----------------------------------------------------------------------- ; 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 15APR93 (aka 15JUL93) *********************************************************************** Changes while 15APR93 was TST: (Note: change 7568, April 28 renames 15APR93 to 15JUL93) *********************************************************************** *********************************************************************** 7097. August 25, 1992 ZABOR2 Chris/Eric Noticed that floating-point errors cause no interupts when debug is invoked even if the particular task is not running debug. The new ZACTV9 means that the parent process will have departed unless it is the debugger and, hence, on non-debug, the parent process id is 1 for root after the real parent is gone. Changed $APLSUN/ZABOR2.C. Also versions in $APLUNIX, $APLBERK, $APLDEC, $APLIBM, $APLBELL, and $APLCRI. Moved to 15OCT92 (NEW) this day. 7098. August 25, 1992 VLBIN Phil The switch cycle change to VLBIN had a minor problem, a divide by zero was introduced.Moved offending line to correct place. Moved to 15OCT92 today. 7099. August 25, 1992 $SYSIBM/FCOPTS.SH Pat Added some optimization features that should improve the performance of AIPS on IBM RS/6000 models. These came from the parameters used by IBM in their RFP (AUI-504) response when they benchmarked AIPS on several models. The -Pv flag makes the compiler use a different preprocessor which according to IBM should result in a 5% speedup. The -ND and -NA flags increase compiler space for variables which apparently has a significant effect on speed of execution. Moved to 15OCT92 today. 7100. August 25, 1992 $APLNOT/TYINI.FOR Pat Inserted a change for Gustaaf Van Moorsel and Gareth Hunt. A change in band during a run was causing NKEY to become too large and (on IBM's at least) was causing DTYP to get overwritten. FILLM then bombed. Moved to 15OCT92 today. 7101. August 25, 1992 tape writing Eric Added new opcode to $APLGEN version of ZTAPE. BEGW causes ZTAPE to BAKF 1, write 2 EOFs, and then BAKF 1 (or 2 if it's at the beginning of tape). Added call to this operation to FITTP. This allows one to write to Exabyte 8200 tape devices even at places other than beginning-of-tape and end-of-information, and at least does not hurt other devices. It does cost some time (56 second in one Exabyte test), and it means that a bad call to FITTP immediately writes on the tape, even if the other adverbs cause the task not to write any data. It does prevent the error of writing in the middle of a file which FITTP allowed (by oversight I assume) previously. Removed the special $APLSUN versions of ZTAPE since the EOM option was removed from BAKTP last May. Moved to 15OCT92 as well from the VLA. 7102. August 25, 1992 DBCON Eric DBCON has an array called JORD that allows the correlators to be rearranged in order. It was dimensioned only 2000! A simple 2 Stokes, 512-channel data set required over 3000. Changed it to 20000 and added an error test on the new limit. Raised all buffer sizes from 5000 to 8192 also. Moved to 15OCT92 as well from the VLA. 7103. August 25, 1992 SPFLG Eric Added new operation - show next baseline. Changed the fortran and also SPFLG.HLP and HLPSPFLG.HLP Moved to 15OCT92 as well from the VLA. 7104. August 25, 1992 PRTTP Eric Added progress messages and record/byte count message. Correct buffer size on each file - allows special TAPIO that limits buffer size on read (for dats). Moved to 15OCT92 as well, from VLA. 7105. August 25, 1992 plotting Eric Corrected the algorithm which interpolates vectors to fit in the plot area. Changed LWPLA, GVEC, TVPL, QMSPL, TXPL, and, CANPL. This caused a serious error in LWPLA where position commands go to the printer. TVPL etc. survived better if the plot program was not too criminal. Also added YHOLD calls to TVPL. Moved to 15OCT92 as well, from VLA. 7106. August 25, 1992 UVLSF Eric Corrected a bug which caused the output continuum values to be determined at ECHAN+1 rather than at the specified channel. Moved to 15OCT92 as well, from VLA. 7107. August 26, 1992 T2DHMS Eric Corrected version of TODHMS - returns the time in sexagesimal form with a sign in a character variable. TODHMS should be gradually replaced. Moved to 15OCT92 as well, from VLA. 7108. August 26, 1992 POSSM, EXTLIST Eric Changed it to plot a line at 0 in phase, real, and imaginary. Fixed it to position inside the plot for blanked channels. Changed it to try to fill the screen when DOTV is true. Changed its handling of times to make it much simpler and to avoid various format overflows. Corrected the exterior character counts. Changed it to use the interactive option in GFINIS when doing multiple plots (SOLINT > 0) on the TV. Also corrected and enhanced AU8A (verb EXTLIST) to describe POSSM plots better, and get correct units in UVRANGE in UVPLT. Fixed POSSM to be more careful to fill in the parameters in the plot header for EXTLIST. Moved to 15OCT92 as well, from VLA. 7109. August 26, 1992 TVMOVI, TVBLNK Eric Changed timing formulae and changed them to skip the ZDELAY call entirely when trying to go their fastest. Moved to 15OCT92 as well, from VLA. 7110. August 31, 1992 $APLOOP/OBJECTMNGR.FOR Bill C. Added UV data class with virtual keyword support for catalog header values. Also added virtual derived keyword 'LREC' for the record size. Added routines OBH2C and OBC2H to deal with character arguments being declared as either a single character string or an array of character strings. This removes the necessity of equivalencing an string array to a single long string when reading or writing strings to an object. Moved nowhere. 7111. August 31, 1992 $APLOOP/GENERIC.FOR Bill C. Added UV data classes to the generic functions. Moved nowhere. 7112. August 31, 1992 $APLOOP/UVDATA,UVUTIL.FOR Bill C. New, rudimentary UV data class. Also added utility library UVUTIL.FOR although there's not much in it. Moved nowhere. 7113. August 31, 1992 $APLOOP/UVDESC.FOR Bill C. New uv data descriptor class. Moved nowhere. 7114. August 31, 1992 $APLOOP/CALEDIT.FOR Bill C. New uv data selection/calibration/editing class. Moved nowhere. 7115. August 31, 1992 $APLOOP/INPUT.FOR Bill C. Added support for logical values passed from POPS. Moved nowhere. 7116. August 31, 1992 $APGOOP/MBDLY Bill C. New task. Fits multi band delays from individual IF fringe fit results. This task is useful for determining multiband delays when the IFs are widely spaced in frequency. Moved nowhere. 7117. August 31, 1992 tape routines Eric Added more explanatory error messages to $APLGEN versions of ZTAPE (for remote tapes), ZTPMIO, and ZTPWAT, to the $APLIBM version of $ZTPMIO, and to the $APLCVEX version of ZTPOPN. Moved to 15OCT92 as well, from the VLA. 7118. August 31, 1992 ASCAL Eric Added deprecatory comments to the inputs header and help file. Moved to 15OCT92 as well, from the VLA. 7119. September 1, 1992 $APGOOP/CL2HF Bill C. New task. Converts AIPS CL table and associated tables and data to a Haystack type HF table. This task is intended to be part of the data path from AIPS to the Haystack/Goddard geodetic/astrometric package. Moved nowhere. 7120. September 1, 1992 VBPLT Phil Added a new option to VBPLT - it will now average spectral channels and/or IFs on the fly before plotting. The labels on the plots let the user know what channels/IFS have been averaged. VBPLT.HLP has a couple of new adverbs. Also added a subroutine (AVGCIF) to APLSUB to do the averaging. Moved nowhere. 7121. September 1, 1992 PRTGA Eric When DOCRT TRUE, it would take a Q to mean loop for the next set of data rather than quit all the way. Changed its error handling to stop this. Moved to 15OCT92 as well, from the VLA. 7122. September 1, 1992 UVINIT Eric Changed it to read only the bytes needed (as it does in UVDISK) rather than a disk sector more than needed. This was taking us off the end of file when the first or second read reached the last visibility. Moved to 15OCT92 as well, from the VLA. 7123. September 1, 1992 CALIB, FRING Eric The call to ZCMPRS was making the file too small. The correct formula for file size is NVIS * LREC + 2 * NBPS. Adding 128 vis's rather than the 2 disk sectors is not right although all other tasks do it. We should add 1 vis for safety plus the 2 sectors. Changed CALIB and FRING which were too small. The rest can wait until I really need something mindless to do. Moved to 15OCT92 as well, from the VLA. 7124. September 1, 1992 labeled tapes Eric If one is reading more than one record then one cannot simply advance file on every tape label record. In fact, it would be wise to read them all and do no file skipping. Changed TPHEAD and TPIOHD to do this. Moved to 15OCT92 as well, from the VLA. 7125. September 2, 1992 $APLSUB/SHIFTP.FOR Glen New subroutine which calculates the U,V and W terms for the phase shift of an image center (or clean component). Modified MX, UVMAP, GRDAT and MAKMAP to use SHIFTP. Moved nowhere yet. 7126. September 2, 1992 SMOSP Phil SMOSP was never dealing with flagged channels - could cause severe problems when interference was present. Now does it correctly. Moved nowhere. 7127. September 3, 1992 FILLM Gustaaf Added several fixes: 1 - FILLM failed to recognize a spectral line data base of ONE channel as such. Instead, it attempted to add the data to an existing (and already open) channel 0 map. In the fix the data bases are labeled as 'channel 0' or 'spectral line'. 2 - Antenna files were labeled with the frequency of the PREVIOUS data base; the first data bases had frequency 0. Fixed updating of the catalog in routine ANTFLM. 3 - On Eric's suggestion made change in FLMCLS. To LSIZE, input parameter to ZCMPRS, an extra term 2*NBPS was added. Its absence has not led to FILLM problems thus far, but other tasks failed when such a term was missing, suggesting potential problems for FILLM, too. Moved to 15OCT92 as well 7128. September 3, 1992 UVGLU Gustaaf When the number of channels in the first data set (nch1) is smaller than that in the second data set (nch2), only nch1 channels in the second set are written correctly to the output; the remaining nch2-nch1 are set to zero. This is for compressed data only. The error was caused by an incorrect parameter of the expansion routine ZUVXPN for the second input data set. This was fixed. Moved to 15OCT92 as well. 7129. September 3, 1992 PTAB.INC Chris New parameter include file. This defines mnemonics for the AIPS table type codes so that we don't have to keep looking them up in "Going AIPS". The mnemonics are as follows. TABDBL double precision floating-point TABFLT single-precision floating-point TABHOL Hollerith (4 characters per word) TABINT integer TABLOG logical TABBIT bit string TABSEL select flag Also defined the following mnemonics for the IRCODE argument to TABIO. TIOROW table row TIOPTR DATPTR/DATYPE record TIOSEL data selection string TIOTTL column title TIOUNT column units field TIOKEY keyword/value pair Moved nowhere. 7130. September 4, 1992 $APLOOP/UVDESC,IMAGEDESC.FOR Bill C. Fixed misspelling of 'DATE-OBS', 'DATE-MAP' and 'CRPIX' in data statment. Moved nowhere. 7131. September 4, 1992 $APLOOP/CALEDIT,VELOCITY Bill C. Fixed bad argument list to SECGET and VELGET. Moved nowhere. 7132. September 4, 1992 SNPLT Phil Suppressed an error message that occurred when plotting variables against elevation/HA and the source column in the table was not filled in. Task finished OK but filled the screen with a particular error message on occasion. Moved nowhere. 7133. September 5, 1992 FITS tapes Eric In order to preserve header keywords and their values when data are written to FITS tapes and read back, invented form HISTORY AIPS HEADERi keyword = value / comment where i is the type code and changed FITTP to write all header keywords out in this form. GETCRD to recognize such cards and return IERR = -1 to indicate their presence. PUTCRD to parse such cards and write keywords into the new header file (new routine). IMLOD to call PUTCRD when needed in subroutine FITHDR. UVLOD to call PUTCRD when needed in subroutine UVFHIS and to handle HI file errors better. FITLD to call PUTCRD when needed in routines IMGHDR and UVFHIS, to handle HI errors better in UVFHIS, to loop to 1000000 header cards everywhere, to call PUTCRD in BTFHIS when needed, to copy the header keywords into the output uv file in the new tables uv format, and to reset the read/write status for uv files so that good files are not deleted on subsequent errors. Moved to 15OCT92 as well. 7134. September 7, 1992 MX, GRDAT, MAKMAP Glen MX modifications to add the History and AN tables to the UVWORK dataset. Modified GRDAT and MAKMAP to use the $APLSUB/SHIFTP subroutine to calculate the phase terms for the field shifts. Moved ot 15OCT92 as well. 7135. September 9, 1992 $APLOOP/OBJECTMNGR.FOR Bill C. Fixed problem with virtual derived keyword LREC in uv data class. Also changed the name to OBJMNGR.FOR to avoid problems with the long name. Also corrected the arguments to LEN to be the first element of the array. Moved nowhere. 7136. September 10, 1992 $APLOOP/TABLE.FOR Bill C. Added routine TBLCOP to copy the contents of a table. Moved nowhere. 7137. September 10, 1992 $APLOOP/TABLEUTIL.FOR Bill C. Added routine OCHNCO to copy the selected contents of an FQ/CH table. Moved nowhere. 7138. September 10, 1992 $APLOOP/UVDATA.FOR Bill C. Modified OUVSCR to honor selection by channel, IF and polarization. OUCREA now sets UV_DESC values for ANAME, FNAME, FDISK. OUVOPN now uses FNAME from UV_DESC when opening a file for write as it might be a scratdh file. Moved nowhere. 7139. September 10, 1992 $APLOOP/UVDESC.FOR Bill C. Added missing class members. Also fixed UVDCOP not to attempt to copy virtual keyword LREC. Added missing items: EPOCH, OBSERVER' Moved nowhere. 7140. September 10, 1992 $APLOOP/UVUTIL.FOR Bill C. Added the following UV class utility routines: OUVDFT - DFT a uv data set to an image; driver routine. UV2SCR - copy a uv data set to a scratch file U2IDES - Copy descriptive info from a uvdata object to an image. UVREFQ - Determine reference frequency of a uv data set with possible averaging in frequency. UV2TAB - Makes a table object associated with a uv data object. Moved nowhere. 7141. September 15, 1992 $QOOP/QUVUTIL.FOR Bill C. New library for OOP UVdata utility routines calling "Q" routines. OUVDFT - DFT a uv data set to an image; driver routine. ODFT - Low level DFT routine OUNFWT - Front to UVTBUN, uniform weighting routine. Moved nowhere. 7142. September 15, 1992 $QOOP/APUTIL.FOR Bill C. New library for OOP utility routines calling "Q" routines. Mostly moving data between an object and the "AP'. AP2IMG : Copy an image plane from "AP" memory to an image object. IMG2AP : Copy an image plane from an image object to "AP" memory. AP2UV Copy a set of visibilities from the "AP" to uv object. UV2AP: Copy a set of visibilities from a uv object to the "AP". Moved nowhere. 7143. September 10, 1992 $APLOOP/IMAGE.FOR Bill C. Fixed IMGOPN to create output files if necessary if OPEN was for write. Moved nowhere. 7144. September 10, 1992 $APLOOP/IMAGEDESC.FOR Bill C. Added missing items: EPOCH, OBSERVER' Moved nowhere. 7145. September 10, 1992 $APLNOT/UVGET.FOR Bill C. Modified to allow using scratch files. Moved nowhere. 7146. September 10, 1992 $APLOOP IBM fixes. Bill C. Made numerous fixes which may have been causing trouble on IBMs. Most modules contained form feeds (^L) which have been removed. Also fixed fortran problems in TABLE.FOR and HISTORY.FOR. Moved nowhere (but similar changes should be made in OCT92). 7147. September 10, 1992 $SYSUNIX/FDEFAULT.SH Pat Modified the default compile option for IBM/AIX so that it now uses -qcharlen=10000. The 5000 previous limit was causing problems for one of the APLOOP routines. Moved nowhere yet, maybe to 15OCT92 soon. 7148. September 10, 1992 IBLED Phil Due to an indexing error the interactive flag option no longer worked - fixed. The select frame option did not do what users expected - it adjusted the selected frame window to put the centre portion in the editing frame if it was too big - changed so that it would downsize the frame but keep the start pixel the same. Also increased the MAXVIS from 8192 to 16384. Moved to 15OCT92 today. 7149. September 10, 1992 UVMAP Glen Small modification to UVMAP to do exact field phase shift calculations. Uses subroutine SHIFTP. Moved to 15OCT92 today. 7150. September 10, 1992 TV common Eric Added parameter TVIMPC = TV images per color to DTVC.INC. This has values 0 for one image only (e.g. XAS on workstations), 1 for one image channel per color (e.g. Deanza), and > 1 for more than one channel displayed at a time in whatever color (e.g. IIS Model 70). Changed $YDEA, $YM70, $YM75, $YIVAS, $YSS, and $YGEN versions of YTVCIN to show the correct value. Added a line to ask for the parameter to SETTVP. TVHUI uses this to do a 3-channel display on TVs which support it. Moved to 15OCT92 as well. 7151. September 11, 1992 TVHUI Eric New task to display two images on the TV using one to control the intensity and the other the hue. This works rather well on all TV types, allowing interactive manipulation using a decremented subimage with the full image updated on button C. An output file with an RGB axis may be written. A special mode for TVs better than XAS (i.e. separate images in RGB) is supported. Also a help file. Moved to 15OCT92 as well. 7152. September 11, 1992 PRTHI Eric AU7 had a limit of 10000 history cards and forgot to issue the error message. Changed to display the warning and to continue when DOCRT is > 0. Moved to 15OCT92 as well. 7153. September 14, 1992 $SYSUNIX/STARTPMON Pat Clarify some error messages, issue errors if there's a new TPMON.EXE and an old TPMONx and the process won't die, etc. More important, use $TST/$ARCH/LOAD instead of $LOAD. This prevents the use of OLD or NEW versions of TPMON. Moved to 15OCT92 today. 7154. September 15, 1992 $APLIBM/ZMOUN2.C Pat Fixed logic that figures out the OS version (3.2 or higher vs. anything lower) and subsequently whether to do a REWIND or UNLOAD. Latter is only available to tctl in AIX 3.2 and higher. Moved to 15OCT92 today. 7155. September 15, 1992 DUVH.INC/UVPGET Phil Added the integration time as a random parameter to uv-data. Needed a small change in DUVH.INC and UVPGET. However a large number of subroutines in APLNOT, APLSUB, APLOOP and QNOT need to be recompiled. Some tasks (see below) need minor work. Moved nowhere. 7156. September 15, 1992 $APLPGM/PRTTP Glen Needed some info about dates on PRTTP PRTLEV = -2 and -1. Now attempts to print observation date, frequency and random axies or number of visabilities. Removed listing tables for PRTLEV = -2. (With DATs with 250 files, any extra lines added up to alot of extra paper.) Moved nowhere. 7157. September 15, 1992 Integration time Phil Modified VLBIN and MK3IN to record the data integration as a random parameter. Also changed DBCON local common to know about the new r.p. Other tasks will need to know about this soon, will let these changes trickle through first. Moved nowhere. 7158. September 15, 1992 $APLOOP/HISTORY.FOR Bill C. Modified how buffers are obtained. OBOPEN and OBCLOS are now used rather than OOPEN and OCLOSE. Moved nowhere. 7159. September 15, 1992 $APLOOP/UVUTIL.FOR Bill C. Fixed typo in U2IDES which caused shifts in declination to be ignored. Moved nowhere. 7160. September 15, 1992 $APGOOP/UBAVG.FOR Bill C. New task to do baseline dependent time averaging of "BT" sorted data. The time averaging is set by a field of view and a maximum averaging time set by the user. Moved nowhere. 7161. September 15, 1992 $APLOOP/UVDATA.FOR Bill C. Fixed bug in OUCFCL which occasionally left MAPCLR trying to clear the status from files on disk 0. Moved nowhere. 7162. September 15, 1992 BS tables Chris Added routines to support baseline fringe solution (BS) tables to APLNOT (BSINI.FOR and TABBS.FOR). Also added OO-AIPS wrappers for these to the routines in APLOOP:TABLEUTIL.FOR. Moved nowhere. 7163. September 15, 1992 EXPLAIN Eric Changed AU1A to use the correct LUN when closing printer files and to branch properly on error so that all things get closed. Moved to 15OCT92 as well. 7164. September 15, 1992 XBASL Eric Corrected one buffer size from 512 to 4096 -- it was the declaration that created the actual core array which was declared 4096 in subroutines to which it was passed! Added DOCONT adverb to the help file and code. If DOCONT is true, the average baseline is restored to the data (this was the only "choice" before). Changed the selection for fitting so that all spectra are fitted for baselines if they have enough unflagged points. Only those with enough high points are displayed for the interactive bits however. Changed the test on "high" points to be high in absolute value. This allows absorption signals. Moved to 15OCT92 this date. 7165. September 17, 1992 $APGNOT/XTRAN Glen Increased maximum image size to 3000 by 3000 pixels. Other minor improvements. Moved to 15OCT92 this date. 7166. September 17, 1992 $APLPGM/STARS Glen Minor help file improvements. Moved to 15OCT92 this date. 7167. September 17, 1992 $HLPFIL/SETJY.HLP Rick/Glen Editted and inserted Rick's explanation of the VLA values of the flux density scale into the HELP file. Moved to 15OCT92 this date. 7168. September 17, 1992 TVRGB Eric New task to load an RGB cube - or 3 independent images, one each for red, green, and blue - to the TV. It uses 3 channels on those TVs which can show 3 independent channels in separate colors. On the lest capable TVs - mainly XAS - it uses a color optimization scheme due to Jef Poskanzer to encode the most popular colors into the (few = about 200) colors available. Interactive enhancement is also offerred. Moved to 15OCT92 today too. 7169. September 18, 1992 MK3IN Bill C./Athol Fixed bugs in differential precession correction to u,v. The apparent rather than mean positions were being used and there was no cos(declination) factor in computing the rotation. Moved nowhere. 7170. September 18, 1992 FRING Phil Minor change to a format statement Moved nowhere. 7171. September 18, 1992 TV tasks Eric Changed TVHUI to release the TV before making the output image. Changed TVRGB to make the number of color levels actually used in color optimization a user-controlled option with 40 as an upper limit (64000 colors). Changed TVRGB to check the RGB axis parameters carefully and use the colors in correct order when they start in some other order. Correct bug limiting windows in TVRGB, made the instructions come out more often, moved the reset of the OFM to just when the picture is loaded, and improved the progress messages. Added the new option to the TVRGB help file. Move to 15OCT92. 7172. September 21, 1992 TVHUI Eric Changed the LUT version of TVHUI coloration to avoid much of the use of LUTs. It was doing a simple multiply, so why bother with the logs and exponentials required by the IIS? The color filters are still computed once into LUTs and the hue used as the index into them. Changed the output routine to do no scaling of the intensity image since the coloring routines no longer require that help and it was leading to some odd biases in the output. Moved to 15OCT92 this date. 7173. September 22, 1992 ISHSRT Eric New service subroutine - does shell sort with two keys on an integer array. SHSORT does it on floating arrays. Changed TVRGB to use the public one rather than a private copy of ISHSRT. Moved nowhere. 7174. September 22, 1992 $APLSUB/PRTLIN Glen New duplex printers were always producing a blank first page, because PRTLIN was inserting a formfeed as the first print character. Added logic to PRTLIN to exclude the very first formfeed. All print tasks will have to be re-linked. Moved nowhere. 7175. September 22, 1992 UVAVG Phil Found and fixed the bug that had been dogging us for a while. On the final call to UVAVDO the call sequence was laced with BUFF2 statements rather than BUFF1 - this caused problems inside UVAVDo when the freqid and source id etc were extracted from these arrays. Moved nowhere - its too late. 7176. September 23, 1992 PP Eric Changed PP.FOR versions from $SYSUNIX and $SYSVMS. Corrected error handling include names like 'MYAREA:DXXX.INC' and revised it to allow include files to include files which include files. That is - three levels are allowed. PP watches for - and skips - any file included twice in a subroutine or function. When it does this, it changes the include statement into a comment with SKIPPED instead of INCLUDE. Now, for example, a local include can include DSEL.INC and DSEL.INC can avoid errors by including its parameter definitions (PUVD.INC). Existing code that already includes these all "by hand" will not be affected. Copied a PP.EXE to $SYSLOCAL for the SUNs. This will have to be done for all architectures as the file is copied by the midnight job. Moved nowhere. 7177. September 23, 1992 UCMPRS Eric This routine should be used to compress all UV data files henceforth. It takes care of things in a "standard" way. Moved nowhere. 7178. September 24, 1992 KNTR Phil Added an option so the user can control what is displayed as a label in each pane. Default is just the channel number as before, but the user can now display the channel velocity if ALTSWITCH has been run. Moved nowhere. 7179. September 29, 1992 $QPGM/RSTOR.FOR Bill C. Fixed routine RESTOR to properly FT the CLEAN components. It was reading an uninitialized scratch file and adding it's contents to the FT of the CLEAN components. I don't understand how this could ever work before. Fixed bug in FILES which caused RSTOR to fail if the various images were on different disks. General cleanup of the handling of disk and catalog slot numbers. Moved nowhere. 7180. September 29, 1992 $INC/ZPBUFSZ.INC Bill C. New system dependent parameter include for defining buffer sizes. 7181. September 29, 1992 PUVD.INC Bill C. Added parameter, MAXCIF, to give the maximum product of the number of channels times the number of IF. This should allow using arrays smaller than MAXIF*MAXCHA. Moved nowhere. 7182. September 29, 1992 DSEL.INC Bill C. Modified to include and use parameter files PUVD.INC and ZPBUFSZ.INC for buffer declarations and arrays whose dimension is the number of channels * the number of IFs. Moved nowhere. 7183. September 29, 1992 Parameterized Buffer sizes Bill C. Modified the following routines/tasks to use parameterized buffer sizes from ZPBUFSZ: $APLNOT/BPASET.FOR, $APLNOT/BPGET.FOR, $APLNOT/SCINTP.FOR, $APLNOT/SCLOAD.FOR, $APLNOT/SDGET.FOR, $APLNOT/UVGET.FOR, $APLNOT/DATBND.FOR, $APLNOT/TABBP.FOR, $APLOOP/UVDATA.FOR, $APGNOT/ACFIT.FOR, $APGNOT/DECOR.FOR, $APGNOT/SDCAL.FOR, $APGNOT/SELSD.FOR, $APGNOT/SPLIT.FOR, $APGNOT/UVFND.FOR, $QPGNOT/BLCAL.FOR, $QPGNOT/BPASS.FOR, $QPGNOT/CALIB.FOR, $QPGNOT/FRING.FOR, $QPGNOT/PCAL.FOR, $YPGNOT/IBLED.FOR, $YPGNOT/POSSM.FOR, $INC/DSEL.INC, $INC/CLASSIO.INC, $INC/OBJPARM.INC, Also changed the "COPYRIGHT" statment in ACFIT to the mealy-mouthed user agreement statment. Moved nowhere. 7184. September 29, 1992 UCMPRS calls Bill C. The following routines were modified to use UCMPRS rather than ZCMPRS to compress uv data files: $APGNOT/ACFIT.FOR, $APGNOT/DECOR.FOR, $APGNOT/SDCAL.FOR, $APGNOT/SELSD.FOR, $APGNOT/SPLIT.FOR, $QPGNOT/BPASS.FOR, $QPGNOT/CALIB.FOR, $QPGNOT/FRING.FOR, $QPGNOT/PCAL.FOR Moved nowhere. 7185. September 29, 1992 Number of channels/IFs Bill C. Modified the following to use MAXCIF as the maximum product of the number of channels and IFs: $INC/DSEL.INC, $APLNOT/DATBND.FOR, $APLNOT/TABBP.FOR, $APGNOT/SPLIT.FOR, $QPGNOT/BPASS.FOR, $QPGNOT/FRING.FOR, $YPGNOT/IBLED.FOR, $YPGNOT/POSSM.FOR. Also added checks that the numbers of channels and IF were OK in these routines and in: $APLNOT/SDGET.FOR, $APLNOT/UVGET.FOR, Moved nowhere. 7186. September 29, 1992 TVCPS Chris New task that reads an image from the TV and creates a PostScript file containing a colour or greyscale copy of the screen image. The PostScript image may be encapsulated for inclusion in a document and may be scaled to fit on any sized "paper". Moved nowhere. 7187. September 30, 1992 $APLNOT/SOUNDX.FOR Glen Seemed to be some problem with selecting a single source after comparision with freqids. Created a buffered list of sources and transfered found/excluded sources to the list. Re-linked SPLIT. Seems that the SPLIT FREQID default may have changed(?). May have to move to 15OCT92 7188. September 30, 1992 ZLASCL and ZLPCL2 in SYSUNIX Chris Modified these scripts to recognise a new printer type PS-CMYK: PostScript with CMYK extensions. CMYK extensions are required to print colour PostScript output from TVCPS and are available on most colour PostScript printers. This is a temporary hack: the printer scripts are extremely clumsy and should be redesigned. Moved nowhere. 7189. September 30, 1992. Gustaaf I added CLCOR to $SYSLOCAL/OPTIMIZE.LIS. The optimized version gave a segmentation fault. Moved nowhere. 7190. September 24 & October 1, 1992 XAS Eric Changed upper limit to allowed intensities to the maximum that can be supported by the machine with our graphics overlays, cursor, foreground, and background (usually 237). The default is still 199. Added explanations of this to XAS.HLP. Moved nowhere - could go to 15OCT92. 7191. September 30, 1992 TVRGB, TVHUI Eric Put back major revisions of these two tasks. Both now offer menu-driven enhancement capabilities and interactive setting of the window used for the interactive subimage. Both also offer a display of the cursor position to help fine tune enhancements and an abort mode that causes the programs to exit without the usual, possibly slow, updates of the displays and, in TVHUI, without writing the output image. TVHUI now offers the option of including an image of saturation as well as intensity and hue images. Created HLPTVRGB.HLP and HLPTVHUI.HLP to provide on-line help imformation and modified the help files. Moved nowhere - could go to 15OCT92. 7192. October 1, 1992 PP Eric Changed $APLUNIX and $APLVMS versions of PP.FOR to have a list of include files that may be included an arbitrary number of times in each subroutine. This is needed for vectorization directives. Moved nowhere. 7193. October 1, 1992 TVHUI Eric Added interactive option to use log or linear transfer functions for the Intensity image while displaying. Linear is of course used when writing the output file. Also updated HLPTVHUI.HLP. Moved nowhere - could go to 15OCT92. 7194. October 1, 1992 FARAD Chris Fixed a call to CALINI which should have been changed when the CALINI arguement list was changed in May '92. Also removed unused variables and reordered variables in a COMMON block. Moved to 15OCT92 on October 2, 1992. 7195. October 2, 1992 SNCOR Bill Erickson/Phil Added another option to SNCOR to transfer phases between frequencies with appropriate scaling to compensate for ionospheric delays. Added two new adverb arrays to the system as a result, SNCORPRM and PHASPRM. Moved nowhere. 7196. October 8, 1992 ZTAPE Eric The beginning-of-medium was tested for in one part of the BEGW operation, but not in the second part. The IBMs at least did then reort this as an error. I tested this a lot, but not, I guess on all architectures. THe SUN, I conclude, does not return beginning-of-medium under all circumstances that it could/should. Moved nowhere - can we still slide this into 15OCT92? 7197. October 9, 1992 $APLOOP/TABLE.FOR Bill C. Allow resetting the number of rows in a table in TABPUT. Moved nowhere. 7198. October 9, 1992 $APLOOP/TABLEUTIL.FOR Bill C. Close input table in COPHED before opening the output table. This is necessary to overwrite the table. Moved nowhere. 7199. October 12, 1992 ANCAL Chris Removed a test that caused caused gain corrections to be flagged if the zenith angle was outside the range 0 degrees to 90 degrees. This made it impossible to calibrate space VLBI data since the formal zenith angle for an orbiting antenna may be much larger than 90 degrees. Moved nowhere. 7200. October 13, 1992 VLBIN Phil Moved a line to avoid a divide by zero. This had not been pulled up from 15OCT92 for some reason. Moved nowhere. 7201. October 13, 1992 TVHUI, TVRGB Eric Corrected TVRGB for errors in computing lookup tables on real TVs and for the option to force real TVs to do the one-channel mode. Fixed the help file to describe it correctly. Changed TVHU and both help files (HLPTVHUI and TVHUI) to add the option to optimize the red-green-blue's instead of using a simple color scheme. This requires using 2 TV channels. The inclusion (or not) of the step wedges in the optimization is an interesting and needed option with this optimization. Also corrected some small errors related to TVHUI on real TVs. Moved nowhere. 7202. October 14, 1992 ZLASCL in SYSUNIX Chris Hacked ZLASCL to get around a serious bug in AIX. The status returned by commands called from a shell script invoked from a FORTRAN 77 program get lost. This causes many if-compounds to execute the true branch in inappropriate circumstances. This may be associated with the spurious system error 10 returned by system() when called from FORTRAN 77 programs. Moved nowhere. 7203. October 14, 1992 TVHUI, TVRGB Eric The logic for resetting zooms neglected the case of small image and zoom high eneough to leave most of the menu somewhere off in space. Changed both to lower the zoom just enough to fit in the larger of the menu and the image. Corrected more bugs in TVHUI with the upper step wedge being altered by the display routine. Moved nowhere. 7204. October 14, 1992 FITLD and cousins Phil Putback all the modified .INC and .FOR routines in the VLBA binary tables software. They were changed to reflect problems in the initial design. Also added 3 new types of table: PC => Phase-cal; CM => CALC model; GC => Gain Curve. These will be used in the near future. Not all aspects of the VLBA interface tested yet, will have to wait for VLBA data. Moved nowhere. 7205. October 15, 1992 $INC/PUVD.INC Bill C. Added MAXCCC as a parameter for the maximum number of columns in a CC table. Moved nowhere. 7206. October 15, 1992 CC table interface. Bill C. Added a pair of interface routines, $APLNOT/CCMINI and TABCCM, to simplify access to CC tables. These are similar in structure and usage to other AIPS tables routines. Also added OOP wrappers OCCINI and OTABCC in $APLOOP/TABLEUTIL.FOR. Moved nowhere. 7207. October 16, 1992 Associative lists Bill C. Added a package of routines in $APLNOT to manipulate associative lists (lists of labeled pieces of information). A list is declared as an integer array and can be used as a simple, memory resident data base system of for passing information around inside programs. Each entry has a unique label and can be of type double precision, real , character, integer or logical. Separate enties may be of different type. Each entry may be an array of dimension up to 100 and the dimensionality of each entry is independent. Entries may be written, retrieved or deleted and a simple minded dump routine is provided. No direct references to the contents of the list are needed or allowed from applications software. There is extensive documentation in $APLNOT/ALSINI.FOR. This package consists of the following: $INC/PALIST.INC Parameters needed to define the structure Note: this is NOT to be included in applications software. $APLNOT/ALSINI.FOR - Initialize $APLNOT/ALSPUT.FOR - Store an entry $APLNOT/ALSGET.FOR - Retrieve an entry $APLNOT/ALSDMP.FOR - Write contents with MSGWRT $APLNOT/ALSDEL.FOR - Delete an entry and compress. The following are needed by the public functions. $APLNOT/ALSKEY.FOR - Look entry by keyword. $APLNOT/ALSC2H.FOR - character array to hollerith $APLNOT/ALSH2C.FOR - hollerith to character array. Moved nowhere 7208. October 17, 1992 FILLM Gustaaf Fixed bug in FILLM. The sky frequency, from which the u,v, and w are calculated, is pointed at by the pointer JLOCF. The value of this pointer is correlator mode dependent, and should be updated with every new type of observation. When the required file does not exist, it is created and JLOCF is determined correctly. When the file already exists, the new data are appended, but JLOCF is not updated. This results in completely different sky frequencies and hence erroneous u,v, and w values. The bug was fixed in that JLOCF is explicitly updated each time a file is appended. Surprisingly enough this bug was present in all existing AIPS versions but evidently not spotted before. 7209. October 20, 1992 VBPLT Bill C. Increased maximum number of model components from 2000 to 10000 and made this a parameter. Also general cleanup of unused variables and inconsistent common usage. Modified to use ZPBUFSZ.INC. Also VBPLT.HLP. Moved nowhere. 7210. October 20, 1992 $APLOOP/VECTOR,ARRAY Bill C. Added functions to copy, fill and add a scalar. Moved nowhere 7211. October 20,1992 VTESS Gustaaf Small bug fix in VTESS. The second output map (the one convolved with the specified beam) was of type POINTS. This was changed in NORMAL. This allows tasks like CONVL to use the beam in the header in an appropriate way. 7212. October 21, 1992 ACFIT Phil After the ZPBUFSZ changes an unnecessary equivalence no longer worked. Removed it. Moved nowhere. 7213. October 21, 1992 FRING Phil Removed the now unnecessary restriction that there be > 1 channel in a database before fringe-fitting is allowed. This was the case several years ago but the code has now been fixed. Moved nowhere. 7214. October 22, 1992 MWFLT Glen Minor improvements to handle subimages correctly and perform an image saturation correction. Moved nowhere. 7215. October 22, 1992 $APLSUB/CCOPY.FOR Glen Character string copy utility for use with soon to be put back PHASE task. Moved nowhere. 7216. October 22, 1992 XGAUS Gustaaf XGAUS did a wrong conversion from the fitted location of the central pixel to the associated velocity. This caused the CENTR1 map to have wrong values. Now the conversion is done properly. Moved nowhere. 7217. October 26, 1992 SQASH Gustaaf SQASH now puts the taskname in the history file in upper case. Moved nowhere 7218. October 26, 1992 XSMTH Gustaaf XSMTH failed when specifying BOXCAR smoothing parameters other than the input values. Rewrote the BOXCAR part of XSMHED. Moved nowhere. 7219. October 27, 1992 ALIST additions Bill C. Added routine $APLNOT/ALSASK.FOR and $APLNOT/ALSADD.FOR to inquire about an entry in an ALIST or to alocate space (but no data) for a new entry. Also modified package description in $APLNOT/ALSINI.FOR. Moved nowhere. 7220. October 27, 1992 XAS Eric Changed it to make sure that it is using the default visual and, if it is not, to create and install its own color table before creating the window. Added code to use a GrayScale visual if a PseudoColor visual is not available. Moved nowhere. 7221. October 27, 1992 KWICK Eric Corrected this major POPS routine. Several tests on stack sizes had hard-coded values related to the old size (60) rather than using the newer parameterized size SLIM. Moved nowhere. 7222. October 28, 1992 $APLOOP/OBJMNGR Bill C. Added a function, OBVHKW, to add catalog header keywords to a class virtual keyword list. Also fixed several bugs causing the reading and writing of catalog header keywords not to work. Moved nowhere. 7223. October 28, 1992 TBTSK Bill C. Fixed adverb name in this paraform task. Moved nowhere. 7224. October 28, 1992 AIPS sequence numbers in OOP Bill C. Rationalized the use of the AIPS sequence number in the OOP package. IMSEQ is now used everywhere to define the aips file sequence number. The table class was using SEQ. Affected in $APLOOP: TABLE, UVUTIL. In $APGOOP: CCGAU, CL2HF, MBDLY, POLSN, TBDIF, TBSUB, UBAVG In $QPGOOP: CXCLN Moved nowhere 7225. October 28, 1992 New OOP uv data derived keyword Bill C. Added derived keyword NCORR for the number of correlations in a uv data set. Changed $APLOOP/OBJMNGR.FOR, $APLOOP/UVUTIL.FOR and $APLOOP/UVDATA.FOR Moved nowhere. 7226. October 29, 1992 $APLOOP/UVDATA.FOR Bill C. Fixed logic error in OUVCLN which caused incorrect output AIPS catalog header to be created when cloning a file. Also improved the error messages. Moved nowhere 7227. October 28, 1992 HINOTE Eric Changed AU7, DAPL.INC, POPSDAT.HLP and created HINOTE.HLP and COMMENT.HLP to create and document a new verb called HINOTE. This is used to enter one or more lines in the HI file. Moved nowhere. 7228. October 30, 1992 $APGOOP/PHASE Glen At long last, I have created a task to estimate the coherence time. This task called PHASE is still experimental, but does several usefull things as well as estimate the coherence time. It will do some automatic flagging of UV-data and also create a coherence time averaged (CTA) version of the input. The successful results are obtained when the CTA data is self-calibrated with CALIB. The SN table produced by CALIB can be copied to the original UV-data and SPLIT used to apply the SN table to the original data. This seems to work well and allows the short spacings to be averaged and used in many CALIB solutions for the longer spacings. All the calculations are done in a new set of "OOP?" subroutines called UVPHAS.FOR Best results are obtained if the input uv-data are first SPLIT into single IFs and these IFs processed one at a time. Much work is still needed before it will be of general use. A test procedure, similar to MAPIT, is being developed. Tested successfully on 43GHz VLBI data and VLA C band data. Moved nowhere. 7229. November 2, 1992 $APLOOP/OBJMNGR.FOR Bill C. Speed up OBNAME by reducing the number of calls to ITRIM. Moved nowhere. 7230. November 2, 1992 $APLUNIX/ZACTV8.FOR Pat The "Could not translate logical" message for TVALT was incorrectly stating that the logical was ARCH. Fixed. Moved nowhere. 7231. November 2, 1992 $APLNOT/GAININ.FOR Glen For single source data, if CLUSE=0, it was set to 1, not the latest version of the SN table. This caused problems for a procedure which created an undetermined number of SN tables, and always wanted to use the latest one. Moved nowhere. 7232. November 2, 1992 $APLSUB/CFILL.FOR Glen New subroutine which is used to fill arrays of character strings in the manner in which FILL and RFILL works. Moved nowhere. 7233. November 2, 1992 $APGNOT/SPLIT.FOR Glen Added a message to tell which SN/CL table is actually being applied to the data. Used new subroutine CFILL to fill source name arrays. Changed GAINUSE=0 for single source data default from SN table 1 to latest SN table. Moved nowhere. 7234. November 3, 1992 $APLOOP/GENERIC.FOR Bill C. Added CLEAN process class. Moved nowhere 7235. November 3, 1992 $APLOOP/IMAGEUTIL.FOR Bill C. New Image class utility package. Has one routine, IM2TAB, to create a table object from an image object. Moved nowhere. 7236. November 3, 1992 $APLOOP/IMAGE.FOR Bill C. IMCREA was not setting the image sequence properly in an image object when if created the AIPS catalogued image. Initilize array name disk in IMGSCR. Added subroutine IMCDES to remove 'Destroy on fail' flags. Modified CFLCLR to compress DFIL lists when file status is cleared. Moved nowhere. 7237. November 3, 1992 $APLOOP/ARRAY.FOR Bill C. Changed call sequence to ARROPN to add a READ or WRITE flag. Add more informative (i.e. some) error messages in ARREAD and ARRWRI. Also moved some of the initialization from ARREAD, ARRWRI to ARROPN to be sure that it was always done. ARROPN now properly sets windows including complex data types. Added better error messages in many routines. Window (BLC, TRC) now cleared on output images for most functions. Array function routines now allow multiple useage of the same object; i.e. the same as input and output by creating temporary objects when necessary. Corrected handling of complex rows in array function routines and ARRIO. Modified open and close to read and possibly write the catblk. Moved nowhere. 7238. November 3, 1992 $QOOP/QARRAY.FOR Bill C. Changed call sequence to ARROPN to add a READ or WRITE flag. Now sets output data type. Fixed bug in logic determining size of hermitian reverse transform in ARRFFT. Moved nowhere. 7239. November 3, 1992 $APLOOP/TABLEUTIL.FOR Bill C. The table interface routines were not closing the table when called with OPCODE='CLOS'. Moved nowhere. 7240. November 3, 1992 $APLOOP/CLEAN.FOR Bill C. New class, a clean process class. Most of the work is done in "Q" routines so is not in this module. Moved nowhere. 7241. November 3, 1992 $APLPGM/SUBIM Glen Added OPCODE to SUBIM. Options are ' ', for old SUBIM, OPCODE='AVE' to average (with blanking) in XINC, YINC box, OPCODE='MIN' to return minimum (with blanking) in XINC,YINC and OPCODE='MAX' to return maximum (with blanking) in XINC,YINC. Updated help file. Moved Nowhere. 7242. November 4, 1992 CGASET,CSLGET Bill C. Fixed a couple of problems that caused data to be incorrectly flagged. The first problem was in CGASET which tried to determine in each time if the data were VLBI data (delay and/or rate corrections) and if so updated the baseline dependent factor table for the fringe rate amplitude correction. The problem occured with a combination of applying antenna position corrections in CLCOR and the problem to be discussed next). The antenna position corrections make delay and rate corrections which, while small for VLA data, trigger the fringe rate correction. If the offending antenna is not present for one update of the current calibration table then the baseline factor table will not be updated and the previous baseline factors, hence flagging, will be used. The second problem is a combination of a change of the time label in CLCOR and the assumption in CSLGET that CL table entries come in groups with EXACTLY the same time. CLCOR doesn't appearing to be overtly changing the time stamp but appears to do so. Since the antenna which required a position change also had a slightly different time causing it to be the only antenna at a single set of CL table entries. The fixes were 1) in CGASET to keep old values of the vlbi logical flag so that if the rate/amplitude corrections ar made once they are always updated and 2) add a tolerance of 0.1 seconds to what CSLGET thinks is the same time. Moved nowhere. 7243. November 4, 1992 CLCOR Bill C. In the routine that corrects the antenna positions (ANTPOS) the total model delay polynomial was being addressed as a real rather than as a double precision value. This was changing the value of the time label. Since it was modifying the lsb the change was probably never large by was enough to cause problems in CSLGET. Moved nowhere. 7244. November 6, 1992 $SYSUNIX/TVDEVS.SH Dave A. Added new AOC workstations to the TVDEVS.SH listing. Moved nowhere. 7245. November 6, 1992 $APLGEN/ZRLR64.FOR Pat JingPing Ge found a bug that affected any IEEE machine with little endian byte order here. The data was not being byte flipped because of a typo. This mostly affects FITTP, although apparently TVMON and VLBIN also refer to it. Affected machines are DecStations or DecSystems running Ultrix. Moved nowhere. 7246. November 6, 1992 PCAL Chris The subroutine PCLSEL, which copies selected data to a scratch file, was taking the dimension of each axis in a visibility record from the original catalogue block for the uv-file (CATUV) rather than the version modified by UVGET (CATBLK). This caused some interesting failures for data sets where the axis order was changed by UVGET. PCLSEL now uses the correct catalogue block. Moved nowhere. 7247. November 7, 1992 PP Eric Added BLOCK DATA to the list of subroutine declarators. Moved nowhere - will need special rebuilding everywhere. 7248. November 10, 1992 $APLOOP/INPUT Bill C. Better error diagnostics. Moved nowhere 7249. November 10, 1991 $APLOOP/UVDESC.FOR Bill C. Made UVDCOP smarter about derived keywords; added a table of such. Moved nowhere 7250. November 10, 1992 $APLNOT/ACTFRQ.FOR Phil New subroutine. It appears that nowhere in the system was it possible to make one subroutine call in order to determine the frequency of a given channel in a given IF for a given source at a give time. ACTFRQ fills that gap - it has a fairly simple calling sequence so should be useful. Moved nowhere. 7251. November 10, 1992 CVEL Neil/Phil The routine that interpolated weights placed the output into the wrong array so the weights were never interpolated properly - not catastrophic because spectrally they are nearly always the same. Also changed to use a real division instead of an integer division to find the interpolation weights internal to the routine. Move nowhere. 7252. November 11, 1992 CSLGET Phil All the dire predictions about AIPS calibrations having some fundamental deep-rooted bug came to naught, the problem was a typo in a go to statement, a one-character fix was all that was needed. Worried us though didn't it? Move nowhere. 7253. November 13, 1992 GETANT Henrietta May/Bill C. Made appropriate changes for the AT compact array (ATCA). It uses earth centered, right handed antenna coordinates. Moved nowhere (sent to Oz). 7254. November 13, 1992 ZLPCL2 and PRDEVS.SH Pat Added BIGPRINT and BIGTHRESH environment variables so that large print jobs (default: large > 1000 lines) can be diverted to another, more suitable printer. This was a problem at the AOC. Moved nowhere. 7255. November 13, 1991 $APLGEN/ZRHPRL Tony Beasley/Bill C. Trapped underflow case and return 0.0. This was apparently causing trouble on Suns althought it shouldn't be necessary. Used in Mk3IN. Moved nowhere. 7256. November 16, 1992 $UPDUNIX/UPDCONTROL Pat/Dave A. (Gee Tony/Bill, did you REALLY do that in 1991? :-) Modified the controlling script for the midnight job so that it now notifies whoever about changes to PP, NEWEST, XAS, SSS, and XVSS in addition to POPSDAT.HLP (cases where something special needs to be done). Also removed reference to old CVAX VPOPS account and replaced with aipssrc. This will prevent the MNJ from doing what the quarterly update does. About to be moved to NRAO1 and the AOC this date. 7257. November 17, 1992 UVPGET Bill C. When the dimension of ILOC was last changed, format 1030 (which prints it on error) was not changed as well; this causes tasks to abort rather than die correctly. Moved nowhere. 7258. November 17, 1992 $RUNSYS/HEX.001 Glen Little procedure to convert decimal to Hex. Moved nowhere 7259. November 17, 1992 $APLSUB/DFILL.FOR Glen Little subroutine to fill Double precisions with a constant double. We should remove excess loops from code, and use the fill and copy routines. Moved nowhere. 7260. November 17, 1992 $APLNOT/STEPCH.FOR Glen Subroutine to take an real number index/year in order to set up to use mark calabretta's coordiante transformation routines. (Name means SeT EPoCH.) Moved nowhere. 7261. November 17, 1992 $APGNOT/STARS.FOR Glen Moved STARS from $APLPGM to $APGNOT so that it could use the coordinate transformation routines, CRDSET, CRDTRN and STEPCH. Now the input file will be optionally transformed to any of the "standard" coordinate systems. Ecliptic is not supported. Updated STARS.HLP Moved nowhere. 7262. November 17, 1992 MAPCLR Bill C. Now traps invalid disk or slot numbers and avoids annoying but meaningless error messages when a task finishes. Moved nowhere 7263. November 17, 1992 $APLNOT/SETGRD.FOR Bill C. Modified not to create scratch files if they already exist. Moved nowhere. 7264. November 17, 1992 $APGNOT/HOLGR.FOR Mark Fixed an error in the computation of the surface deviation map in routine SURDEV as reported by Jim Morgan, UMd. 7265. November 19, 1992 FILLM Gustaaf Added an adverb DOTABLE to FILLM. If 1, it will create an OT (over-the-top table). This table is needed for a new baseline fitting task (BASFT) soon to be introduced to AIPS. As a result, two new subroutines were added to $APLNOT, namely OTTINI and TABOTT. These two routines will also be called by BASFT. Obviously DOTABLE defaults to 0 and most users will probably not even notice the difference. 7266. November 23, 1992 $SYSUNIX/OPTIMIZE.LIS Pat Added several settings for DEC that seem to help make AIPS work under Ultrix 4.2 and the DEC fortran compiler (not the Mips one). These were reported by Jingping Ge at Brandeis. The -O2 setting appears to cause quite a few problems. Moved nowhere. 7267. November 23, 1992 $APLOOP/UTILITY.FOR Bill C. New module of general utilities: OBSCNF - Fills in file descriptive info for scratch files created outside of the OOP system but attached to objects. OBFEXS - Determines if an object is fully instantiated (underlying files exist). Moved nowhere 7268. November 23, 1992 $APLOOP/IMAGE.FOR Bill C. Adds keyword/value SCRCNO to scratch images giving the DFIL.INC scratch file number needed by numerous AIPS utilities. IMGCRE now does not create an extant object. Added routine IMBTNF to update the tables info from one CATBLK to another. Moved nowhere. 7269. November 23, 1992 $QOOP/QUVUTIL.FOR Bill C. Rearrainged OUNFWT to put more of the information lookup into this routine rather than having to pass it. Also added front to AIPS routines: OUMSUB - Subtract an image model from uvdata object. OUMDIV - Divide an image model into a uvdata object. OUVIMG - Make an image or beam from a uvdata object. Moved nowhere. 7270. November 23, 1992 $APLOOP/CLEAN.FOR Bill C. CLNOPN now open the beam 'WRIT' for UV data to force it's creation. Moved nowhere. 7271. November 23, 1992 $APLOOP/UVUTIL.FOR Bill C. In UV2SCR explicitly copy CATBLK produced from the calibration package to the scratch file. UV2SCR will now implement the 'VLBI' weighting (1/4 power) if the appropriate value is in a keyword 'UVWTFN' attached to the input uvdata. Added routines to copy tables and select a portion of an FQ table with the appropriate update of the uv CATBLK. Renamed old UVTCOP to UVCALT. Modified U2IDES to make "RA" reference pixel NX/2 rather than NX/2+1. Fixed bug in UVREFQ computing reference frequency pixel. Moved nowhere. 7272. November 23, 1992 $APLOOP/IMAGEUTIL.FOR Bill C. Added routine IMTCOP to copy tables with update of the image catalog header. Moved nowhere. 7273. November 23, 1992 $APLOOP/TABLE.FOR Bill C. Save updated catalog header in TBLCOP. TABCRE now does not create an extant object. Moved nowhere. 7274. November 23, 1992 $APLOOP/TABLEUTIL.FOR Bill C. Now saves updated catalog header whenever a table could have been created. Changes numerous routines. Moved nowhere. 7275. November 23, 1992 $APLOOP/UVDATA.FOR Bill C. Correctly fill in name, class etc in scratch file CATBLK in a probably vain attempt to keep CATDIR from reaming me out again. OUVSCR now check to see if the output object exists and simply returns if it does. OUVCRE now does not create an extant object. AIPS catalog flags are not set for READ as UVGET does this alreads and a double set of 'READ' flags completely disables TABINI from reading tables attached to this object. Another wonderful feature of AIPS catalog flags. Added routine OUVTNF to update the tables info from one table to another. Use this routine in OUVOPN to get table info for the uvdata. Moved nowhere. 7276. November 23, 1992 $APLOOP/ARRAY.FOR Bill C. ARRCHK now includes consideration of the dimensionality of the array and random numbers for undefined axes are not considered. Also ARRWIN now properly takes notice of the total number of dimensions. Moved nowhere. 7277. November 23, 1992 $QOOP/QCLEAN.FOR Bill C. This module contains most of the working code for the CLEAN class. It contains APCLN-like (CLNIM) and MX-like (CLNUV) CLEAN deconvolutions. Moved nowhere. 7278. November 23, 1992 $APLOOP/CCSEL.FOR Bill C. New task: CCSEL is intended to filter out weak isolated CLEAN components which can cause a signifigant bias. This task sums the flux of all components within a specified distance of each component and then selects those components for which this sum exceeds a specified threshold. Also CCSEL.HLP Moved nowhere. 7279. November 23, 1992 $QPGNOT/UVADC.FOR Bill C. New task. This task is like UVSUB except that it adds components to a uv data set and optionally corrects for w-term distortion, bandwidth smearing and an incorrect assumed center frequency. Also UVADC.HLP Moved nowhere. 7280. November 23, 1992 $QPSAP/QPTADC.FOR Bill C. New routine. performs arithmetic for UVADC. Moved nowhere. 7281. November 23, 1992 $SYSUNIX/PP Pat Modified sed commands so they don't do the wrong thing when there is a directory with a dot in its name as part of the current directory (or the AIPS_VERSION). Moved nowhere. 7282. November 25, 1992 $SYSUNIX/FC Glen Modified FC to remove all reference to FCOPTS.SH. For some time FC has only used FDEFAULT.SH to determine compilation Options. Updated FDEFAULT.SH for varios architectures and and removed all FCOPTS.SH files in the various directories. Also updated the $SYSLOCAL/OPTIMIZE.LIS Moved nohwere 7283. November 25, 1992 IBLED Phil Minor fix to get around the occasional problem of the task loosing track of the sample separation interval. Moved nowhere. 7284. November 30, 1992 POSSM Phil A major rewrite of POSSM with the following features: (1) Can now produce multiple plots/page controlled by the adverb NCOUNT (2) All labelling routines and plot generation routines tidied up and rationalized. It will be much easier to maintain from now on. (3) Added the BASELINE adverb array to conform to other data selection tasks. Also now deal correctly with antenna selection and deselection. (4) fixed a bug in the x-axis labelling that has been there from day 1. Moved nowhere. 7285. November 30, 1992 $APLNOT/CSLGET Glen PCAL set the DXTIME to 1 day, which caused CSLGET to make time tolerence too large. Arbitrarily added a DXTIME upper limit of 2 hours to the time tolerance logic. Changed seconds per day and milli-seconds to parameters. Moved nowhere. 7286. November 30, 1992 $APGNOT/AFILE Phil Removed the small section of code that worried about duplicate entries for R/L type 52's. According the to MPI group that was the wrong thing to do. Moved nowhere. 7287. December 1, 1992 $SYSSUN/COMLNK,COMRPL,LIBR,LINK and $SYSUNIX/AIPSEXEC Pat Extensively modified these shell scripts to support two sets of libraries: one for nondebug, optimized code, and the other for debug code. There are also two load area hierarchies supported. $LIBRDBG points to $AIPS_VERSION/$ARCH/LIBRDBG and $LOADDBG to a similar load area. The whole scheme depends on the existence of a single file $SYSLOCAL/DOTWOBIN. If this isn't present, all bets are off. -- In addition, moved location and name of shared libraries. From now on, these will be in $LIBR/.so instead of the old $LIBR//SUBLIB.so; this should make a binary distribution of a shared library installation possible. -- Also, if the staging directories $LIBR// do not exist, they are created. Likewise for $LIBRDBG//. -- Finally, if environment variable AIPSRCS exists and is set to the string "NO", the AIPS RCS transaction files on baboon (or wherever) are NOT updated. Moved nowhere, but I disabled the AOC main MNJ to prevent riled astronomers from breaking Dave's door down tomorrow morning while he's in the AIPS meeting... 7288. December 2, 1992 $QPGNOT/PCAL Phil Changed the default value of DXTIME from 1 day to 1 sec. Moved nowhere. 7289. December 3, 1992 $APGOOP/TBDIF Bill C. PRTHED was trying to access 'SEQ' which has been changed to 'IMSEQ'. Moved nowhere. 7290. December 3, 1992 $APGNOT/BPCOR Phil The call seq. to CHNDAT was never changed in this when we modified it. Moved nowhere. 7291. December 3, 1992 $APLPGM/TACOP Gustaaf INVERS=0 in TACOP now defaults to the highest version number. This was already claimed in the help file, but in practice INVERS=0 defaulted to INVERS=1, creating much confusion in spectral line calibration. 7292. December 5, 1992 $QYPGNOT/MX Glen Fixed problem found by Jim Condon for restarting MX to restore a different sized beam. Also made the use of the FLUX adverb much more convoluted. (Sorry, will add a new adverb later) IF Flux > -99 and < 0 then clean stops at first negative clean component. If FLUX < -99 then FLUX is interpreted as the minimum MILLI-Pecent change in the Flux density allowed during a MAJOR CYCLE. (Ie FLUX = -1000 => MX stops if difference in FLUX is less that 1 % between major cycles.) This feature is used in the COMAP procedure to automate estimation of coherence time and self-calibration. Moved nowhere. 7293. December 7, 1992 $YPGNOT/POSSM.FOR Phil Fixed minor bug in the estimation if what baselines to plot when all are asked for. Added ability to plot 2 new data types - the auto-correlation functions and cross-correlation functions. Also modified .HLP to reflect these changes. Moved nowhere. 7294. December 7, 1992 $APLNOT/ALSDMP.FOR Bill C. This routine was blowing up if a character string entry in the ALIST was too long. Moved nowhere 7295. December 7, 1992 $QPGNOT/CVEL.FOR Phil Sometime in the rewrite of CVEL it forgot how to flag data. Since I no longer use UVGET I had to clone DATFLG and NXTFLG for CVEl's use. Moved nowhere. 7296. December 8, 1992 $SYSUNIX/CC (now AIPSCC) Pat With the increased presence of C++ compilers, the conflict between the standard name for these (CC) and the AIPS script of the same name is increasing. I've renamed $SYSUNIX/CC to be $SYSUNIX/AIPSCC and have modified all scripts that call it: COMRPL, COMLNK (generic and Sun4 versions of these), and a bunch of CCOPTS.SH files ($SYSUNIX, $SYSSUN, $SYSIBM, $SYSDEC, $SYSCVEX, $SYSCVCV, $SYSVLAC1, $SYS3090, $SYSALLN, $SYSCRI). Moved nowhere. 7297. December 11, 1992 $SYSUNIX/OPTIMIZE.LIS Pat Added a line for COMB on architecture CVEX (convex). Received some complaints that it needs -O1 or less. Moved nowhere. 7298. December 11, 1992 $YPGNOT/VBPLT.FOR H.May/Phil Added the AT as a specific type of array so the model can be plotted correctly. Also updated the help file. Moved nowhere. 7299. December 11, 1992 FRING Bill C. FRING was writing the wrong source IDs to the SN table when there were multiple source in a file and only one source was selected. Now there is no attempt to get the source ID from the iNdeX table. It is either 1) 0, or 2) obtained from the visibility data (multiple sources) or 3) obtained from SOUWAN(1) if a single source was selected. Moved nowhere 7300. December 11, 1992 $SYSAIPS/AREAS.DAT Pat Added APLHP, APLSOL, QHP, QSOL, INCHP, INCSOL, and SYSHP, SYSSOL definitions. Removed obsolete YXAS area. Have yet to create the RCS and actual areas. Moved nowhere. 7301. December 11,1992 FILLM Gustaaf Changed FILLM in two places: 1) I got rid of the offensive DOTABLE adverb. Since all existing adverb positions were taken I had to use the following ugly solution: when CPARM(1).GE. 100000, FILLM subtracts 100000 from CPARM(1) and creates the OT table. While such a solution may be frowned on for general use, the very incidental and specialized need for OT tables may justify it. 2) Several error reports have come in from users who specified TIMERANGE but left REFDATE blank. Taking Gareth's advice no time was spent on solving this problem. Instead, FILLM issues a warning message in case TIMERANGE is specified, and REFDATE is left blank. Moved nowhere. 7302. December 14, 1992 ISPEC Gustaaf Following frequent requests from users, I added adverbs DOCRT and OUTPRINT to ISPEC. Moved nowhere. 7303. December 14, 1992 APCLN Gustaaf Fixed bug. The PHAT option did not work; in fact PHAT = 0 and PHAT = any other value gave identical results. In BMSHOV a relevant pointer was updated too early, which caused the new beam centre value to be put in a unused location. Furthermore, I isolated the hard-coded parameters (size of area, and cut-off) used in fitting the beam. This in preparation of future attempts to improve on the beam-fitting algorithm. Moved nowhere. 7304. December 15, 1992 XASERVERS Pat Made the logging messages about what XAS or TEKSERVER is running a bit smarter. They now say who is running XAS or TEKSERVER, and where the DISPLAY is set. These messages go to a log file in /tmp (AIPS.TVTEK..LOG) unless you have the variable AIPS_VERBOSE set to something non-blank. Note: you have to set this in your .cshrc or .bashrc (or .kshrc?) so that it gets defined when "rsh" is done. Also made it accomodate hosts where the hostname command returns a fully qualified name (e.g. orangutan.cv.nrao.edu instead of orangutan). Moved nowhere. 7305. December 16, 1992 $YPGNOT/POSSM Phil A couple of minor changes to the multiple plots/page mode. (1) The global labels are now written below the plot rather than the clumsy 'at-the-side' approach (2) Tidied up and fixed the antenna/baseline selection (3) Occasionally POSSM would create a blank plot at the end of the data - will now only create plotfiles when data has been collected. Moved nowhere. 7306. December 17, 1992 BLING Chris First version of the baseline-based fringe-fitting program. This is still experimental (and pretty useless until I write the program that applies the results). Moved nowhere. 7307. December 18, 1992 BSINI in APLNOT Chris Changed the units field for rate and acceleration to agree with what BLING writes (ie. Hz and Hz**2). These units are somewhat more user oriented than the sec/sec and sec/(sec**2) used originally. Updated the header comments for OTABBS (in APLOOP/TABLEUTIL.FOR) and TABBS to match. moved nowhere. 7308. December 18, 1992 $YPGNOT/POSSM.FOR Phil In the multiple plots/page section the number of IF's would be lost for plots > #2, this caused initialization problems. Fixed. Moved nowhere. 7309. December 18, 1992 $YSUB/STARPL.FOR Glen Fixed a bug about ellipse sizes pointed out by J. Condon. Moved nowhere. 7310. December 18, 1992 ZLASCL in SYSUNIX Chris Fixed a typo that prevented colour PostScript printers being recognised unless they where the default printer. Moved nowhere. 7311. December 18, 1992 TVCPS Chris TVCPS now writes a ``%%Requirements: color'' comment in the header if writing a colour image. It should have done this all along. Moved nowhere. 7312. December 21, 1992 $YSUB/STARPL.FOR Glen Fixed a bug introduced during fix of 92-dec-18. Moved nowhere. 7313. December 21, 1992 $YPGNOT/POSSM.FOR Phil Tidied up the antenna/baseline selection into a separate subroutine. A couple of variables were in the wrong common and so the velocity in the printed output was all zero. Moved nowhere. 7314. December 21, 1992 $APGNOT/AVSPC.FOR Phil Added a new option requested by the VLBA correlator people. AVOPTION = 'SUBS' will now do a boxcar average on a spectrum reducing it from n to m (< n) channels. Any extraneous channels from the input spectrum are thrown away. One little twist is that the average is done in the direction away from the LO so that if channels are thrown away they are those at the roll-off end of the filters. Also modifed AVSPC.HLP. Moved nowhere. 7315. December 21, 1992 $QPGNOT/CVEL.FOR Phil Fix # 7295 did not work if no FG table was present. A minor oversight now fixed. Moved nowhere. 7316. December 21, 1992 $APLNOT/DATCAL.FOR Glen DATCAL was randomly corrupting a single source data set when applying a STOKES='I' solution (ie 1 polarization solution) to a standard VLA (RR,LL,RL,LR) uv-dataset. Logic now checks for a single solution and does the appropriate thing for my case. I hope it still does the correct thing for all other cases.... (Before it was accessing an array out of bounds.) Relinked SPLIT. Moved nowhere. 7317. December 22, 1992 $APGNOT/UVAVG.FOR Phil Changed the conditions under which a warning message was issued about the possibility of a bad average. It was untrue. Moved nowhere. 7318. December 22, 1992 $SYSUNIX: HOSTS.CSH, HOSTS.SH, TVDEVS.SH, START_AIPS Pat Changed the behaviour of these shell scripts so that they use the simple hostname, e.g. orangutan, instead of the fully qualified hostname, e.g. orangutan.cv.nrao.edu. Moved nowhere, but Dave and I should merge them with AIPS_ROOT versions in CV and at the AOC. 7319. December 22, 1992 CHANGE.DOC Pat The formatting of some items near the beginning (before Sept. 29) was screwed up. Fixed. Will no doubt play havoc with the "diffs" step in the MNJ's. Moved nowhere. 7320. January 4, 1993 $APGNOT/PRTUV.FOR Craig/Phil Added 2 new options to PRTUV. CPARM(10) will scale the amplitudes listed at the users request. DOEBAR will print the 'error' on the data rather than its weight. Also changed PRTUV.HLP. Moved nowhere. 7321. January 4, 1993 $YPGNOT/POSSM.FOR Phil Finally got a foolproof subroutine for creating lists of antennas/baselines to plot. Moved nowhere. 7322. January 4, 1993 $QPGNOT/FRING.FOR Phil There was always a lot of frustration over the sometimes small amounts of data that FRING would deal with in a fringe-fit. It is a function of the # antennas, # channels & the integration time. The array size was parameterized at 300000 - and this was just too small for moderat applications, I increased it to 1,000,000 and that is somewhat better. For wide windows the user will still have a problem and may have to narrow the windows but at least more data can be fringe-fit at a time. Moved nowhere. 7323. January 4, 1993 IAND, IEOR and IOR in APLUNIX Chris Added the return type (int) to these functions. All three will now compile cleanly with both K&R and ANSI compilers even if all warnings are enabled. Moved nowhere. 7324. January 4, 1993 ZABOR2 in APLUNIX Chris Made changes so that ZABOR2 will compile with ANSI C compilers on POSIX.1 operating systems as well as with K&R C compilers. NOTE: if compiling with an ANSI compiler make sure that it is operating in strict ANSI mode rather than some half-baked compatibility mode (ie. use the -Xc flag on Suns). NOTE 2: this version should replace the system-specific versions for POSIX.1 systems. Moved nowhere. 7325. January 6, 1993 BPASS Phil Removed a leftover debug msgwrt in the rereferencing subroutine. Moved nowhere. 7326. January 7, 1993 TVFLG, SPFLG, IBLED Eric Changed handling of FLAGVER so that 0 => 1, but FLAGVER = -n (n > 0) => ignore flags on input, but write any new ones to FLAGVER = +n on output. Also help files of course. Moved nowhere. 7327. January 7, 1993 CONVL Gustaaf Increased size of scratch files created by CONVL. The old size was insufficient for e.g. 512 * 1024 data. Moved nowhere 7328. January 11, 1993 $YPGNOT/VBPLT.FOR Phil Major changes: (1) I stripped out the closure phase plotting, it is going into its own task (CLPLT - following soon), in the course of this many subroutines had to be placed in the APLNOT area so both VBPLT and CLPLT can see them. The relevant routines are: MODPLT.FOR, PLTSET.FOR, PLOTDA.FOR, BASLIN.FOR, SIDERL.FOR, VBPREC.FOR, NUTATE.FOR, POLCHK.FOR nad POLVIS.FOR. (2) I added the ability to average the data on the fly before plotting it - essential code for this is in $APLNOT/BASAVG.FOR. (3) I also added the ability to plot several more quantities: Source #, FREQID, Weight, RMS, int. time. Most of these are useful for debugging the VLBA correlator, but they are there anyway. Moved nowhere. 7329. January 11, 1993 $YSUB/STARPL Glen/Jim C. Jim C. requested a way of creating overlays for the Palomar Sky survey using AIPS and the Guide star catalog. This is possible although painful. The stars were being covered up by the plot mark and a new type (20) now is a cross with the central third removed. Comlnked CNTR, GREYS and PCNTR Moved nowhere. 7330. January 11, 1993 NEW TASK: CLPLT Phil A new task to plot closure phases and the closure model. Based on VBPLT so many of the same inputs: Moved nowhere. 7331. January 11, 1993 $APGNOT/MULTI Phil MULTI had no check in it to see of it was trying to multi an already multi'd database. Inserted the check. We hit this because the CIT correlator now writes multi-source FITS files, hurrah! Moved nowhere. 7332. January 12, 1993 $APGNOT/UVAVG Phil Now writes out the correct integration time after averaging in the r.p. slot Moved nowhere. 7333. January 13, 1993 $APLPGM/SUBIM Glen Seemed to have input an indexing bug to SUBIM during the addtion of mathmatical function upgrade. I believe the YINC bug is fixed. Moved nowhere. 7334. January 13, 1993 $SYSLOCAL/NOSHARE.LIS Phil This turned out to be the problem with POSSM - God knows why? Can no longer use shared libraries with it. Moved nowhere. 7335. January 17, 1993 $APGNOT/ANCAL.FOR Phil Added a new option to ANCAL. If GAINVER = -1 then the CL table is not updated, all that happens is that the TY table is created and filled. This option is used by spectral liners and will be explained in the new version of ACFIT. Watch this space. Also updated the helpfile. Moved nowhere. 7336. January 18, 1993 $YPGNOT/VBPLT Phil Fixed a long-standing bug in the baseline selection in VBPLT. When the user specifies antennas/baselines and the same number appears in both arrays, the autocorrelation is counted in the # of baselines to be plotted, however it does not exist internally and so some baselines are repeated. Used the method now used by POSSM and IBLED - this is more general. Moved nowhere. 7337. January 19, 1993 $YPGNOT/CLPLT Phil Under some circumstances CLPLT would not form the closure phase because the timeaveraging trap did not catch all cases. It now does. Moved nowhere. 7338. January 20, 1993 $APLNOT/DGETSD.FOR Bill C. Fixed a bug in the logic involving the selection of beam caught by the HP compiler. Now calls SCLTD to determine if the beam is selected. Also fixed declarations of arrays used for access of the index table to use MAXNXC rather that the hardcoded and now incorrect value of 6. This routine is used in the single dish package and with any luck none of the problems fixed had noticable effects. Moved the usual places (nowhere). 7339. January 21, 1993 $QNOT/VISDFT.FOR Glen Fixed a bug found by the HP compiler. Should not have previously affected execution, since buffer size are now large. Also added a trap for message about "110 percent complete". Moved nowhere. 7340. January 21, 1993 Phase rereferencing routines Bill C. Added a number of routines for doing polarization coherent rereferencing of phase, delays and rates in an SN table to a common reference antenna. All valid entries in the table are processed. REFCNT: Count reference antenna usage. REFDLY: Rerefererence delays REFRAT: Rereference rates REFFAZ: Rereference phases Moved to the usual places (nowhere). 7341. January 21, 1993 $APGOOP/CCSEL Bill C. The image sequence number was called 'SEQ' rather then 'IMSEQ' in the table object. Moved nowhere. 7342. January 25, 1993 $YPGNOT/IBLED.FOR Phil IBLED has 2 menus, one for a wide screen TV, the other for a narrow screen. The narrow one was incorrect in one entry, it did not allow the user to switch polarizations. Fixed. Moved nowhere. 7343. January 25, 1993 $HLPFIL/SPLIT.HLP Phil Changed a paragraph in the help file to reflect what is actaually done when applying already applied SN tables. Moved nowhere. 7344. January 25, 1993 MK3IN Chris Fixed a bug introduced in July 1992 which caused the tape I/O structures not to be reset when skipping over a file. Because of this data was being missed on some Mk3 archive tapes. Moved nowhere. 7345. January 25, 1993 $APLNOT/SOURNU.FOR Gustaaf The subroutine SOURNU.FOR did not treat the case: blank source name, specified qualifier, correctly. This was discovered in UVCOP. COMLNK'ed the seven tasks that call SOURNU (BLFIT, QUACK, UVCOP, UVFLG, SNPLT, VBPLT, and CLPLT). Moved nowhere. 7346. January 26, 1993 $YPGNOT/CLPLT.FOR Phil There was no check on whether SOLINT = 0, default is now 5 minutes and a warning is printed. Also the user scaling was lost after the first closure triangle was plotted, fixed that. Moved nowhere. 7347. January 26, 1993 OOP catalog flag handling Bill C. The handling of AIPS catalog status flags cause problems when the same image or uv data files were being accessed for read and write simultaneously. Now if a disk and catalog slot are entered in the DFIL.INC common a new catalog flag is not atempted. The affected routines are: $APLOOP/UTILITY.FOR - OBACF Determines if the catalog status is set in DFIL.INC. $APLOOP/IMAGE.FOR - CFLSET No new status is set if program has set one. - CFLCLR Checks to see if flag set before clearing one. - IMCREA Enters file in DFIL.INC $APLOOP/UVDATA.FOR - OUCFST No new status is set if program has set one. - OUCFCL Checks to see if flag set before clearing one. - OUCREA Enters file in DFIL.INC Moved nowhere. 7348. January 26, 1993 $APLOOP/ARRAY.FOR Bill C. The buffer was not always being flushed when an array object opened for write was closed. Routine ARRIO was modified so that this is always done. Moved nowhere. 7349. January 26, 1993 TAPIO in APLSUB Chris Added an 'ADVF' (advance to next file and reset I/O for reading) to TAPIO. This provideds a way to skip files while reading from a tape without first closing and reopening the tape to reset the I/O structures. Moved nowhere. 7350. January 26, 1993 MK3IN Chris Modified ATSKIP subroutine to use the new 'ADVF' feature of TAPIO rather than using ZTAPE directly. This gets rid of the extra 'increased receive buffer size' messages when using remote tape drives. Moved nowhere. 7351. January 26, 1993 ZFRE2.C Pat Moved $APLBERK/ZFRE2.C to $APLUNIX/ZFRE2.C, blowing away the old module there. The old one relied on the output of "df" which is distinctly non-portable. $APLBERK/ZFRE2.C compiled and worked as is on HP-UX, a Posix and SVR4 variant of Unix, so I'm optimistic it will work elsewhere too. Moved nowhere. 7352. January 26, 1993 $APLOOP/HISTORY.FOR Bill C. Don't copy history files in OHCOPY if the input and output are the same. Moved nowhere. 7353. January 26, 1993 MK3IN Chris Added code to issue a message and abort if REFDATE is badly formatted. Without this MK3IN assumed that a badly formatted REFDATE was equivalent to JD 0.0 and silently messed up the timestamps in the output file. Moved nowhere. 7354. January 27, 1993 $APLHP/various Pat Yesterday I put back a bunch of files in the new Hewlett- Packard APLHP area. This is a brief overview of the files and why they are needed. ZABOR2.C SIG_DFL on SIGCLD instead of SIG_IGN (needed?) ZDCHI2.FOR Floating Point defs, System type. ZGNAME.FOR Use HP-specific IGETARG to get command line args ZSSS*.* (7 files) exact copies of $APLBERK code (socket) ZSTRTP.C Use , setresuid in place of seteuid. ZTPWAD.FOR Catch HP-specific IOSTAT-EOF codes (two of 'em) ZTXMA2.C Use POSIX reg. expression functions regcomp etc ZWHOMI.FOR Extra sanity-checking code (may be redundant) There's a bunch more that has to go in $SYSHP and other areas. Later... Moved from JPL yesterday. 7355. January 27, 1993 $INSUNIX/INSTEP1 Pat Made to work under POSIX (grep and echo work differently) and added HP support. Also rationalized the XAS making a bit so it tries to do the right thing vis-a-vis the MIT shared memory extension to X11. Moved from JPL yesterday. 7356. January 27, 1993 UVCOP.HLP Phil Added a sentence about what happens when FREQID's are selected. Moved nowhere. 7357. January 27, 1993 POSSM Phil Added the ability to perform scan averaging if an NX table is present. Fixed a LL only bug caused by an array overflow that I recently added. Fixed a problem that when plotting multiple bandpasses/page and a blank entry occured the program would exit. Also updated the .HLP file Moved nowhere. 7358. January 27, 1993 BLING Chris - Added code to pick a sensible default for the STOKES adverb. - Corrected a type mismatch problem between BLING and the library subroutine BASLST that caused the ANTENNAS and BASELINE adverbs to be ignored. - Removed the assumption that the mean amplitude of the test patch used to estimate the noise is zero: this was leading to excessively pessimistic estimates of the SNR. Moved nowhere. 7359. January 29, 1993 $APLOOP/UVUTIL.FOR Bill C. Added routines for selfcalibration: SLFCAL: Determine self cal solutions SLFREF: Reference self cal solutions to one antenna UVCOPY: Copy one uvdata to another NXTAVG: Average data for next time interval SLFPA: Lower level selfcal solution routine SLFSOU: Write selfcal solution to SN table object. Moved the usual places (nowhere). 7360. January 29, 1993 $QOOP/QUVUTIL.FOR Bill C. Since OUMSUB and OUMDIV use standard AIPS routines they now explicitly mark the output uv data as valid on successful completion. Also clear any AIPS status flags on output, copy relevant tables to the output object and update visibility number in output. Open and close output object to update disk. Moved nowhere. 7361. January 29, 1993 More catalog flag bugs Bill C. OOP routines creating scratch objects were not entering then in the DFIL.INC arrays. Affected routines are IMGSCR in $APLOOP/IMAGE.FOR and OUVSCR in $APLOOP/UVDATA.FOR Moved nowhere. 7362. January 29, 1993 $APLOOP/UVDATA.FOR Bill C. OUVATT now fill in default EIF and ECHAN. Added routine OUCDES to clear destroy-on-fail flag. Moved nowhere. 7363. January 29, 1993 $APLOOP/TABLE.FOR Bill C. In TABZAP DELEXT is now called with status 'WRWR' rather than 'WRIT' to avoid CATIO problems. Moved nowhere. 7364. January 29, 1993 TABBS in APLNOT Chris Added code to increment the row number on read or write. Moved nowhere. 7365. January 29, 1993 BLING Chris Added a missing subroutine call to close the BS table. Moved nowhere. 7366. February 1, 1993 PASTE Bill C. New task in $APGOOP. Inserts or adds a subimage of one image into another. Also PASTE.HLP. Moved nowhere. 7367. February 1, 1993 LWPLA Glen/Bill C Added an option to LWPLA for 35 mm Slides. Moved nowhere. 7368. February 1, 1993 AU7.FOR / ALTSWTCH Gustaaf Added warning to user when doing an ALTSWTCH from FREQUENCY to FELO-HEL on a uv data set. Mover nowhere. 7369. February 4, 1993 $APLNOT/AG...*.FOR Phil Generalized the axis offset column in the AG table. Moved nowhere. 7370. February 8, 1993 $APLSGI, $SYSSGI, $INCSGI, $QSGI Pat Created empty areas for the Silicon Graphics port. Also modified $SYSAIPS/AREAS.DAT to reflect these new areas, and ran AREAS. Finally, threw away the old RCS versions of AREAS.CSH and AREAS.SH in $SYSUNIX; the AREAS script doesn't do anything about these. Checked in the newly created versions instead. Moved nowhere, but may need repeated on all MNJ sites... 7371. February 8, 1993 UVPGET Gustaaf UVPGET gave an obscure error message when one or more axes were missing. The new version prints out the name of the missing axis. I relinked all tasks using UVPGET in $APLPGM and $APGNOT. Moved nowhere. 7372. February 8, 1993 $APLNOT/OTBSRT.FOR Bill C. Removed a hidden death trap. This routine was writing to a scratch file version 1 related to the output table. If the associated catalog entry was a scratch file this caused it to destroy the main data file. It now uses scratch file version 2. Moved nowhere. 7373. February 8, 1993 REFCNT, REFFAZ, REFDLY, REFRAT Bill C. These routines in $APLNOT were not properly handling a flagged SN record as the last record of an SN table and were incorrectly returning error code -1. Moved nowhere. 7374. February 8, 1993 $APLOOP/ARRAYSTAT.FOR Bill C. Added class member functions DATAMEAN and NUMPIXEL and updated documentation. Actual implementation of these function is using ARRAY class member function ARSSET. Moved nowhere 7375. February 8, 1993 $APLOOP/ARRAY.FOR Bill C. Added member function ARSSET to determine ARRAY_STAT members. Any selected windowing is honored. Moved nowhere 7376. February 8, 1993 $APLOOP/CALEDIT.FOR Bill C. Added routines SECSLT and SECSAV to fetch and store the IF , channel and Stokes selection criteria. Moved nowhere 7377. February 8, 1993 BLING Chris Corrected the direction of the Fourier transforms into the search grid. Delays and rates now have the correct sign. Also changed and inappropriate call to the FLOAT intrinsic to a call to REAL (this ANSI violation was missed by almost all of our FORTRAN 77 compilers). Moved nowhere 7378. February 8, 1993 IRING Gustaaf Fixed bug which caused variable to be overwritten for illegal ring number. Also added option to handle non-astronomical images without arcsec - pixel conversion. Moved nowhere. 7379. February 8, 1993 CLPLT Phil Increased the max # triangles plotted to 90. Fixed an array overflow problem and also fixed a bug which did not close down the final plot under some circumstances. Moved nowhere. 7380. February 9, 1993 FILLM Gustaaf Installed new FILLM with following improvements: 1 - improved holography code added by Tim Roberts 2 - added APARM(1) specifying creation of new files. Can be changed interactively using TELL. Moved nowhere. 7381. February 9, 1993 BLAPP Chris New task to apply the results of BLING. Moved nowhere. 7382. February 10, 1993 OPTIMIZE.LIS in SYSUNIX Chris Added BLAPP to OPTIMIZE.LIS. BLAPP will be optimized at level 6 on all machines except Suns. The least- squares fitting routine is taken from LAPACK and should respond well to vectorization and parallelization. Optimization is reduced to level 3 on Sun's since the program is too long for version 1.4 of Sun's compiler to optimize at any higher level. Moved nowhere. 7383. February 10, 1993 BLING Chris Corrected an error in interpreting NX table time entries that was causing scan times to be skewed by half of the scan duration. Moved nowhere. 7384. February 10, 1993 UVPLT Phil UVPLT will now print the channel number being plotted in the information at the top of the plot. Moved nowhere. 7385. February 11, 1993 AIPSUPD, UPDCONTROL, UPDCOMRPL/LNK Pat Fixed some bugs and made these procs verbose if done in interactive mode. (They thought BLAPP.FOR was PP.FOR among other things). Generalized grep (-s/-q) in some scripts. Moved to AOC (aguila, yucca, zuni) a moment ago. 7386. February 12, 1993 $QNOT/UVMSUB,UVMDIV Bill C. Remove prohibition of using gridded model computation methods for TB sorted data. The gridded method routines have been able to handle TB data for some time but this prohibition has never been removed from these routines. This should make CALIB, FRING and UVSUB MUCH faster under the appropriate circumstances. UVMDIV now deletes it's scratch file if it creates one. Moved nowhere. 7387. February 12, 1993 $QSUN/QMTYP.FOR Bill C. New Sun version of the Q routine that decides whether DFT or gridded subtraction is faster ('DFT' or 'GRID' in MX). The appropriate constants were determined on a SUN IPC using a 1024x1024 image. This should speed up MX as the old version was for Vax+AP and under estimated the time for a DFT and over estimated the time for the gridded subtraction. Moved all the usual places (nowhere). 7388. February 12, 1993 $SYSSUN/LIBR.DAT Pat Added QSUN to the QPSAP library list so the previous change can be compiled. Think I got it right :-/ Moved nowhere. 7389. February 12, 1993 $APGNOT/SPLIT Phil Ensured that the message about which SN/CL table is being applied is only printed when DOCALIB is true. Moved nowhere. 7390. February 15, 1993 $SYSUNIX/OPTIMIZE.LIS Pat Integrated the SUN4-specific version with the generic one and blew away the former. In the process, made POPSGN and PCAL non-optimized (former on suns only), made BLAPP the same for SUN3 and SUN4, and made DATCAL non-optimized for suns too. Moved nowhere. 7391. February 15, 1993 $INSUNIX/INSTEP3 Pat Added TPMON and TEKSERVER to the list of routines built. Moved nowhere. 7392. February 16, 1993 $QOOP/QUVUTIL.FOR Bill C. Modified logic of OUMSUB and OUMDIV to create a second temporary uv object to use for allocating a buffer if the input and output objects were the same. This allows updating the input file. Also under these circumstances the uv data object is not opened for write. In OUMDIV and OUMSUB copy only AN table 1 from input object to output. Also if MODCCEND = 0 use the number of CC in the table as advertized. Also numerous other fixes in OUMSUB and OUMDIV. In OUMDIV fixed call sequence to UVMDIV. Moved nowhere. 7393. February 16, 1993 $QOOP/QCLEAN.FOR Bill C. CLNUV and CLNIM were not restarting the CLEAN properly with a single component. CLACLN now writes the total number of CC in the table header. CLRHIS now opens and closes the array for each box, it was failing on multiple boxes. Moved nowhere. 7394. February 16, 1993 $APLOOP/UVUTIL.FOR Bill C. Routines UVTCOP and UVFCOP now check if the input is the same. Added routine UVAMPS do determine the average amplitude as a function of baseline length. In SLFCAL the SN table is now opened after the open call to NXTAVG to avoid LUN conflicts. Also the sort order is marked in the SN table. Fixed several routines that failed if there was no IF axis. Moved the usual places (nowhere). 7395. February 16, 1993 $APLOOP/UVDATA.FOR Bill C. OUVINI was not looking up Status before using it. OUVCLO was not releasing the LUN when the object was opened WRIT. OUVOPN now expands the file if is VO larger than the current number of visibilities. Added class member OLDFILE which is true iff the file existed prior to object instantiation. Moved nowhere. 7396. February 16, 1993 $APLOOP/UTILITY.FOR B. Cotton OBFEXS now returns IERR=0 rather than 1 if the underlying files do not exist. Also it's smarter about using return codes from OBNAME and OBCLAS. It still sometimes returns IERR=1 when EXIST=.FALSE.. Moved nowhere. 7397. February 16, 1993 $APLOOP/TABLE.FOR Bill C. Added routine TBLMRG to merge tables. Added routine TABRMV to delete any underlying AIPS tables. Added routine TBLHIV to find the highest numbered version of a table type. Modified TABOPN to be more careful about rejecting tables with version=0. Moved nowhere. 7398. February 16, 1993 $APLOOP/INPUT.FOR Bill C. Added member function INTELL to obtain AIPS TELL adverbs and pass them as an INPUT object. Made OUTKEY in IN2OBJ of unknown length. Moved nowhere. 7399. February 16, 1993 $APLOOP/HISTORY.FOR Bill C. History routines force sync of memory resident version of CATBLK with the one on disk. This keeps tables which have been deleted from reappearing the the catalog header. OHCOPY now no longer copies the input history to the output history if either 1) the output history already exists or 2) the input history doesn't exist. If the history is not copied an time/task label is inserted in the output history. This keeps restarts with the same output file from copying the entire history each time. OHLIST now lists two dimensional arrays in the correct order (first dimension fastest). Moved nowhere. 7400. February 16, 1993 FRING Bill C./Ed Fomalont When fringe rates were being converted to sec/sec the reference frequency was being used rather than the actual IF frequency. Moved nowhere 7401. February 17, 1993 BLING Chris Now fits a two or three dimensional elliptical Gaussian model to the amplitudes in the neighbourhood of the peak amplitude in delay-rate space to refine the estimate of the coordinates of the peak. Also fixed a bug in the AP addressing in the code section that estimates the noise that produced bad (and sometimes negative) estimates. of the variance. This was previously masked by a test for small values but caused serious problems for the non-linear least-squares model fitting routines. Moved nowhere. 7402. February 17, 1993 ZLASIO Chris Changed the semantics of the close operation so that files are not automatically spooled to a printer if an explicit output filename is given. This should prevent superfluous plots being generated by laser printer tasks (CANPL, QMSPL and TVCPS). Moved nowhere. 7403. February 19, 1993 MAPCLR Bill C. MAPCLR now longer gives all those stupid messages when the file no longer exists. After years of annoyance I drove an axe thru the beast's head. Moved nowhere. 7404. February 19, 1993 CALIB Bill C. Now CALIB is allowed to pick the DFT or GRID method whichever is faster. This should make CALIB much faster for large CC tables. The VLAC and DDT test data are no longer on line in CV so I can't run the standard tests. Also sets FACGRD. 7405. February 19, 1993 $UPDUNIX/AIPSUPD, UPDCONTROL, and UPDERROR Pat Cleaned up some error reporting code, and inserted some where it was missing. Hopefully this will clear up the MNJ problems. Moved nowhere yet; may move to AOC hosts if it works. 7406. February 19, 1993 $YSERV/XAS.SHR Pat Fixed memory leak by making the closedown function delete any resources, _especially_ shared memory, allocated on startup. Also installed I/O error handler so that killing/quitting XAS will cause the closedown() function to be called. Moved (XAS binary) to IBM load area in CV, will move Sun version at midnight via an at-job. 7407. February 19, 1993 STARS Gustaaf STARS only plotted the horizontal part of star positions. Reason was that the Y-size defaulted to zero. Changed code in STCARD to fix this. Moved nowhere. 7408. February 19, 1993 BLING Chris Fixed problems found while processing a single-IF data set. The multiband delay search is now automatically disabled for single-IF data, the AP index calculations in PEAK2D have been fixed and the time selection criteria have been relaxed so that all scans that overlap TIMERANG are used. Also removed a vestigial mention of SOLINT from the HELP section of the help file. Moved nowhere. 7409. February 22, 1993 PASTE.FOR Bill C. Modified to complain and trim the BLC and/or TRC to make the subimage fit into the output image rather than dying. Moved nowhere. 7410. February 22, 1993 HGEOM Bill C. Modified HGEOM to do something sensible when the output image is smaller than the second input image. In this case the output image is centered on the input image rather than being the bottom left corner of the secondary input image. Also made corresponding changes in HGEOM.HLP. Moved nowhere. 7411. February 22, 1993 $QPGNOT/UVADC Bill C. Removed subroutine VSCDFT and moved it to $QNOT. Moved nowhere 7412. February 22, 1993 $APLOOP/ARRAY.FOR Bill C. ARRCLO now simply rteturns if the input object is closed rather than giving an error return. Moved nowhere. 7413. February 23, 1993 Sorting routines Bill C. Made a set of uv data sorting routines which use the "AP" memory to do the presorting of the data. If this presort does not include all the data then a disk based merge sort is done. Also a rotation of the u and can also be applied. The following routines were added to $QNOT: UVSORT - High level sort routine SRTIN - Read data, add sort keys and presort SRTOUT - Strip keys from scratch file and write output. The following were added to $QPSAP: QSORT - High level "Q" sort routine QOSORT - "Q" copy of $APLSUB/OSORT quicksort routine. QPERMA - "Q" copy of $APLSUB/PERMAT permutation routine QLSORT - "Q" copy of $APLSUB/LSORT for when all else fails. The last 3 routines are added to avoid potential problems with linkers and may not be really necessary. Moved nowhere 7414. February 23, 1993 BLING Chris Fixed a number of bugs and made some minor enhancements. - Scans that lie partly outside of TIMERANG are now modified so that the data processed by BLING lies wholly within the TIMERANG limits. - Corrected the times and durations written to the BS table in the case where BLING subdivides a scan to fit into the AP. - Fixed the weight calculation in MBFIT. - Added the missing weight calculation in SBFIT. Moved nowhere. 7415. February 23, 1993 BLAPP Chris Corrected the NRHS parameter to SGELSS in SOLVSB. This was causing bad single-band results under some circumstances. Also changed the calls to TBLSRT that sort the SN and CL tables to use the full 'ANTENNA NO.' string rather than just 'ANTENNA'; using the shorter string now fails for some reason (has a routine in TABLE.FOR changed?). Moved nowhere. 7416. February 24, 1993 POSSM Phil Fixed a couple of annoyances: (1) When NCOUNT = 1 and only one baseline was plotted, POSSM claimed that several baselines were averaged in the label. (2) The refreshing of the page when plotting many pages was still failing under obscure circumstances, think I have them all now. (3) Made sure when plotting scan averages of many sources, the source name appears on the plot label. Moved nowhere. 7417. February 25, 1993 TVFLG, SPFLG, IBLED Eric Changed all 3 (and their help files) to keep the all channel and all IF flags true at all times if the input data set has only one channel or one IF (resp.). In IBLED, they also stay true if all input channels are averaged together or all input IFs are averaged together. Moved nowhere. 7418. March 1, 1993 $AIPNOT/SETPAR Pat Increased limit of 2nd ap size from 32768 to 327680. Some (lucky) people have up to 4 Gbytes of physical RAM to play with. Moved nowhere. 7419. March 2, 1993 $APLNOT/FGSEL Gustaaf FGSEL, only called by UVCOP, failed to update the channel numbers in the flag table when BCHAN and ECHAN were specified explicitly. Code was added to do so. At the same time, a bug was fixed which potentially could lead to a zero second IF in the flag table, which would have flagged ALL IF's for that row. Moved nowhere. 7420. March 2, 1993 $APGNOT/UVCOP Gustaaf UVCOP did NOT update channel entries in the flag table when BCHAN .NE.1. Also, a potential bug when copying entries flagged by IF was fixed. Main code change was in $APLNOT/FGSEL (see above), but UVCOP needed some modifications, too. Moved nowhere. 7421. March 3, 1993 MK3IN Chris Increased the number of CL table entries allowed per scan from 40 to 500. This was necessary for data from the TDRSE SVLBI experiment that requires a phase correction entry for every integration. Moved nowhere. 7422. March 3, 1993 Division of compressed data by model Phil The division and subtraction of a model image from compressed data was not working and had obviously never been working. The problem was 2-fold, (1) the scratch files should never be compressed, unfortunately they were all operations were being performed on compressed data and so were incorrect; (2) various parameters such as the record length and number of random parameters are played with in this process and these changes were not being taken into account. The changes I made to fix this were designed to minimize the number of changes to the code in order not to screw things up. So a new include (DSCD.INC) was needed - it carries the parameters necessary to describe the true state of the scratch file, and a new routine PRMUPD to correct the DUVH.INC parameters after conversion from compressed -> uncompressed data. This was necessary because the CATBLK carried around reflects the header of the input data, not the scratch file. (1) new include DSCD.INC (2) new routine $APLNOT/PRMUPD (3) $QNOT/ALGSTB.FOR - trap changes to LREC, NRPARM (4) $QNOT/BUFGET.FOR - better check on data compression (5) $QNOT/BUFPUT.FOR - same (6) $APLSUB/UVCONJ.FOR - remove assumptions on position of variables within data arrays (7) $APLNOT/UVDOUT.FOR - trap changes to LREC, NRPARM - will compress output if input was compressed (8) $APLNOT/UVDPAD.FOR - trap changes to LREC, NRPARM - deal with compressed input data. (9) $QNOT/UVMSUB.FOR - trap changes to LREC, NRPARM (10) $QNOT/UVMDIV.FOR - same (11) $QPGNOT/UVSUB.FOR - minor changes to deal with LREC etc. Fixed another major bug, the output file from UVSUB could not be plotted with UVPLT because the filename in the output catalogue header was wrong - fixed this. (12) Minor changes to ASCAL, BLCAL, BPASS, CALIB, FRING, MX, PCAL, VSCAL to initialize variables in DSCD.INC. Moved nowhere. 7423. March 4, 1993 MSGKILL Eric Changed the pseudoverb MSGKILL to accept a numeric argument instead of just TRUE and FALSE. If MSGKILL is set to n, then all messages of priority 2 through n are totally suppressed (i.e., do not appear on the terminal or the file). If it is set to TRUE, then no messages go to the file, but all of them go to the terminal. Setting to FALSE restores normal operation. User beware! Changed $APLSUB/MSGWRT.FOR, $AIPSUB/PSEUDO.FOR, and MSGKILL.HLP. Moved nowhere. 7424. March 4, 1993 $QOOP/QUVUTIL Bill C. Added routine OUCDFT to add the DFT Fourier transform of a "W" corrected image model to a uv data set. Only point models are processed. Can also optionally make minor corrections for bandwidth smearing and correct the values of u,v, and w for errors in the reference frequency. Added routine OUVSRT to sort and optionally rotate uv data. Routines OUMSUB and OUMDIV were modified to use UVDTCO to copy descriptive tables and to copy the sort order to the output. Also put in Phil's changes for compressed data. Added a "fast" imaging option to OUVIMG using the standard AIPS routines IMGMEM et al. Moved nowhere. 7425. March 4, 1993 $APLOOP/UVUTIL.FOR Bill C. Added routine UVTAVG to time average time-ordered UV data. Added routine OUVCLP to copy data flagging corelations with an excessive visibility amplitude. Added routine UVDTCO to copy descriptive (AN, SU, FQ) tables. Modified UVCOPY and UVSCR to use UVDTCO rather than UVFCOP and UVTCOP. Routines UVCOPY and UV2SCR now copy uv descriptors. U2IDES no longer changes selected Stokes from ' ' to 'I' in the uvdata object. Moved nowhere 7426. March 4, 1993 QCLEAN.FOR CLEAN.FOR Bill C. Modified to allow round CLEAN boxes. If the first element of the box is -1 then the second element is the radius and the third and fourth elements are the x and y pixel of the center. Note: this doesn't affect MX or APCLN. Fixed problem in algorithm deciding where to stop. There was some confusion as to the meaning of NBMHIS; it was set to the dimension of BMHIS rather than the number of possible levels (i.e. the dimension of RESHIS). The input restoring beam was never being used. Now it is read and if zero then the fitted beam is used. Streamlined CLRHIS not to determine RESMAX before determining histogram unless the initial guess is in considerable error. Also added CLRMAX to determine abs max pixel value. Moved nowhere. 7427. March 4, 1993 $APLOOP/UVDATA.FOR Bill C. OUVCLN now initializes the new object's CALEDIT member. OUVATT now sets new UV_DESC member ISCOMP to indicate if underlying disk file is in compressed format. Moved nowhere. 7428. March 4, 1993 $APLOOP/UVDESC.FOR Bill C. Added new class member ISCOMP to indicate if the underlying disk file uses compressed format. Moved nowhere. 7429. March 4, 1993 $QNOT/UVTBUN Bill C. Implemented superuniform weighting by reducing the grid size by a factor of 1/(1+UNFBOX). This should allow using UVBOX>0 on TB sorted data. Moved nowhere. 7430. March 4, 1993 $QNOT/MAKMAP Bill C. Fixed incompatability with UVGRTB/UVGRID. These routines were scaling SCL*G(1) by the ratio of beam size to first field size and MAKMAP was setting SCL*G(1) the the appropriate value for the beam. Thus if the beam and first field size were different the beam was misgridded. I don't think there are any current applications that hit this bug. Moved nowhere. 7431. March 4, 1993 $APLNOT/REFFAZ Bill C. A bug in the test for bad solutions caused bad single polarization solutions to slip through. There were other bugs related to single polarization. Moved nowhere. 7432. March 4, 1993 $QNOT/IMGMEM Bill C. New package of routines to do an in memory, gridding, FFT and gridding corrections. On the first pass of the FFT rows containing all zeroes are not transformed. This package is much faster for cases where the entire grid and work arrays fit into the "AP" memory. The following routines are used: - GRDMEM - read uv data and grid. - FFTMEM - do 2-D FFT - GRCMEM - do gridding correction and write image to disk. Moved nowhere. 7433. March 4, 1993 $QPGNOT/SCMAP Bill C. New task to do multiple iterations of self calibration especially intended for VLBI use. On successful completion it produces a final, deconvolved image and a set of calibrated data. Also SCMAP.HLP. Moved nowhere. 7434. March 4, 1993 $APOLNOT/GRDFLT Bill C. Changes message level to 2 so that MSGKILL could suppress them. Moved nowhere. 7435. March 4, 1993 $YPGNOT/KNTR Gustaaf KNTR refused to label with velocities on VELO- type axes. It required FELO- type axes only. Now VELO- axes (radio defini- tion) are permitted, too. Moved nowhere. 7436. March 4, 1993 FNDPOL Bill C. Change advisory messages from level 8 to level 2 so that MSGKILL can get rid of them. Moved nowhere. 7437. March 4, 1993 TAPQL Chris TAPQL is a new task that allows AIPS table data (minus the keywords --- I haven't figured out what to do with them yet) to be exported to the POSTGRES database management system. This probably won't be needed by anyone else but has been checked into the system as a template task for anyone who wishes to export AIPS table data to a RDBMS. Moved nowhere. 7438. March 4, 1993 TAPQL Chris Found out that source tbales have full stops in one of the column titles (ID. NO.); this upsets POSTGRES. Added code to replace full stops with underscores in attribute names. Moved nowhere. 7439. March 5, 1993 POSSM Phil New version did not deal with subarrays when scan averages were asked for, now does. Also when plotting scan averages of BP entries a blank plot file was created for every scan in the NX table even though they did not correspond to BP entries. Now just runs until there is something to plot and then creates the plot table. Moved nowhere. 7440. March 5, 1993 BLING Chris Gave the user the option of increasing the search resolution. This works well for 2D searches but will cause the 3D search to fail if the narrow-band delay resolution is increased to a degree where the delay resolution is similar to the difference between narrow-band delays for different IFs. Moved nowhere. 7441. March 6, 1993 QUVUTIL Bill C. Routine OUNFWT declared and used variables JLOCF and JLOCIF which were also declared in the DUVH.INC include. The Sun compilers don't catch this but Convexes do. Changed variable names. Moved nowhere. 7442. March 6, 1993 QCLEAN Bill C. Modified trap for CLEAN gone bad to only work for IPOL as it gets upset at normal behavior of polarized CLEANs. Moved nowhere. 7443. March 6, 1993 $APLOOP/UVUTIL Bill C. In a number of routines the UV_DESC was being copied before the output was opened eliminating it's effect. These were: UVCOPY, UV2SCR, UVCLIP, UVVCLP and UVTAVG. Now calls UVDSCP rather than UVDCOP. Moved nowhere. 7444. March 6, 1993 $APLOOP/UVDESC.FOR Bill C. Made a version of UVDCOP (UVDSCP) which doesn't copy GCOUNT, NDIM, NAXIS, NRPARM, ISCOMP, ANAME, FNAME, FDISK and VISOFF members. This was causing confusion about how large a file is and which file it is. Also modified UVDCOP not to copy ANAME, FNAME, FDISK. Moved nowhere. 7445. March 6, 1993 $APLNOT/DGINIT.FOR Bill C. DGINIT was not finding Qpol data if that was all that was in the file. Likewise for Upol. Moved nowhere. 7446. March 6, 1993 $APLOOP/TABLEUTIL Bill C. Added routine OSUPAN to determine the parallactic angle of a source at an antenna at a given time. Moved nowhere. 7447. March 8, 1993 BLING Chris Fixed an indexing bug in the 3D search routine. This caused subtle errors that were not noticable until the FFT interpolation code was added. Also fixed the calculation of the frequency spacing on the IF axis for the case where IFs are not evenly spaced. Moved nowhere. 7448. March 9, 1993 MX,QCLEAN Bill C./Martin Shepard The loading of residuals had a couple of problems. First the beam patch was allowed to be sufficiently large that there wasn't room for 256 residuals which is the size that the residual loading routines want to process. Second, in the residual loading routine if the last buffer of residuals would not ALL fit in memory NONE were loaded. These problems could result in no residuals loaded and the CLEAN quitting under some circumstances. Moved nowhere. 7449. March 9, 1993 BLAPP Chris Fixed a design bug in BLAPP. Delays and rates are now forced to close by adding or subtracting the delay or rate ambiguities from the baseline-based quantities before performing the least- squares solution. This required adding the ambiguities to the BS table (changed PUVD in INC, TABBS.FOR and BSINI.FOR in APLNOT, TABLEUTIL in APLOOP and BLING in QPGOOP). Moved nowhere. 7450. March 10, 1993 CATDIR Eric Corrected it to check for empty slots on CSTA and return error code 6 (as INFO does) on empty. Moved nowhere. 7451. March 10, 1993 MAPCLR Bill C. Changed to deal appropriately with changing behavior of CATDIR. CATDIR error 6 now means the file didn't exist and 10 means it's been asked to clear a flag that wasn't set. Moved nowhere. 7452. March 10, 1993 UVLIN Gustaaf Put NPARMS to right number. Juan U. was concerned that too high a number was responsible for a reported problem. Moved nowhere. 7453. March 11, 1993 $SYSUNIX/OPTIMIZE.LIS Dave Changed DATGET and DATFLG to IBM OPT 0; Removed reference to PBASET and added BPASET IBM OPT 0; Also added UVLIN IBM OPT 1 as per Gustaaf's request. Moved nowhere. 7454. March 11, 1993 UVLIN Gustaaf COMLNK'ed UVLIN in view of change in OPTIMIZE.LIS which prevented UVLIN from running on the IBM's. Moved nowhere. 7455. March 11, 1993 BLING Chris BLING now filters the scan list so that it doesn't try to read data for sources that were not requested by the user. This speeds up processing and gets rid of some ugly-looking error messages. It also appears to cure a problem where BLING would occasionally abort prematurely after examining a large number of scans without finding any selected data. Moved nowhere. 7456. March 11, 1993 BLING Chris Entries in the BS table now carry times determined from the beginning and end of the actual data rather than the midpoint of the scan. Also corrected to output units of the residual phase to agree with the units field in the BS table and widened the time selection window by 1/10th of a second at each end of a scan to avoid data being lost at the ends of a scan due to accumulated floating-point rounding errors. Moved nowhere. 7457. March 12, 1993 IMAGE Bill C. IMGCLN now uses IMGATT to create the new file and attach to it. In the old scheme the file name etc. could be left pointing at the original file. IMGOPN no longer marks an object as invalid when it opens it for write. This was running afoul of the dastardly CATIO. If an object is to be updated (read AND written) it has to be opened 'WRIT' first. The image class only sets the catalog status on the first open of an AIPS file. CATIO refuses to update the catalog header (required on close for write) if the catalog status is "READ". If it weren't for CATIO I wouldn't have any problems (he said deliriously). Modified a number of member functions to open the output object first in case it was also an input object. Also added better error messages to these routines. Member functions IMGSUB and IMGDIV were missing; added them. Moved nowhere. 7458. March 12, 1993 $APLOOP/ARRAY.FOR Bill C. Added missing routines ARRSUB and ARRDIV to subtract and divide array objects. ARRCLO now saves the CATBLK if it modified it so IMGCLO will not do something bad. ARRCWI now sets the first TRC element to zero to cover the case of complex data. Improved error messages in ARRIO on attempt to access data after file exhausted. Moved nowhere 7459. March 12, 1993 $QOOP/APUTIL Bill C. Reset access to row in AP2IMG and IMG2AP. Moved nowhere. 7460. March 12, 1993 $APLOOP/CXUTIL.FOR Bill C. New package of utility routines related to the complex image (CX_IMAGE) class. These mostly convert between CX_IMAGE objects (the only type of complex image allowed permanently in AIPS) and IMAGE objects of type complex (the useful ones for image arithmetic). IMG2CX - Convert image of type complex to a CX_IMAGE. CX2IMG - Convert CX_IMAGE to image of type complex. IMGR2C - Convert real image to image of type complex with zero imaginary. Moved nowhere 7461. March 12, 1993 $QOOP/QINTER.FOR Bill C. New package of image interpolation routines. The top level is IMGIMT and it interpolates one image to another. This routine only does flat geometries and positions are relative to the image reference pixel. Unlike the *GEOM tasks this routine will interpolate over blanked pixels and can fill in small blanked regions. At the edges of the image being interpolated the interpolation kernal is adjusted to reduce edge effects. Public functions: IMGINT - Interpolate with shift and rotate. Private functions: IMINST - Set up for interpolation, read input plane. IMINTR - Returns interpolation value at (x,y) ININFN - Return interpolation function Moved nowhere 7462. March 12, 1993 $YPGNOT/VBPLT.FOR Phil When plotting time-averaged phases the plots were labelled in radians. Fixed. Moved nowhere. 7463. March 12, 1993 BLAPP Chris Made a number of changes in the handling of residual phases. - Phases in the BS table are now recognized as being in degrees. - Phases are forced to close before solving for antenna-based phases (I missed this when forcing delays and rates to close). - Phases for the same scan are rotated to a common reference time using the baseline-based residual rates. The following changes were also made. - Rates are now converted to sec/sec using the reference frequency for each IF rather than a single frequency for all IFs. - The value of the SOLINT adverb does not get lost in cyberspace any more. Moved nowhere. 7464. March 12, 1993 BLAPP Chris Fixed a whole slew of stupid mistakes that I made in the previous change. Moved nowhere. 7465. March 12, 1993 ZDCHIN Eric Changed $INC/DZCH.INC to add a pointer to a non-file "file descriptor" for Pat to use with the message window. Changed $APLGEN/ZDCHIN to init this and pass it into ZDCHIC. Changed $APLCRI and $APL3090 versions of ZDCHIN to match $APLGEN, including adding the network-wise code from Mark. Changed $APLUNIX/ZDCHIC.C and $APLGEN/ZDCHIC.FOR (stub) for new call sequence and, in the former, to set set both of the non-file parameters into external variables for C use. Changed $APLIBM and $APLCRI to match $APLUNIX - neither had been fixed the last time the call sequence was changed! Removed the $APLDEC version which was exactly the $APLUNIX version. Moved nowhere. 7466. March 12, 1993 ZACTV8 Eric Message levels are not some game where you can choose anything you want. Informative messages are 2 through 5 with 5 only for the real answers. Errors and warnings start at level 6 with warnings, 7 for minor errors (suppresible), 8 for serious errors and higher for the building burning down. Changed level 5 error messages to levels 6 and 7 in $APLUNIX/ZACTV8.FOR. Moved nowhere. 7467. March 15, 1993 GAL Gustaaf After nine years, a completely new version of the velocity field fitting task GAL is now added to AIPS. Improvements include: 1) fewer restrictions on image size, 2) a weight map (IN2NAME) can be specified, 3) a wider choice of residual maps is offered, 4) plot option supports DOTV. Moved nowhere. 7468. March 16, 1993 TABED Bill C. Output tables were not being marked as unsorted. Moved nowhere. 7469. March 16, 1993 UVFIT Phil Added the option to fit to the real part of the data only, useful if have calibrated and want to avoid the noise bias problem inherent in amplitude only fits. Also updated UVFIT.HLP Moved nowhere. 7470. March 16, 1993 MK3IN Chris Added source selection using the SOURCES adverb. Also added a capability for defining substitute names for antenna names in the input data as part of the experiment file. This also allows multiple antenna names to be mapped to a single antenna. Moved nowhere. 7471. March 17, 1993 MX Bill C. Fixed a bug in computing beam histogram; sidelobes exterior to the maximum beam patch were being ignored. This probably only affects VLA widefield snapshots. In the DDT large MX clean test this change caused zero differences. Moved nowhere. 7472. March 17, 1993 BLING Chris Changed the order of a calculation in MRQCOF to reduce the chances of an arithmetic overflow. Moved nowhere. 7473. March 18, 1993 UVUTIL.FOR Bill C. Corrected message level from 6 to 4 in SLFREF for an informative message. Moved nowhere. 7474. March 18, 1993 VLAPROCS, VLACALIB, VLACLCAL Dave A. Added QUAL adverb to $HLPFIL/VLACALIB.HLP; added SOUCODE adverb to $VLACLCAL.HLP; took QUAL default setting out of $RUNSYS/VLAPROCS.001. Moved nowhere. 7475. March 18, 1993 $SYSUNIX/FDEFAULT.SH Pat Changed default options for IBM compiler to include "-u" which disables implicit typing (i.e. turn on "implicit none). Moved nowhere. 7476. March 18, 1993 $APGNOT/SETJY.FOR Glen Since 3C295 is actually the VLA flux density calibrator, and I needed to calculate the flux density at several frequencies, the 3C295 Baars etal coefficients have been added. Moved nowhere. 7477. March 19, 1993 BLING Chris I have noticed that the fit to a 3-D peak often fails due to a loss of precision due to gradual underflow. The peak fitting procedure has been modified to work in double precision to reduce the chances of this happening. Also added a test on the value of lambda (which controls the amount by which the fitted parameters are changed) to the convergence criteria since the result of underflow is that lambda become very large while chi-squared does not change significantly; without the test on lambda BLING was accepting parameters from fits that were diverging. Moved nowhere. 7478. March 22, 1993 BPASET Phil Added weight arrays to the code for DOBAND = 1. This is useful in the case of half a BP entry being flagged and being averaged with an entry which is not flagged. This was never dealt with before but now is. Moved nowhere. 7479. March 22, 1993 BLING Chris After some experimentation I found that the round-off instability in fitting a 3D Gaussian model to the peak could be avoided by permuting entries in the curvature matrix. Also added very small values of the lambda parameter as a disjunct criterion for convergence and switched to the LAPACK expert driver for solving linear equations (DGESVX) which allows the augmented curvature matrix to be equilibrated and gives some additional information useful for debugging. Moved nowhere. 7480. March 22, 1993 COMB Dave A. COMB was blanking all pixels with Q**2+U**2 < 1E-09. This apparently was throwing away too much good data. Reduced the above restriction to 1E-15. Moved nowhere. 7481. March 23, 1993 MK3IN Phil/Walter Alef 3 minor fixes: (1) In its concatanation mode MK3IN would expand the file but then not shrink it down at the end, fixed this. (2) Also in its concatanation mode it was possible for the AN table to be screwed up, trapped this. (3) There was a small bug in deciding the time interval between CL entries, fixed. Moved nowhere. 7482. March 23, 1993 SETJY Glen Still trying to sort out a proper amplitude calibration. When calculating the error due to the Baars etal cooefficients, I discovered that the flux densities estimated are Very Large. Added error baars to SETJY output. Moved nowhere. 7483. March 24, 1993 CALIB.FOR, .HLP Bill C. Added adverbs CMETHOD and CMODEL to control the model computation step. Moved nowhere. 7484. March 24, 1993 POSSM.FOR, .HLP Phil Some minor changes: (1) NCOUNT = 1 now means 1 plot/page and multiple pages; NCOUNT = 0 will average data to produce a single plot. (2) Made the y-axis labels flotaing point and accurate again, but as a result made the maxium NCOUNT = 9. (3) removed the restriction one plotting single point spectra but only in the case of plotting multiple IF's Updated the .HLP file to reflect these changes. Moved nowhere 7485. March 24, 1993 DGINIT Bill C. If STOKES was 'V' and the data set had only Vpol this fact escaped DGINIT and it pronounced the data defective and died. Moved nowhere. 7486. March24, 1993 tapes Eric Made a bunch of changes in an attempt to get tapes to work correctly, especially with end-of-tape and the like. Changed: AU4 $AIPSUB: changed BAKF operation on TPHEAD error from a TAPIO call to a ZTAPE call with 1 file (or 2 on an EOF); also changed FNDEOT call to ZTAPE AEOI call and changed to allow super-high density 22500. AVTP $APLPGM: changed FNDEOT call to ZTAPE AEOI. FITTP $APGNOT: changed FNDEOT call to ZTAPE AEOI. PRTTP $APLPGM: changed error handling on header read to ZTAPE BAKF call with 2 files for EOF and 1 for all other errors. TAPIO $APLSUB: added tests on number of bytes written. TPHEAD $APLSUB: added test for end-of-medium and reworded the error messages. ZMOUN2 $APLSUN: added code to open the tape device to determine its type and then to use the density input to alter the device name appropriately by tape type; 22500 => 5 Gbyte Exabyte. Also changed to return message about the type. Changed to force _POSIX_SOURCE and do file locking with fcntl rather than flock. ZMOUNT $APLGEN: changed to handle messages on success, adding the host information, and to avoid changing tape names in common until the mount is successful (it was wiping the common when a user tried the same mount twice, making the dismount not work). Changed call sequence to pass back the message and system error code for TPMON. ZSTAIP $APLGEN: changed call sequence to ZMOUNT. ZTAP2 $APLSUN: added BOT, EOT, EOF error codes from Xylogics and scsi system include files; added BEGW and AEOI opcodes; added status checking to ADVR (BAKR could use it too); BEGW and WEOF have full status checking and write only 1 EOF except for Xylogics (1/2inch reel) for which we write 2 and back up over all EOFs and then forward if we weren't at BOT to begin with; dropped MEOF opcode; added AEOI to EOM opcode and added ZMSGWR call to tell where the tape was positioned after AEOI. ZTAPE $APLGEN: dropped MEOF and added AEOI and removed all references to MOUNT in precursor comments; moved WEOF, BEGW, and AEOI operations into ZTAP2. ZTPMI2 $APLUNIX: changed to leave ierr = 0 ZTPWA2 $APLSUN: new version does Sun-specific status checking needed to tell the difference between EOM and EOF. ZTAPE $APLVMS: copy of old $APLGEN/ZTAPE with code to do the AEOI in the way FNDEOT did it (but without recursion) including a message about what it did. ZTAP2 $APLVMS: updated precursor comments only. ZTAP2 $APLGEN: updated precursor comments only. ZTPOP2 $APLIBM: new version, gets tape type from system. ZMOUN2 $APLIBM: gets tape type from lsdev command, sends message back to ZMOUNT about the type, makes the actual device assignment depend on density (reel tapes only for 6250). ZTAP2 $APLIBM: added call to force blocksize to 0 for Exabytes; changed WEOF to do 1 or 2 EOFs with appropriate tape positioning; added BEGW to the WEOF part and a simple-minded AEOI (no IBM opcode for this is available! so I did FNDEOT in C inside ZTAP2) including an appropriate message. ZTAP2 has to generate this message since on Suns it is an absolute file number and on others it is a file number relative to the starting position. MOUNT $HLPFIL: changed comments and limit on DENSITY. DENSITY $HLPFIL: changed remarks about accepted values and how they are used. TPMON $AIPNOT: changed call sequence of ZMOUNT to pass back the message and system error correctly. GRITP $AIPNOT: fixed ZMOUNT call sequence. GR2TEX $AIPNOT: fixed ZMOUNT call sequence. ZMOUN2 $APLCVEX: added test for density 22500 (= 6250 here) and improved error and good return messages. ZTAP2 $APLCVEX: add code to do full WEOF and BEGW and AEOI, add further status checking. ZTPWA2 $APLCVEX: add status checking on error. ZMOUN2 $APLVMS: changed to return message on good return. ZMOUNT $APLVMS: updated call sequence and handling of returned strings. Moved nowhere. 7487. March 26, 1993 tapes Eric More changes: ZMOUNT $APLGEN: was not unsetting the logical if the mount fails. We can't unset this on an attempt to mount twice; so added test on tape name in common before trying the mount as well as code to unset on error. ZMOUN2 $APLCVEX: added code to test tape type and report. ZTAP2 $APLCVEX: corrected code for AEOI and others. Dropped test on fileno = recno = 0 since they can get off. Tried this extensively on yucca. Moved nowhere. 7488. March 29, 1993 BPCOR Phil Ensured that the diameter of and the support width of the smoothing functions had reasonable defaults. The task failed if these were left at zero. Moved nowhere. 7489. March 29-30, 1993 tapes Eric Changed the $APLIBM versions of ZMOUN2.C and ZTPOP2.C to give the full path of the function used to find the tape type. Changed ZMOUNR to check logical DISMOUNTED rather than the obsolete UNMOUNTED. Changed $APLGEN ZMOUNT to make SYSERR equivalent to IERR on MOUNTS (ZMOUNR can return IERR = 0 meaning no communication error, but some real SYSERR), and to reinitialize all common parms and the logical whether or not dismount has an error. These should make mounting a little less ``sensitive.'' Moved nowhere. 7490. March 30, 1993 LWPLA.FOR/LWPLA.HLP Glen/Fred Fred suggested I add the slide maker 4x5in mode to LWPLA options. Moved nowhere. 7491. March 30, 1993 $SYSUNIX/STARTPMON Pat Made it a bit more bullet proof and made the messages more informative. It now tells you (a) if the TPMONx files in $LOAD are out of date, and (b) if it can't do something because of ownership differences (process, file). To be moved to the AOC in a few minutes time. 7492. March 30, 1993 $APLCVEX/ZTT* Pat Removed ZTTYIO.FOR and ZTTOP2.C. There are no message terminals used anywhere in NRAO (or ATNF according to Eric) and the MSGSERVER works fine with the generic $APLUNIX versions to be put back momentarily. (re)moved nowhere else. 7493. March 30, 1993 BLING Chris Switched to an explicit calculation of the variance in estimating the noise rather than an approximation. The approximation produced negative variances on rare occasions. Also removed the artificial SNR degradation prior to the 3D peak-fitting; this is no longer necessary after improving the stability of the fit. Moved nowhere. 7494. March 30, 1993 MSGSERVER Pat Finally got MSGSERVER and its bits and pieces finished. In the process I got rid of the convex (see above) and IBM versions of some routines that were unnecessary as the new generic versions work fine. Also replaced XASERVERS, TVDEVS.SH and START_AIPS. The last two need to be moved to $AIPS_ROOT at the three AOC sites as soon as the MNJ is done!!! Changed in $APLUNIX: ZTTOPN.FOR and ZTTYIO.FOR. New routines in $APLUNIX: ZTTOP2.C and ZTTYI2.C. Tested on Sparc, IBM RS/6K, Convex. Moved nowhere. 7495. March 31, 1993 $QOOP/QCLEAN.FOR Bill C. CLBHIS had a maximum beam patch size with an incorrect, hardcoded upper limit. Also fixed bugs in computing beam histogram. Sidelobes exterior to the maximum beam patch were being ignored. CLCCRM was not initializing variable JCOUNT; multiple calls caused erratic behavior. Changed all calls to OUMSUB to UVSUBM to allow more generic modeling routines. Modified the "AP" memory usage. Calls to QCLNSZ are made to see the amount of "AP" memory rather than using only KAPWRD. The beam is allowed to use any of this memory necessary but there is a limit of 20,000 residuals. This allows using much larger dirty beams; on suns a full 1024 beam is possible. Moved nowhere. 7496. March 31, 1993 QCLNSZ.FOR Bill C. New routine to tell the size of the "AP" for the Clark CLEAN algorithm. A generic version in $QPSAP gives the traditional size to allow for vector implementations. Another version put into $QSUN and $QIBM tell the full memory available. 7497. March 31, 1993 CLEAN,QCLEAN.FOR Bill C. Added option to suppress making a new beam when starting a CLEAN. This should be used only if an appropriate beam has been made since the CLEAN object was opened. Moved nowhere. 7498. March 31, 1993 $QOOP/QUVUTIL.FOR Bill C. Modified OUNFWT to use UVTBUW rather than UVTBUN for doing the uniform weighting. Added routine OUFDFT to do first order bandwidth smearing corrections in adding a model to a uv data set. Added generic routines: UVSUBM - subtract the FT of a model from a uv data UVDIVM - divide the FT of a model into a uvdata set. The routines allow more general modeling routines and in particular support DFT imaging and imaging with frequency dependent primary beam corrections. Moved nowhere. 7499. March 31, 1993 $QOOP/PBUTIL.FOR Bill C. New utility module for uv model calculations including a frequency dependent relative primary beam correction. The model of the antenna pattern is that of a uniformly illuminated disk which appears to be a good approximation for the VLA antennas. Public functions are: PBFSUB - Subtract FT of a model with corrections from uvdata PBFDIV - Divide FT of a model with corrections into uvdata Private functions: PBFCCT - Make temporary, frequency scaled CC table object PBFSCI - Make temporary, frequency scaled image object PBFACT - Returns relative primary beam correction. Moved nowhere. 7500. March 31, 1993 $QNOT/UVTBUW.FOR Bill C. This is a new routine to do uniform weighting of arbitrarily sorted data. This is a "more correct" uniform weighting as it uses the sum of the weights rather than the counts in a uv cell to determine the correction. Moved nowhere 7501. March 31, 1993 $QPSAP/QWTCNT.FOR Bill C. New routine. Like QHIST but is given a vector of elements rather than incrementing by 1.0. Used by UVTBUW. Moved nowhere. 7502. March 31, 1993 $QNOT/VSFDFT.FOR Bill C. New routine to add the DFT of a model with first order corrections for bandwidth smearing and frequency error corrections to a uv data set. Moved nowhere. 7503. March 31, 1993 $APLOOP/UVUTIL.FOR Bill C. NXTAVG had an error in handling UVRANG; it was not being converted into wavelengths. Moved nowhere. 7504. March 31, 1993 SCMAP.FOR, .HLP Bill C. Since the first CLEAN is likely to be deeper than following CLEANs the RMS of the first CLEAN is now prohibited from being the best for purposed of determining convergence. Changed the convergence criteria to RMS/(sum of CC fluxes) to deal with the case of very bad initial calibration. Also now expliticly state when the solution is diverging and force abnormal termination. Moved nowhere. 7505. March 31, 1993 $APLUNIX/ZTTOPN.FOR Pat If the MSGSERVER dies, one wants messages to revert to the main terminal or xterm. Modified this routine to check status of the ZTTOP2 call and if it reports problems on the socket (connecting, for example), it SILENTLY reverts to normal use. Moved to the AOC today. 7506. March 31, 1993 AIPSB Eric Changed a hard-coded limit of 51 on the job number and added better error messages. Job numbers can now go up to 64 (since the code overhaul). Moved nowhere. 7507. March 31, 1993 $APLUNIX/ZTTOPN, ZTTOP2, ZTTCLS, ZTTCL2 Pat Fixed bug that was causing zillions of open file descriptors not to be closed and an eventual collapse of the MSGWRT process. Moved to the AOC Sun cluster just now. 7508. April 1, 1993 BLING Chris Fixed several instances where xxxGET routines from the OOP package were called with parameter TYPE arguments. TYPE is an output argument. Apparently we have no machines other than the Convex that catch this error. Moved nowhere. 7509. April 1, 1993 BLAPP Chris Eradicated more calls to xxxGET routines with constants in the wrong places. Moved nowhere. 7510. April 2, 1993 BLAPP Chris Found that a counter was not being reset after each scan leading to possibly erroneous results and fixed it. Also found that the column index was not being cleared for multiband delay solutions and fixed it. Moved nowhere. 7511. April 2, 1993 BLAPP Chris Made CL table handling safer. BLAPP now has GAINVER and GAINUSE parameters and writing to CL version 1 is no longer permitted. Also added code to delete the scratch SN table if BLAPP dies unnaturally. Moved nowhere. 7512. April 5, 1993 XAS Chris XAS was positioning the icon off screen when no icon position was specified on the command line or in the resource database. XAS no longer sets the IconPosition IconPositionHint flag unless the user explicitly sets an initial icon position. This is makes XAS much more sociable as it allows the window manager to decide what the default icon location is. Also made the X11R4 interfaces the default rather than X11R3 (no one ships R3 any more). This required a slew of bug fixes (it is apparent that the X11R4 code had not been tested at all). - Added a missing declaration for old_handler in SetupWindow(). - Made sure that the PPosition and PSize flags are not set outside the X11R3 #ifdef branches in SetupWindows() since the corresponding fields of the size hints structure are not set. - Corrected calls to XStringListToTextProperty(): text property structures are passed by reference. Also. - Declared the buffer correctly for XGetErrorText() in XAS_IOhandler(); it was previously a pointer with no allocated storage. - Made sure that XGetErrorText() is not called if the server connection is terminated; the XErrorEvent field is not filled in in this case and core will be dumped if it is dereferenced to get the error number for XGetErrorText(). - Corrected comments about -I and -L flags for Sun OS in Makefile. - Removed (bad version of) AIPSSHAR from XAS.SHR. Moved nowhere. 7513. April 5, 1993 BLAPP Chris Fixed a bug in setting up the simultaneous equations that caused the reference antenna to be disconnected from the rest of the array. Moved nowhere. 7514. April 6, 1993 TVRGB TVHUI Eric Changed TVRGB and TVHUI help files to explain use of 2nd channels. Changed TVHUI to enter 2nd image in image catalog. Moved nowhere. 7515. April 6, 1993 TVCPS Eric Corrected addressing into the LUTs and OFMs. Fortran arrays begin at 1, but the data values begin at zero. Changed it to control the gamma correction, allowing the user to set gamma correction values for all channels. The previous default was to use no gamma correction and hence to produce a rather faint picture or one in only the 3 primary colors. Also changed the help file of course. Moved nowhere. 7516. April 6, 1993 TVPHLAME Eric Changed POPSDAT.HLP and AU6.FOR and created help file to make new verb TVPHLAME - yet another pseudo-coloring scheme. Moved nowhere. 7517. April 6, 1993 VLBA.001/VLBA.HLP Phil New procedure placed into the $RUNSYS area to ease the examination of VLBA correlator data. It will automatically read the VLBA FITS file, generate a CL table, plot the total power data, fringe-fit, plot the delay/rate solutions, apply the delay/rate solutions and plot the calibrated visibility amplitudes and phases. Moved nowhere. 7518. April 6, 1993 BLAPP Chris BLAPP now writes the frequency ID to the temporary SN table correctly. SNPLT, CLCAL etc. will now handle the SN table when it is preserved. Moved nowhere. 7519. April 6, 1993 CALIB/VISDFT Glen Added an option to allow CALIB to scale the clean components file when using a clean model with a multisource dataset. This allows clean models to be used to calibrate the antenna gains. Moved nowhere. 7520. April 6, 1993 COMAP.001 Glen Minor errors cleaned up. Moved nowhere. 7521. April 6, 1993 BLAPP Chris BLAPP now writes blanks the the wide-band delay column of the SN table rather than zeros if the wide-band delay is not solved for. Previously, SN2CL was applying rate corrections to the zero wide-band delay producing non-zero multiband delays in the output CL table. Moved nowhere. 7522. April 6, 1993 BPASET Phil The change I put in on March 22 screwed up for 2IF data, it did not initialize the internal weight array on the 2nd IF when further antennas were read in from the BP table. Fixed. Moved nowhere. 7523. April 6, 1993 XASERVERS in SYSUNIX Chris Modified XASERVERS to allow users to choose which terminal emulators to use (if any) for TEKSERVER and MSGSERVER by setting the variables AIPS_TEK_EMULATOR and AIPS_MSG_EMULATOR. The default is to use xterm for the Tek window and (by popular request from AOC users) to disable the message server. This change also allows XASERVERS to run on machines that do not have xterm installed. Also removed "hard-wired" resources to prevent annoying icon placement under olwm and mwm (if the mwm icon manager wasn't turned on the AIPS icons had a nasty habit of appearing underneath open windows) and to allow the user more freedom to customize the windows. Moved nowhere. 7524. April 7, 1993 UVGET Phil There was a rounding problem in selecting the UVRANGE of the data. Normally it doesn't matter but Craig had some cross-correlated data with u,v,w=0 (looking at phase-cals) and AIPS refused to select this. So I changed the selection criterion by -1.0e-10, and it now works. Moved nowhere. 7525. April 7, 1993 Server documentation Chris Added help files giving information about the Tektronix server (TEKSERVER.HLP) and the message server (MSGSERVER.HLP). These contain instructions for selecting terminal types and setting selected resources. Also added some information about the servers to the AIPS man page (AIPS.L in SYSUNIX). Moved nowhere: someone should install the new man page in Charlottesville. 7526. April 8, 1993 $APLOOP/ARRAYDESC,IMAGEDESC Bill C. Added new member function to copy descriptors (ARDCOP, IMDCOP) and to copy descriptors not related to size (ARDSCP). Moved nowhere. 7527. April 8, 1993 $APLOOP/UVUTIL.FOR Bill C. UVREFQ was creating a temporary table object and never deleting it. Moved nowhere. 7528. April 8, 1993 $QOOP/QCLEAN.FOR Bill C. The default was changed from 200 interations to 0. Fixed several problems related to image position shifts. Added a new routine CLNIIF to check and obtain image information after the initial dirty/residual images are made. Moved nowhere 7529. April 8, 1993 $QNOT/MAKMAP,IMGMEM Bill C. Added calls to IMHCOR to force agreement of the catalog header with what was done to make the image. This should fix some problems with phase position shifts in the OOP imaging routines. IMGMEM now sets NGRDAT to true. Moved nowhere. 7530. April 8, 1993 $APLNOT/IMHCOR Bill C. New routine to force agreement between the image header and the imaging parameters in the various commons. Moved nowhere. 7531. April 8, 1993 XASERVERS in SYSUNIX Chris Forgot that X programs may not be in the search path if XASERVERS is run from an rsh session :-(. Likely places for X clients to hide are now appended to PATH in a less over-engineered scheme than before. Moved nowhere. 7532. April 8, 1993 VLBA.001/HLP Phil Made the default fringe solution interval 15 minutes. Added an option to switch off the lsq fit in fring, for when the correlator group just want a quick look. Moved nowhere. 7533. April 13, 1993 $APLOOP/HISTORY.FOR Bill C. OHCOPY was not correctly resetting MSGSUP. Moved nowhere. 7534. April 13, 1993 $QOOP/PBUTIL.FOR Bill C. PBFCCT was not deleting temporary input table. Moved nowhere. 7535. April 13, 1993 $APLOOP/TABLEUTIL.FOR Bill C. OCHNCO was making unnecessary copies of the input table objects. Moved nowhere. 7536. April 13, 1993 $QOOP/QCLEAN.FOR Bill C. CLACLN modified so that it requires the field flux to be dropping more and the minimum absolute flux to have increased more before giving up on the CLEAN. At end of CLACLN the minimum flux is now set to the minimum of this pass. Also, CLACLN no longer assumes asynchronous "AP" operation and this eliminated finding an extra 32 components. Moved nowhere. 7537. April 13, 1993 $APLOOP/IMAGE.FOR Bill C. IMGATT now checks for compatability of requested file with existing file if an existing file was found and IMGATT was called with DOCREA = true. Added routine IMCOPY to copy one image to another. Added routine IMCLNX to copy descriptors from one object to another if the output object is not fully instantiated (the AIPS file exist). Added calls to IMCLNX to most image operations (add, multiply etc.) These routines can now be provided with output objects with only the names set. Moved nowhere. 7538. April 13, 1993 $SYSLOCAL/NOSHARE.LIS Phil Added SCMAP and IM2UV to this list, needed for the AOC. Moved nowhere. 7539. April 13, 1993 VLBA.001 Phil Added the options to not read a FITS file, i.e. if want to restart the procedure; and to scale the time range, amplitude and phase ranges plotted by VBPLT. Moved nowhere. 7540. April 15, 1993 $APGNOT/FILLM Gustaaf Made a small change to FILLM which allows now source names up to 12 characters instead of the previous 8. This was prompted by the survey observations where sourcenames of length 11 are customary. Moved nowhere. 7541. April 15, 1993 $APGNOT/ANCAL Phil Major changes to ANCAL. ANCAL had trouble dealing with MkIII data, the TSYS columns had to be in the same freq. order as the data, it could not deal with the case when there was only one TSYS value for all IF's, it needed 2 passes to calibrate dual polarization data etc. Added three new keywords in the text file: INDEX enables the user to specify the polzn and/or freq order of the TSYS columns that follow; ALLCHAN deals with the case where an antennas has only one value of TSYS for all IF's BOTHSIDE deals with the case where an antenna has TSYS measurements for all IF's but only one sideband of the pair. Also made the DPFU and OFFSET parameters arrays of size 2 so that the user can specify a value for each polzn (R/L) and so calibrate dual polzn data in one pass. Made extensive changes to the explain file, to clean the whole thing up. Moved nowhere. 7542. April 15, 1993 $YPGNOT/SNPLT Phil In its PLIF mode, SNPLT was not dealing with LHC polzn data when 2 polzns were present due to an indexing error. Fixed. I also got fed up having to type OPTYPE='AMP ', instead of OPTYPE = 'AMPL' every time, fixed ot so that only the first 3 characters are compared. Moved nowhere. 7543. April 15, 1993 $APGNOT/FITLD Phil Many changes to reflect the debugging of the VLBA - AIPS interface. (1) We were breaking FITS rules in using the OBJECT keyword to define the VLBA FITS format. Now if NAXIS = 0 and EXTEND = T the task expects all the information to lie in binary tables and will invoke the VLBA tables. (2) Added a secret print level for debugging, if OPTYPE = 'DE', then debug printing is turned on. (3) Increased the unit of file size created and added to to be 10000 visibilities. (4) FITLD would fail if trying to concatanate and some files were in write status, fixed so it will only fail if possible targets for concatanation or in the wrong status. (5) Will reverse the baseline direction if Ant2 > Ant1 (6) Will normalize the VLBA data by the weight (7) Has the ability to shuffle the data in freq order so that it sits in AIPS in increasing freq. order. (8) Made the error message more informative if current VLBA quantum header disagrees with the current AIPS file header. The task is not perfect yet, and many more changes will be necessary, however it can read the current state of the VLBA correlator's output. Moved nowhere. 7544. April 16, 1993 $QOOP/QUVUTIL Bill C. UVSUBM was incorrectly handling the default for MODFAC effectively setting it to zero. Several related bugs fixed. In ODFT the center pixel was changed to (nx/2, ny/2+1) to agree with FFT conventions. Also modified to tell the max and mins. Moved nowhere. 7545. April 16, 1993 CLEAN.FOR,QCLEAN.FOR Bill C. Added MAXPATCH as a parameter that can be set like MINPATCH. Moved nowhere. 7546. April 16, 1993 $QNOT/IMGMEM Bill C. If the output image channel is 1 or there is only one frequency channel in the image the max and min are reinitialized. Single channel residual images were not having the max and min properly set. Moved nowhere. 7547. April 16, 1993 $QPGOOP/WFCLN.FOR Bill C. New task incorporating much of the development for widefield and wideband imaging techniques developed for the VLA D array survey in an MX like task. It pretends to do up to 16 fields but with the current object manager it hits the limit of 20 simultaneous objects before it can do 16 fields. The exact limit depends on the options. Amoung the features of this task are: 1) Frequency dependent primary beam corrections in the CLEAN subtractions from the visibility data. A model for the primary beam is used to correct the CLEAN components as a function of frequency before subtracting them from the visibility data. The correction is made to the average frequency and does not remove the overall primary beam. 2) Simple correction for source spectral index. The visibility amplitudes are scaled to the average frequency using a single spectal index. MX does this implicitly with a spectral index of 0.0. 3) Correct errors in assumed center frequency. The u, v and w terms are scaled. 4) Uses double size beam if possible for CLEANs. MINPATCH of up to 500 is possible. 5) Uniform weighting is by the sum of the weights in the box rather than just to counts in the box. 6) Labels images with average zenith and parallactic angles so that snapshot VLA images can be corrected with OHGEO for misaligned coplanar array geometric distortion. 7) Supports 3-D DFT imaging to avoid noncoplanar array image distortion. (THIS IS REALLY SLOW!) Moved nowhere. 7548. April 16, 1993 $APGNOT/TABED Bill C. TABED was hanging up opening the same table twice when writing table header keywords when INVER = -1; i.e. loop over all tables of that type. In TEDIN SAMOUT is now set true for this case. Moved nowhere. 7549. April 16, 1993 ZCRLOG Eric $APLBERK/ZCRLOG.C was copying the input environment into its own array, but neglecting to set the end in its own array to NULL on that first copy. If core happened to have no zeros in the rest of that array, an abort could/should occur. Raised the size of the array (1000 environment variables limit), added test for going over that in creating new ones, and added that important initial NULL. Moved nowhere. 7550. April 16, 1993 $APGNOT/SDTUV Bill C. SDTTAB had an incorrect test to determine how many entries were in the baseline table. This was causing zero divides. Moved nowhere. 7551. April 17, 1993 $RUNSYS/VLBA.001 Phil Default outname is now the VLBA job number, i.e. the portion of INFILE after the : and before a . Makes it easier for the correlator group to know what data came from what correlator run. Moved nowhere. 7552. April 17, 1993 Z routines Eric In reviewing things, I have made minor changes. Deleted all ZXAS* subroutines since XAS now uses the SSS routines. Changed $APLGEN/ZWHOMI to incorporate the minor changes present in the $APLHP version and then deleted the $APLHP version. Corrected a minor error in the C code of $APLUNIX/ZFRE2.C. Moved nowhere. 7553. April 19, 1993 $QPGOOP/WFCLN Bill C. The intrinsic LEN function was being called with an array argument. Convexes don't like this but the Suns don't notice. Moved nowhere. 7554. April 20, 1993 $APLOOP/IMAGEDESC Bill C. IMDCOP now copies Convolving BEAM information. Moved nowhere. 7555. April 20, 1993 $APLOOP/TABLE.FOR Bill C. TABDPT now updates the maximum row number each call. TABDGT now write the old row if necessary on read. Moxed reads and writes now work as advertised (I hope) using TABDPT and TABDGT. Moved nowhere. 7556. April 21, 1993 $APLNOT/IMHCOR Bill C. Modified only to change descriptive parameters necessary to reflect what the related imaging routines had actually done. Moved nowhere. 7557. April 21, 1993 $APLOOP/ARRAYSTAT Bill C. Cleaned up documentation Moved nowhere. 7558. April 21, 1993 $APLOOP/ARRAY Bill C. Added member function ARHIST to compute a histogram. Moved nowhere. 7559. April 21, 1993 $APGNOT/FILLM Gustaaf FILLM prints out full source name. Removed non- functional line of code. Moved nowhere. 7560. April 23, 1993 $QPGOOP/OHGEO Bill C. New task. Like HGEOM except with options to correct for 3-D distortions of a misaligned, coplanar array (VLA snapshots) and a radial scaling correction to correct for variations of the antenna primary beam over the IF bandpass. Also OHGEO.HLP. Moved nowhere. 7561. April 23, 1993 $APLOOP/POSITION Bill C. Added new class functionality. Routine PSNCVT converts the pixel position in one image to the corresponding pixel position in another and PSNCV3 does this with corrections for array plane misorientation and radial scaling. Moved nowhere. 7562. April 23, 1993 $QOOP/QINTER Bill C. Added new routine IMGHGE to do HGEOM like interpolation of one image to the grid and geometry defined by another. Also has option to correct for geometric distortions caused by the "w-term" in misaligned, coplanar arrays (e.g. VLA snapshots). This routine will also do radial scaling using a parameterized model. IMGHGE requires a new interpolation routine IMINTP whose input argument is in pixels. Modified IMINST to call ARREAD directly rather than use IMG2AP so that the window could be honored. Moved nowhere. 7563. April 23, 1993 $APLPGM/XMOM Gustaaf Added adverb OPTYPE. If OPTYPE is blank, functionality is as before. If OPTYPE='NBIV' (No Blanking Illegal Velocities), output pixels are not blanked if their first moment falls outside the range of input channels. This prevents ugly holes in the zero'th moment map. 7564. April 23, 1993 $APGNOT/BLWUP Gustaaf Added this simple minded task which borrows heavily from the SUBIM code. It replaces each input pixel by a square grid of pixels of the same value. The dimension of this grid is specified by the adverb FACTOR. 7565. April 26, 1993 $SYSUNIX/TVDEVS.SH Dave A. Added some more AOC machines; took out references to the evil and long-since-departed CHOLLA. Moved nowhere. 7566. April 26, 1993 $APGNOT/FETCH Bill C. Modified to allow blanking of the input image. Also FETCH.HLP. Moved nowhere. 7567. April 27, 1993 $RUNSYS/CROSSPOL.001 Bill C. Removed line setting version to 'wdc' left over from testing. Moved nowhere. 7568. April 28, 1993 $APLSUB/GETRLS.FOR Pat Changed release string from 15APR93 to 15JUL93 in accordance with decision made at yesterday's AIPS meeting. I won't change the directory name or the version name the shell scripts know about yet as that would cause downtime at the AOC as a complete rebuild would be needed. Moved Nowhere. 7569. April 29, 1993 $APGNOT/FILLM.FOR Gustaaf Added new version of FILLM. It now contains the same logic as the online version in 15APR92 at the VLA site for interactive use. This in anticipation of installing a later version than 15APR92 at the VLA site as soon as the fast T1 link is opera- tional. Also Mike Kestevens improvements for handling holography were put in. Added adverb CALCODE. Moved nowhere. 7570. April 29, 1993 $APGNOT/FITTP.FOR Gustaaf Finally got rid of warning lines about old version of tables. Moved nowhere. 7571. April 29, 1993 $QPGNOT/FRING.FOR Phil The division by a CC table model in FRING could never have worked. The FACGRD variable (same as FACTOR in UVSUB) was not set and therefore defaulted to zero. Obviously nobody runs through FRING for a 2nd iteration. Set FACGRD to 1.0 Moved nowhere. 7572. April 30, 1993 $APLOOP/UVDATA Phil/Bill C. Modified OUVATT so that it always created the output file if the sequence number was 0 and DOCREA was true. Moved nowhere. 7573. April 30, 1993 $APGNOT/AVSPC Phil The freq increment in the header was not being updated fully when averaging IF's together. Moved nowhere. 7574. April 30, 1993 $APGNOT/FILLM Gustaaf Forgot to take APARM out when APARM was turned into a hidden 'TELL' parameter. Fixed now. Moved nowhere. 7575. May 5, 1993 $APLNOT/SELSMG Phil Fixed up so that it will now compile on machines that previously had problems. It is never called by any routine so I did not go into the effort of ensuring that it works, anyone who wants to use it should do so themselves. Moved nowhere. 7576. May 3, 1993 $APGNOT/SDTUV Bill C. Fixed bug in routine MEDIAN that caused problems if there was insufficient data in the current window. Also fixed bug that occurred when a point to be dropped from the list wasn't found. Moved nowhere. 7577. May 3, 1993 $APGNOT/FILLM Gustaaf The Convex compiler choked on a Hex constant which the Sun let pass. Fixed this. Moved nowhere. 7578. May 4, 1993 $APLOOP/UVDESC.FOR Phil/Bill C. Routines UVDCOP and UVDSCP now copy associated VELOCITY and POSITION members. Moved nowhere. 7579. May 6, 1993 $APGOOP/SPECR Phil New task. SPECR is a spectral regridding task. It will read a user's input line format database and, based on the number of channels requested in the o/p data, will FFT the spectra to correlation functions and either pad them or truncate them before FFT'ing back to the spectral domain. The FQ and SU tables also have the channel widths updated. Moved nowhere. 7580. May 6, 1993 IMEAN Eric The number of input parms was stated to be 45, while only 41 variables were provided. This caused the number of bytes per aips byte to be set to 0 for use by ZMIO (!) at least as linked under Solaris. I wonder how many more such things we'll find. Moved nowhere. 7581. May 6, 1993 ACOUNT, PRTLIN, PRTALN Eric Changed CLOSE parms don't match message to say more in ACOUNT. Changed PRTALN to avoid printing a blank page at the start of print jobs inside AIPS itself and corrected Glen's similar change in PRTLIN. Moved nowhere. 7582. May 6, 1993 YTVCIN Eric Changed initial character size to 7 x 9 - the 14 by 22 is too large normally. Moved nowhere. 7583. May 6, 1993 TVFLG, SPFLG Eric With a message terminal, MSGWRT's go to the message server, but ZTTYIO messages and prompts go to the users' main window. So we need to be more careful which we use. Changed a little at the close down section in TVFLG and SPFLG. IBLED should get this too but Phil has it out. Moved nowhere. 7584. May 11, 1993 Upgrade for standard C and Solaris Eric Changed most C routines in AIPS to test for ANSI standard C and to use prototypes when it is present. Added includes to many C routines so that everything is prototyped or declared. Created an AIPS prototype file called $INC/AIPSPROTO.H which prototypes in standard and old C ways all AIPS routines called by some C routine. In $APLBELL, deleted ZABOR2.C, ZCREA2.C, ZLOCK.C, ZTACT2.C, and ZTKILL.C and changed: ZDELA2.C New - uses setitimer which is a non-Posix routine to get accurate delays. Sometimes the 1st delay was missed so also gave it an interval timing. When the timer is finally caught, setitimer is turned off. ZTQSP2.C To standard C. In $APLHP, deleted ZABOR2.C and ZTXMA2.C, and changed ZGNAME.FOR Cleaned up the handling of characters to return only that part after the last "/". In $APLSOL, created ZDCHI2.FOR Gives 'SVR4 Sol' as the system version. ZFRE2.C Differs from generic in that Solaris renamed statfs to statvfs. ZMOUN2.C SUN-like, but the tape names have changed for giving the densities. Also DATs are (so far) seen as Exabyte 8200s by Solaris. ZTAP2.C Rewritten (!) since DATs do not do BAKF. Changed call sequence to pass in the count and pass it back on AEOI. ZTAPE.FOR Generic version changed to pass COUNT on to ZTAP2 rather than do anything itself. ZTPOP2.C O_NONBLOCK added to open parameters. ZTPWA2.C Added ioctl call to forward space the tape over an end-of-file. Solaris does not go past EOFs on reads. ZTPWAD.FOR SUN version copied over. In $APLBERK, moved all Berkeley socket routines (ZSSS*, ZVTV*, ZVTP*, most of the ZR*) up to $APLUNIX since they seem to be portable to most Bell systems. Also deleted ZABOR2.C, ZCMPR2.C, ZCRLOG.C, ZDATE.C, ZTACT2.C, ZTAPIO.C, ZTIME.C, ZTKILL.C, and ZTXMA2.C and changed: ZCREA2.C Rename MAXPATH to MAXAPATH to avoid IBM-standard name. ZDELA2.C Added the interval to the setitimer call and the set all of setitimer to 0 at the end. This might explain some jerkiness in TVMOVIE and the like. ZLOCK.C Rename MAXPATH to MAXAPATH to avoid IBM-standard name. Also force it now to Posix fcntl locks rather than flock. ZSTRTP.C Rename MAXPATH to MAXAPATH to avoid IBM-standard name. In $APLSUN, added ZTACT2.C, ZTKILL.C, ZTXMA2.C all of which are the generic UNIX codes except that the inode number is not "d_ino" but "d_fileno". In $APLIBM, deleted ZDATE.C, ZDCHIC.C, ZTACT2.C, ZTIME.C, ZTKILL.C, and ZTXMA2.C, and changed ZCREA2.C, ZLASC2.C, ZLPCL2.C, ZTPOP2.C, and ZTPOPD.C to rename MAXPATH to MAXAPATH to avoid an IBM-standard name (PP used to do this (!!!), but of course could not handle C including AIPSPROTO.H). In $YSERV, changed XAS.SHR as: Makefile Added comments and STDLIBS parameter. xas.h Do ugly kluge to get around MIT's error in prototyping XSetIOErrorHandler(). Change the icon and the data rows to unsigned char. Add prototyping and declarations of all routines. xas.c Add correct casts, add returns, and function type statements. colors.c Add return statements and function typing. comm.c Add return statements and function typing. Change function index to strchr. cursor.c Add return statements and function typing. image.c Add return statements and function typing. init.c Add return statements and function typing. Corrected use of toupper (which does 1 character at a time) to check the display assignment. screen.c Add function typing, change bcopy to memcpy, and add prototype for XShmGetEventBase. Moved to test areas in CV and the VLA in advance. 7585. May 11, 1993 Upgrade for standard C and Solaris Eric In the $APLUNIX area, changed ALL C routines to be ANSI standard when the compiler sets __STDC__ to be greater than 0. The Solaris compiler when it is allowed non-Ansi extensions still defines __STDC__ but sets it to 0. Other changes are: ZABOR2.C Changed to be more careful in its selection of signals and to call a function "signel" which uses the Posix function sigaction (preferred over old function signal). the code for signel was copied from Stevens' book. ZABORS.C Changed to use signel rather than signal, changed pid type declaration. ZACTV9.C Changed to use MSGWRT for error messages. ZCMPR2.C Replaced old copy file and rename with function ftruncate which is wipe spread although not basic Posix. ZCPU.C Change to reference sysconf(_SC_CLK_TCK) for the number of ticks per second in the Posix way when possible, and assume 60 otherwise. ZCRDIR.C Change to use Posix functions S_ISDIR to determine if a file is a directory and mkdir to make a directory directly (setting permissions with standard symbols) rather than assuming a format for mode, the default permissions, and calling system to do the mkdir. ZCREA2.C Set permissions with standard symbols, not 0666. ZCRLOG.C Since putenv was explicitly not included in Posix (so far), adopted and corrected the Sun version. This moves the environment into a local large array (of pointers) and replaces or adds definitions as needed. I found a serious bug in the Sun version which did not null terminate the initially copied list. ZDCHIC.C Added test for NBBY and defaulted it to 8 and added call to umask to force new files to have unlimited access except write is prohibited to "other". This means our users can have any umask they want, but their AIPS files will be created the way we want. ZDELA2.C Lowered round-off, but I hope no one uses this version since sleep takes integer seconds as its agrument and that is just too long. ZDICC2.C Changed to use MSGWRT. ZDICIO.C Changed to use MSGWRT. ZDICO2.C Changed to use MSGWRT. ZEXPN2.C Corrected serious error (== when = needed) in filling the new space with 0's. ZFREE.FOR Changed to drop early parts of long disk area names if needed, so that the useful part is displayed. ZGNAME.FOR Changed to check the task's first argument for full path names rather than simple task names and to drop the path part. ZLOCK.C Created new routine adapted from Posix part of BERK. ZLPCL2.C Added check for file's existence. ZMKTMP.C Rewritten. Dropped non-standard mktemp function and a pipe to get the current working directory and replaced them with the standard functions getcwd, getpid , and access to build our own unique file name anywhere we want. ZPRPAS.C Replaced call to getpass which was deliberately omitted from Posix with 12-character (finally) code adapted from W. Richard Stevens' book. ZROPN2.C Changed bcopy to memcpy and clean up typing. ZSSSO2.C Changed index, bcopy, bzero to strchr, memcpy, memset. ZTACT2.C New, adaptation of BERK version with standard ways to test for file type and empty inode. ZTKILL.C New, adaptation of BERK version with standard ways to test for file type and empty inode. ZTKOP2.C Changed bcopy, bzero to memcpy, memset and fixed cast of sock_name. ZTPOP2.C Changed strncasecmp (which no longer exists) to 2 strncmp calls. A tape must be "on-line" or "ON-LINE" if it is to be treated as a VLA real-time input. ZTXMA2.C Rewritten and corrected. Now uses standard functions opendir and readdir rather than open and read. It checks the characters both before and after the * in the (simple) wildcard pattern. This generic routine can replace all the wild variations that have used various non-standard (but full) wildcard pattern analysis and comparison. ZVTPGC.C Correct cast of fromb_in. ZVTPO2.C Change bzero, bcopy to memset, msmcpy. ZVTVGC.C Correct cast of fromb_in. ZVTVO2.C Change bzero, bcopy to memset, msmcpy. ZVTVO3.C Change index to strchr. Moved to test areas in CV and the VLA in advance. 7586. May 11, 1993 Upgrade Message and TEK servers Eric/Pat Changed $AIPNOT/TEKSERVER.C into $APLUNIX/ZTKSRV.C making the code pass ANSI standards and adding sleep (5) calls after fatal error messages before exiting. Also changed $AIPNOT/MSGSERVER.C into $APLUNIX/ZMSSRV.C making it pass ANSI standards. Created stubbed routines $APLGEN/ZMSSRV.FOR and $APLGEN/ZTKSRV.FOR. Then wrote simple programs to call the Z routines $AIPNOT/MSGSRV.FOR and $AIPNOT/TEKSRV.FOR. All this keeps our directory structure clean and would allow some VMS fanatic to develop Z routines to support his obsession. Speeded up the Z routines which talk to the message server (and made them into ANSI standard, etc.). Changed $APLUNIX versions: ZTTCL2.C Reset TTDEV to TTDEV00 on close error and reset "remembered " parameter. ZTTOP2.C Created global parameters "remembered" and ms_sock_name to avoid calling ztrlog, gethostbyname, and getservbyname more than once. Also resets logical TTDEV to TTDEV00 on error. Did the usual changes to memset and memcpy. ZTTOPN.FOR Avoids additional ZTRLOG calls with parameter MSGRMB in the MSG common. ZTTYI2.C Resets logical TTDEV to TTDEV00 on error. ZTTYIO.FOR Changed error message formatting and changed to turn off MSGRMB on error. $INC/DMSG.INC added parameter MSGRMB replacing dummy parameter to keep lengths fixed. Moved to test areas in CV and the VLA in advance. 7587. May 11, 1993 More tape changes Eric Changed the COUNT parameter so that it is an In/Out parameter to ZTAPE. This means one can no longer hard code numbers into that argument (e.g. 1). Changed TAPIO, PRTTP, MK3TX GR2TEX, GRITP, and RDFITS to remove such hard-coded numbers. Changed AU4, FITTP, and AVTP to use the returned COUNT to imply how many files were skipped to end of tape (> 0 => eoi after COUNT files, < 0 => ski[pped -COUNT files to eoi). Changed BAKTP to use ZTAPE operation BEGW rather than the WEOF, BAKF combination it had used before and to avoid hard-coded COUNTs. Restored all of the old changes to the $TST version of FITTP since the $NEW version had overwritten them in the previous FITTP change by Gustaaf. Changed $APLGEN/ZTAPR.FOR to pick up a returned COUNT and $AIPNOT/TPMON.FOR to send that COUNT back. Since there are 13 versions of ZTAP2 (!!!), I cannot change them all to take COUNT as an argument and return the file position on AEOI. So changed $APLGEN/ZTAPE.FOR to force the returned COUNT to 0 on AEOI. The only version at this moment that does not do this is the Solaris version which has the new ZTAP2 call sequence and its on version of ZTAPE. Deleted the obsolete $APLSUB/FNDEOT. Also changed $APLUNIX/ZBKTP1.FOR which called FNDEOT to call ZTAPE of AEOI, to use proper variables for COUNT, and to report the returned value if any on AEOI. Found a bug in $APLSUN/ZMOUN2.C and changed the formatting of the tape name based on density. It was stuffing it all in one character but Sun uses a decimal format. Moved nowhere so far. 7588. May 11, 1993 BLING Chris Found and fixed more cases where constants were being stored in return parameters for APLOOP routines. Moved nowhere. 7589. May 11, 1993 $SYSUNIX/* Pat Modified many -- most -- of the shell scripts in $SYSUNIX and elsewhere to implement the Solaris port. The code in these scripts was very heavily dependent on Berkeley specific features of Unix. Many of the scripts now key off either $ARCH or the presence of a system variable $TZ which is present on SYSV flavor Unixes such as HP-UX and Solaris. Among the files changed are: AIPSASSN.CSH/SH: Solaris/hp additions AIPSCC: put unix-style options first AIPSEXEC: remove LOADDBG, correct $TZ test (was backwrds) CCOPTS.SH: add -I$INC to get COMLNK: Skip subroutines, look in $PREP for .f, .o, etc. COMRPL: look in $PREP, unix options first FDEFAULT.SH: Solaris 2.x compiler settings MAKEAT: Many bug fixes, sanity checks PWD: Use $PWD if it's defined STARTPMON: various improvements (ps, etc) START_TVSERVERS: more rational OPENWINHOME usage TVALT: Add Solaris, HP support; improve DEC support TVDEVS.SH: remove dead hosts WHICH: Remove dot from wildcard expression XASERVERS: Many changes to get SYSV flavor ps to work (msg terminal enabled by default, but this is disabled for the NRAOAOC site). There is more to come (SYSSUN, SYSIBM, SYSSOL, INSUNIX) but this is a fair share of the changes. Moved nowhere. 7590. May 13, 1993 $SYSUNIX/STARTPMON Pat This script is now almost bulletproof and should give slightly better warning messages and has a greatly enhanced debug mode. The code was overhauled to make more sense and not repeat things. Moved nowhere (at least not by me). 7591. May 12, 1993 UVPLT Eric Removed an obsolete loop that caused all points to be plotted twice! Also added FACTOR as an adverb to UVPLT to change the size of the plotted dots. For many uses they are too small and need scaling up. Moved nowhere. 7592. May 13, 1993 tapes Eric In one move, put the $APLSOL version of ZTAPE up into $APLBELL and changed the APLCRI (non-functional!) and APLMASC versions of ZTAP2 to expect COUNT in their call sequence and to know about the new operations such as BEGW, AEOI, and doing all of WEOF. I'll bet several other ZTAP2's do not know about these. Corrected the naming algorithm for densities in $APLSUN/ZMOUN2. Corrected the passing of parameters in $APLGEN/ZTAPR and the corresponding bits of TPMON. Moved nowhere. 7593. May 14, 1993 LWPLA.{FOR,HLP} Mark Added the option to specify shades of grey for line drawing via DPARM(9). It's sometimes useful to have white contours superimposed on a greyscale image for example. 7594. May 17, 1993 $APGNOT/SDTUV Bill C. Fixed bug in routine SDTRED in which it didn't close the table if the data was too large for the internal array. Moved nowhere. 7595. May 18, 1993 VBPLT Neil Killeen/Bill C. Fixed problem with subarray selection. SUBARRAY=0 was advertised to mean all and in fact it defaulted to 1. Also fixed a problem with selection by timerange when there were multiple subarrays with 5 day offsets. Various parts of the program were selecting data either with or without the offsets using the same value of TIMERANG. Also added clarification to VBPLT.HLP. Moved nowhere. 7596. May 19, 1993 $APLDEC/ZR64RL.FOR Pat/JingPing Ge There is a known problem with the generic unix version of this routine when used with DEC Fortran (not MIPS?) in that double precision numbers get corrupted in the FQ tables on a FITTP-FITLD cycle. JingPing Ge at Brandeis found an empirical workaround which is codified in this module. Moved nowhere. 7597. May 19, 1993 $QPGOOP/FRCAL Bill C. New task to do Faraday rotation self calibration. Also FRCAL.HLP. I tried to get this to work using the GRID Fourier transform routines but the are riddled with bugs and structural errors and so I gave up. Moved nowhere. 7598. May 19, 1993 $QNOT/VISDFT.FOR Bill C. If a point model was passed throught the commons there was no way of specifying the polarization. Under there circumstances the common (DGDS.INC) variable KSTOK with a value of 2 now indicates Qpol, 3 Upol and 4 Vpol. The common variable VOFF is also modified to give the correct visibility offset. Moved nowhere. 7599. May 19, 1993 $QNOT/UVPDIV.FOR Bill C. New routine to divide the polarized portion of a uvdata set by a polarized model. Moved nowhere. 7600. May 19, 1993 $APLOOP/UVDATA Bill C. OUCREA was not saving the correct image sequence number in an object if it already existed. Added function OUVCGT to returen the catalog header as suitable for passing to standard aips routines (tell the truth about compressions). Moved nowhere. 7601. May 19, 1993 $APLOOP/UVUTIL.FOR Bill C. Added self calibration for Faraday rotation functions: UVFRSC high level function to determine Faraday rotation corrections. FRSOLN low level routine to compute Faraday rotation. Also modified NXTAVG not to set the STOKES on the UV data unless averaging in polarization is requested. Suitably modified routine (SLFCAL) calling NXTAVG. Moved nowhere. 7602. May 19, 1993 $QOOP/QUVUTIL.FOR Bill C. Added function UVPDIV to divide a Q + U polarized model into the LR and LR portions of a uv data set. Modified numerous routines to use OUVCGT to fetch the CATBLK to pass to standard AIPS routines. In routines OUMSUB and OUMDIV the CC tables are not checked if a point model was specified. Moved nowhere. 7603. May 20, 1993 BS table Chris Fixed a bug in the format of BS tables. Previously there was only one weight for each record rather than a separate weight for each IF. Files changed were: BSINI in APLNOT - changed column dimension and incremented table format version number. TABBS in APLNOT - changed WEIGHT dummy variable to be an array TABLEUTIL in APLOOP - changed call to OTABBS BLAPP in APGOOP - changed to use IF weights BLING already tried to write out a weight for each IF. Moved nowhere. 7604. May 20, 1993 BLING Chris Overhauled BLING to make it easier to use. Specific changes are as follows. - Removed the code to show the results of each step of the fit to the peak in the search grid and eliminated the APARM that controlled this. - Removed the option to replace valid data points with data points of unit amplitude and zero phase (used to verify the FFT and grid search). - Removed the uniform weighting option (which never made any significant difference). - Reduced the volume of message output by printing the results in tabular form. - The oversampling is now calculated automatically instead of being set by the user. - Integration times are set automatically from the INTTIM random parameter if it is present. The integration time is assumed to be constant for each baseline within a scan. - The fringe acceptance criterion is now a minimum SNR rather than a minimum confidence level. The latter proved unworkable. - If the fit to the peak in the search grid fails the raw results from the search are used instead of discarding the results as before. The unrefined results are better than nothing. I also discovered that there was no test for missing IFs when searching for separate narrow-band delays (which lead to divide-by-zero problems) and added one. Moved nowhere. 7605. May 21, 1993 $APLNOT/SNSMO.FOR Leonid/Phil When smoothing VLBI solutions in the SN table a problem occurred if the real/imag columns were blanked. When there was a good delay/rate solution no check was made to ensure that the real/imag were unflagged before a phase rotation was applied. This problem only showed up in multi-IF data because for single IF data the phase rotation was zero. Inserted the check on flagged columns and relinked CLCAL. Moved nowhere. 7606. May 21, 1993 TVCPS Chris TVCPS now adds a label containing the AIPS usernumber to the plot as demanded by AOC management. This can be turned off using APARM(8). Moved nowhere. 7607. May 24, 1993 BLAPP Chris BLAPP now determines which BS table records belong to the same scan without requiring SOLINT adverb. The SOLINT adverb has been deleted. Moved nowhere. 7608. May 24, 1993 $QPGNOT/HORUS.FOR Gustaaf Fixed controversy 'channels averaged/channel increment', which caused wrong header information in output. This fix required changes to $APLNOT/IMCREA, $QNOT/IMGMEM, $APLNOT/IMHCOR, $QNOT/MAKMAP, and $QOOP/QUVUTIL. Moved nowhere. 7609. May 26, 1993 UVSUB.HLP Bill C. Added warning about using CMETHOD='GRID' for RL, LR data. This is due to a fundamental error in UVCONJ. UVCONJ doesn't know what type of data it has and therefore cannot swap RL and LR data. Moved nowhere. 7610. May 26, 1993 $APGNOT/GETJY.FOR Gustaaf Added test to see if a gain table entry is blanked. This was necessary since CALIB does not reset the weight to zero when it deals with the reference antenna. Blanked solutions then give rise to crazy flux values. CALIB will be changed later so that it always puts the weight to zero when the real or the imaginary part is blanked. After that the current fix is not absolutely necessary anymore, but does no harm either. Moved nowhere. 7611. May 27, 1993 $QPGOOP/FRCAL Bill C. Modified to allow passing a data set previously divided by FRCAL. In this case an output file is not created. Also changed FRCAL.HLP. Moved nowhere. 7612. May 27, 1993 $APLOOP/UVUTIL.FOR Bill C. Changed FRSOLN to subtract all of the R-L phase difference from the L gain rather than splitting it between R and L. This change allows smoothing the solutions. The values printed are now the right-left difference rather than the Faraday rotation angle. Moved nowhere. 7613. May 27, 1993 MX.FOR Bill C. Fixed indexing bug in MDECID that sometimes caused it to try CLEANing too deeply in a given major cycle. Also in MXBHIS the last valid beam value is now propagated to the end of the table. Formerly, the beam histogram was effectively zero outside of some region. Moved nowhere. 7614. May 27, 1993 $APLGEN/ZDCHIN.FOR Pat The code that read the NET0:NETSP file was seriously flawed in that it never stopped reading, despite having a limit of 100 entries for the in-core version of the file. This caused all sorts of problems on the IBMs at the AOC as there are currently 138 valid entries. Who knows how come it still worked on the suns... Fixed so that the limit is now 512 non-comment entries and it gives up reading once the limit is reached. Because of this change, EVERY task will need to be rebuilt. Moved to the AOC sites (aguila, yucca, zuni, kiowa) today. 7615. May 27, 1993 $QOOP/QCLEAN.FOR Bill C. Propagate the last valid beam histogram value to the end of the histogram in CLBHIS. This prevents CLEANing too deeply. Moved nowhere. 7616. May 31, 1993 BLING Chris Added a check for grid axes longer than 4096 elements (the maximum FFT size for QXFOUR). BLING was occaisionally trying to perform longer FFTs than this and overwriting state variables. Moved nowhere. 7617. June 1, 1993 CALIB Phil Added a check in routine CLNSOU to avoid writing SN entries with blanked real/imag columns but non-zero weights. Moved nowhere. 7618. June 2, 1993 $QPGOOP/WFCLN.FOR Bill C. Restoring beam parameters were not in the correct units (seconds rather than degrees) as passed to the CLEAN routines. Moved nowhere. 7619. June 2, 1993 CALIB Chris Fixed a bug in applying channel selection criteria to single- source files. In this case CALIB will now derive the solutions from the selected channels and apply them to all channels. The user will be warned about this. Moved nowhere. 7620. June 3, 1993 TVCPS.HLP Dave Added a paragraph (marked for AOC only) describing the AIPS logical that will route output to the holding directory. Moved nowhere. 7621. June 4, 1993 $APLOOP/UVDATA.FOR Bill C. In OUVCRE the default initial value of OLDFILE is now true. The value will be reset in OUCREA if the file is atually created. In the old scheme OLDFILE was left FALSE if it in fact existed. Moved nowhere. 7622. June 4, 1993 $APLOOP/UTILITY.FOR Bill C. Added a routine FNDKSC to lookup the disk and catalog slot numbers of an object. It uses name, class etc. so, unlike OBDSKC it can work on an object which has not been opened before. Modified OBFEXS to use FNDSKC rather than OBDSKC. Moved nowhere. 7623. June 4, 1993 $APGOOP/MULIF Bill C. New task, adds IF axis or changes its dimensionality. The intent of this task is to either add an IF axis into data sets which do not contain one or to change the dimensionality in data sets which do. This is normally not necessary but may be needed if a particular task insists on having an IF axis or in order to combine datasets (DBCON) with different IF axis dimensions. Moved nowhere. 7624. June 4, 1993 $APGOOP/IMTXT Bill C. New task, writes AIPS image to external text files in form used by FETCH. Moved nowhere. 7625. June 4, 1993 $APGOOP/UV2MS Bill C. New task to append a single source file to the end of a multisource file. The output will be created if necessary. The visibility geometries must be identical (except for the source random parameter) in the input and output data sets. This task avoids using MULTI and DBCON although the resulting dataset will need to be sorted (UVSRT) and indexed (INDXR). Also UV2MS.HLP. Moved nowhere. 7626. June 4, 1993 $QOOP/QCLEAN.FOR Bill C. In CDECID there is now a trap so that JLEV can never exceed NBMHIS. CLBHIS had an indexing error which caused particularly nasty problems for non square beams; the peak would appear very far from the center. Moved nowhere. 7627. June 4, 1993 $QPGOOP/WFCLN.FOR Bill C. Fixed typo which caused IMSIZE(1) to be used instead of IMSIZE(2) for the second dimension of the output images. Moved nowhere. 7628. June 4, 1993 $APGNOT/SNCOR.FOR Phil Added an option to set the residual delays in an SN table to zero. Moved nowhere. 7629. June 6, 1993 New task: $APGOOP/DTSUM Phil A new task, I kept promising a data summary task earlier this year and I finally needed it for the VLBA Summer School. It took 4 hours to write in the OOP package - a testament to the ease of programming in that package once one is familiar with it. DTSUM provides the user with a matrix listing of the number of visibilities per baseline per scan, it also provides a little scan information and lists the names of the antennas. I can see myself expanding this task as time goes on. Moved nowhere. 7630. June 7, 1993 $APGOOP/UV2MS.FOR Bill C. Modified to update the output times to reflect the reference date of the output file. Adding data before the output reference date is disallowed. Moved nowhere. 7631. June 7, 1993 $QPGOOP/SCMAP.FOR Bill C. RASHIFT and DECSHIFT were being stored as the wrong data type (character rather than real). Modified SCUVST to set the uv range for full weight not to include baselines longer than those used in imaging. Modified self cal portions to loop over subarrays and use new call sequence to SLFREF. Moved nowhere. 7632. June 7, 1993 Subarrays and smoothing Bill C. A number of the calibration smoothing routines didn't properly handle subarrays: - $APLOOP/UVUTIL(SLFREF) Added desired subarray to call sequence; fixed call sequences to REFCNT, REFFAZ. Also removed extra call to UVDTCO in UVCOPY. - $APLNOT/REFCNT.FOR Added subarray number and column pointer to call sequence, now checks subarray. - $APLNOT/REFFAZ.FOR Added subarary number to call sequence and subarray now checked. - $APLNOT/REFDLY.FOR Added subarary number to call sequence and subarray now checked. - $APLNOT/REFRAT.FOR Added subarary number to call sequence and subarray now checked. Moved nowhere. 7633. June 7, 1993 $YPGNOT/IRING Gustaaf Enlarged various arrays to allow much bigger maps and more rings. Also added long overdue checks on these numbers to prevent overwriting of arrays. Moved nowhere. 7634. June 9, 1993 Shift and rotates Bill C. Fixed numerous problems involved with using the "Gridded FFT" modeling method with images that had been rotated or rotated and shifted. The principle problems were that: 1) the model wasn't being rotated about the correct position, 2) the model CC's weren't being constrained to be on grid cells causing bad model values. Problem 1 is fixed in ALGSTB by rotating the u's and v's before interpolating the model and then unrotation them after the model. For obvious reasons ALGSUB can't do this and GRDSUB now uses ALGSTB for this case. Problem 2 was fixed by putting logic in GRDCRM to force the "shift" corrections to be an integral number of cells. In ALGSTB and ALGSUB there was a corresponding fix to include the effect of the fractional pixel ignored in GRDCRM. Modified routines: $QNOT/GRDSUB,ALGSUB, ALGSTB, GRDCRM. These bugs mostly affected UVSUB, CALIB and SCMAP, MX and WFCLN(?) should not have been affected. Moved nowhere. 7635. June 8, 1993 POSSM Phil Fixed a POSSM bug that has been there for years, when plotting the amplitude of the bandpass the reals rather then the amplitudes were plotted, this was never noticed because the imaginary values are often close to zero. Moved nowhere. 7636. June 10, 1993 ANCAL Phil Had to increase the size of the array allocated for reading Tsys values, its too small for even modest multi-IF data from the VLBA. Moved nowhere. 7637. June 10, 1993 $APGNOT/JMFIT.FOR Gustaaf The conversion (error in R.A. pixels => error in arcseconds) was wrong by a factor 2. IMFIT did it right all along. Moved nowhere. 7638. June 11, 1993 $APGNOT/UVFIX.FOR Phil Caltech now send us multi-source files from their correlator, these can quite happily be run through UVFIX because they only contain one source and the RA/Dec are present in the catalogue header. However if they are DBCON'd and then run through UVFIX the uvw's can be screwed up if there are more than one source in the DBCON'd file. Added a very noticeable warning message if this might occur. Also added same to the helpfile. Moved nowhere. 7639. June 13, 1993 $APGNOT/ANCAL Phil There was a problem with the new index keyword, if a column was ignored via the Xn keyword and n=1 then the number of columns to be read from the ascii file was incorrect. Moved nowhere. 7640. June 14, 1993 BLING Chris Added the BADDISK adverb. Moved nowhere. 7641. June 14, 1993 BLAPP Chris Removed SOLINT from the internal parameter list. As this was no longer declared in the help file later input parameters were being misaligned. This became evident when BLAPP ignored GAINUSE/GAINVER. Also added a BADDISK parameter. Moved nowhere. 7642. June 14, 1993 $APGNOT/SNSMO.FOR Phil SNSMO had real problems in the subroutine that rereferenced phases when dealing with multi-IF data. It thought it new if dual polzn data was coming when the pointer for the 2nd polzn column was > -1, but the calling routine incremented that as the IF number increased. Changed the logic so that REFFAZ did the pointer incrementing itself. Moved nowhere. 7643. June 15, 1993 $SYSUNIX/FC Pat Modified the code so that it checks the status return of the compiler _as well as_ seeing if there's a nonzero-size object file. Previously it only checked for the object file and this caused the IBM compilers to accept very questionable compiler errors as "ok". Same story on the HP. Moved nowhere. 7644. June 15, 1993 $APLNOT/CALADJ.FOR Phil This routine will not have worked properly for multi-IF data, the call to CHNDAT was made using the same LUN with which the SN table had just been opened and so no FQ/CH table would ever have been found - the correct error message was also supressed down inside ISTAB, which was why we never noticed before. Switched the order in which operations are done. Moved nowhere. 7645. June 16, 1993 $HLPFIL/CLCOR.HLP Gareth Added equations to the explain section for the SSLO option to make it clearer to Elias Brinks (maybe others - who knows?). Moved nowhere. 7646. June 17,1993 $APGNOT/SNCOR.FOR Dave A. Changed some writes to refer to SN instead of CL tables. Moved nowhere. 7647. June 18, 1993 $APGNOT /VLBIN Phil Under some circumstances the time span in the decode file could be 0, if this was the case made the default 1 second. Moved nowhere. 7648. June 21, 1993 $YPGM/GREYS.FOR Mark Removed the frame drawing commands from routine GLAB and put them after the greyscale commands in the main program. This will prevent the grey image from painting over the frame when the plot is translated to PostScript by LWPLA. 7649. June 21, 1993 $HLPFIL/LWPLA.HLP Mark Added a warning about the need to specify the -K0 option to 'dvips' in some circumstances when LWPLA output is included in a TeX or LaTeX document. 7650. June 19, 1993 $YPGNOT/CLPLT Phil Changed the maximum number of closure phases that can be plotted from 90 to 200, to keep VLBA users happy and to kill more trees. Moved nowhere. 7651. June 22, 1993 BLING Chris BLING now uses the default integration time specified in APARM(1) if the value of the INTTIM random parameter is zero as has occurred in some files written by VLBIN. Moved nowhere. 7652. June 22, 1993 $YPGNOT/POSSM Phil Added the option to label corr fn plots in (micro)seconds if the user so desires. While doing this discovered that channel selection for corr fn plotting was not what I intended. Changed it so that the whole spectrum is FFT'd but only the selected lag range is plotted. Also updated POSSM.HLP. Moved nowhere. 7653. June 23, 1993 KWICK Eric Added tests on the value of the stack pointer. User errors such as calling a procedure with a number followed by a minus could cause the internal value of true to be reset to 0! I am really suprised that I did not build such checks into everything before. Oh well, they are there now. Moved nowhere --- so to speak. 7654. June 24, 1993 plot tasks Eric Tasks SNPLT, GAPLT, and TAPLT all contained vestigial loops J = 1,2 which caused the points to be pltted twice exactly. These must have been left over from UVPLT when the loop also included something useful altering the second plot. Anyway, dropped these now. Moved nowhere. 7655. June 24, 1993 $SYSUNIX/MAKEAT Pat Fixed MAKEAT so that it no longer duplicates the OOP areas in the APLSUB listings. This was causing compilations to be done twice in all installations. Moved nowhere. 7656. June 25, 1993 plotting Eric In checking a gripe I found some unrelated problems . The plot routines were plotting RA's over 24 hours as 100 (meaning 1 day, 00 hours). Changed CTICS, ITICS, TKTICS, PROFL, and CHNTICS to plot 1/00 for times over 1 day and to leave off the 1 entirely for right ascensions. Changed STARPL to clean up the typing and to have it plot strings on FACTOR < 0 as advertised (rather than > 0). Fixed KNTR, PCNTR, PROFL to accept STFACTOR < 0 and corrected help files for CNTR, GREYS, PCNTR, KNTR, and PROFL. Also created new help file STFACTOR. Made repairs June 27, 28. Moved nowhere. 7657. June 28, 1993 TVFLG, SPFLG Eric Added code to make sure things get initialized even if the opening TV load has no valid data. Added warning message on the undo command (it takes a while). And added warnings to the help files about undo and suggesting smaller master grids. Moved nowhere. 7658. June 28, 1993 CHANGE.DOC Eric Corrected 2 numbering errors (Phil) in which 7319 was followed by 7220 last Christmas and (old) 7528 was followed by 7259 on June 4. Moved nowhere, but diff will be big tomorrow. 7659. June 28, 1993 POSSM Phil When fixing up the cross-correlation channel selection a few days ago I had neglected to ensure that the default of bchan = echan = 0 did the correct thing. Now it does. Moved nowhere. 7660. June 29, 1993 $APLPGM UV tasks Eric Changed all $APLPGM UV tasks to use UCMPRS rather than the messy ZCMPRS structure (pioneered in FUDGE) and to use the parameterized buffer sizes from include ZPBUFSZ.INC. Changed CLIP, CORER, DIFRL, FUDGE, TBAVG, UVBAS, UVLSF, UVMTH, and UVPOL in both these ways. Changed buffer sizes only in CORFQ, UVCMP, and UVDIF. Changed CORER to handle compressed data and to copy table files (it was still trying EXTCOP on AN files!). Changed UVDIF to allow slight differences in the number of UV samples although the comparison will fail when the missing sample is found. CORFQ and UVMTH cannot handle compressed data, but at least test for it. Moved nowhere. 7661. June 29, 1993 $UPDUNIX/UPDOBSLT.c Pat Modified the reading mechanism so that it ignores stuff in the input transaction files beyond the last "important" field. That way, a "reason" can be inserted in the PUTBCK.UPD file without sending this program into a tizzy. Moved to rhesus (CV IBM), aguila (AOC sun) today. 7662. June 29, 1993 task timings Eric Changed $AIPSUB AU2 and TASKWT to reduce the time delays associated with starting and waiting for tasks. The total time allowed for a task to start is sharply reduced (from 303 sec to about 30) and the intervals between checks are also reduced to less than half the old values. Machines are now faster so these times have to be reduced accordingly. I debated puting this as a SETPAR parameter but that is a lot of work and would be very subject to abuse and misunderstanding. Moved nowhere 7663. July 1, 1993 $APGNOT UV tasks Eric Changed the following $APGNOT tasks to use UCMPRS instead of the messy code around the call to ZCMPRS and to use parameterized buffer sizes (ZPBUFSZ.INC size parameter UVBFSS). They are ASCOR AVER AVSPC BLOAT DAYFX DESCM FILLR MK3IN MULTI PHSRF SBCOR SDTUV UVAVG UVCOP UVDGP UVFIL UVFIX UVGLU UVLOD UVMOD UVNOV UVSIM VBCAL VBMRG VLBIN WTMOD Also changed AVSPC to use MAXCIF rather than MAXCHA * MAXIF and added a test on the resulting buffer size. Note that ASCOR, AVER, BLOAT, UVDGP, UVFIL (!!), UVMOD, VBCAL, VBMRG, and WTMOD cannot handle compressed data although all of them know enough not to try. Changed $YPGNOT/IBLED to call UCMPRS and $QPGNOT/IM2UV to call UCMPRS and to use parameterized buffers. Changed $APLPGM/CLIP to use MAXCIF rather than MAXCHA * MAXIF. Moved nowhere. 7664. July 1, 1993 $APGNOT/FITLD Phil Changed to write TY tables, and the model into the IM table. Also fixed so that weights are dealt with on an IF-by-IF basis for VLBA data. Moved nowhere. 7665. July 2, 1993 More UV tasks Eric SBCOR: corrected Fortran error - one is required to compare logicals with .EQV. and .NEQV., not .EQ. and .NE. which compare all bits rather than just the true/false state of the variable. The IBM compiler forces this. FITLD: made many buffers parameterized sizes, change file compression call. I could not get rid of the MAXCHA*MAXIF buffers; they are too deeply buried in the system and Phil or Bill will have to fix this. $APGNOT tasks to which I added parameterized buffers: DBCON FILLM FITTP INDXR PRTUV SWPOL USUBA UVFLG UVSEN and minor typographic changes to DECOR SDCAL SELSD SPLIT UVFIX UVFND Also removed excess buffers (MAXCHA*MAXIF) making them MAXCIF in $APGNOT/FITTP and SWPOL. Corrected COMMON alignment error in DBCON. Replaced a hexadecimal constant in FILLM. Also parameterized buffers in $YPGM/UVHGM, $YPGNOT/DFTPL, $QYPGNOT/MX, and $QYPGNOT/UVMAP. The last one particularly was badly underutilising its buffers. Moved nowhere. 7666. July 4, 1993 UVMAP Eric Corrected error in buffer sizes in UNIF. I had which was bigger switched in 1 case. It then worked on SMALL but not LARGE. It turned out after some experimentation that something does not work if the buffer is not a power of 2! Therefore, I changed the included UVBFSS to parameter UVMAPB = 16384. This is not the way we meant to do such things. Sigh... Moved nowhere. 7667. July 5, 1993 ASCAL, VSCAL Eric Changed to parameterized buffers. These 2 had very small buffers and called one routine with a buffer of size MAXIF when it should have been much larger (the 4th buffer to UVMDIV). Moved nowhere. 7668. July 6, 1993 UVLIN Juan Uson/Pat Murphy (entry inserted for Juan by PM; all PM did was add GETNUM and do the checkout/putback; Juan did all the important stuff). Cleaned up UVLIN. Moved all input stuff to subroutine UVLIIN, deleted extra variables. Implemented a reporting option for VLA data (checked using TELTYP). Report goes to message file, user controls with DOOUTPUT. Overhauled the explain file. *********************************************************************** Changes while 15JUL93 was NEW: *********************************************************************** 7669. July 9, 1993 IBM Fortran problems Bill C. Fixed a couple of minor FOrtran problems which gave trouble on the new IBM compilers: $APGOOP/CL2HF.FOR Arguments to MOD function of different type. $APGOOP/TBDIF.FOR Logical comparison of LOGICALs requires .EQV. or .NEQV. Moved from 15JAN94. 7670. July 13, 1993 MINSK, UVMAP Eric Changed MINSK to force the buffer size used to be a power of 2 less than or equal to the buffer size given in the call. Wrote a new $APLSUB/POWER2 to compute this. Changed UVMAP to use the standard UVBFSS buffer size again. This will fix some of the mysterious bugs in MX. Moved to 15JUL93. 7671. July 13, 1993 TPMON Eric Found that several of the error messages for checks on the number of words read/written in the handshakes were wrong. Corrected the error messages. Note that 15JUL93 TPMON does not work with 15OCT92. Moved to 15JUL93. 7672. July 13, 1993 COMAP Eric Changed COMAP.001 to set FACTOR=0 ahead of UVPLT calls and changed UVPLT.HLP to allow this value. UVPLT sets all FACTORs under 0.1 to 1 anyway. Moved to 15JUL93. 7673. July 15, 1993 Various Pat Went through the entire NEW hierarchy with an automated checker shell script, comparing rcs text of files with what was in the clear text area. Found and fixed the following discrepancies (some of which were only found in TST and fixed in NEW): $AIPSUB/VERSTR.FOR 6 char string, not 8 $APLNOT/DVDMIN.FOR Remove obsolete hack for suns $APLSUB/LMDER.FOR Remove obsolete hack for suns $APLSUB/LMSTR.FOR Remove obsolete hack for suns $INSUNIX/INSTEP1 version date $SYSUNIX/AIPSPATH.CSH,SH OLD/NEW/TST definitions $SYSUNIX/FDEFAULT.SH SUN4: OPT8,9="-O4 -fast" $SYSSUN/CCOPTS.SH PURGE=FALSE, no -pipe $SYSSUN/LDOPTS.SH PURGE=FALSE $APGNOT/BLFIT,JMFIT Remove obsolete hack for suns Some of these were moved from TST (15JAN94). 7674. July 16, 1993 IBLED Eric Changed it to spread the main plot out some when the frame size is small (i.e. small amounts of data or a small selected frame). This will make it easier to see things. Made first frame, next frame, last frame support the selected small frame size as well. Changed IBLED to accept negative values of DPARM(5) as initial guesses for the interval which it will then set. Before, if the actual interval was > 50 seconds the program blew up. Also added test to avoid the blow up when the interval is not determined. Changed the help file for this one. Moved 15JAN94 which is a bit ballsy. 7675. July 19, 1993 Shift and rotation bugs. Bill C. Undo some of the damage I did in attempting to fix the problem with rotation in the gridded uv data modeling and fix other problems. There were previously numerous bugs affecting the computation of the Fourier transform of set of CLEAN components with shifts AND rotation on the sky. These only affected the gridded-FFT modeling method (CMETH='GRID'). The recent incarnation of this series of bugs affected both deconvolution (MX, WFCLN) and modeling (CALIB, FRING UVSUB etc.). -GRDCRM: Add all of XPOFF, YPOFF to the model positions; if these values are correct (not previously true) then the positions in the CC table plus XPOFF and YPOFF should correspond to grid cells. Added a test to check that component positions are on grid cells and to complain loudly (but continue) if not. -ALGSUB,ALGSTB: corresponding changes to those in GRDCRM; i.e. shift is that given in DXCG, DYCG. -GRDAT: XPOFF, YPOFF were being rotated by MAPROT with the wrong sign. -MX: in MXIN MAPROT was not being converted into radians before being fed to the sine and cosine routines. This affected the ra and dec shifts being put into the catalog header. The DFT model calculation in CALIB/FRING was not sensitive to this error but the gridded-FFT routines were. $APLNOT/IMCREA,IMHCOR: same problem with argument to sin/cos as MX. $QYPGNOT/UVMAP: same problem with argument to sin/cos as MX. Moved from 15JAN94 this date. 7676. July 19, 1993 MX restore problem Bill C. Equivalences of local arrays with task buffers in MXCCRM caused the restore step to fail when more than 2048 components were restored. This routine now checks that these arrays are not over flowed. Moved from 15JAN94 this date. 7677. July 19, 1993 ALGSUB Bill C. In the test to see if a row was finished, a floating point ".LT." was used rather than a ".LE." causing erratic behavior in those rare cases where there was an EXACT floating point match. Symptoms observed are the introduction of a series of ripples into the residual image. Moved from 15JAN94 this date. 7678. July 19, 1993 $QOOP/QCLEAN.FOR Bill C. Now uses BEAM size in determining how big to make imaging scratch files. WFCLN was having problems with beams larger than could be made in memory. Also fixed CLCCRM so that it wouldn't overflow internal arrays for ra, dec etc of components. Moved from 15JAN94 this date. 7679. July 19, 1993 $APLIBM/ZMOUN2,ZTPOP2 Pat Added code that supports Exabyte 8500 drives via the same DENSITY setting as on suns (22500 => high density 5gb mode, anything else => 2gb low density mode). This includes a "hack" that permits the 22500 density to be used on 8200 drives, as some 3rd party 8500 drives (TTI, for example) will identify themselves as 8200's. Also inserted code that will report the pid of the owner of the lock on the lockfile, i.e. the process number of any process that has the drive already mounted. This change should probably be moved to the other versions of ZMOUN2 someday. Moved from 15JAN94 just now. 7680. July 19, 1993 $APLDEC/ZABOR2.C Pat The DEC Fortran compiler doesn't need a signal declaration for SIGFPE exceptions. Commented out. Moved from 15JAN94 just now. 7681. July 19, 1993 MK3TX Chris Added code to deal with a new oddity in HP text files: negative record lengths. Examination showed that large negative record lengths were actually the true record length (in 16-bit words) minus 32 769 while small negative record lengths denote the end of file. Moved from 15JAN94 on this date. 7682. July 21, 1993 $APLBERK/ZLOCK.C Pat Forced it to use the "posix" [sic] file locking via routine fcntl() instead of the old flock() call; latter can't tell if a NFS daemon has the file locked for a process on a different host. Moved to 15JUL93 this date. 7683. July 21, 1993 shift bug Eric Old versions of UVMAP and MX put the shift parameters incorrectly into the header but did shift the reference pixel correctly. Unfortunately, modeling routines use the shift parameters. Wrote a new routine to compare the 2 header entries and to correct the header shift parameters for the old bug if that gives agreement between the 2 header representations. It is $APLSUB/FXSHFT.FOR. Changed GRDAT to call it. Moved from 15JAN94 this date. 7684 August 16, 1993 Last-minute fixes Pat As usual, the installation testing on the various architectures revealed many minor but irritating bugs. These are the last changes made to 15JUL93 prior to release. Any additional changes will have to be implemented as patches. Summary: $APLDEC/ZDCHI2.FOR Changed SYSVER to 'ULTRIX 4' $APLDEC/ZTPOPD.FOR key off SYSVER instead of SYSTYP $INSUNIX/INSTEP1 Many hacks, mostly for convex/ultrix brain-dead bourne shells, uname/convex. Probably lots more too. $SYSUNIX/TPDEVS.SH Remove "unset" $SYSUNIX/XASERVERS Ditto $SYSUNIX/MAKEAT Extra sed filter to remove "ENV" in reverse "logical" lookup $SYSUNIX/COMRPL Ditto $SYSUNIX/START_AIPS watch out for sh5 shell on ultrix, remove "echo -n". $APLNOT/STEPCH.FOR Make '' empty string into ' '. $APLDEC/ZTXMA2.C Need __POSIX and _XOPEN_SOURCE defined $APLDEC/ZTKILL.C Ditto $APLDEC/ZTACT2.C Ditto $SYSUNIX/SYSETUP Remove "unset", also HP fix $SYSUNIX/FC Ditto $QYPGNOT/MX.FOR Typo in equivalence statement $APLSOL/ZDAOPN.C APLUNIX version causes error in SunOS 5.2 $SYSSOL/LIBR Shared Library fixes $SYSSOL/LDOPTS.SH Various $SYSSOL/CCOPTS.SH OPT0 set to blank, not "-O". $QYPGNOT/VTESS.FOR Wrong INPARM count (was 198, correct=195) $SYSSOL/ZLPCL2 Use "postprint" as a filter, also asa. $APLHP/SIGNEL.C Split from ZABOR2.C $APLSOL/ZTPWAD.FOR EOF code differs. Same as Sun4 I think. There are other modules changed that are not worth mentioning in detail: $SYSSUN/COMRPL, $SYSUNIX/OPTIMIZE.LIS, $SYSUNIX/AIPS.L, $SYSSOL/COMRPL, $SYSUNIX/FDEFAULT.SH, $APLUNIX/ZMSSRV.C, $SYSSOL/CDVER.*, $SYSSOL/LIBR.DAT, $SYSSUN/LDOPTS.SH. Moved from TST and to the master 15JUL93 tar tapes today. = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =