----------------------------------------------------------------------- ; Copyright (C) 1996-1998 ; Associated Universities, Inc. Washington DC, USA. ; ; This program is free software; you can redistribute it and/or ; modify it under the terms of the GNU General Public License as ; published by the Free Software Foundation; either version 2 of ; the License, or (at your option) any later version. ; ; This program is distributed in the hope that it will be useful, ; but WITHOUT ANY WARRANTY; without even the implied warranty of ; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ; GNU General Public License for more details. ; ; You should have received a copy of the GNU General Public ; License along with this program; if not, write to the Free ; Software Foundation, Inc., 675 Massachusetts Ave, Cambridge, ; MA 02139, USA. ; ; Correspondence concerning AIPS should be addressed as follows: ; Internet email: aipsmail@nrao.edu. ; Postal address: AIPS Project Office ; National Radio Astronomy Observatory ; 520 Edgemont Road ; Charlottesville, VA 22903-2475 USA ----------------------------------------------------------------------- This file must be used to record any changes made to the the master version of AIPS kept in Charlottesville. An entry is: line 1: Date Program name Your name line 2: What was done (may span multiple lines). MAX 72 COLUMNS. line n-1: Where has this change been moved (e.g. 15APR97, nowhere) line n: Blank PUT A TAB IN THIS FILE, AND YOU WILL BE WHIPPED WITH A WET BANANA PEEL! The details of line 1 are important since programs are used to find specified strings and do conversions on them. It MUST be typed as follows: 1 blank, 4 digits, a period, two blanks, then the Month, day, comma, year, some space, a title, some more space, your name. Likewise, the n'th line should be COMPLETELY blank (no spaces). Changes to 15OCT97 *********************************************************************** Changes while 15OCT97 was TST: *********************************************************************** *********************************************************************** 9374. July 9, 1997 $YPGM/PCNTR Phil I got fed up never knowing what p.a. rotation had been applied to a particular image so I added the ROTATE value used to the information in the plot file. Moved to 15APR97 today. 9375. July 9, 1997 $YPGNOT/KNTR Phil While in the mood I decided to remove the irritating message concerning the block being plotted in KNTR. This is not really necessary. Moved to 15APR97 today. 9376. July 9 1997 INDXR Chris Increased the size of the buffer used to sort the IM table and shuffled LUN assignments to avoid clashes with TABSRT. Moved to 15APR97 on this date. 9377. July 9 1997 GPSDL Chris Disabled delay corrections (which are not reliable), leaving only the Faraday rotation correction, and made some changes to the diagnostic output. Moved to 15APR97 on this date. 9378. July 9 1997 $APGNOT/UVMOD Ketan UVMOD now will process compressed data. Moved to 15APR97 on this date. 9379. July 10 1997 LPSEC in APLNOT Chris Split LPSEC from UVFIX to make it easier to find out where leap seconds should be inserted. Also made it more compact added the leap second for July 1 1997. Moved to 15APR97 on this date. 9380. July 10 1997 $APLNOT/BACOOR.FOR Leonia Recently I implemented the subroutine BACOOR which evaluates coordinates of the both ground based and orbiting antennas at the equatorial coordinate system. This subroutine is used at VPLOT and CLPLT for calculating U, V in model for the orbiting VLBI. Having started working on adopting FRMAP for orbiting VLBI I found that speeds of change of the equatorial coordinates need to know at this case. That is why I have modified BACOOR to have the projections of the velocity at the output as well. Moved to 15APR97 on this date. 9381. July 10 1997 $YPGNOT/VPLOT.FOR Leonia COMLNK VPLOT with the new version of BACOOR. At the same time the small bug found by Athol has been fix: In the subroutine SCALVB the points with negative weights were skipped. That is correct if there is no averaging in frequencies. But if there is averaging in frequencies then the data for all selected frequencies were skipped if the weight of the wirst selected frequency is negative. It has been fixed. Now if the first selected frequency has the negative weight it is skipped only if there is no averaging in frequencies (the first selected frequency equals the last one). The individual frequencies which have a negative weight are skipped inside of the routine providing averaging in frequencies (AVGCIF). Moved to 15APR97 on this date. 9381. July 10 1997 $YPGNOT/CLPLT.FOR Leonia COMLNK CLPLT with the new version of BACOOR. Moved to 15APR97 on this date. 9382. July 10 1997 $YPGNOT/FRMAP.FOR Leonia Calculation of the speed change of U, V was correct only for ground based antennas in the system of the coordinates fixed at the Earth. The sky equatorial coordinate system is more preferable if orbiting antennas participate at the VLBI set. The new modified subroutine BACOOR (see #9380) has been used as a part of the routine UUUV to provide operation for both ground based and otbiting antennas. A small bug was found additionally in the case SUBARRAY>1. It is fixed. A relevant change in HELP file has been done. Moved to 15APR97 on this date. 9383. July 11, 1997 $QNOT/GRIDUV Rick/Eric/Pat There was no lower bound check on the chunk of the UV data read into the AP in cases where it would not all fit, and the next visibility was so far down in the second chunk that the conv. fn. did not overlap into the first chunk. This caused negative offsets into an array in such cases with disastrous results. Spotted by Rick, fixed and tested by Eric, implemented here by yours truly. Moved from CVX and to 15APR97 this date. 9384. July 14, 1997 Large file support Athol Implement AIPS support for 64-bit file systems, as currently offered by SGI/XFS and Dec Alpha, primarily for support of Space VLBI data reduction, where files exceeding the current 2.1 GB limit are not uncommon. SUN architectures will be able to support this with the release of Solaris v2.6 later this year. The underlying AIPS I/O model has not been changed. The modifications are confined to: i) using off_t arithmetic in Z C-routines for file size computations; ii) use FCBSP3 to store the file size in the FTAB in extended precision (total of 4 AIPS bytes when taken together with FCBSIZ); iii) use integral double precision for long int type in the FORTRAN parent routines; iv) all access to the extended precision file sizes encapsulated in all C- and FORTRAN routines. Specific changes: 1) encapsulate access to extended precision file size in FCB: $APLUNIX: ZFILSZ.C, ZSETSZ.C, ZCPYSZ.C 2) encapsulate file size computations in FORTRAN routines $APLGEN: ZBLKS.FOR, ZTOTSZ.FOR, ZCMPSZ.FOR, ZADDSZ.FOR 3) Z-routines supporting the new file size types: $APLGEN: ZCREAT.FOR, ZEXIST.FOR, ZCMPRS.FOR, ZEXPND.FOR, ZMSGXP.FOR. $APLUNIX: ZDAOPN.C, ZEXPN2.C, ZFI2.C, ZMI2.C, ZEXIS2.C, ZCMPR2.C, ZCREA2.C (all architectures excluding (BELL/CRI, IBM/3090, VAX VMS)). 4) Modified ZCREAT file size input parameter. Called by: $AIPPGM: FILAI2, FILAIP, FILINI, AIPSC $AIPSUB: AU2, CATCR, SGLAST, SGLOCA, STORES $APLPGM: TAFLG, MOVE $APLNOT: CALCOP, OTBSRT $APLSUB: CATDIR, CATOPN, EXTINI, FILCR, HICREA, MCREAT, RESCAL, SCREAT, TABCOP, TABINI, UVCREA $YSUB: GINIT $APLCONTR: WSLOD 5) Modified ZEXIST file size input parameter. Called by: $AIPPGM: FILAI2, FILAIP, FILINI, AJAX, RECAT $AIPSUB: AU3A, AU8, CATCR $APLDEV: ZEXPND, ZWHOMI $APGNOT: FILLM, MK3IN, FITLD, OTFUV $APLPGM: DISKU, TAFLG, MOVE $APLNOT: CALCOP $APLOOP: UVDATA, UTILITY 6) Modified ZCMPRS file size input parameter. Called by: $AIPSUB: PRTMSG $APGNOT: FILLM $APLNOT: CALCOP $APLSUB: UCMPRS, TABIO $QPGNOT: CVEL $YSUB : GFINIS $APLPGM: PRTAC 7) Modified SCREAT file size input parameter. Called by: $APGNOT: LTESS, STESS, UVSRT, FITLD, IMLIN $APLPGM: PRTPL, TRANS, XBASL, XGAUS, XMOM $APLNOT: CALCOP, GRDSET, ITBSRT, SETGRD, UVDPAD $APLOOP: IMAGE, UVDATA $APLSUB: MAPSNC $QPGNOT: BPASS, CONVL, FFT, GRIDR, IM2UV, IMERG, PCAL, SDGRD, LPCAL, CPASS $QPGM : RSTOR $QNOT : MAKMAP, VISDFT, VSCDFT, UVSORT, VSFDFT, OLDMAP $QSUB : APROLL, UVWAIT $QYPGNOT: APGS, APVC, MX, SDCLN, STEER, UTESS, UVMAP, VTESS $QYPGM : APCLN $YPGNOT: SPFLG, TVFLG $YPGM : BLSUM 8) Modified MAPSIZ file size output parameter. Called by: $APGNOT: LTESS, STESS, IMLIN $APLPGM: XBASL, XGAUS, XMOM $APLNOT: GRDSET, SETGRD $APLOOP: IMAGE $APLSUB: MAPSNC, MCREAT, RESCAL $QPGNOT: FFT, GRIDR, IM2UV, IMERG, PCAL, SDGRD $QPGM : RSTOR $QNOT : MAKMAP, OLDMAP $QOOP : QCLEAN $QSUB : UVWAIT $QYPGNOT: MX, SDCLN, UTESS, UVMAP, VTESS $QYPGM : APCLN $YPGNOT: SPFLG, TVFLG $YPGM : BLSUM These changes require re-linking of AIPS. Moved nowhere. 9385. July 16, 1997 $APLNOT/CALREF Ketan CALREF accepted the FREQOF parameter as a REAL when it is actually sent in as a DOUBLE PRECISION. This effectively rendered some parts of the code in CALREF non-functional. The declaration of FREQOF was changed, and the relevant code was commented out. Moved to 15APR97 this date. 9384. July 16, 1997 $QPGOOP/IMAGR Ketan $QPGNOT/CALIB When operating on a multi-source file, if a single source name was not specified, IMAGR and CALIB did baddd things. Now they warn that they will do baddd things. Moved to 15APR97 this date. 9385. July 16, 1997 $APGNOT/MSORT Ketan In the case when there are no records to sort [the file _is_ in the proper sort order] MSORT incorrectly left an empty file behind. This has been fixed. Moved to 15APR97 this date. 9386. July 18 1997 CL2HF Chris for David Gordon Merged in changes allowing CL2HF to read items from the interferometer model (IM) table for VLBA data. Moved to 15APR97 this date. 9387. July 22, 1997 $HLPFIL/FRMAP.HLP Leonia Change the range of BASELINE to allow negative values. Moved to 15APR97 this date. 9388. July 23, 1997 $YPGNOT/APCAL.FOR Athol/Leonia The task has not provided output for orbiting antenna in some cases. If elevation calculated for orbiting antenna is negative, such an antenna was rejected. Because there is no sence in elevation for an orbiting antenna we excluded the elevation calculation for the orbiting antennas, forcing it to be equal 90 degrees. Moved to 15APR97 this date. 9389. July 25 1997 POSSM Chris Corrected a manifest real constant in a call to BPCOEFF where a double-precision value was expected. Moved to 15APR97 on this date. 9390. July 25, 1997 $APGNOT/SNSMO.FOR & .HLP Athol Corrected a re-referencing bug affecting multi-IF calibration tables for the second secondary reference antenna and above. Subroutine REFFAZ was modifying the table column pointers directly (RF1KOL etc) thus causing problems when REFFAZ was called for secondary reference antennas other than the very first one encountered. Use local variables in REFFAZ to avoid this. Also corrected a typo. in the ref. antenna selection in REFRAT, and the update condition in RFFXFG. Added new subroutine RFFIX to clean up after re-referencing, but this should not be required in normal processing. All the above are long-standing problems. Only the REFFAZ error is likely to have been significant, causing outliers in the smoothed output table. The effect of the other bugs is limited. Also changed BPARM(6), introduced recently to DOBLANK. Moved to 15APR97 on this date. 9391. July 25 1997 UVSUB Chris Rearranged a common block to avoid data alignment problems. Moved to 15APR97 on this date. 9392. July 28, 1997 $APGNOT/UVNOU Ketan At user request, allowed UVNOU to select a rectangle of U and V baselines to flag, instead of just a rectangle of infinite length in V as previously coded. Moved nowhere. 9393. July 28, 1997 $APGNOT/MSORT Ketan In the case when the input file is already sorted, MSORT should not have created an output file, this has now been remedied. Also, the default sort order is now TB [instead of XY]. Moved nowhere. 9394. July 28, 1997 $APGNOT/DBCON Ketan If the user requested a position shift, and the positions agree to better than 3 microarcseconds in RA _and_ in DEC, DBCON silently turned off the position shift - the user is now notified in this event. Also, in the case when position shifts were requested and the input files were compressed, DBCON would refuse to run saying, "CANT SHIFT COMPRESSED DATA" even though it would later turn off the position shift - DBCON now only refuses to run under these conditions _if_ the position shift passes the 3 micro- arcsecond threshold. Moved nowhere. 9395. July 28, 1997 $APLPGM/UVLSF Ketan CHANSEL in UVLSF incorrectly ignored the user settings if it was sent CHANSEL = 0 470 1 670 1000 1 because one of the beginning or ending channel were not greater than zero. This has been changed so that selecting either one results in the other defaulting to 1 or NFREQ when not set. Eg in a 2049 channel data set, setting 0 470 1 670 1000 1 results in 1 470 1 670 1000 1 and 1 470 1 670 0 1 results in 1 470 1 670 2049 1 while 0 0 ... results in CHANSEL being ignored - as before. Moved nowhere. 9396. July 28, 1997 $APGNOT/BPASS Ketan When division by channel 0 is selected using an internally computed channel 0, BPASS says that it wants to disable BCHAN and ECHAN selection - this has been fixed. Previously, BPASS only disabled channel selection if both BCHAN.NE.1 AND ECHAN.NE.NFREQ . Also in this case BPASS incorrectly updated NUMFRQ, potentially screwing up calculation of the 'inner 75% of the band'. Moved nowhere. 9397. July 29, 1997 $APLNOT/FLINI,FLTAB,FLIO Phil As a first step in calibration transfer the VLBA correlator distribution system is now capable of writing flagging and weather tables. The flag tables generated are a little more general than those used with AIPS and require a minor translation. The VLBA FL table format was slightly modified from the 1992 design and so it's handling routines and it's definition routine required updating. Moved nowhere. 9398. July 29, 1997 New task: $APGNOT/SNDUP Phil SNDUP is a new task that will take an SN table from a single polarization uv file and copy it to a multiple polarization file. In doing so it copies the single-pol entries into both R and L columns in the output file. This step is necessary when doing polzn calibration on sources in which Stokes V is not negligible. Moved nowhere. 9399. July 29, 1997 New task: CLFIX Chris for David Gordon Added CLFIX to the APLCONTR (contributed) area and updated ASTROMET.HLP to refer to it. CLFIX adds geometric and atmospheric delay information and clock offsets from IM and MC tables to CL table 1. This is only needed as a stop-gap until INDXR is able to read information from VLBA MC tables. Moved to 15APR97 on this date. 9400. July 30, 1997 FITLD Phil At last, calibration transfer is coming. Since the structure of the FL tables changed slightly FITLD needed minor modifications to handle the translation to FG tables. Hopefully TY and GC tables should come through with few or no changes once they are ready. Moved nowhere. 9401. July 29, 1997 $APGNOT/MSORT Ketan The routine used to accumulate the sort keys mis-counted the number of visibilities in the data file. Fortunately, this bug was only triggered in the case when the last call to UVDISK requested exactly one visibility. Moved nowhere. 9402. July 30, 1997 $APGNOT/UVMOD Ketan A typo in the test of the positivity of a weight sometimes caused an FPE when QUAL = -1. this has been fixed. Moved nowhere. 9403. August 1, 1997 $QPGNOT/SDGRD Leonia The flagging has not worked because DSEL.INC was not incuded in SUBROUTINE SDGRDP. This has been fixed. Moved nowhere. 9404. August 4, 1997 New FFT routines Chris Added a new set of Fourier transform routines to the system. The new routines are mixed-radix transforms and are not restricted to lengths that are a power of two. The first dimension of the transform is unrestricted while the second and subsequent dimensions are limited to FFTSIZ defined in INCS:PFFT.INC due to the possibility that a scratch array is required to gather and scatter values for these dimensions. All of these routines are for complex-to-complex FFTs. The new routines are QCFT1I - initialize 1D FFT QCFT1D - 1D FFT QCFT2I - initialize 2D FFT QCFT2D - 2D FFT QCFT3I - initialize 3D FFT QCFT3D - 3D FFT Only generic versions (in QPSAP) are currently available. The generic versions use the FFTPACK routines CFFTI - initialize FFT CFFTB - inverse FFT CFFTF - forward FFT (all in APLNOT) and are approximately twice as fast as QXFOUR. Adding machine-specific versions that exploit performance libraries supplied by compiler vendors will provide even better performance (eg. the SunSoft perflib FFT routines are 3.5 times faster than QXFOUR for an 8192-point FFT). Moved nowhere. 9405. August 4, 1997 BLING Chris BLING has been reworked to use dynamically allocated memory rather than the AP. This allows it to search extremely large delay-rate spaces as may be required during the HALCA in-orbit checkout. It should be noted, however, that large searches cause excessive page-thrashing and should be avoided unless absolutely necessary (even the Origin 200 can be brought to a crawl for several hours by a full-range HALCA search). The data structures used to accumulate the data will have to be thoroughly reworked in order to address this problem. FRSRCI has been temporarily reintegrated with BLING in anticipation of interface changes being necessary as part of this reworking. Also made the default fringe-detection thresholds less stringent and changed the default behaviour to be NOT to stack data from different baselines (which is dangerous where there are different integration times on different baselines). Moved nowhere. 9406. August 4, 1997 $QPGOOP/IMAGR Ketan Change #9382 was improperly implemented in that IMAGR printed dire warnings even if sources had been selected. The warnings are now only printed IF more than one source makes it past the OOP data selection system. Moved to 15APR97 on this date. 9407. August 6, 1997 $QPGOOP/OMFIT Ketan OMFIT has been rewritten in many ways. Many new models have been included and subjected to user testing. The error-analysis has been considerably improved. This is still development code and problems should be pointed directly to me. Moved to 15APR97 on this date. 9408. August 6, 1997 $QPGOOP/IMAGR Ketan $QPGNOT/CALIB Change 9384 [which merely printed out warnings] is to be back'ed out of NEW pending further testing of the claimed effects. CALIB revision 1.37 therefore backs out the changes in revision 1.36. Moved to 15APR97 on this date. 9409. August 8 1997 FILLM Chris Fixed a bad test condition that would turn off solar mode. Moved to 15APR97 on this date. 9410. August 8 1997 Single-dish OOP Chris/Eric Merged single-dish OOP support from CVX. The following source files are changed: UVDATA.FOR in APLOOP UVUTIL.FOR in APLOOP QUVUTIL.FOR in QOOP The following were added DSDG.INC in INC SDGRID.FOR in QNOT NB: OOP baseline selection in UV2SCR needs to be tested --- this was the most complex part of the merge and I can't find anything that actually uses it. Moved nowhere. 9411. August 8, 1997 $QPGOOP/OMFIT Ketan Fixed a typo spotted by the DEC Alpha and by the IBM compilers. Moved to 15APR97 on August 7th. 9412. August 8, 1997 $SYSUNIX/OPTIMIZE.LIS Ketan The IBM compiler cant handle the number of subroutines in OMFIT. Therefore, the optimization level in OMFIT has been turned way down [off, in fact] for the IBMs. Moved to 15APR97 on August 7th. 9413. August 12, 1997 $YPGNOT/POSSM.FOR Leonia POSSM has not worked if POLPLOT .NE. '' when both requred polarisations do not exist for some antennas. The subroutine POLVIS provides modification of input visibilities to the relevant polarisation ratios (RR/LL, LL/RR....). If one of the required input polarisations absent (flagged), POLVIS has not regected such a visibility as UVGET does at the case of POLPLOT.EQ.''. As a result the logic does not work at MULTPL and POSSM dies with abort. I have added several codes lines after POLVIS to reject the visibility which does not have both required polarisations. Moved nowhere. 9414. August 14, 1997 $APGNOT/JMFIT.FOR, .HLP Leonia Several bugs and problems which have been discovered recentely are fixed. 1. Sometimes solution for position of the gaussian was given as integer. It happenned because the algotithm stoped itteration process after a few itteration probably founding a too small gradient in comparison with given eps=E-14. The problem was by changing type of the solution from real to double precision and changing eps to E-15. 2. Having provided a deep test of the errors offered by Miller a mistake at the error of position for the case DOWIDTH=-1 have been discovered. It fixed and a good aggreement of predicted by JMFIT errors and measured rms was acchieved. 3. Miller asked to give more digits at the position. It's done. 4. The frequency channel number is printed now in the case of multy frequency data (cube). 5. Explanation of TRC, BLC in the HELP file was not clear for the multy frequency data. It is improved. Moved nowhere. 9415. August 15, 1997 FILLM Gustaaf Recently Ken Sowinski changed the scaling definition of data written to tape to allow more dynamic range, and earlier this year I modified FILLM to make proper use of this. Where I should have modified the scaling only for revisions 23 and up I did it for revisions 20 and up (i.e. for all data in the new - post 1988 - format). This caused problems in the autocorrel- ations which were undefined until revision 23. Moved nowhere. 9416. August 17, 1997 $QPGNOT/FRING Ketan A DFILL call should have had a double precision zero but didnt. Moved nowhere. 9417. August 17, 1997 $QPGNOT/CALIB Ketan The CALIB implementation of DOFIT didnt satisfactorily treat the case when the REFANT was also included on the DOFIT list. Now, the REFANT, if it appears on the DOFIT is removed since it is understood by the user there should be no gain corrections for the REFANT. Moved nowhere. 9418. August 18, 1997 FITS readers and writers Chris The FITS readers and writers were applying an optimization that reduced the number of data transfer operations by fusing adjacent columns of the same type (performed in FITTP and R3DTAB). This optimization was performed incorrectly for character data leading to data being lost under some circumstances (eg. the frequency group code in the HF table). The column-fusion optimization has been removed to avoid this problem; in practice, it only led to a small saving in time (a few seconds on slower machines). Moved to 15APR97 patch area on this date. 9419. August 21, 1997 AIPS interoperability Athol First put back of subroutines required for AIPS/AIPS++ interoperability. $APLUNIX/ZXDRST.C - XDR encoding/decoding of strings. $APLUNIX/ZXDRFP.C - XDR encoding/decoding of FP values. $APLNOT/EVTHDR.FOR - Encode/decode AIPS event headers. $APLNOT/EVTPOP.FOR - POPS commands as XDR events. $APLNOT/EVTRDY.FOR - XDR READY events Moved nowhere. 9420. August 25, 1997 FARAD and GPSDL Chris Added code to fix up the azimuth at the intersection of the line of sight with the ionosphere when the source is in the southern sky. Also fixed a transcription error (SIN for COS) in FARAD. Moved nowhere. 9421. August 27, 1997 AIPS++ interoperability Athol Implement an AIPS daemon DAIP.FOR, which can be used to execute the POPS interpreter in a remote/distributed environment, to facilitate interoperability with AIPS++. Communication is via standard IPC sockets with XDR network encoding. DAIP shares the available interactive POPS numbers with AIPS. Specific changes: $SYSUNIX: START_AIPS - implement new parameter NOEX, which is passed to AIPSEXEC and defers AIPS execution. Can be used to start up only the AIPS servers. AIPSEXEC - see START_AIPS. $APLUNIX/ZACTV9.C - treat AIPSx and DAIPx equivalently. $APLGEN/ZWHOMI.FOR - allow AIPSx or DAIPx. $AIPGUNIX: ZSTRTA.FOR - check for DAIPx in selecting a free POPS no. ZSTRTD.FOR - analog of ZSTRTA to start DAIP.EXE. $AIPPGM/DAIP.FOR - analog of AIPS.FOR with modifications in obtaining the user no. and reading POPS commands. $APLSUB: Move EVTPOP.FOR, EVTHDR.FOR and EVTRDY.FOR to $APLSUB from $APLNOT (see #9419). Moved nowhere. 9420. August 28, 1997 ZMOUN2.C for Alpha Pat Changed the density logic so that it conforms with the OSF/1 4.x definitions of Exabyte tape drives. The medium density device (/dev/nrmt#m where # is 0, 1, 2, ...) is what should be used for 8500 mode on Exabytes and the low density device (/dev/nrmt#l) for 8200 mode. The other modes may be compressed, and such compression may be non-portable (definitely if user doesn't have a compression-enabled tape at their home institution). Moved nowhere, will move to CVX after testing. 9421. August 29, 1997 $AIPSPUBL/AIPSMEMO95.* Athol Note on AIPS/AIPS++ interoperability. Moved nowhere. 9422. September 2, 1997 $RUNSYS/NEWPARMS.001 Ketan $HLPFIL/POPSDAT.HLP $HLPFIL/SPARM.HLP Introduced a new adverb SPARM(10) which is an array of STRING*8 s. This adverb is used in ANTAB for the implementation of #9423. Moved nowhere. 9423. September 2, 1997 $APGNOT/ANTAB.FOR Ketan $HLPFIL/ANTAB.HLP ANTAB crashes if it encounters information in the Tsys file for an antenna that does not appear in the AN table. The new adverb SPARM allows the user to specify antennas for which there may be entries in the Tsys file but which may not appear in the AN table. Tsys information for such antennas will be ignored. This was introduced for SVLBI data reduction. Moved nowhere. 9424. September 2, 1997 $HLPFIL/ALIAS.HLP Ketan When ALIAS was introduced way back in #9338, I forgot to add a HLP file for the new adverb ALIAS. Moved nowhere. 9425. September 2, 1997 $APLSUB/TABINI.FOR Ketan Athol pointed out that the variable DLSIZE was uninitialized in TABINI. This will necessitate relinking large parts of AIPS [which I found by using WHOUSES TABINI] Moved nowhere. 9426. September 2, 1997 $QPGNOT/FRING.FOR Ketan Jim Ulvestad pointed out that the exhaustive fringe-search mode silently discards solutions when the chosen REFANT does not appear in the solutions; this did not happen in the traditional modes of FRING. This was a one-liner. Moved nowhere. 9427. September 5, 1997 $APGNOT/OBTAB.FOR Leonia The task OBTAB having been tested with simulated orbiting VLBI data has not worked with actual orbiting VLBI data at the mode of updating antenna table. The reason of this is because the correlator has not labeled orbiting antennas at the AN table (MNTSTA=2), has not numerated antennas in OB table and has not provided the space for 6 orbiting parameters at AN table. While the correlator has these problem, I modifyed OBTAB to circumwent the problem for the case of the only orbiting antenna - the actual case for a while. The new parameter APARM(8) determines the number of the only orbiting antenna. The task put MNTSTA=2 for this antenna at AN table, put this number at OB table together with main modification of the tables. Moved nowhere. 9428. September 8, 1997 TABCS in APGNOT Chris Corrected the calculation of the number of IFs to use column 6 (gain factor) rather than column 5 (subarray) which is not dimensioned according to the number of IFs. Moved nowhere. 9429. September 8, 1997 MC Table Routines Chris Added MCINI to initialize a VLBA model components (MC) table and GETMC to locate atmospheric group-delay and clock offsets in an MC table; both APLNOT routines. Moved nowhere. 9430. September 8, 1997 TABLEUTIL in APLOOP Chris Added new routines. OIMINI is an object-wrapper for IMINI. OGTDEL is an object-wrapper for GETDEL. OMCINI is an object-wrapper for MCINI. OGETMC is an object-wrapper for GETMC. Moved nowhere. 9431. September 8, 1997 INDXR Chris Added the capacity for merging atmospheric delay and clock offset information from VLBA model-components (MC) tables into newly created CL tables when MC tables are present. This required a thorough rewrite to bring the table handling logic to a higher level to allow coordination of LUN usage when sorting tables. INDXR is now an AIPS OOP program (although this turns out not to have helped as much as I expected). The logic governing CS table generation has also be slightly modified to ensure that CS table records are generated whenever new beams are encountered during a scan. Moved nowhere. 9432. September 8, 1997 $APGNOT/UVFIX.FOR, HLP Leonia The task UVFIX has been modifyed to allow handling orbiting VLBI. The calculation of X,Y,Z at the RH sky coordinate system (RA=0 at X) is provided separetely for each antenna: ground based and orbiting. The baseline formed as the coordinate difference is recalculated to the U, V, W of the given source. X,Y,Z of the orbiting antenna is calculated by fitting polynoms of 4-th degree to the X,Y,Z having taken from OB table. The OB table data are devided for intervals of 100 sec each. It is supposed the time interval at the OB table is 10sec. Different polynoms are fitted at the different intervals. The coefficients of the polynoms as well as the begins and ends of the intervals are stored for following calculation X,Y,Z of the satalite at the arbitrary time. Two new parameters are added: UVFIXPRM(15)-the OB table number; UVFIXPRM(16)-orbiting antenna number. The relevant change of the .HLP file is carried out also. The old .HLP file erroniously informed that UVFIXPRM(2)=0 corresponds to subarray=1. Indeed it corresponds to all subarrays. It's fixed as well. Moved nowhere. 9433. September 9, 1997 $QPGNOT/FRING Ketan Fixed a typo from #9426 and also introduced a new feature. DPARM(8) can now be used to zero rates, single-band delays, or phases _after_ FRING computes its solutions. This obviates a step in the data reduction involving running SNCOR and is useful for the GEODOSY/ASTROMETRY folks. Moved nowhere. 9434. September 9, 1997 UTESS Chris Added code to protect the CATBLK for the output image file when opening the default image. This prevents confusion over the existance of history files during file creation. Moved nowhere. 9435. September 10, 1997 CLBOX help file Chris Rewrote help file for CLBOX adverb to cover the use of circular boxes in IMAGR, SCMAP and WFCLN. Moved nowhere. 9436. September 15, 1997 $APGOOP/VCOPY Ketan VCOPY did not respect the compression state of the input file. This was due to my not properly understanding Chris' implementation in OOP. Moved nowhere. 9437. September 15, 1997 $QPGNOT/FRING Ketan When using the exhaustive search mode of FRING (APARM(9)>0), the SNR of the first solution found in the FFT stage was improperly down-weighted by a factor of 4. This was caused by a typo, I divided by 2 instead of multiplying by 2. This had no effect if APARM(9)=0 [the usual case]. Moved nowhere. 9438. September 16, 1997 $QPGOOP/IMAGR Leonia IMAGR has not worked if number of averaging points exceeded 28. It occured because array PBFREQ was declared with dimension MAXIF=28. It must be declared with dimension MAXCIF=8192. The two subroutines PBFSUB and PBFDIF are located at the module $QOOP/PBUTIL.FOR. I fixed the error at the module. Moved nowhere. 9439. September 16, 1997 BLING Chris Fixed a stupid error that caused the FFT array not be cleared between acceleration steps and therefore causing the acceleration search to fail. Moved nowhere. 9440. September 16, 1997 FILLM Chris Increased the number of FQ entries allowed to 250. Moved nowhere. 9441. September 16, 1997 SGI Printing Chris Removed vestigial APLSGI versions of ZLASC2 and ZLPCL2 that were leading to a glut of undeleted files in /tmp. Moved nowhere. 9442. September 18, 1997 REGISTER Pat/Ernie Added auto-version-detecting logic instead of a fixed VERS= string; also added a new keyword about the aipsletter in the code to help auto-generate the AIPSletter mailing list (paper). Moved nowhere. 9443. September 18, 1997 $SYSUNIX/OPTIMIZE.LIS Pat Bumped FITTP down to OPT0 for ALPHA systems. On some datasets it was generating spurious "floating invalid" errors that vanished when the task was recompiled and relinked with no optimization (instead of the default OPT2 for that architecture). This change will only affect people with DEC Alpha systems. Moved nowhere. 9444. September 19, 1997 $APGNOT/UVFIT Pat A report from Michael Bietenholz at York University showed a bug in UVFIT in 'SPHE' mode: "It doesn't zero the counter used to calculate the rms/point in routine UVRMS2, so you get the right answer at the start but the wrong one at the end". Fixed. Thanks to Michael for providing the fixed code. Moved nowhere. 9445. September 22, 1997 $YPGNOT/POSSM Ketan POSSM did not respect timerange selection/solint specification. If the specified SOLINT was longer than the chosen timerange, POSSM effectively extended the timerange to cover the SOLINT. Now, if the selected timerange is too short, SOLINT is shrunk to fit. Moved nowhere. 9446. September 22, 1997 FITLD Chris Added code to transfer atmospheric delays and clock offsets from the VLBA MC table to the initial CL table. Also reordered table translation so that frequency tables are processed before antenna table so that the output AN table reference frequency is consistent with the data. Moved nowhere. 9447. September 22, 1997 Low-Level Networking Chris Added a set of Z-routines for low-level networking in anticipation of splitting SLIME into a GUI portion and a data server. These routines are thin bindings to the core UNIX socket routines (which are also part of the win32 API). The new routines are ZSACPT - wait for a connection from a client (accept) ZSBIND - bind a socket to a network address (bind) ZSCLOS - close a network connection (close) ZSCONN - establish a connection to a server (connect) ZSCREA - create a socket (socket) ZSLISN - put socket into "listen" mode (listen) ZSREAD - read data from network (read) ZSWRIT - write data to network (write) Versions are in APLUNIX and APLGEN (the latter stubbed). These routines are more general than those previously available in AIPS in that they will allow AIPS tasks to function as network servers where previous routines always assumed that AIPS tasks would be clients of some external server. Moved nowhere. 9448. September 22, 1997 $APGNOT/PCCOR Ketan The maximum number of frequency channels processed in subroutine PCCAL in PCCOR was hard-coded to 256. This has been changed to MAXCHA. Moved nowhere. 9449. September 24, 1997 DATGET and DGETSD Chris Modified DATGET and DGETSD not to ignore fully flagged data records when flagging is turned off using FGVER = -1. Data could still be flagged using negative weights in the data and this was causing problems when records needed to be counted (eg. in INDXR). Moved nowhere. 9450. September 24, 1997 $QPGNOT/FRING Ketan The subroutine WANSRC in FRING did not properly implement source matching when the user was trying to deselect source using a '-' symbol in the CALSOUR list. This caused problems when trying to determine scan boundaries in the presence of an NX table. Moved nowhere. 9451. September 29, 1997 $SYSLINUX/UPDCONFIG Pat Fixed it so that it can be used both on power.aoc and orangutan.cv. Any other Linux midnight job sites need to let me know so I can include them too. Moved to AOC this date, tested, works. 9452. September 30, 1997 $APGNOT/UVFIX.FOR Leonia In the situation of three different vectors: ground based antennas, orbiting antenna, the source, it is necessary to have all of them at the same coordinate system. The orbiting antenna's position is given at the OB table at the standard (J2000) frame. The source can be recalculated to the standard frame also. So we need to recalcu- late the vectors of ground based antennas. The routines provided the recalculation take care of rotation due to precession and nutation as well as other effects such as abberation (annual and diurnal), the general relativistic bending of light by the Sun. I applyed the routine to the ground based antennas to provide only rotation and applyed the routine to the source to correct its position due to other (not rotation) effects. To implement this method I modified the precesion routine to split rotation and not rotation effects. Moved nowhere. 9453. October 3, 1997 $QPGOOP/GPHAS Ketan GPHAS is finally getting a work-out on real SVLBI data. The first problems were: a sign error on the phase of the stacked baseline, user complaints about the help file, and complaints about the implemented method of refant and special antenna selection. Moved nowhere. 9454. October 7, 1997 $APLDEC/ZFRE2.C Pat Jeff and Rob actually uncovered this bug while porting to the Linux Alpha system. The size parameter in the statfs call was never initialised, causing abrupt terminations (and probably lots more other nasties as well). Modified the routine to remove the calloc call, just use a local buffer for the contents of the statfs call, and get the size parameter right. Moved nowhere. 9455. October 9, 1997 $APGNOT/UVFIX.FOR Leonia The effect of differential aberation (Lorentz contraction) has implemented in the UVFIX. This effect is visible in a precise astrometry. Long time this effect has been treated by an outside AIPS routine after coordinates of the sources had been found by JMFIT. It was inconvenient. Now this effect is taken into account by UVFIX and the astrometry experiments can be consummated in AIPS. Moved nowhere. 9456. October 14, 1997 Mk3 VLBI Utilities Chris/Walter Alef Merged format updates into MK3IN. Also added the new tasks M3TAR, which reads Mk3 data in UNIX tar archives, and TFILE, which sorts and edits Mk3 AFILEs for M3TAR. Moved nowhere. 9457. October 21, 1997 $APGNOT/PCCOR.FOR Leonia PCCOR.FOR has been modifyed to handle both USB and LSB. The phase of PC tones of PC table are minused for the IFs with LSB. The tones itself were rearanged in encreasing order of frequencies for the IFs with LSB. The task is improved to handle the cases when some antennas are not presented at the PC table or at the calibrator's scan for the selected time range. The relevant add to the help file has been made. Moved nowhere. 9458. October 28, 1997 System stuff Eric (1) Add HP2 and SUL to system files: UPDREMOVE, LIBR, START_TVSERVERS, TVALT, XASERVERS, ZLASCL, ZLPCL2, START_AIPS (2) Changed $SYSSOL CCOPTS.SH and LDOPTS.SH to test for $ARCH and do Ultra-specific if SUL. (3) Replaced 15OCT97 version with 15APR98: UPDCONFIG Previous was bare bones, new is good example. UPDCONTROL Fixed order of UNQ building versus special testing and use MasterArch. FDEFAULT.SH Adds HP2, SUL, corrects HP. LINK Corrects export of LDOPTS. Remove $SYSHP/LINK OPTIMIZE.LIS Corrects HP list, adds SUL, HP2. Moved nowhere. 9459. November 7, 1997 AREAS etc. Pat Modified AREASSH.SED to produce less baroque AREAS.SH when the AREAS script is run. Also modified UPDAREAS for the midnight jobs, and the putbck script in the checkout system to make the business of regenerating AREAS.SH and .CSH from the master AREAS.DAT file more rational. When you putback the AREAS.DAT file henceforth, it will automatically check out the AREAS.CSH and .SH files from SYSUNIX, discard the clear text copies, regenerate them from the freshly put back $SYSAIPS/AREAS.DAT, check these back into the RCS system and put entries in the transaction file for them. When UPDAREAS in a midnight job kicks in and sees the change in a job where UpdDoCopy is set to YES (i.e. not accessing files via NFS), it checks that AREAS.SH and .CSH were put back also. If not, it tries to regenerate the files locally via AREAS (and issues dire warnings). Moved from 15APR98 this date. 9460. November 7, 1997 Sundry Pat I've been a bit lax about reporting what I've done lately. These items have been addressed in the past couple of weeks. $SYSSUL/INCS.SH - use INCSOL and INC, nothing more. $UPDUNIX/MAKE.MNJ - change to use UPDUNIX/UPDCONFIG as the master config file for all MNJ's. $UPDUNIX/AIPSUPD - ditto, use UPDUNIX/UPDCOPYSUB too. $UPDUNIX/UPDSRTUNQ - use older sort flags for SUN4. $UPDUNIX/UPDCONFIG - add all known MNJ sites. $SYSUNIX/ALLRSH - allow ssh in place of rsh. $SYSUNIX/COMRPL,LNK- RCS changes to use kochab, not baboon. $SYS*/CDVER.*SH - fix CVX references. Moved from 15APR98 on various dates. 9461. November 7, 1997 FITLD Chris Prevented FITLD from incrementing day numbers in the OB table (which uses an MJD instead of a relative day number) when midnight boundaries are crossed. Any time greater than an arbitrary limit of 1000 days is now assumed to be an MJD and is left alone. Moved from 15APR98 on this date. 9462. November 12, 1997 PCCOR Leonia Put all recentely modifications added to TST to the NEW. is left alone. Moved from 15APR98 on this date. 9463. November 18, 1997 BLING Chris Removed a superfluous test on the dump interval that was causing time differences between data points on different baselines to be interpreted as unusually short dump intervals on a single baseline under rare circumstances. Also added a check for dump intervals less than APARM(1); this condition was not caught before and would lead to memory access errors. Moved from 15APR98 on this date. 9464. November 20, 1997 M3TAR Chris Modified M3TAR to handle null-terminated strings in from tar headers (reported by Michael Wunderlich with a different solution). Moved from 15APR98 on this date. 9465. November 30, 1997 PCLOD Leonia Put all recentely modifications added to TST to the NEW. is left alone. Moved from 15APR98 on this date. 9466. December 2, 1997 INDXR Chris INDXR now allows the user to supply a text file that specifies times at which scan boundaries must occur. This is mainly intended to force scans to be broken when clock setting events occur at Haruka ground stations. Moved from 15APR98 (with subroutine APLSUB:DATEST) on this date. 9467. December 2, 1997 $YPGNOT/IRING Ketan The various buffers used inside IRING did not have consistent lengths and were restrictively small so that IRING failed on large [>2048x2048] images. This has now been fixed. Moved from 15APR98 [#9625] on this date. 9468. December 3, 1997 UVINIT Chris for Athol Modified UVINIT to handle files larger than 2 AIPS Gbytes when appending. Uses new APLGEN:ZMODSZ routine. Moved nowhere. 9469. December 4, 1997 VLBDDT Chris Set INFILE to blank before calling INDXR. Uses DTSIM from specified VERSION rather than TST. Moved from 15APR98 on this date. 9470. December 4, 1997 BPASS Chris Changed test on NSIZE to use two polarizations irrespective of how many are actually present since BASOLV always partitions its arrays for two polarizations. Added warning messages that solutions may be ditributed over several BP records when solutions can not be obtained in a single pass. Added code to override user specified BCHAN/ECHAN and put out a warning message since BCHAN > 1 leads to BP solutions being shifted in frequency. Corrected the beginning and ending correlator numbers for NORMBP in the case where only LL data is present. Moved nowhere. 9471. December 4, 1997 COHER Leonia The new version of COHER in contrast to the old one can determine coherence time in the case of non zero fringe rate. That allows to use it to predetermine the solution interval for FRING. Coherence time is determined by comparing vector and scalar averaged amplitudes. Vector averaging is provided by fitting a complex exponent by non lenear least square method. The first approach for the least square is given by FFT. The output as tables of found coherence times of different baselines and antennas for different time intervals and sources, can be displayed at the terminal or printed at printer or sent to the file under control of DOCRT and OUTPRINT. Moved nowhere 9472. December 8, 1997 $HLPFIL/ASTROMET.HLP Ketan $APLCONTR/CLFIX.HLP $ALPCONTR/CLFIX.FOR $APGOOP/CL2HF.FOR $APGOOP/HF2SV.FOR The long-awaited geodesy update is in. In the absence of agreement between the CfA, GSFC, and NRAO geodesy 'experts' on tap, the above routines have been checked in as is. ASTROMET.HLP has been changed to reflect new capabilities in FRING, the new procedure REFREQ, and changes to the frequency handling in FITLD. CLFIX.HLP now mentions that INDXR supplants it. CL2HF.FOR and maybe CLFIX.FOR need improvement. HF2SV changes are cosmetic. Improvements will have to wait until the next AIPS release. Moved to 15APR98 as of this date. 9473. December 9, 1997 ZMOUN2 for SOL, ALPHA Pat Added support for DLT 7000 drives. CAUTION: make sure you have the device set up with densities as described in the comments of $APLDEC/ZMOUN2.C and $APLSOL/ZMOUN2.C; there are three different possible densities, added to which compression gives six possible configurations, with only four of these accessible at one time via different device names. Moved from 15APR98 today. 9474. December 9, 1997 Compile/Link Scripts Pat/Chris Ported part of Chris's 15APR98 change to these (#9590) so that the STRIP option is now supported. This involved all LDOPTS.SH scripts, as well as COMLNK and LINK. Moved from 15APR98 today. 9475. November 19, 1997 TVFLG, SPFLG Eric Found a bug that caused the rolling buffer average to average across source names. This affected the AMP V DIFF and similar modes, causing excess values at the ends of scans. Sigh... December 7, 1997 TVFLG, SPFLG Eric The gridding subroutine $APLNOT/GRIDTB had an error causing it to misgrid data one time row too high under some circumstances when doing averaging. In particular, 5-second samples with a 10-second averaging would encounter this problem. Then, the times appearing in the flag table would not necessarily apply to the bad data. This error only arose when averaging on input, not when averaging for loading to the TV. Fixed a GT test to GE to correspond to a later test on row number. Also fixed the handling of times - they were rounded to integer intervals of 1/11 of the averaging time. This caused some times in the above case to appear to be in the next row (exactly at the bottom) when their times actually placed them in the previous row. Dropped all the interval business. SPFLG had the GE bit right, but did make the error of integerizing the data times. Moved (corrections only) 8 December to 15OCT97 versions. 9476. December 8, 1997 AIPS.FOR Eric Fixed copyleft date to 1997 in displayed copyright. Moved from 15APR98. 9477. December