of the antenna beam from 8 (2 from subsampling data and 4 from larger grid) to 2 (all from larger grid), (3) drop the use of UVMAX < 0 to imply a fixed radial taper which was applied to the data but not the weights, and (4) to add a choice of tapers which apply to the data and to the weights. Rearranged and equivalenced the arrays to reduce the memory required from 30 to 16 Mbytes. Changed it to offer a choice of interpolation factors in the antenna beam (option 10) output, up to 2048/NPIX only and corrected the axis increment in the output header. Moved to 15JUL94 this date. 8195. July 10, 1994 FILLM.HLP Eric FILLM actually loads all observation codes when VLAOBS is blank. Changed help file to say this and to recommend setting VLAOBS. Moved to 15JUL94 this date. 8196. July 10, 1994 SETJY Eric Changed it to allow CALCODE = '----' to mean reset the calcode to blank. Also added test to refuse to set calcode to '*'. Moved to 15JUL94 this date. 8197. July 12, 1994 $APLULTRX/ZCREA2 Pat I don't know how this ever worked... the reference to the data structure member giving the available free space was incorrect. Fixed (compiler no longer complains). Moved to 15JUL94 this date. 8198. July 12, 1994 $APLBELL/ZDELA2 Pat Minor modification to permit timing to be correct. Moved to 15JUL94 this date. 8199. July 12, 1994 CALIB, SNSMO, CSLGET Eric The parameter used to keep the average gain scaled to 1 was (a) miscomputed in CALIB and SNSMO since null solutions (1,0) were included (for sure in the former, probably in the latter) and (b) misapplied in CSLGET by using the SQRT of the average antenna gain rather than the average. Repeated self-cals with gain normalization would gradually lose flux. Relink all. Moved to 15JUL94 this date. 8200. July 12, 1994 GRIDTC, SPFLG Eric Subroutine GRIDTC is used by SPFLG alone and calls SPFBLT which is in SPFLG only. Therefore copied GRIDTC.FOR into SPFLG.FOR and deleted the $APLNOT version. It was pulled out initially for vectorization, but now causes unresolved reference messages in NEW. Moved to 15JUL94 this date. 8201. July 13, 1994 WTMOD Eric/Craig A minus sign was left out and a LT should have been an LE.0 when setting weights to already flagged data. This had the effect of unflagging data in the non-reweight mode and unflagging data that had weight 0 in the rewieight mode. Moved to 15JUL94 this date. 8202. July 13, 1994 UVFND Eric/Craig When the task was changed to use UVGET it was forgotten that correlators with weight <= 0 can be returned; the routine used before did not return flagged data (GETVIS). Changed the flux checking loop to check weights. In the single-dish subroutine, changed the flux loop to check that the flux is not exactly 0.0 which is what GETVIS returns if the correlator is flagged. Moved to 15JUL94 this date. 8203. July 13, 1994 $APLALPHA/LDOPTS.SH Pat Modified to attempt to force -fpe0 option so that underflows do not cause a f.p. exception. Unfortunately this appears to be innefective. Was going to move it to 15JUL94 but missed the deadline. 8204. July 15, 1994 FITTP Eric Put in things to report the data scaling to the user and to count and report the number of vis's and image pixels written as pure zero. Perhaps this will warn people who are destroying their data by writing them as 16-bit integers. (The case that just came up had a least bit of about 0.48 Jy!) Moved by the grace of pmurphy to 15JUL94 8205. July 18, 1994 $APLGEN/ZOINTD Pat This module failed to compile under some K&R compilers because there was no #include (see Jan 1994 patches). Inserted same. Moved to 15JUL94 this date. 8206. July 18, 1994 $SYSUNIX/XASERVERS Pat The flags for "ps" (berkeley mostly) used in this script were arbitrary and inconsistent. They didn't even work on OSF/1. Rationalised the "is this process running" logic and in the process somewhat simplified the script. It gets the name of the process owner for XAS, TEKSRV, and MSGSRV if running, and also the display variable if possible. Moved to 15JUL94 this date. 8207. July 18, 1994 $APGNOT/UVFND.FOR Athol Minor typographical error corrected in UVFND fix of 13 Jul (# 8202). Weights referenced as VIS(K,3) instead of VIS(3,K). Request to move this to 15JUL94 pending with Pat. 8208. July 19, 1994 ZIGNAL Eric Corrected $APLUNIX signal handler ZIGNAL.C. I had copied Stevens' book incorrectely, using two != rather than |= 's. Thanks to Martin Shepherd (mcs@phobos.caltech.edu) for pointing this out. I do not understand why the C compilers accepted the old statements without complaint. Moved by the grace of pmurphy to 15JUL94 8209. July 19, 1994 ZSSSX2 Eric Corrected $APLUNIX/ZSSSX2.C for the bad assumption that all words to be byte-flipped appear initially to be positive. Stole the "union" code from XAS itself. Moved by the grace of pmurphy to 15JUL94 8210. July 20, 1994 ZTPMID Eric Changed the $APLIBM version of ZTPMID (the only one that uses C for this) to loop until all bytes are done for reads and writes. This will mostly just allow proper trapping of errors. When nothing is wrong, only 1 read or write should be needed for a mere 2880 bytes. Moved nowhere. 8211. August 1, 1994 $APGOOP/DTSUM.FOR & $HLPFIL/DTSUM.HLP Athol Added a print-level option to DTSUM to allow the printing of a condensed summary of visibiltiy counts for the entire uv-data file in the same antenna-matrix format. Useful for finding gross discrepancies in VLBA correlator output files. Moved nowhere. 8212. August 2, 1994 $APGNOT/ANCAL.FOR Phil ANCAL hit a problem when being used on S/X data in the same file. User was using the INDEX keyword (incorrectly as it turned out - but that was fixed) and ignoring the first 4 columns of the input text file. This resulted in the first 4 columns of the internal Tsys array being blanked - correctly. However when applying the Tsys information to the CL table no check was performed to ensure the Tsys column was blanked and so the CL column was filled with numbers scaled by the sqrt(fblank). Moved nowhere. 8213. August 3, 1994 $APGNOT/ANCAL.FOR Phil Modified slightly to not blank valid CL columns based on blanked Tsys columns. They are now passed through untouched. Moved nowhere. 8214. August 3, 1994 New task: $APGNOT/BASRM.FOR Phil New task to remove spectral baselines from total power spectra and write out a new uv-file. Any cross power data in the file are passed through untouched. This first version only works on uncompressed data, the next version (next week?) will work on compressed data also. Moved nowhere. 8215. August 3, 1994 $YPGM/ISPEC.FOR Dave A user pointed out a minor problem: while ISPEC calculates the "average over area" and correctly says this on the plot, the printout read "sum over area". Changed the printout to have the correct header. Moved nowhere. 8216. August 4, 1994 $YPGM/GREYS.FOR Dave The combination of DOWEDGE=1 and LTYPE .LE. -2 was causing some plotting problems; the filename would be plotted on top of the wedge coordinates. Changed a couple of if statements to take the absolute value of LTYPE. Moved nowhere. 8217. August 9, 1994 Misc Eric Changed: AU5 Added prompting messages to TVPOS, IMXY, IMPOS verbs PRTTP Corrected format for positions in Export format display, the 0 fills were offset 1 column Moved nowhere. 8218. August 9, 1994 Stars Eric Changed: CCNTR Corrected (updated) the column numbers for MF files and corrected the plotting of the components to use the correct position angle wrt North. STARPL Changed names of width columns - they are not delta RA and delta DEC, but rather major and minor axes. Changed meaning of position angle to be E from N. STARS Removed any division by COS (Dec) and changed column titles for major and minor axes. STARS .HLP file: corrected spelling and changed it to describe major and minor axes rather than position uncertainty. MF2ST NEW TASK: convert MF entries to ST entries. Moved nowhere. 8219. August 9, 1994 $AIPSPUBL/COOK11.TEX & .PS Phil It was rumoured that Eric was beginning to doubt the existence of the revised version of Chapter 11 (VLBI) of the Cookbook. In order to allay his fears I have checked it back in. There are numerous changes, the most importnant of which is that the Chapter now concentrates on describing the reduction of VLBA data. This is only a first draft, we need sections on polarization and phase-referencing, however it will serve as a guide for most observations currently being reduced. Moved nowhere. 8220. August 10, 1994 $APGNOT/UVCOP.FOR Athol Improve table handling and FQ ID selection in UVCOP. Expanded table handling to include re-numbering FQID/IF/FREQ for BP tables; corrected FQID selection and re-numbering for CL, SN, SU and FG tables. Still scope for further work to reflect time, source and antenna selection in the output tables. Moved nowhere. 8221. August 10, 1994 $APLNOT/BPFSEL.FOR Athol New subroutine to re-number FQID/IF/FREQ parameters in BP tables. Analoguous to CLSEL/FGSEL/SNSEL/SUSEL. Called by UVCOP (see #8220). Moved nowhere. 8222. August 10, 1994 $APLNOT/CLSEL FGSEL SNSEL SUSEL Athol Corrected FQID selection and re-numbering in these routines. All input FQID entries were being copied to the output table and labelled as FQID 1 without selection. Improved error handling if table routines fail; did not previously report all errors. Moved nowhere. 8223. August 11, 1994 $APGNOT/SKYVE.FOR,$HLPFIL/SKYVE.HLP Mark New task to regrid a Digitized Sky Survey image onto a coordinate system and projection recognized by AIPS. 8224. August 14, 1994 $APGNOT/FITLD.FOR Phil Minor changes. 1) added the correlator supplied scaling value to the history records. 2) Changed some message priorities to those defined by Eric. 3) Added the code to prepare for the digital corrections. 4) Print out the IF sky frequencies before they are re-ordered. Moved nowhere. 8225. August 14, 1994 TVCPS Chris Added an ASPMM adverb to TVCPS. Moved nowhere. 8226. August 15, 1994 $APGNOT/SHOUV.FOR $HLPFIL/SHOUV.HLP Athol Corrected array boundary error in scanning XANT(50) in subroutine CLOSET. Caused an FPE in attempting to process elements beyond the end of the array. Also removed a reference to optypes MATX and LIST in SHOUV.HLP. Moved nowhere. 8227. August 15, 1994 UVLOD, IMLOD, FITLD Eric Changed all 3 help files and UVLOD and IMLOD source code to do no tape movement when NFILES = 0. Sent a code fragment to Phil to fix FITLD too. Linux will not allow one to BAKF at the start of the tape followed by anything besides a close. Moved nowhere. 8228. August 15, 1994 $APGNOT/SETJY.FOR Athol SETJY was adjusting the source velocity in the wrong direction for a non-zero alternate ref. pixel (APARM(1)) when using the RADIO velocity definition. Was also making the sideband correction twice in all cases. Moved nowhere. 8229. August 16, 1994 $SYSUNIX/OPTIMIZE.LIS Pat For reasons as yet unknown, if FILL is optimized with Gnu C on Linux systems, FITTP cannot even write to disk. Reduced optimization to squat. Moved nowhere, but may put in patch area... 8230. August 16, 1994 $QPGNOT/CVEL.FOR & CVEL.HLP Phil Made the antenna based LO-offset array double precision. Also removed the code that switched the direction of the total-power spectra. It was under the control of the user and was superfluous since I now understand the sign conventions and have it right (famous last words!). Moved nowhere, will need to go into the patch area. 8231. August 19, 1994 Shifting routines Phil In preparation for the bandpass correction for VLBA data I have moved all the velocity shifting routines from CVEL into the $APLNOT library. Routines are : ACSHFT, XCSHFT, WTSHFT - freq. shift total power, cross power and weight spectra DOPLR, CVLDOP - calulate velocities and interpolate in time DAYNUM, GEODLL, GETTIM, GETFRQ, VELPIX - utility routines to get various pieces of useful information SPINTP - a spectral interpolation routine TRTOCF, TRTOCS - FFT spectra and cross-correlation function to and from each other. DSHIFT - control the shifting Moved nowhere. 8232. August 19, 1994 $APGNOT/ANCAL.FOR Athol Made the following changes in ANCAL: 1) mark negative system temperatures read from the KEYIN file as indefinite (would previously generate an FPE in attempting to apply these values); 2) prevent interpolation with blanked values; 3) increased parameter MAXNX in TYCONS and introduced a check for array boundary errors (would previously generate an FPE); 4) constrain the task to update only the TY table it has created (would previously create a new TY table with the KEYIN file data but then consolidate TY 1 by adding source information). Outstanding problems: interpolates across source and fqid scan boundaries. Checked out again for further work. Moved nowhere. 8233. August 23, 1994 Single dish Eric Changed to allow for compressed data (which was not checked and did not even try to work!): PRTSD Cleaned up looping constructs, changed it to honor the requested first component (it used to subtract 1). SDGET Init for uncompressing DGETSD Changed to uncompress the data and to fix the addressing. It did not select channels and IFs correctly. Moved to the patch area. 8234. August 23, 1994 SDGRD Eric New task to select and project single-dish data, sort them, weight them, and grid them into a cube. It is a rewrite of SELSD, UVSRT, and GRIDR with some revised options. Moved to the patch area. 8235. August 24, 1994 Shifting routines Phil Added the necessary variables to DCVL.INC to allow use of PSAP Q routines. Updated TRTOCF, TRTOCS, XCSHFT & DSHIFT to initialize and use the Q routines if FFT is a power of two. Moved them to the $QNOT area. Added the routines ISPRIM & PAD to $APLNOT to determine if a spectrum is a prime number and, if so, pad it to the next power of two so the PSAP routines can be used. Changed the level of routine messages in GETFRQ & FRQUPD so they can easily be switched off if desired. Moved nowhere. 8236. August 24, 1994 $QPGNOT/CVEL Phil A much stripped down CVEL. SInce I've moved most of the routines out I had to change the way some of the internal bookkeeping is handled. Moved nowhere. 8237. August 24, 1994 $QPGNOT/BPASS Phil Added a new ability to BPASS. If the data is from the VLBA correlator (recognized by array name string in AN table) then BPASS will remove the effects of the diurnal aberration introduced in the VLBA correlator model. The effect is to shift all bandpasses in frequency space back to the RF frequency at which they were originally recorded. This enables the bandpasses to be averaged in time before further shifting. Also modified the BPASS.HLP file. Moved nowhere. 8238. August 24, 1994 SDGRD inspired Eric Changed: DIRCOS It did not handle RA-RA0 > PI for full-sky systems like AIT and GLS. LSTHDR Changed format of ref pixel adding 4 digits. MSGHDR Changed format of ref pixel adding 4 digits. PRTTP Changed format of ref pixel adding 4 digits. SDGRD Changed to do 16 planes at a time or fewer if the rows are long and the support wide. Changed to use DEC 0 as the reference value for GLS, AIT, and MER systems. Moved SDGRD and DIRCOS to patch area. 8239. August 25, 1994 $APGNOT/FXPOL Phil FXPOL will now reconstruct the IM table based on the changes in the polarization structure of the data. We needed this because the correlator group are currently unable to correct the output FITS data for dual polarization without a major rewrite of the software. Also updated the .HLP file. Moved nowhere. 8240. August 26, 1994 Single dish Eric Changed PRTSD to display 0.1's of a second. Installed Jeff Mangum's submitted task UVF, revised heavily by me, in $APLCONTR. It has a C main program which can run stand-alone or as an AIPS task using a set of subroutines in UVFAIP.FOR. It does require the fitsio package which is available on anonymous ftp from Bill Pence. Moved these to the patch area. 8241. August 28, 1994 $APGNOT/UVFLG Leonia The new parameters in APARM have been installed. These parameters limit the range of amplitudes and weights of a visibility to be flagged or unflagged. This new option works only for a single source file. Moved nowhere. 8242. August 29, 1994 PRTUV, UVDIF Eric Changed PRTUV to set phase to 0 for autocorrelations Changed UVDIF to check that antenna numbers are in range 1-MAXANT, not 1-28 and to skip that test for singledish data. Moved nowhere. 8243. August 29, 1994 OTFUV Eric New task to read singledish data SDD format for on the fly mapping and convert it directly into an aips uv file. Moved to patch area. 8244. August 29, 1994 $APGNOT/SETJY Phil Added the newly determined flux values for the source 1934-638. It is the primary flux calibrator for the ATCA. Also updated the .HLP file. Moved nowhere. 8245. August 31, 1994 $QPGNOT/BLCAL.FOR Athol BLCAL did not set FACGRD, the factor used by VISDFT from DGDS.INC for scaling CC components. BLCAL could not divide by a CC model as FACGRD was zero. Set FACGRD to 1 in BLCMOD. Moved nowhere. 8246. September 1, 1994 $QPGNOT/FRING.FOR Phil Changed the way in which delay searching is turned off. IFf the user does not want a delay determined then DPARM(2) should now be set to a negative number. Also updated .HLP file. Moved nowhere. 8247. September 1, 1994 $APLNOT/TABBP,BPINI,BPREFM Phil In preparation for dealing with VLBA bandpasses I have renamed the IFFREQ column in the BP table - which I never used anywhere (that's foresight, eh!) - to CHN_SHIFT. It nows contains the number of channels the BP entry has been frequency shifted from its RF value. TABBP and BPINI needed to know about the change and BPREFM needed to update older tables. Moved nowhere. 8248. September 2, 1994 $INC/DANS.INC $APLNOT/GETANT Phil Added the antenna numbers into an array (TELNO), they do not always correspond to the row number. Updated GETANT to get the information from the AN table. Moved nowhere. 8249. September 6, 1994 $APLNOT/BPREFM Phil The error code from the call to TABBP was not being checked correctly. Only non zero positive values are real errors, negative values indicate some degree of flagging within the BP table. Taught BPREFM to deal with this and also return a more sensible error message. Moved nowhere. 8250. September 6, 1994 $YPGNOT/FRMAP Leonia The new task FRMAP installed in $YPGNOT to provide fringe-rate mapping of spectral line sources. Moved nowhere. 8251. September 6, 1994 $QPGNOT/BPASS Phil Some of the BP tables generated by VLBA data were too big for the buffer sizes allocated to the sorting routines. Changed. Moved nowhere. 8252. September 7, 1994 $YPGNOT/VBPLT Phil The baseline selection was a real mess. I removed the flawed subroutine FILANT and replaced it with a much neater routine BASLST. This now works nicely. In addition, when averaging in time, the list of baselines to be averaged was being set up incorrectly, also fixed this. Moved nowhere. 8253. September 8, 1994 $APGNOT/UVFLG Leonia Default value of new parameters APARM meant not to make flagging (unflagging). I change it to the opposite. Now if APARM = 0, the task does not pay attention to amplitudes or weights and works identicaly to old version. 8254. September 9, 1994 $INC/DSEL.INC Phil Added 3 new variables to serve as BP table keywords. Needed for the VLBA bandpass corrections. Moved nowhere. 8255. September 9, 1994 $APGNOT/SNSMO Phil The amplitude in the SN table for the 2nd polzn was being doubled when smoothing. Removed the offending code. Bill says there is a lot more trouble lurking in here so no doubt we'll be back. Moved nowhere. 8256. September 9, 1994 BP table changes Phil Added 3 new keywords to the BP table, needed for the vlba bandpass corrections. Necessitated a change to the call seq of BPINI which of course meant changes to the following routines: BPREFM, BPFSEL, BPASET, TABLEUTIL, BPCOR, UNCAL, FITLD, BPASS, CVEL and POSSM. Moved nowhere. 8257. September 9, 1994 $YPGNOT/POSSM.FOR Phil With VLBA bandpass tables I've added a 3rd dimension and so that requires an additional adverb to select the bandpass required when plottong the BP table. Added the adverb FQTOL to select the required shift, it is not invoked if data are non-VLBA, if VLBA data and FQTOL = -999, all bp anetries for an antenna are added together. FQTOL = 0 is a sensible choice. Also modified .HLP file. Moved nowhere. 8258. September 12, 1994 UVDIF Eric Found a test with an OR in it that caused it to ignore differences in flagging! Changed the OR to an AND. Moved to patch area. 8259. September 12, 1994 UVLSF, UVLIN Eric Changed the pre-cursor remarks and categories for UVLIN.HLP. Added to UVLSF the option to flag data based on the peak residual in the selected channels (for fitting a baseline) and on the rms in those channels. UVLSF does not need the shift option of UVLIN since UVFIX now does it correctly. Moved nowhere. 8260. September 12, 1994 SPLIT Eric Removed statement which kept SPLIT from using an FG table on a single-source file. Other cal programs will use one. Moved to patch area. 8261. September 12, 1994 $APGNOT/UVAVG.FOR Athol UVAVG did not average random parameter INTTIM correctly; output values were substantially overestimated due to an initialisation error. Moved nowhere. 8262. September 13, 1994 CALIB, FRING, HORUS Eric All 3 refused to apply flagging to single-source data sets and the firsrt two also turned off bandpass and gain calibration for such data sets. This is pretty arbitrary, especially since dozens of other tasks do not do this. So removed the3se arbitrary prohibitions. Moved to the patch area. 8263. September 13, 1994 $SYSUNIX/LINK Pat The use of separate versions of LIBR.DAT for shared or non- shared libraries was not working correctly. For Sun4, there are two files LIBR.DAT and LIBR.DAT.SHARED supplied, and if the installer selects the use of shared libraries, the former is renamed LIBR.DAT.UNSHARED and the latter to LIBR.DAT. The generation of the list of libraries by LINK was not taking this into account. It does now. Moved to the patch area. 8264. September 13, 1994 $APGNOT/SPCAL.FOR $HLPFIL/SPCAL.HLP Athol New task for spectral line polarization feed calibration; assumes circular feeds and moderate linear polarization. Moved nowhere. 8265. September 14, 1994 $APGOOP/IMCLP Bill C. New task to clip the range of values in an image. Values outside of the range are replaced with the closest value inside the range. Also IMCLP.HLP. Moved nowhere. 8266. September 14, 1994 SDGRD Eric Changed help and code to allow abs(weight) cutoff as well as weight cutoff. Moved to patch area. 8267. September 14, $INC/DSEL.INC Phil Added a new array to keep track of the current frequency shift for each antenna. Moved nowhere. 8268. September 15, 1994 $INC/DCVL.INC Phil Added some new variables to this include for use in another new subroutine that will soon be back in the system. Alos changed the name of some variables to avoid clahses with those in DSEL.INC. This necessitated changes in the following routines: CVLDOP, FRQUPD, GETFRQ, DSHIFT & CVEL.FOR. Moved nowhere. 8269. September 16, 1994 $APLNOT/GETTIM, $QNOT/DSHIFT Phil Added CATBLK to GETTIM's call sequence, caused change in DSHIFT. Moved nowhere. 8270. September 16, 1994 $QPGNOT/BPASS Phil Tidied up the BPASS code, had been experimenting with fanning out the bandpass entries over channel shift space and had left the inactive code in BPASS. Removed it and so BPASS now operates as described in change entry # 8237. Moved nowhere. 8271. September 16, 1994 $APLNOT/XCSHNQ, NQTOCF, NQTOCF Phil Versions of $QNOT/XCSHFT, TRTOCF & TRTOCS that do not use Q routines. SInce VLBA bandpass correction is going to require FFT's and complex shifting (only occasionally in time) we need non Q versions of these routines to avoid having to shift the whole calibration system to $QPGNOT. Moved nowhere. 8272. September 16, 1994 $YPGNOT/POSSM.FOR Phil Since I've changed slightly the implementation of the VLVA bandpass correction I no longer need to allow POSSM to select among the fanned out BP entries. So I've removed the FQTOL adverb added a while ago. Moved nowhere. 8273. September 16, 1994 ZFIO, OTFUV Eric To enable ZFIO to read non-AIPS files it must be able to read the last bytes in a file even if they are < 1024 bytes. Changed ZFI2.C to return error code 1000+n (n = # words read okay) in this case and changed ZFIO to zero out the remainder. Changed OTFUV to use this return. Moved to patch area for SD people. 8274. September 16, 1994 SDGRD Eric Changed and corrected the blanking test to allow no blanking when there is no division (weight or convolution outputs) and to do blanking on the abs(weight) if so specified. Also fix up help file to explain. Moved to patch area for SD. 8275. September 16, 1994 CONPL Eric New task: plots AIPS convolution functions in the upper half. In the lower it can plot the convolution of the upper function with a Gaussian. Or, in the lower, it can plot the Fourier transform of the convolving function or the expected signal to noise for a uniformly sampled aperture. These last two can be plotted as X, 1/X, or log(X) forms. Moved to patch area for fun. 8276. September 16, 1994 DBCON Eric DBCON insists on adding -SIN to the UU and VV parts of the random parameters. This is not good for single dish data. So fixed the addition to limit it to real uv data. Moved to patch area. 8277. September 17, 1994 TVFLG, SPFLG, IBLED, UVFLG Eric Changed all 4 tasks to use FLAGVER even for single-source files, creating a flag table if needed and FLAGVER > 0. They will write in a flag table if one already exists. Only if fLAGVER = 0 and no FG table exists will they actually flag the single-source data. All help files too of course. Moved nowhere. 8278. September 19, 1994 $APGNOT/FITLD Phil Added the ability to remove the FFT artifacts from total power spectra (they are washed out in the correlator from the cross-power spectra). At the moment that correction spectra (supplied by Chuck Broadwell) are hard coded into data statements inside FITLD, but after the correlator software revision is over they will be passed through the correlator as FA tables and applied by FITLD. A minor change was to skip to the end of a file being read if the time detected was > the timerange specified by the user. Moved nowhere. 8279. September 20, 1994 $YPGNOT/FRMAP Leonia Installed the blank option for the source selection. Now if the source is identified as blank, the first source from the data is mapped. Additionnaly the information about maximum possible window size is printed now. Moved nowhere. 8280. September 20, 1994 $APGNOT/INDXR.FOR Athol Increased array GEODLY in subroutine WRITCL. Subroutine GETDEL, which gets the VLBA correlator model from the IM table, was writing over the array boundary and corrupting other fields in the CL table created by INDXR. Also, now warns the user that CPARM(3) and CPARM(4) are ignored if CL#1 already exists. Moved nowhere. 8281. September 21, 1994 $APGNOT/FITTP.FOR Dave FITTP was claiming to report the number of unflagged records in the CL table when in fact it was reporting the flagged records. Changed so it correctly reports the unflagged records. Moved nowhere. 8282. September 22, 1994 $YPGNOT/FRMAP.FOR Leonia In a single source file the task (using standard subroutine GETSOU) did not recalculate RA and DEC on the observation time. As a result there was a small discrepancy between the single- and multi source files. I've written a new version of GETSOU - GETSO which recalculates RA and DEC to the apparent position. So now we have RAAPP and DECAPP for both single and multi source files. Moved nowhere. 8283. September 22, 1994 $APGNOT/SPLIT.FOR Phil Stopped SPLIT from copying the IM, CT, MC & AT tables to its output file. They are not needed at that point. Moved nowhere. 8284. September 22, 1994 UVPRT Eric Added more flagging of data automatically in Holography mode: if W = 40000 or the previous W = 40000 and the current is 40001, then delete the sample. Moved to the AOC $LOAD area. 8285. September 22, 1994 ZDAOPN Eric Changed $APLUNIX ZDAOPN.C to attempt a read-only file open when the read write open fails for access privilege and non-exclusive use is requested. This will fail later if a write is attempted but will allow read only privilege for reads. Moved to the patch area. 8286. September 22, 1994 DBCON Eric Changed it to skip the scaling of "U" and "V" for single-dish data. This was shifting positions in an attempt to make channel 1 the ref channel. Moved to the patch area. 8287. September 23, 1994 WFCLN Eric In the SUM mode it set NCHAV = BCHAN-ECHAN+1 (sigh..). This negative number of channels was used to dimension arrays etc and things failed only down at the primary beam correction function which summed over no channels and divided by the 0 sum. Also changed $QOOP/PBUTIL.FOR to avoid thhis 0 divide. Moved to the patch area. 8288. September 27, 1994 various $SYSUNIX scripts Pat COMRPL and COMLNK were failing for routines compiled in the $APLCONTR area, as they assumed it was a valid AIPS area (it's under $AIPS_VERSION) but was not found in LIBR.DAT. I've modified (hacked) COMRPL, COMLNK, FC, AIPSCC, SEARCH and LINK so they now recognise $APLCONTR as "different" and requiring an options file. Moved nowhere (yet). 8289. September 28, 1994 $APLPGM/PRTAB & $APGNOT/FITTP Phil Both of these tasks had problems with the large spectral sizes of VLBA databases. One particular project has 4 IFs and 512 spectral channels and was overflowing the declared buffers of 4096 words. Increased them where approrpiate to 8192 words. Moved nowhere. 8290. September 30, 1994 $APGNOT/FITTP Phil/Leonia The size of a buffer in subroutine EXTWRT was not enough for writing especially large tables. The parameter MXLREC has been doubled. Moved nowhere. 8291. September 30, 1994 UVFLG Eric Corrected it to handle compressed data and to flag single-dish rather than always using an FG table. Changed GETNAN to suppress the missing AN file message for single-dish data. Moved nowhere. 8292. October 1, 1994 TVCPS Eric Upgraded the help file and added an option to force the program to use portrait mode. Changed the program to use landscape when that will fit the image better except when prohibited by the user. Moved nowhere. 8293. October 3, 1994 FITS tables Eric Changed DTHD.INC to allow for 128 columns - it limited us to 60. Changed TABHDR and TABHDK to check for the column limit intrinsic to AIPS of 128 and to return an error if too many requested. Changed UVLOD, IMLOD and GSCAT to skip too wide tables. Phil will get FITLD. Moved nowhere. 8294. October 3, 1994 $APGNOT/SNCOR Phil Added a new option, CPRT, to copy the rates from one IF to all other specified IFs. Also changed SNCOR.HLP. Moved nowhere. 8295. October 3, 1994 $APGNOT/FITLD.FOR Phil Added the option of selecting spectral channels from each IF. Ensured that, if the file expansion failed due to lack of disc space, FITLD dies elegantly. Added Eric's change from #8293 above. Updated FITLD.HLP to reflect channel selection change. Moved nowhere. 8296. October 4, 1994 $QPGNOT/FRING.FOR Phil If no SBD delay search requested, force the solutions to really be zero so that phase offsets do not creep in. Moved nowhere. 8297. October 4, 1994 $APGNOT/SNCOR Phil The fix I put in yesterday had a minor logic error in that it did not overwrite flagged rates in specified IFS. Fixed. Moved nowhere. 8298. October 5, 1994 $APGNOT/FITTP Leonia This is a repeat of the same thing which me and Phill made three days ago. With misunderstood reason the version of FITTP which had existed before Phil correction (8289 Sept. 29) appeared in the CFA system. Moved nowhere. 8299. October 4, 1994 SDGRD Eric Changed $QPSAP QGRIDA to recognize a type 4 visibility and call a new routine QGRD5 to grid visibility data that have 0 for the imaginary part and for which the beam grid is returned in the imaginary part. Changed SDGRD to invoke this, shaving about 20% off the run and cpu times for the test data set. Moved nowhere so far - I intend to experiment more. 8300. October 5, 1994 TOUCH Eric Some major computational tasks may not read a SCratch file header often enough to keep its catalog time updated. The VLA automatic TIMDEST was deleting SC files that were actually open and "locked" by another computer. Since the AOC does not use a file locking that works over NFS, the lock did not prevent the deletion. TOUCH will silently update all SC files in the DFIL common if the last call to it was more than 10 minutes ago. Changed APCLN, SDCLN, MX, and ($QOOP) QCLEAN to call it. Moved nowhere. 8301. October 6, 1994 SDGRD Eric Corrected bug introduced in 8299 leading to bad rows at the bottom of each frame. Added code to do the problem in-core where possible. Created new grod routine $QPSAP/QGRD6.FOR and revised help to explain new uniform weight options. Moved to patch area with 8299. 8302. October 7, 1994 UVMOD Eric/Arie Grossman Subroutine BESJ was giving floating point errors for planet visibilities. Replaced with BESJ1 from Numerical Recipes. There was a missing factor of 2 in the calculation of disk size which was also fixed so that BMAJ and BMIN are correctly the major and minor diameters. Code submitted by Arie Grossman of the University of Maryland. Moved nowhere. 8303. October 11, 1994 $APGNOT/SHOUV.FOR $HLPFIL/SHOUV.HLP Athol VLBA tests required closure phases to higher than 1 deg precision. Modified SHOUV to print all phase quantities in decimal form if DPARM(3), the no. of digits per column, is sufficient (allowed range: 5-9). Toned down the max. allowed column width for amplitudes from 10 to 7 as the the single precision real rounding calculations were breaking down. Both amplitude and phase can now be printed to a maximum precision of ~6 significant digits. Changed the format of the scaling factor to avoid scaling information of the form "1000 = 0.0000 Jy". Moved nowhere. 8304. October 11, 1994 $QPGNOT/BLCAL.FOR Athol Add a warning message when appending solutions to an existing BL table; in response to a gripe. Moved nowhere. 8305. October 11, 1994 $APGNOT/SNCOR Phil Added another option to SNCOR. This one is called 'CPSN' and is a generalization of 'CPRT', it will copy the whole solution from one IF to another within the SN table. Also modified .HLP to reflect this change. Moved nowhere. 8306. October 12, 1994 IMVIM Eric The call to GINIT requires the input parameter array returned by GTPARM, preferably with the defaults filled in by actual values. Fixed this plot task - it called GINIT with the character form of the first (HOLLERITH) variable in the parameter array. Also changed it to fill in more of the defaults. Moved nowhere. 8307. October 13, 1994 SETJY.HLP Eric The last person to edit SETJY.HLP must have used an old version of the file - eliminating the changes I put in to allow users to reset a calcode to blank. Put my changes back. Moved nowhere. 8308. October 13, 1994 UVLSF Eric Changed UVLSF and its help to allow for shifting the phase center before fitting the continuum and then shifting both back before writing them out. It takes twice as long but it could be worth it. UVLIN is still not obsolete since it does cutoffs with a different algorithm and offers a complete flagging report if so desired. Moved to the patch area since the CookBook will mention this. 8309. October 13, 1994 NEWPOS Eric Changed NEWPOS to handle checking of large angles differently. It had a test that made no sense and limited MER, GLS, and AIT. Also added a test that is sensible to MER. Moved nowhere. 8310. October 17, 1994 FRMAP Leonia Add a new option in FRMAP. Now it can do mapping using absolute values of fringe rates without referencing to a reference channel. Moved nowhere. 8311. October 19, 1994 EHEX, REHEX Pat Two new shell scripts added to $SYSUNIX. These convert from decimal to "extended" hex (base 36) and back, respectively. Moved nowhere. 8312. October 20, 1994 $APGNOT/UVAVG.FOR Phil Added the adverb ZINC to the task. If ZINC is non-zero it specifies the pre-average time of the data thus enabling UVAVG to estimate how large an output file need be created. If ZINC is zero it reverts to its old ways. Also updated the helpfile. Moved nowhere. 8313. October 20, 1994 SDCLN Chris Fixed a bug in calling MAPOPN (name and class were passed as Hollerith rather than CHARACTER) that caused sporadic failures. Moved nowhere. 8314. October 20, 1994 $APGNOT/UVCOP Phil Added the adverb NCOUNT to UVCOP. When selecting AC data only UVCOP would create an output file that could hold all selected AC and XC data, in these days of huge filesizes this is impractical. NCOUNT specifies the fraction of AC data to all data in order to limit the size of output file created. Moved nowhere. 8315. October 23, 1994 Number base Eric Changed the basic numbering scheme in aips from hexadecimal to extended (base 36) hex. This means we can now have 35 disks, 1295 TVs, 46655 users (1, 2, and 3 digits, resp.). Changed: To allow more tapes and disks - $INC/DDCH.INC To generate and understand base 36 - $APLGEN/ZEHEX (new) to encode extended hex $APLGEN/ZHEX10 changed to do base 36 rather than 16 The file name is to change, adding another digit in the version field and droppin the 1 of the ;1. The following required simple changes for name structure and to use ZEHEX rather than ZHEX: $APLSUN/ZBKLD2 $APLSUN/ZBKTP2 $APLIBM/ZBKLD2 $APLUNIX/ZBKLD2 $APLUNIX/ZBKTP2 $APLVMS/ZBKLD2 $APLVMS/ZBKTP2 Limit checking changes with the change in base as well as calling ZEHEX rather than ZHEX: $APLVMS/ZMOUNT $APLGEN/ZMOUNT $APLGEN/ZMOUNR $SYSCVEX/SPACE Basic changes: $APLGEN/ZPHFIL changed to do new names $APLGEN/ZPHOLV changed to do either old or new names Changed to call ZEHEX rather than ZHEX were: $AIPGUNIX/ZSTRTB $AIPNOT/TPMON $APLCVEX/ZTPOPN $APLUNIX/ZBKLD1 $APLUNIX/ZBKLD3 $APLUNIX/ZBKTP1 $APLUNIX/ZDIR $APLVMS/ZBKLD1 $APLVMS/ZBKLD3 $APLVMS/ZBKTP1 $APLVMS/ZBKTP3 $APLVMS/ZDIR $APLVMS/ZDOPRT $APLVMS/ZFRE2 $APLVMS/ZMOUN2 $APLGEN/ZIVSOP $APLGEN/ZSTAIP $APLGEN/ZTPOPN $APGNOT/FILLM $APLSUB/LPOPEN $APLSUB/OFMDIR $APLVMS/ZSHCMD (only partly changed - rest real hex) Changed to call ZEHEX rather than do their own hex conversions were: $AIPGUNIX/ZSTRTA $AIPGVMS/ZSTRTA $AIPSUB/AU2 $AIPSUB/PRTMSG $APLCRI/ZDCHIN $APLCRI/ZIVSOP $APLDEC/ZTTOPN $APL3090/ZDCHIN $APL3090/ZTTOPN $APLVMS/ZACTV8 $APLUNIX/ZACTV8 $APLUNIX/ZTTOPN $APLGEN/ZDCHIN $APLGEN/ZWHOMI $APLSUB/BLDTNM $APLSUB/MSGWRT $APLSUB/WHOAMI $APLCONTR/UVFAIP $APL/DEV/ZIVSOP $APL/DEV/ZSTAIP $APL/DEV/ZMOUNR $AIPNOT/TPMON $APGNOT/LWPLA (drop large dim on ICHBUF too) Change version limits from 255 to 46655 in: $AIPSUB/AU3A $AIPSUB/AU8 $AIPSUB/AU8A $AIPSUB/SGLOCA $APLPGM/DISKU $APLSUB/EXTINI $APLSUB/MADDEX $APLSUB/MDESTR $APLSUB/TABINI $QPGOOP/SCMAP $QPGOOP/WFCLN $QPGM/RSTOR $QYPGNOT/SDCLN $QYPGM/APCLN Change device limits from 255 to 1295, user limits from 4095 to 46655 in: $AIPPGM/FILAI2 $AIPPGM/FILAIP $AIPPGM/FILINI $AIPPGM/SETSP $AIPNOT/SETPAR $AIPSUB/CATCR $APLSUB/CATDIR $APLSUB/CATOPN $APLSUB/IAMOK $APLPGM/PRTAC also parameterize and check counter sizes Correct sizes of DASSGN and TIMEDA and their support buffers in $AIPPGM/AIPSC $AIPSUB/AU3A $APL3090/ZDCHIN $APLCRI/ZDCHIN $APLGEN/ZDCHIN $APLPGM/DISKU $APLGEN/ZFREE up buffers of disk values Moved to CIT and CfA inadvertently. 8316. October 23, 1994 New header parms Eric Changed DHDR.INC and VHDRIN to define pointers: KRCOK Code to indicate WCS parameters are usable KDLON LONGPOLE value KDPRJ PROJP1...PROJP9 KRPCM PC pixel conversion matrix (7 x 7) There are still 17 free words in the header! Moved nowhere. 8317. October 23, 1994 UPDAT Eric Dusted off the old format conversion program. Changed it to convert from C to D with modern release names (none before the overhaul). Uses ZPHOLV to find old and new names and then does a rename. Added code to open all CB files, read record 1 and set the header values for the coordinates words. Moved nowhere. 8318. October 23, 1994 CLBOX Eric New adverb CLBOX - 50 clean boxes. Changed POPSDAT.HLP and NEWPARMS.001 to define the adverb and created CLBOX help file. Changed help files named: APCLN CXCLN MX SCMAP SDCLN WFCLN GREYS REBOX TKBOX TKNBOXS TVBOX COMAP COMAP_UV DDTSAVE HYB MAPIT MAPIT_UV POPSDAT Changed procedures to use CLBOX in $RUNSYS DDTLOAD.001 NEWPARMS.001 COMAP.001 HYB.001 MAPIT.001 Changed dimensions and names of BOX in fortran routines: $AIPSUB/AU5C $YSUB/GRBOXS $QOOP/QCLEAN $APLOOP/CLEAN $QYPGM/APCLN $QYPGNOT/SDCLN $QYPGNOT/MX $QPGOOP/CXCLN $QPGOOP/SCMAP $QPGOOP/WFCLN $YPGM/GREYS Moved nowhere. 8319. October 23, 1994 LIBR.DAT Eric On SUNS at least LIBR.DAT needed updating for $APLPGM. A final pass of APLSUN/SUBLIB is now required - apparently I changed something that used to pick up zptbit and zgtbit earlier in the link. Sigh... Moved nowhere. 8320. October 24, 1994 LIBR.DAT Eric The link lists for YPGNOT, QYPGM, QYPGNOT, QPGNOT, and QPGM required another reference to APLSUN in the $SYSSUN version for some - not all - programs. Moved nowhere. 8321. October 24, 1994 CATINI Eric Created a subroutine CATINI which 0's and blank files an image header and puts 1's for the PC matrix diagonals, coordinate increments, reference pixels, and number of ppixels. Changed UVLOD, IMLOD, FILLM, MANDL, PATGN, UVFIL, and CANDY to use it. Moved nowhere. 8322. October 24, 1994 Unused ... Eric Changed due to warnings in mass compile: MCUBE A common was - 1 place - not the same length as the rest - not really a problem. TAFFY Put the OLD header in a common without regard for the alignment (which was off). This was bad. Also removed unused variables in AHIST HISEQ PRTCC TASAV TRANS UVLSF XSMTH XSUM Moved nowhere. 8323. October 25, 1994 More unused Eric Removed unused variables from: $AIPGUNIX/ZSTRTB $AIPSUB/TBHEAD $QPSAP/QCSQTR $QPSAP/QFINGR $QPSAP/QGASUB $QPSAP/QGRD1 $QPSAP/QGRD2 $QPSAP/QGRD3 $QPSAP/QGRD4 $QPSAP/QINTP $QPSAP/QMAKMS $QPSAP/QMENT $QPSAP/QPTDIV $QPSAP/QSPDIV $QPSAP/QSPSUB $QPSAP/QUVIN $QPSAP/QVINDE $QPSAP/QVTRAN $QSUN/QCLNSU $QNOT/APWRIT $QNOT/CONV1 $QNOT/CONV2 $QNOT/CONV4 $APLNOT/BLREFM $APLNOT/BPINI $APLNOT/FILSWP $APLNOT/VMBLKD $QPGNOT/CALIB $YPGM/BLANK $YPGM/BLSUM $YPGVDEV/TVMON $YPGNOT/IMLHS $YPGNOT/TVHLD $AIPGUNIX/ZSTRTB $APGNOT/ADDIF $APGNOT/ASCOR $APGNOT/CCMOD $APGNOT/GEOM $APGNOT/GLENS $APGNOT/IMFLT $APGNOT/IMMOD $APGNOT/NINER $APGNOT/NNLSQ $APGNOT/POLCO $APGNOT/REMAG $APGNOT/RGBMP $APGNOT/SHADW $APGNOT/STFUN $APGNOT/SUMSQ $APGNOT/TBIN $APGNOT/TBOUT $APGNOT/WARP Changed the include file DWXV.INC and the call sequences of $APLNOT WXINI WXIO and WXTAB to eliminate a pile of unused variables. Of course, it looks as if WX itself is unused. Raised buffer sizes in BLANK and BLSUM. ASCOR and SHADW both had a CATBLK-like variable ina common, not on a correct boundary. Moved nowhere. 8324. October 25, 1994 CALIB Eric Corrected a bug: the loop that averaged the observed gains was done over all frequency channels (ECHAN-BCHAN+1) rather than only over all IFs. This caused addressing problems. Moved nowhere. 8325. October 27, 1994 LIBR.DAT Pat Essentially the same changes as Eric made in #8320 above, but for the Solaris and Linux versions. Had to hit APLPGM as well. Moved nowhere. 8326. October 26, 1994 UPDAT Eric Added the ability to handle missing and 0-length CB files, fixing the CA file entry for them and closing all open files to allow the program to continue. Moved to IBM immediately. 8327. October 26, 1994 IBM complaints Eric TVCPS Error in FORMAT - X with no number. PLCUB Lower case in code BLANK Comma missing in FORMAT. VLBIN TAB characters. UVNOU TAB characters. UVLIN TAB characters. SBCOR TAB characters. MODVF TAB characters. AFILE List-directed IO to string. FRMAP Comma missing in FORMAT. CLPLT Comma missing in FORMAT. HLPB String length on initialization. SWPOL Comma missing in FORMATs, removed SAVEs of COMMONs and declared them in the MAIN (the SAVE is required if the COMMON is not at a high level, but upsets the IBMs) SETAN String length on initialization. PRTAN Comma missing in FORMAT. FFTV2 Commons initialized in DATA. SNINI String lengths. Moved nowhere. 8328. October 26, 1994 HOLGR, FFT Eric Changed for the 2nd time the subroutine named FFT. We cannot have one named the same as a task. Its new name is FFTMC and changed HOLGR to call it by that name. Moved nowhere. 8329. October 27, 1994 IAND, IOR, IEOR Eric Fortran compilers think they know what IAND, IOR, and IEOR are and that is illegal for the way we use these routines (which we actually provide). So change all references to them to ZAND, ZOR, and ZEOR, create $APLGEN Z routines of these names (in C), and $APLVMS routines in Fortran. Subroutines changed: $AIPSUB/AU1A $AIPSUB/AU5 $AIPSUB/AU6 $AIPSUB/AU6B $AIPSUB/AU6C $APLCRI/ZR64RL $APLSUB/DECBIT $APLSUB/DWRITE $APLSUB/RANDUM $APLSUB/TKCURS $YDEA/YGRAPH $YDEA/YIMGIO $YDEA/YLUT $YDEA/YMKCUR $YDEA/YSCROL $YDEA/YZOOMC $APL3090/ZR32RL $APL3090/ZR64RL $APLVMS/ZSHCMD $APLVMS/ZTAP2 $APLVMS/ZTKFI2 $APLVMS/ZTPWA2 $YGEN/YCHRW $YGEN/YCOVER $YGEN/YCUCOR $YGEN/YCURSE $YGEN/YFIND $YGEN/YLOCAT $YGEN/YLOWON $YGEN/YSLECT $YSUB/TVFIND $YIVAS/YBUTON $YIVAS/YCHRW $YIVAS/YCUCOR $YIVAS/YCURSE $YIVAS/YGRAPH $YIVAS/YIMGIO $YIVAS/YLUT $YIVAS/YOFM $YIVAS/YSCROL $YIVAS/YSPLIT $APLGEN/ZGTBIT $APLGEN/ZR32RL $APLGEN/ZR64RL $YM70/YCUCOR $YM70/YCURSE $YM70/YSCROL $YM75/YSCROL $YIIS/YGRAPH $YIIS/YIMGIO $YVTV/YFIND $YVTV/YLOCAT $YVTV/YSLECT $YSS/YIMGIO $YSS/YSCROL $YSS/YSPLIT Moved nowhere. 8330. October 27, 1994 IAND, IOR, IEOR Eric Fortran compilers think they know what IAND, IOR, and IEOR are and that is illegal for the way we use these routines (which we actually provide). So change all references to them to ZAND, ZOR, and ZEOR and changed tasks: $APGNOT/FILLM $APGNOT/IMLOD $APLPGM/PRTTP $APLPGM/SAD $APLPGM/CPYRT $APLCONTR/TVFLM $APLCONTR/WSLOD $YPGM/BLANK $YPGM/BLSUM $YPGM/TVRGB $YPGNOT/IBLED $YPGNOT/TVCPS $YPGNOT/SPFLG $YPGNOT/TVFLG MOved nowhere. 8331. October 27, 1994 unused Eric Suns think that variables that are declared and DATAed are "used" even if they are not. Linux knows better. Inconsistent call sequences are also pointed out. These are usually okay, but a nuisance worth fixing usually. Changed only those files that seemed to have IOR, IAND or IEOR in them: $AIPSUB/AU1A $AIPSUB/AU6B $AIPSUB/PRTMSG $APLSUB/TKCURS $QOOP/QCLEAN $YGEN/YCOVER $YSS/YIMGIO $AIPPGM/BATER $AIPPGM/GRIPR $APGNOT/CANPL $APGNOT/DBCON $APGNOT/FILLM $APGNOT/GEOM $APGNOT/HGEOM $APGNOT/IMLOD $APGNOT/LGEOM $APGNOT/LWPLA $APGNOT/MK3IN $APGNOT/PGEOM $APGNOT/SLICE $APGNOT/TXPL $APGNOT/STRAN $APLPGM/PRTPL $APLPGM/PRTTP $APLPGM/QMSPL $APLPGM/TKPL $APLPGM/XSMTH $APLPGM/SAD $APLPGM/CPYRT $QYPGNOT/MX $QYPGNOT/SDCLN $QYPGM/APCLN $YPGM/BLANK $YPGM/BLSUM $YPGM/TVRGB $YPGM/TVPL $YPGNOT/IBLED $YPGNOT/SPFLG $YPGNOT/TVFLG MK3IN: One of the inconsistent call sequences was an error, leading to bad seconds in a time string. Dropped the SAVEing of labeled commons in MK3IN, declaring them in the MAIN. SAD: changed a typo in a column name and added a REVISION keyword with current value 3. MX: one of the header block saves and restores only did the first half - a holdover from the code overhaul! Moved nowhere. 8332. October 27, 1994 GREYS Gustaaf/Eric The slice plotting routine never worked. It used parts of the name advers for the corners. Moved nowhere. 8333. October 28, 1994 $APGNOT/CLCOR.FOR Leonia As I've discovered just recently VLA and VLBI have different system of coordinates for their antennas. X axis is located in Greenwich meridian for VLBI for VLBI antennas and in local meridian for VLA antennas. The subroutine ANTPOS calculated local hour angle before February 1994 when I changed it for the Greenwich hour angle. So the programm calculated correctly correction for VLA's antenna position and incorectly for VLBI's before the February and vice versa after the February. I've instaled the new parameter CLCORPRM(7) which distinguish VLBI and VLA. Moved nowhere. 8334. October 31, 1994 $YPGM/UVHGM.FOR Leonia The task has been terminated with arithmetic esception in all options. I fixed it. Moved nowhere. 8335. October 31, 1994 MX Eric Certain summing arrays were too small for averaging across a lot of channels/IFs. Raised them to MAXCIF size (from 128 - a factor of 64). Moved nowhere. 8336. November 1, 1994 $APGOOP/DTSUM.FOR Leonia I have increased the number of digits in integration time print and added print of number of 131072 microsec intervals producing the integration time. Moved nowhere. 8337. November 1, 1994 ZCREA2 Eric None of the versions of this cleared the buffer that they wrote to disk. So it has been non-zero for a long time. Changed adding a memset call in versions $ALPUNIX, $APLBERK, $APLIBM, $APL3090, $APLDEC, $APLULTRX, and $APLSGI. If massive updates are to be done INSTEP2 should not be used. It caused the COMRPL.UPD file to point at APLSUN which is then ignored by IBM's. So ZDCHI2.FOR was not recompiled and hence set NTAB3 = 0 (no maps allowed)! Whole system needs relinking for ZCREA2 fortunately. Moved nowhere. 8338. November 1, 1994 SDGRD Eric Corrected the gridding - it forgot to rezero the array processor summing areas for later passes. Therefore, if more than one pass was required to do all the channels, the later channel images contained ghosts of the earlier channels. Halloween I guess. Moved to the patch area. 8339. November 2, 1994 KNTR Eric Changed the labeling to put numbers on all left edge vertical axes but letters only once and to put numbers on every other bottom axis with letters only once. Moved nowhere. 8340. November 3, 1994 CALIB, SNPLT Eric The mean gain modulus was being computed wrongly by CALIB. It was computed separately for each sub-array and the last sub-array's value was written to the table. Moved the summimg parameters to common, set them to zero in the init routine, and let it keep summing throughout. SNPLT was multiplying by the mean gain when it should have been dividing. Moved to the patch area. 8341. November 4, 1994 CookBook Eric Rewrote Chapter 10 of the CookBook using Gustaaf's draft for guidance. Added $AIPSFIG figures KNTR.PLT and TVCUBE.PLT. Updated the table of contents COOK0.TEX for chapter 10 and to add tables of figures and recipes. Moved to the WWW area. 8342. November 4, 1994 $APGNOT/VBGLU Phil VBGLU was telling the users that parameter MAXBAS neede to be increased. It was and VBGLU then worked....except when I originally wrote this beast I forgot to glue together the SU tables from the input files. Fixed this up and tidied up the GLUTAB routine to deal with tables without a time axis. Moved nowhere. 8343. November 4, 1994 $APLNOT/SPINTP.FOR Phil SPINTP was overwriting the input weight array with the corrected weights rather than overwriting the output weight array. Fixed. Moved nowhere. 8344. November 4, 1994 $APGNOT/MFPRT.FOR Leonia The task printed an additional wrong collumn and as result there was there was a confusion in output table. The reason is that there is an additional column (after column #2 in the table MF) which is not needed to print. It is not clear whether this extraneous column had existed ever or it s a result of recent modification. Moved nowhere. 8345. November 7, 1994 OTFUV Eric Changed help and program to allow for concatenation, say of multiple IFs. Moved to patch. 8346. November 8, 1994 TVDEVS.KLUDGE,START_AIPS,START_TVSERVERS Pat Created a "hook" version of TVDEVS.SH called TVDEVS.KLUDGE that will call the appropriate (TST/NEW/OLD) version of TVDEVS.SH depending on VERSION. Modified START_AIPS so that when it calls START_TVSERVERS via a remote shell, it passes the version (TST/NEW/OLD) to it, and finally modified START_TVSERVERS to use this version. This is to ensure that a TST version of AIPS does not start the TV servers via a shell script that defaults to NEW. Moved nowhere, but info mailed to the AOC this date. 8347. November 8, 1994 SETPAR Pat Modified some of the limits on the parameters, in particular the limit of 15 TV's. Moved nowhere. 8348. November 8, 1994 $APGNOT/UVCOP.FOR Athol UVCOP was not always copying the tables it has to correct for frequency selection due to a bug in calling MULSDB in subroutine ESTSIZ. This bug exists in NEW but not OLD so has been around for a while. Moved nowhere. 8349. November 8, 1994 ZMOUNT Eric Changed an error message in TPMON and an index error in ZMOUNR which was causing the wrong tape number to be passed on dismounts. Moved nowhere. 8350. November 8, 1994 SAD Eric Added keyword REALRMS to MF table. Moved nowhere. 8351. November 8, 1994 OTFUV Eric Corrected setting of times to base them on the date in the output file header (when concatanating) and to allow an optional addition of integer days to the time. Moved to patch. 8352. November 8, 1994 MFPRT Eric Fixed it to handle missing MF files with some grace, to get the fit channel from the table column 1. Changed adverb names in help file. Moved nowhere. 8353. November 9, 1994 SNSMO in APGNOT Chris SNSMO now smooths phases rather than real and imaginary components. This should preserve the phase differences between IFs better than before. Moved nowhere. 8354. November 13, 1994 SDGET, UVFLG Eric When I tried flagging single-dish data with a FG table there were problems. Someone added parms to DSEL.INC and employed them in flagging without adding them to SDGET (just UVGET). Copied a bunch from UVGET. And UVFLG decided to add 300 lines to the FG table for each flag command to allow for 300 sources all number 0. Changed this too. Moved to patch area. 8355. November 14, 1994 SYSETUP Pat Fixed the bug in SYSETUP that caused it to add a gazillion IC and ID files unnecessarily. It now works as advertised. Moved to the patch area today. 8356. November 14, 1994 VLBA bandpass subsystem Phil Numerous changes: 1) DCVL.INC: Removed unused variables, declared AVDELI correctly. 2) CVLDOP.FOR: tidied up variable useage. 3) FRQUPD.FOR: added CATBLK to call sequence, determined #IFs internally from CATBLK. 4) GETFRQ.FOR: added FRQSEL to call sequence, determined #IFs & reference frequency internally from CATBLK. 5) DSHIFT.FOR: added FRQSEL to call sequence, changed call seqs to GETFRQ & FRQUPD, call TPSHFT now instead of ACSHFT. 6) New routines: $APLNOT/DETRAT - determines natural fringe rate wrt Earth centre. $APLNOT/TPSHFT - shifts total power spectra with optional hanning smoothing. $APLNOT/BPSHFT - shifts bandpass spectra a given amount, to be used in DOBAND routines. 7) BPASS.FOR: removed clumsiness with catalogue headers in shifting routines, ensured source number available for first record, changed call seqs for routines above, removed the initially very clumsy code to determine frequency shifts - replaced with simple routine DETRAT, use TPSHFT with smoothing instead of ACSHFT, write max shift for given antenna into the BP table in CHN_SHIFT columns. 8) CVEL.FOR: ensured shift info from BP tables passed through cleanly, ensured source number available before first call to DETRAT, recognize VLBA data, add all the bandpass shifting routines as implemented in DATBND. Moved nowhere. 8357. November 15, 1994 IMEAN Eric Added code to fit and display the noise bump in the histogram, attempting to get the noise (true rms and noise mean). Added comment to the help file on this. Moved nowhere. 8358. November 15, 1994 SDGRD Eric Fixed bug that led to 0 coordinate increment on the frequency axis. Basically I set BIF to 1 if there is no IF axis. Moved to patch area. 8359. November 15, 1994 OTFUV Eric It would be nice if single dish data formats were documented correctly (or even at all). The "RA offsets" recorded with on-the-fly data turn out to be real angles on the sky, not RA offsets. Actually, even that is not quite true. To get a real RA one adds the ra offset divided by the cos of the central DEC to the central RA. Sigh.... All data handled through OTFUV (and UVF) are hereby declared null and void. Moved to the patch area. 8360. November 16, 1994 $APLNOT/DATCAL.FOR Athol/Bill BLPNT was not correctly incremented in loop 300 for the case of single-polarization, multi-IF data. Baseline-based correction factors in this case were always taken from IF.1 and were incorrect for higher numbered IFs. Moved nowhere. 8361. November 16, 1994 MANDL, etc Eric Moved the CATINI call in MANDL, PATGN, and CANDY to after the GTPARM call. The user number was 1 otherwise. Moved nowhere. 8362. November 17, 1994 $APGOOP/DTSUM.FOR Phil DTSUM was not working on the Solaris machines here at the AOC. The problem was caused by the call sequence to UVDGET At one point in DTSUM one of the output parameters contained a DATA'd variable and so UVDGET could not overwrite it. Obvioulsy SunOS and other compilers allow this. Moved nowhere. 8363. November 17, 1994 $APLSOL/ZMOUN2.C ERic Added some new device type codes to recognize for better monted messages. Moved nowhere. 8364. November 18, 1994 SETSP Pat Made two significant changes and a few minor ones. The minor ones involved changing some limits that needed it. The first major change was to have it read HOSTS.LIST intelligently to retrieve the list of hosts to process, and also to use that to figure out the names of the SP files. Done with backwards compatibility; if this is compiled with 15JUL94 or older, it should work with the "C" format data. The second major change was to allow looping through the hosts, changing a given item to a different value for each host. This is very useful when you want to make an "almost" global change where many of the hosts will have the same value, but not all. There are still some bugs (not *all* mine I think either...) For example, if you run it on a host with a built-in 2nd AP size of 1216, you won't be able to exceed this for a host with a larger value (running on a sun trying to set it for a big IBM/AIX system). Also, it does not elegantly handle TIMDEST parameters, but that bug manifested itself in a portion of the code that I don't think I changed, and it's not too relevant since those parameters are overridden by the info in NETSP. We should now be able to can the SPLIST file. Moved nowhere, thinking about a patch... 8365. November 18, 1994 BLING and BLAPP Chris BLING and BLAPP have been substantially rewritten to improve their reliability and additional information has been made available in the BS table. The operation of the new programs is described in AIPS memo 89. PUVD.INC, BSINI.FOR, TABBS.FOR and TABLEUTIL.FOR have been updated to support the changes. Moved nowhere. 8366. November 21, 1994 $YPGNOT/VPLOT Leonia VPLOT is a new faster version of VBPLT. The code of the task has been tidied up to simplify the original one written by many people and as a result being too complicated. A new option (plotting both amplitude and phase) has been added. The sign size depends now upon the number of plots on the page. This size is inversly proportional now to square root of the number of plots on the page. The labeling of Y axis had been missed during modifications of VBPLT. It has been restored. Moved nowhere. 8367. November 21, 1994 $QPGNOT/BPASS.FOR Phil The shifting routine installed for VLBA data had a bug in that if any spectrum for an IF was totally flagged the average weight was divided by zero. Fixed. Also tidied up a little to avoid repeating operations that only needed doing for the first visibility. Moved nowhere. 8368. November 21, 1994 $APGNOT/FXPOL.FOR Phil FXPOL did not attempt to deal with FG tables - 'twas an oversight. It now does the correct thing to them. Moved nowhere. 8369. November 21, 1994 $APGNOT/UVCOP.FOR Phil Speeded up UVCOP considerably for some cases. In the ESTSIZ routine if no source was selected then an output file of the same size as the input file was created for copying, ignoring any other selection criteria. Now if timerange or FREQID is selected and no particular source is selected then ESTSIZ used the NX table to determine the size of the output file. Also UVCOP now takes advantage of the fact that data are sorted TB. If that is so and timerange is selected then if the current sample time is > than the stop time it jumps to the end of the copying routine. Moved nowhere. 8370. November 21, 1994 BLING Chris Drastically reduced the size of the executable by sharing some of the scratch arrays in COMMON. Also replaced some inappropriate calls to RFILL with more appropriate calls to DFILL. Moved nowhere. 8371. November 22, 1994 Pulse cal tables Phil Extended the definition of the pulse cal table to include the state counts. That is the time that the sampler spends in one of the 4 states when 2-bit mode is in use. Updated DPCV.INC, PCINI and PCTAB. Will be using these routines soon. Moved nowhere. 8372. November 22, 1994 BLING Chris Added missing TYPE arguments to QPUT. Rather surprisingly, missing these off caused no problems most of the time but gave rise to sporadic segmentation faults. Moved nowhere. 8373. November 22, 1994 MOVE Eric New task to move or copy a catalog slot's worth of file to yourself or some other user. Move is done on the same disk, deleting your copy. Fixed error message in CATDIR. Moved nowhere. 8374. November 23, 1994 misc Eric Changed MOVE to default OUTSEQ = 0 to highest+1 and use -1 to mean INSEQ. Added messages on each file moved or copied. Fixed it to loop over all extensions. Fixed OTFUV to compare axis types correctly for concatenation. It had left out KHCTP and hence did the first words of the header including date-obs and the like. Moved OTFUV to the patch area. 8375. November 27, 1994 MOVE and ZAP Eric Corrected MOVE to set the new user number into the image header. Images were still partly owned by the old user number! Changed AU3 and ZAP.HLP to use the user number in file deletion. Otherwise, files stored under odd user numbers cannot be deleted. This caused my 1000000+ APCLN to be zapped! Changed MDESTR as well - it also prohibited this. Now it prohibits this only on public catalogs. Moved nowhere. 8376. November 28, 1994 FRING Chris Corrected a logic error in the handling of scans from the index table. FRING was setting the scan duration to the larger of SOLINT and the duration from the NX table where it should have retained the latter. This lead to errors in calculating the scan boundaries and would cause scans where the source number changed to be split unexpectedly. Moved nowhere. 8377. November 30, 1994 PC routines Phil Changed the cable-cal variable from s.p -> d.p. We store it as seconds but its typically a picosecond value so want to keep full precision. Moved nowhere. 8378. November 30, 1994 SUBIM Eric Changed it to warn about but not delete constant images. It still deletes ones that are fully blanked. Moved nowhere. 8379. November 30, 1994 INDXIN, DATGET Eric When an index table entry has no samples, the loop range over vis becomes N to N-1. This causes UVGET to quit when doing an INIT if that is the first scan otherwise selected and is a silly waste of time in all cases. Changed INDXIN and DATGET to skip empty index entries. Moved nowhere. 8380. November 30, 1994 GETDEL Tony Beasley/Phil The original version of GETDEL took IM values and interpolated them onto the finer time grid of the CL table by looking for the nearest IM entry to the CL entry and perofroming either a forward or backward shift of the IM polynomial. John Benson tells us that only forward shifts should be done so the routine was modified to do this. Also avoids the majority of the problems that were causing the CL table to have some zeroed entries for the total model. Some cases still occur but this is due to the correlator not providing an IM entry before the first data point. Benson is looking into this. Moved nowhere. 8381. November 30, 1994 IBLED Eric Added a bit of time (min of 0.5 sec and 0.05 * DELTIM) =- to the time range for flagging. Round offs were getting us otherwise. Moved nowhere. 8382. December 1, 1994 BPASS Phil Ensured that the solution interval written into the BP table was never zero. BP values will not be applied if this is the case. Should never happen but we had a user trying to generate a BP table with data from one time stamp only and the time range for that is, of course, zero. Moved nowhere. 8383. December 1, 1994 $APLNOT/GETFQ.FOR Phil GETFQ relied upon UVGET having been run before it to set up some catalogue indicies. Fixed so that it calls AXEFND internally to avoid this problem. Moved nowhere. 8384. December 2, 1994 $APLNOT/PCTAB Phil Fixed an indexing error in the state counts loop. Moved nowhere. 8385. December 5, 1994 $YPGM/PCNTR Pat Changed INPRMS in the GTPARM call from 87 to 85 as per bug report from David Titterington. Moved nowhere. 8386. December 5, 1994 New task: $APGNOT/PCLOD Phil PCLOD is a new task to read the ascii format pulse-cal tables generated from the VLBA monitor system. It reads in the pulse-cal, cable cal and state count data into a PC table. Also a helpfile. Moved nowhere. 8387. December 5, 1994 $APGNOT/SNCOR.FOR Dave The XFER option was not working correctly, and likely never has. Changed some array arguements in subroutine XFER. It also looks like antennas to be changed have to be specified individually in ANTENNAS (ie ANTENNAS=0 does not do all of them), but I'll look into that later. 8388. December 5, 1994 SAD Eric The histogram computation had a problem when the untrue rms was too large compared to the true one (i.e., the peak was too narrow). Widened the histogram and covered up the failure. The routine to gues more than one peak in an island did not check for blanked pixels. This led to an island with 4 INDE components. Moved nowhere. 8388. December 5, 1994 POSSM Phil Added a couple of informative lines to the plots, re bandpass corrections and application of CL tables. Removed the annoying message that appeared when plotting a subset of IFs from a dataset. Moved nowhere. 8389. December 5, 1994 IMEAN Eric Changed IMEAN's handling of the fit for a noise rms. There are too many divergent cases to do the blind histogram. Added 2 adverbs to let the user estimate mean and rms in advance (IMSTAT, say). If PIXSTD is 0 it does two passes reading the data using the rms found in the first pass to set the parms of the histogram for the second. If PIXSTD < 0, it skips the noise estimation. Moved nowhere. 8390. December 5, 1994 GALPOL eric This subroutine put out a message about epoch at every call - say 10**6 times in a big hgeom! Killed the stupid message, relinked HGEOM. Called by COORDT. Moved nowhere. 8391. December 6, 1994 SPLIT Phil Quick change to ensure it does not copy the PC table through to the output single source file. Moved nowhere. 8392. December 6, 1994 WFCLN Tony Beasely/Eric Changed WFCLN and its help to add an INFILE option to define clean boxes for all 16 windows as needed. Needs standardizing. Moved nowhere. 8393. December 6, 1994 MOVE Eric This tar baby won't go away. Fixed 2 places that needed user number set to output user. Moved from the AOC. 8394. December 6, 1994 PCLOD Phil Corrected the reading of the phases. Moved nowhere. 8395. December 6, 1994 $YPGNOT/SNPLT.FOR Phil Added the ability to plot the cable-cal and amplitude & phase of the pulse-cal tones from the PC table. Also tidied the task up a little and added to code to recognize and forbid plotting variables that don't exist in the different table types that SNPLT can handle. Should really tidy this up more to handle the tables differently - one day. Updated the helpfile. Moved nowhere. 8396. December 7, 1994 SN2CL, CLCAL John Conway / Phil Added Conway's extensions to the interpolation modes. He has added 3 new modes and we have ensured that the previous modes are untouched and perform as before. New modes are SIMP: performs a classic linear interpolation; AMBG: uses the mean residual rates to resolve phase ambiguities and then does a SIMP; CUBE: fits a 3rd order polynomial to fit the rates and phases and then does a SIMP. Added a warning message about doing smoothing for VLBI data - suggest users do that in SNSMO, but carry on anyway. Updated the helpfile. Moved nowhere. 8397. December 9, 1994 $APLSUB/FPARSE.FOR Phil When NAXIS = 0 and GROUPS = T in a FITS header FPARSE printed an error message about NAXIS and GROUP structure not matching. In fact this is a legal construct and is used by the VLBA FITS format when constructing the base header - no data follows the header, only tables. Made a minor change to FPARSE to recognize this construct and set up the keywords accordingly. Moved nowhere. 8398. December 9, 1994 $APLPGM/PRTTP.FOR Phil VLBA data is recognized by the values of 4 keywords (for redundancy), one of these keywords changed as a result of the previous change to FPARSE. Fixed PRTTP to recognize this. Moved nowhere. 8399. December 9, 1994 Holography Eric Added a new phase routine for subreflector geometries done by Mike Kesteven and changed file naming conventions in HOLGR. In UVPRT, altered the automatic flagging algorithm to make it simpler, added the option to have multiple reference antennae for a single output antenna, and changed the output file name convention, adding the IF number. Also both help files. Moved from Perley's area. 8400. December 9, 1994 Misc Eric IMVIM read the wrong number of adverbs and so ignored GRCHAN. UVGLU made a last call to the "DIDDLE" routine which had been changed not to recognize a last call! Deleted it. Moved nowhere. 8401. December 9, 1994 CookBook Eric Rewrote chapters 7 and 8 on display, making it one chapter called COOK7. Renamed and adjusted chapter 9 into COOK8, moving some Slice info from the old chapter 8. Revised the spectral line chapter 10 to add info and renamed it COOK9. Corrected and converted AIPS Memo 76 to be Appendix A and acquired a PostScript copy of the spectral-line outline from Andrea Cox to be Appendix B (COOKB). Created 7 more Figure plots in support of these. More will need to be done with TOC and cross references and renumbering chapters 11-15. Moved nowhere. 8402. December 9, 1994 VLBA bandpass stuff Phil Added the last routines that provide us with the structure to correct VLBA bandpasses. We still have some algorithmic problems, in that the sharp edges we see in the bandpasses cause Gibb's type ringing when we do the shifting. We are investigating alternate smoothing algorithms to remove the ringing yet not change the shape of the bandpass in subtle ways. Changes are: BPASET: Check if VLBA data, if so set up needed arrays and variables. BPGET: When setting up the bandpass buffers, if VLBA data make a copy of them for shifting purposes - save having to go back to the BP table each time we shift. DATBND: Determine how much to shift bandpass to match program source, shift the data in the secondary buffer. If the shift is > 0.1 channels from the previous shift then load the original buffer and shift it. Apply flagging to the channels that are wrapped around as a result of the shift. DSHIFT: Call ACSHFT not TPSHFT. Deal with the AVDELI array slightly differently, it now carries information on the shift applied in BPASS so I can keep track of which channels to flag due to wrap around. SPINTP: Gave this routine much more intelligence on dealing with the ends of the spectrum, will now average for 5 channels of good data and use that to extrapolate, rather than just taking the first or last good point. ACSHFT: Added the option of Hanning smoothing while shifting. CVEL: Minor change to deal with the AVDELI arrays. Moved nowhere. 8403. December 12, 1994 $UPDUNIX/AIPSUPD Dave Changed a couple of /AIPS to $AIPS_ROOT. It was keeping me from running the midnight job on the special large-AP version of AIPS on Pilabo. Moved nowhere. 8404. December 13, 1994 holography Eric Fixed the parameter count to HOLGR which should have been increased for the 2 new adverbs. Rearranged the file open and close flags in UVPRT. It was getting confused into thinking the file was closed and was then not able to open the next one. Moved nowhere. 8405. December 13, 1994 CookBook Eric Changed chapters 1, 3, 4, 5, and 6 to reflect the new chapter numbering and the nomenclature that a chapter is not a \Sec. Changed Chapter 4 to allow the large plot to be optional. Changed the table of contents (COOK0) for the revised chapters (7 and 9) and chapter numbering and the new appendices. Also made the front picture optional. Reinserted COOKB.PS. Apparently the chkout system had put it in correctly to the rcs area, but made a 0-byte file in the plain text area! Moved to the anonymous ftp area and updated cook.html for WWW. 8406. December 14, 1994 UVCOP Eric UVCOP, when adjusting for channel, IF, ... selection, took itself too seriously. Changed it to be forgiving if there are no AN files and to copy FQ tables even if the file is not multi-source. Moved nowhere. 8407. December 14, 1994 aips man page Eric The man page, repeated in AIPS.HLP and COOK2.TEX, should say tvhost once where it says tvdisp. Did not fix the man page itself - I do not know where that is. Moved nowhere. 8408. December 15, 1994 SDGRD Eric The task was creating two SC files at all times. It only needs the second if it is forced to sort the data. Moved that creation into the sort routine, which is seldom called. Moved to the patch area. 8409. December 15, 1994 OTFUV Eric Added adverb CHANSEL to the task to let the user specify channels to be flagged as they are read in. CHANSEL is more general than the current need (which is a short list of single channels) but it is the channel swelection adverb in AIPS and it allows for future generality. Moved to the patch area. 8410. December 19, 1994 FRING Leonia Changed the warning text in the case of too long SOLINT. If SCAN > SOLINT > SCAN/2 the task forced SOLINT = SCAN and if the SCAN interval exceeded the allowed threshold the warning text confused the user because it informed him about too long solution interval and at the same time recomended him to take this interval even larger (SCAN > SOLINT). Decreased the additionnal intervals at the beginning and end of a scan from 8.6 sec (1E-4 day) to 2 sec. Added some explanation of SOLINT and its limitation to HELP. Moved nowhere. 8411. December 20, 1994 Delay and rate smearing corrections Athol Add a facility to the calibration system to correct the ampl. of VLBA data that have been pre-averaged in frequency in the presence of residual delays, usually as spectral averaging in the VLBA correlator. Trigger this correction if the array name in the AN table is 'VLBA' and the keyword 'SPEC_AVG' as stored in the AN table is greater than 1. This correction is applied to all polarization correlation pairs independently and assumes alignment of polzn. delay solns in the same frame (ie. no uncorrected offsets between the R and L solutions). Also improve the application of rate smearing corrections by extending them to all polzn. correlation pairs rather than only RR and LL. Allow two new options for DXTIME on input to UVGET, namely DXTIME=0 and DXTIME < 0. The former will use the integration time for each u-v record as specified by random parameter INT_TIM (failing that; 1 sec) and the latter will suppress any rate smearing correction. Changed the following routines: $INC/DSEL.INC Add new internal variables to DSEL.INC DODSM, DORSM L Do delay, rate corrections ? NXDSM I No. of pre-avgd. freq. chan. IANLUN I LU for AN table ANBUFF I(512) Buffer for AN table $APLNOT/SELINI.FOR Initialise DXTIME=0; NXDSM=0; DODSM=F; DORSM=T $APLNOT/UVGET.FOR Set IANLUN=45 $APLNOT/GAININ.FOR Check AN table for array name and SPEC_AVG keywords. Set NXDSM; set DODSM if NXDSM > 1 and array name = 'VLBA' $APLNOT/DATGET.FOR Before each visibility record is passed to DATCAL and lower calibration routines, determine the integration time to be used for the rate smearing correction as described above. Pass this value to DATCAL; set DORSM false if DXTIME < 0. $APLNOT/CGASET.FOR Remove all previous scaling of the BLFAC array to accomodate rate smearing corrections. $APLNOT/CSLGET.FOR Use new integration time as calculated in DATGET in deciding time tolerance in CL table entries. $APLNOT/DATCAL.FOR Calculate delay and rate smearing corrections on the fly in DATCAL if DODSM and/or DORSM are set and delay/rate solutions are being applied. Do a discrete Sinc correction for delay smearing where only a small number of channels may have been pre-averaged. Correct all polzn. pairs independently. $APLNOT/DSMEAR.FOR New subroutine to extract the SPEC_AVG and ARRNAM keywords from the AN table. Moved nowhere. 8412. December 20, 1994 GAL Gustaaf Implemented a number of bug fixes suggested by Martin Shepherd. Introduced improved array bounds checking, and better handling of the PL extension file. Moved nowhere 8413. December 20, 1994 VPLOT Leonia The task did not work with the symbol '|'. It's been fixed Moved nowhere. *********************************************************************** Changes while 15JAN95 was NEW: *********************************************************************** 8414. December 23, 1994 TVFLG, SPFLG Eric Changed TVFLG, SPFLG, and GRIDTB to try hard to avoid bad beating between the data times and the Master grid cell times. The programs all integerize the times on a 0.1 of a Master grid cell (DPARM(6)) and look through the first 10000+ times to see what regularity there is in those sample times. Then the Master grid cells are set to be centered on that preferred 0.1 of a cell. Also made fixes that will allow data sampled < 1 sec apart (i.e. OTF) including raising averaging buffers to hold all the data allowed (16384 rows in the master grid). Changed DSMEAR to suppress error messages and clean up its typing. Fixed SPFLG and TVFLG to average across blank rows so long as the time range and source ID are ok. A blank row stopped it before. Also fixed a quirk with the menu and help display. Changed the smooth and scan times displayed to be the integer number of basic intervals rather than try to show seconds. Moved from 15JUL95. 8415. December 23, 1994 OTFUV Eric Changed it to make the beam be 257 (i.e. baseline 1-1). The old scheme confused SPFLG etc. Added code to keep track of data ordering and to put SORT = 'T' in the header iff its true. Added message at the end about it too. Moved to the patch area and from 15JUL95. 8416. December