XAS release 1.2 BUG LIST - this file opened 10 Apr 1996 this file closed 19 Feb 1997 ------------------------------------------------------------------------ Bug # : 1 Date : 15 Apr 1996 Status : Fixed with modifications of PCF and packetcap file Module : PCF and packetcap file for MECS HK Title : (NOT A BUG) finalised support to FOT HK (MECS) The MECS $XASTOP/calib/sax/mecs/mecs.pcf has been updated adding the description of all MECS FOT HK parameters, inclusive of the new cali- bration curves supplied by G.LaRosa (which should go in Laben SUM issue 3 when released). Contextually the GROUND HK parameters have been renamed (names with a G prefix) and a minor bug fixed (po#4 -> po#6 for GRM1/2/3). THEY SHALL BE RENOVED IN THE FUTURE IN THE EXTERNAL DISTRIBUTION. The MECS $XASTOP/calib/sax/mecs/saxmecshk.packetcap file has been modified accordingly for support to FOT ratemeter and HKD packets (also to support changes in bug #2 below) ------------------------------------------------------------------------ Bug # : 2 Date : 15 Apr 1996 Status : fixed with program and routine upgrade/addition Module : program SAXHKACCUM and routines in FOTLIB Title : (NOT A BUG) finalised support to FOT NFI HK Support to ENG ratemeters and payload HKD for NFI has been added. This complements XAS 1.1 bug #23. The modifications are as follows : In program SAXHKACCUM 0.6 - added support to calibration curves given by points (and described in PCF ; this is the case of NFI temperatures, also with an adhoc INTERPOLATE routine - extended some buffers to cope with consequent longer PCF entries - added elements to HKCOMMON (for secondary type 4, see below) - the latter two apply also to routines SAX_PCF_LOAD and SAX_PCF_LOOKUP - added PCF lookup for parameters of secondary type 4) - other changes for support of pseudosupercommutation. Also the following routines in fotlib have been modified : a) SAX_ACC_BT_3 0.2 to support new secondary types 3 and 4 b) secondary types 1 and 2 are now reserved to GROUND data c) SAX_ACC_B3S3 added to support FOT ENG ratemeters d) SAX_ACC_B3S4 added to support FOT payload HKD e) SAX_WHICH_DATA 0.4 now compliant with FOT packet names Note that (e) will apply to all accumulations ------------------------------------------------------------------------ Bug # : 3 Date : 10 Apr 1996 Status : **provisionally** fixed with routine version 0.2 Module : routine SAX_OPEN_DIR in library FOTLIB Title : no check on slew code A provisional "message warning and return" handling is taken in the case a wrong "slew code" is given via XASSET SLEW to a code other than one of the legal ones (Normal, Initial or Final) ------------------------------------------------------------------------ Bug # : 4 Date : 15 Apr 1996 Status : fixed with routine upgrade Module : program SAXHKACCUM and routines in FOTLIB and XASLIB Title : (NOT A BUG) streamlining of (HK) accumulation a) routine PKTCAP_LOAD (in XASLIB) now gets a free logical unit via FREE_LU and b) same applies to SAX_PCF_LOAD within SAXHKACCUM c) compliance of SAX_PKTCAP_LOAD with FOT packet names verified d) created include file HKCOMMON.INC e) modified all SAX_ACC_B3S* to support it. Note that(a) will apply to all accumulations ------------------------------------------------------------------------ Bug # : 5 Date : 19 Apr 1996 Status : fixed with program and routine upgrade/addition Module : program SAXHKACCUM and routines in several libraries Title : (NOT A BUG) support to times in UT Program SAXHKACCUM 0.7 now supports dialogue for time limits and bin size in UT and seconds, and creation of time profiles with times as elapsed seconds from 0 UT of day of observation start. This has involved also modifications or additions of new subroutines of general use. We list here the ones of specific nature, and in the next bug entry the ones of general nature. a) SAX_HK_RANGE 0.1 remains in source saxhkaccum.f so far. Has support to UT via new calls only for FOT data. Still contains the routine for ground data GET_FIRST_AND_LAST (to be removed a' regime) b) GET_START_END (in FOTLIB) is a new routine to get start and end defaults from instrument directory c) SAX_PCF_LOAD and SAX_PCF_LOOKUP moved to library FOTLIB d) SAX_OPEN_DIR 0.3 (in FOTLIB) has been streamlined e) SAX_ACC_B3S3 0.1 and SAX_ACC_B3S4 0.1 contain the minimal change necessary to handle a single time recycle (in FOTLIB) f) ASKTIME and ASKBIN (in XASLIB) are the routines handling the dialogue for start/end time and bin size in UT/seconds. g) REARRANGE_INSTREC (in FOTLIB) is a routine to make instrument directory records suitable for Fortran list-directed READ h) include file HKCOMMON.INC is now a self standing file (used for HK accumulation definitions_ i) new include file OPCOMMON.INC contains definitions for observation or observing period start and end times ------------------------------------------------------------------------ Bug # : 6 Date : 19 Apr 1996 Status : fixed with routine additions Module : several routines in various libraries Title : general purpose routines moved to library a) INTERPOLATE (in GENERAL) is a new routine for interpolation with included search in arrays of X and Y data of the interval to be used. b) HEXI4 (in GENERAL) is an hex to integer conversion routine c) SAX_DF_KEYWORDS (in FOTLIB) writes standard keywords (so far SATELLIT INSTRUMENT, OBSERVER and OBJECT) ------------------------------------------------------------------------ Bug # : 7 Date : 22 Apr 1996 Status : fixed with LB_NUMBER version 0.2 Module : routine LB_NUMBER in library XASGRAPH Title : format error using H format The routine crashed with a conversion error when formatting a time of the form H M S (with S real) using all integer I formats. The error affected format labelled 103 (vestigial of HP1000 installation). The format for seconds has now been been replaced with an F format. ------------------------------------------------------------------------ Bug # : 8 Date : 22 Apr 1996 Status : fixed with TPLOT 0.2 Module : program TPLOT Title : (NOT A BUG) better look and handling of UT in TPLOT Program TPLOT has been improved (also inserting in it two dedicated variants of pre-existing XASGRAPH routines) in several senses and is now controlled by some additional environment variables> a) the time axis is now labelled correctly (no asterisks) in elapsed seconds, or, if XAS environment variable TIMEAXIS is set to Hours, in elapsed hours (in form hh:mm:ss) b) the datastyle can now be selected or defaulted in a more sensible way (error bars if errors present in the file, solid or histogram otherwise). Still there is no gap handling. c) the choice of the quantity to be plotted in Y (defaulting to the only data band) may be set to deadtime (if present), to the "other" data column, or to the n-th data band setting the XAS variable TQUANTITY to Deadtime, Other or number n> ------------------------------------------------------------------------ Bug # : 9 Date : 26 Apr 1996 Status : *** partially *** fixed with new programs and routines Module : various programs and routines Title : (NOT A BUG) support to time windows The following programs and routines have been added for support to time windows (creation and handling) : a) TWINDOW, program to create time windows b) IWINDOW, program to create intensity windows c) MERGEWINDOW, program to merge time window files d) WPLOT, program to overplot time windows *) it is intended that programs (a) and (b) are called via a dispatcher MAKEWINDOW, and (d) via DISPLAY and OVERTRACE but this is not yet implemented e) added routines OPENWINDOW and LOADWINDOW to librarty XASLIB f) added routine Y_FILL to graphical library GRAPHSERV ------------------------------------------------------------------------ Bug # : 10 Date : 26 Apr 1996 Status : fixed moving routine to library Module : routine OPEN_TIME in library XASLIB Title : (NOT A BUG) routine OPEN_TIME ,moved to library The routine OPEN_TIME, formerly contained within the source of TPLOT, has been moved to library XASLIB without change. ------------------------------------------------------------------------ Bug # : 11 Date : 26 Apr 1996 Status : fixed with program upgrade Module : program SAKHKACCUM and related routines Title : (NOT A BUG) support to time windows in SAXHKACCUM Program SAXHKACCUM now supports accumulation via a time window file specified as XASSET TIMEWINDOW file. Related modifications include the new XASLIB routine INIT_TIMEWINDOW and modifications to the FOTLIB routines SAX_ACC_B3S3 and B3S4. ------------------------------------------------------------------------ Bug # : 12 Date : 26 Apr 96 Status : ** fixed ? on May 10-16 with version 0.2-3 of MERGEWINDOW Module : program MERGEWINDOW Title : merging of complex windows does not work MERGEWINDOW fails in merging correctly two series of time windows when they overlap in a complex way (i.e. a window in one series overlaps with more than one in the other series). Correction by Stefano Signorile of SDC, replaced sorting routine and done some fixes in subroutine WORKHORSE ------------------------------------------------------------------------ Bug # : 13 Date : 03 May 96 Status : (MG) fixed with version 0.1 of REARRANGE_INSTREC Module : routine REARRANGE_INSTREC in library FOTLIB Title : instdir record rearrangement fails on SunOS The original version of REARRANGE_INSTREC used "in place" substring assignments with left-shifts (e.g. like INSTREC(75:)=INSTREC(71:) ),i.e. an assignment where the left and right-hand sides (lhs, rhs) are substrings of the same string. This kind of assignments works on Ultrix and HP-UX but not under SunOS. In order to have a common version, a copy of the original string is made, and the rhs of the assignment is taken from such copy. ------------------------------------------------------------------------ Bug # : 14 Date : 07 May 1996 Status : (DDF) fixed with version 0.1 of MERGEWINDOW Module : program MERGEWINDOW Title : Typo in subroutine call One of the calls to the subroutine OUT had a spurious comma in the argument list. This has now been removed. ------------------------------------------------------------------------ Bug # : 15 Date : 07 May 1996 Status : (MG) fixed with version 0.1 of GET_START_END Module : routine GET_START_END in library FOTLIB Title : wrong (partial) read of instdir records on SunOS The second internal READ in this routine used dummy integer arguments to skip the start time of the last observation and to read only the end time. However one of the list items (hex string) is a character string. This mismatch on SunOS causes the remainder of the record to be unread. A dummy CHARACTER variable is now correctly used instead of an INTEGER ------------------------------------------------------------------------ Bug # : 16 Date : 07 May 1996 Status : fixed with SAXHKACCUM 0.9 Module : program SAXHKACCUM and library XASLIB Title : HISTORY keyword was wrong or incomplete The filename was missing from the HISTORY reconstructed run string. It has been added. Also the start and end times in HISTORY were the "default" ones and not the user supplied ones. A new XASLIB routine UPDATE_START_END has been written to update the correct values in OPCOMMON. Also SAX_ACC_HKRANGE caused the drop of the last bin, now fixed. The program has also been aligned to the new pointer ACCUMCOMMON_TINDEX described below. ------------------------------------------------------------------------ Bug # : 17 Date : 07 May 1996 Status : fixed with program/routine upgrades Module : program SAXTACCUM and several routines Title : (NOT A BUG) major upgrade of SAXTACCUM Program SAXTACCUM 0.3 has been modified similarly to SAXHKACCUM for the support of times in UT and time windows This implied also the following changes to library members : a) added a variable in ACCUMCOMMON to point to the TIME field in the event. b) support to UPDATE_START_END routine c) XASLIB routine ASKBIN has been modified to make bin size calculation as real and not integer (necessary for bins lesser than 1 sec) d) FOTLIB routine SAX_ACC_RANGE now calls ASKTIME and ASKBIN for the time range dialogue in UT. Also the drop of the last bin was fixed. e) FOTLIB routine SAX_WHICH_DATA now returns an extra argument (the number of "eligible" packets), used in main for the HISTORY kwd. NOTE THAT THIS CHANGE REQUIRES A MODIFICATION TO THE CALLING SEQUENCE IN PROGRAMS SAXxACCUM with x=G,H,HK,P. f) FOTLIB routine SAX_ACC_PRELOAD also sets up the new pointer TINDEX in ACCUMCOMMON (described at item a). g) FOTLIB routine SAX_ACC_B1S1_Y has support for single time recycle and time windows ------------------------------------------------------------------------ Bug # : 18 Date : 09 May 1996 Status : fixed with new version of programs Module : programs TPLOT, WPLOT and IWINDOW Title : (NOT A BUG) time profile programs aligned to SAXTACCUM changes Program TPLOT 0.3 has been aligned to the fact that SAXTACCUM may produce files with INTEGER "elapsed times" not in seconds, but with 2**-n sec resolution. If the TIMERES keyword is missing the old 1 s resolution is assumed. Note that the program has NOT been aligned to the fact that the keyword BINSIZE may now be REAL, since this is virtually never used !! Also IWINDOW 0.1 and WPLOT 0.1 are able to handle elapsed times in fraction of seconds ------------------------------------------------------------------------ Bug # : 19 Date : 10 May 1996 Status : fixed with new version of programs Module : programs SAX*ACCUM Title : changed calling sequence of SAX_WHICH_DATA Refer to entry #17 item e above. The call to SAX_WHICH_DATA has now an extra argument. This implied the update to the following programs (besides SAXTACCUM) : SAXHKACCUM 0.10 (09 May 96, see entries #17 and previous) SAXGACCUM 0.3 (10 May 96, see entry #20 below) SAXHACCUM 0.7 (10 May 96, see entry #20 below) SAXPACCUM 0.3 (10 May 96, see entry #20 below) ------------------------------------------------------------------------ Bug # : 20 Date : 10 May 1996 Status : fixed with new version of programs/routines Module : SAX*ACCUM programs and some library routines Title : (NOT A BUG) support to UT times etc. in all accumulators Programs SAXGACCUM, SAXHACCUM and SAXPACCUM (see entry #19 for version numbers) are now aligned with the other accumulators in support of selection of time range (and binsize when applicable) in UT, and in the usage of time windows. Also the reconstruction of the HISTORY keyword is now done correctly in all cases. Contextually the following library routines has been upgraded : a) INIT_TIMEWINDOW 0.1 in XASLIB (more general handling of TIME fields) b) SAX_ACC_OTHER_RANGE 0.6 in FOTLIB (calls ASKTIME) to handle the dialogue for TIME quantity c) SAX_ACC_RANGE 0.7 in FOTLIB handles the dialogue for TIME quantity as interesting quantity also for accumulations other than time profiles These changes induce relinking also of SAXTACCUM and SAXHKACCUM. ------------------------------------------------------------------------ Bug # : 21 Date : 15 May 1996 Status : fixed with new version of programs/routines/libraries Module : programs SAX*ACCUM plus several routines Title : (NOT A BUG) support to instrument dependent corrections Programs SAXGACCUM 0.4, SAXHACCUM 0.8, SAXPACCUM 0.4 and SAXTACCUM 0.4 now include support to instrument dependent corrections (to direct modes). Such corrections are implemented provisionally only for MECS. A number of changes/additions to library routines have been necessary (as a side effect also other SAX*ACCUM may be relinked), i.e. a) ACCUMCOMMON.INC has been modified adding new variables b) a new MECOMMON.INC has been introduced for MECS corrections c) SAX_WHICH_DATA 0.6 now pre-caches instrument and packet codes d) SAX_ACC_B1S1_I 0.6 (for ground data) and SAX_ACC_B1S1_Y 0.4 (for FOT data) now call the correction routines e) new FOTLIB routines INIT_CORRECT, CORRECT and INSTR_KEYWORDS have been introduced acting as dispatchers for instrument dependent routines f) a new library MECSLIB has been introduced to host the MECS specific (provisional !!) correction routines ------------------------------------------------------------------------ Bug # : 22 Date : 15 May 1996 Status : fixed with modification to packetcap Module : MECS direct mode packetcap file saxmecsdir.packetcap Title : (NOT A BUG) change to ground calibration names Since newer versions of SAX*ACCUM support indication of time ranges in UT but this is not supported by (MECS) ground calibration data, the dialogue may be inconsistent and cause malfunctions. To prevent this the "TIME" quantity for ground data has been renamed "GTIME" effectively disabling the wrong UT dialogues. This is preferred to the other solution of making ad-hoc "ground" changes to SAX_ACC_RANGE and SAX_ACC_OTHER_RANGE. ** this modification has no effect on FLIGHT data ** ------------------------------------------------------------------------ Bug # : 23 Date : 15 May 1996 Status : fixed with new routine CHECK_PACKET Module : programs SAXIACCUM and SAXSACCUM Title : ACCUMULATE builds wrong runstring when only one eligible packet The "intermediate front-end" programs SAXIACCUM and SAXSACCUM build the runstring for SAXGACCUM and SAXHACCUM respectively (when called by ACCUMULATE IMAGE and ACCUMULATE SPECTRUM). They assumed that the second argument shall always be a packet name (the first being a filename). This is incorrect in the case there is a single eligible packet for the given accumulation. A new FOTLIB routine CHECK_PACKET is now used by both programs to verify how many eligible packets there are and interpret/build the second runstring argument accordingly. New versions are SAXIACCUM 0.3 and SAXSACCUM 0.1. ------------------------------------------------------------------------ Bug # : 24 Date : 16 May 1996 Status : fixed with HEADER_EDIT 1.1 Module : program HEADER_EDIT Title : header editing allowed only for files in current directory HEADER_EDIT did not call BUILDPATH, therefore it could not process a file other than in the current directory, or giving the full path. Now the program is aligned to HLIST and TLIST and recognizes the usual XAS paths. ------------------------------------------------------------------------ Bug # : 25 Date : 20 May 1996 Status : fixed with version 0.2 Module : program SAXSACCUM Title : ACCUMULATE SPECTRUM does not recognise packet argument The command ACCUMULATE SPECTRUM passed the wrong run string to SAXHACCUN when more than one eligible packets exists (like for PDS where science HK packets along with DIR mode packet are eligible !). This was due a typo in the call to CHECK_PACKET in SAXSACCUM (incorrectly copied from SAXIACCUM). The second argument of the call has now been corrected from the wrong value 'Image' to 'Spectrum'. ------------------------------------------------------------------------ Bug # : 26 Date : 23-30 May 1996 Status : fixed with program and routine addition Module : MECSMACCUM plus several routines (most in MECSLIB) Title : (NOT A BUG) MECS response matrix program A preliminary release of the MECS response matrix generator MECSMACCUM has been made available. Besides the main program, this release includes (a) version 0.3 of ACCUMULATE, which supports the front-end syntax ACCUMULATE MATRIX (b) the following new general purpose routines in XASLIB : MULTIPLY_RMFARF and XASMATOUT. They could be used to handle XAS matrices also for other instruments (c) a sequel of new routines in MECSLIB (list omitted, all but those mentioned in previous entries) (d) some calibration files in $XASTOP/calib/xas/mecs have been updated and a sequel of new ones has been added ------------------------------------------------------------------------ Bug # : 27 Date : 24 May 1996 Status : fixed with Z_CHANNEL 0.3 (Unix), 0.2 (VMS) Module : routine Z_CHANNEL in library VOS Title : createserver crashing with POSSIBLE ATTEMPT TO MODIFY CONSTANT The problem showed up only on Suns with the newer compiler (it was already partially cured on VMS), and was caused by the fact that the argument "operation" in a call like CALL Z_CHANNEL(pipe,operation) is usually a constant in the caller (like 'OPEN', 'DELETE') but is manipulated within the routine. Manipulation now occurs on a local copy. ------------------------------------------------------------------------ Bug # : 28 Date : 30 May 1996 Status : fixed with SAXHKACCUM 0.11 and PCF updates Module : program SAXHKACCUM Title : (NOT A BUG) PCF field zp must be REAL for PDS The intercept parameter "zp" for linear calibration curves of HK is a REAL number for PDS. Therefore SAXHKACCUM is modified to support this instead of the previous INTEGER value. The following PCFs are updated (syntax of zp changes from zp#val tp zp=val) : MECS, PDS, HPGS. Contextually at IFCTR also the PDS and HPGS packetcaps are loaded. ------------------------------------------------------------------------ Bug # : 29 Date : 03 Jun 1996 Status : fixed with FOTFILE 0.2 Module : program FOTFILE (pass 2) Title : added support to EXPCONF files The original version of FOTFILE did not handle EXPCONF files as "ASCII", but as "binary files". All such6 filetypes (leexconf to w2exconf) have been added to the list of ASCII files ------------------------------------------------------------------------ Bug # : 30 Date : 03 Jun 1996 Status : fixed with packetcap upgrade Module : MECS direct mode packetcap Title : added support to DIR001 packets The MxDIR001 packets (used in SVT3 and possibly commissioning) were missing in the packetcap file. They have been added. ------------------------------------------------------------------------ Bug # : 31 Date : 03 Jun 1996 Status : fixed with program upgrade Module : programs SAXG/H/T/PACCUM Title : overflow error adding HISTORY keyword The generation of the HISTORY keyword may overflow the variable BUFFER in the case of packets with many fields per event (a case is the MECS DIR001 diagnostic). Therefore variable buffer is now longer. Programs affected are SAXGACCUM 0.5 SAXHACCUM 0.9 SAXTACCUM 0.5 and SAXPACCUM 0.5 ------------------------------------------------------------------------ Bug # : 32 Date : 11 Jun 1996 Status : fixed with program upgrade Module : programs LOCALIZE, IWINDOW, TPLOT, TWINDOW, WPLOT, X[GHPT]ACCUM Title : Constant parameter in CALL H_READ_[JR]KYWORD The fifth parameter in the CALL H_READ_[JR]KYWORD was a constant, while it might have been modified by the subroutine themselves. A fix was achieved by substituion of the constant with NELEM, where NELEM=constant value Programs affected are LOCALIZE1.1, IWINDOW0.1, TPLOT0.4, TWINDOW0.1, WPLOT0.1, XGACCUM1.1, XHACCUM1.1, XPACCUM0.1, XTACCUM0.3 ----------------------------------------------------------------------- Bug # : 33 Date : 13-18 Jun 1996 Status : provisional fix with TOOGIP 0.5 Module : program TOOGIP Title : newer XSPEC wants extra keywords... TOOGIP has been provisionally modified (FG, IFCAI) to add a sequence of FITS kewyords apparently *now* required by the recent XSPEC standards. This includes some which have to be supplied interactively (in the future this part is likely to be modified) ------------------------------------------------------------------------ Bug # : 34 Date : 18 Jun 1996 Status : provisional fix with TOOGIP 0.6/0.7 Module : program TOOGIP Title : ...but not all the keywords that were asked for in 0.5 TOOGIP has been provisionally modified to avoid some not mandatory keyword value (ARFFILE, RESPFILE) are asket to the user via runstring. TOBS has been temporarily fixed to a dummy value (10000 s); it must be read from the header of the spectral XAS file when the relevant quantity will be calculated ----------------------------------------------------------------------- Bug # : 35 Date : 18 Jun 1996 Status : fixed with version 0.3 of MERGEWINDOW Module : program MERGEWINDOW Title : merging of windows doesn't work MERGEWINDOW fails in merging correctly of time windows when considered AND case. Correction by Stefano Signorile (SDC), rearranged method of the time windows Choice in subroutine WORKHORSE ----------------------------------------------------------------------- Bug # : 36 Date : 19 Jun 1996 Status : fixed with new issue of routines Module : routines SAX_ACC_B*S* and SAX_ACC_OPEN_TLM in library FOTLIB Title : no protection if packet does not exist for one obs in chain In the case one concatenates several observations, and then performs one accumulation requiring a given packet type, and no file with these packets exists for one of the intermediate observations, the results of the program are unpredictable (the error opening the non-existing file is ignored, and the program reads data from nowhere). Routine SAX_ACC_OPEN_TLM 0.6 has been modified to return ACCUMCOMMON_NREC=0 in such cases (it won't even try to open a non-existing file). The SAX_ACC_B*S* routines are instructed in this case to skip the observation (routines affected are b1s1_i 0.7 b1s1_y 0.5 b1s2_y 0.3 b2s1_y 0.4 b3s1 0.4 b3s2 0.3 v3s3 0.3 and b3s4 0.3), which implies reload of all accumulation programs ------------------------------------------------------------------------ Bug # : 37 Date : 19 Jun 1996 Status : fixed with FOTFILE 0.3 Module : program FOTFILE (pass 3) Title : added support to EXPCONF files I forgot to apply the change in entry #29 to fotfile_3 as well. As a result a spurious "incorrectly filed" message was generated in pass 3 for expconf files. This has now been fixed. ------------------------------------------------------------------------ Bug # : 38 Date : 27 Jun 1996 Status : fixed with program and routine upgrades Module : program TOOGIP and routine XASMATOUT in library XASLIB Title : miscellaneous errors converting XAS to OGIP matrices When converting a XAS matrix created by MECSMACCUM using TOOGIP, a spurious error complaining about a wrong REFHISTO keyword was generated. This was due to the fact that routine XASMATOUT in MECSMACCUM padded the keyword with extra blanks. XASMATOUT 0.1 cures this (relinking MECSMACCUM) Nevertheless TOOGIP contained two little bugs, one in the name of the variable receiving the value of REFHISTO, and another one in the position of a runstring argument in an X_READ instruction> TOOGIP 0.8 cures also these bugs. ----------------------------------------------------------------------- Bug # : 39 Date : August 14, 1996 Status : fixed with routine update Module : pktcap_lookup.f Title : FIELD variable redifined In PKTCAP_LOOKUP the variable FIELD was defined as CHARACTER*2. That might create problems when the would-be-accumulated quantities have a two-digit identificator. The definition has been therefore moved to the more standard CHARACTER*(*). PKTCAP_LOOKUP 0.2 cures these problem. ----------------------------------------------------------------------- Bug # : 40 Date : August 14, 1996 Status : fixed with routine update Module : sax_acc_range.f, sax_acc_other_range.f Title : Accumulation range reac from packetcap file The would-be-accumulated quantities range were calculated from the dimension of the relevant byte field in FOT. Since such field are expanded in FOT, that results in meaningless range for HPGSPC and PDS. Now the filed dimension in read in the packetcap file if the words m# (lower limit) and M# (upper limit) are found, otherwise they are calculated as before. SAX_ACC_RANGE 0.9 and SAX_OTHER_RANGE 0.8 cure such problem. Contextually the new PDS packetcap implementing such feature has been loaded. For packetcap without the new fields (e.g. MECS) the old behaviour will safely be assumed as default. ----------------------------------------------------------------------- Bug # : 41 Date : 14 Aug 1996 Status : fixed with TPLOT 0.5 Module : program TPLOT Title : wrong x-axis label for fractionary binsize time profiles Some time profiles (those produced by saxtaccum if binsize is less than one second, and the attitude ones produced by saxauxaccum with 0.5 s binsize) use internally as timeunits a fraction (2**-n) of a second (as indicated in keyword TIMERES). Although TPLOT correctly plots the data renormalizing them to either seconds or hours (according to global variable TIMEAXIS), the LABEL of the time axis is annotated in a wrong and misleading way (as "0.xxx s since yyyy-mmm-dd"). This is now fixed modifying some IF statements in routine TIME_AXES ----------------------------------------------------------------------- Bug # : 42 Date : 26 Jul 1996 Status : fixed with new program Module : program XASUNSET Title : (NOT A BUG) new program to unset (delete) a XAS variable Since so far clearing (unsetting, deletion) of a XAS variable could only occur setting it to blank, which, for shell reasons, could be done ONLY using XASSET in an INTERACTIVE way and not from the runstring, the new command XASUNSET has been added ----------------------------------------------------------------------- Bug # : 43 Date : 16 Aug 9996 Status : fixed with new program Module : program CHECK_EXPCONF Title : (NOT A BUG) new program for checking experiment configuration A new program to produce summaries of experiment configuration files under control of an instrument parameter file in $XASTOP/calib has been implemented. Contextually a service routine to collapse duplicated blanks has been added to the GENERAL library (routine COLLAPSE). Also the MECS parameter file MECS.EXPCONF has been inserted in the relevant subdirectory $XASTOP/calib/sax/mecs. ----------------------------------------------------------------------- Bug # : 44 Date : 19 Aug 1996 Status : fixed with version 0.2 Module : program CHECK_EXPCONF Title : (NOT A BUG) added target coordinates to listing Two additional rows have been added with the target RA and Dec taken from the INSTDIR file. ----------------------------------------------------------------------- Bug # : 45 Date : 21 Aug 1996 Status : fixed with version 0.7 Module : sax_acc_preload.f Title : Side effect of FIELD variable changed declaration Since the variable FIELD has now been declared as a CHARACTER*3, the PKTCAP_LOOKUP routine searches for packetcap fields (://FIELD) of the type ':nf ', when the fields are constitued by two characters only (as usual). It was therefore necessary to pass to the PKTCAP_LOOKUP subroutine the input parameter FIELD(:TRUE_LENGTH(FIELD)) instead of the row FIELD. SAX_ACC_PRELOAD 0.7 cures such a problem ----------------------------------------------------------------------- Bug # : 46 Date : 22 Aug 1996 Status : fixed with version 0.8 Module : sax_acc_preload.f Title : Bug in CALL PKTCAP_LOOKUP fixed The call to PKTCAP_LOOKUP when the byte size of the packecap fields is looked up is wrong. The dummy output paraneter ID is used insted of the correct ACCUMCOMMON_SIZE(I). SAX_ACC_PRELOAD 0.8 cures such a problem ----------------------------------------------------------------------- Bug # : 47 Date : 02 Sep 1996 Status : fixed with version 0.11 Module : toogip.f Title : Some bugs/features corrected 1. The call to the function LEFTNUMBER within a formatted internal WRITE left the program waiting for an input. The function is now called before the WRITE instruction and the output stored in a DUMMY variable which is effectively read; 2. spectra are now converted in count/rates from counts before writing in OGIP file; 3. RESPFILE and ANCRFILE keyword added. TOOGIP 0.11 contains all the above new features. ----------------------------------------------------------------------- Bug # : 48 Date : 04 Sep 1996 Status : fixed with version 0.9 Module : sax_acc_preload.f Title : Follow-up of #46 The correction in buglist #46 was uncorrectly implemented for field 'sf': the variable FIELD (and not FIELD(:J)) was passed to PKTCAP_LOOKUP. Moreover, the #46 fix has been extended to indirect modes (no possibility of testing since they are not currently supported in XAS 1.2.10. Hope it works) SAX_ACC_PRELOAD 0.9 cures such problems ----------------------------------------------------------------------- Bug # : 49 Date : September 16, 1996 Status : Fixed Module : saxtaccum.f Title : Instructions to write TUNITS keyword moves The TUNITS3 and TUNITS4 keywords were written by SAXTACCUM as a 80-characters string. That produced an error condition in H_READ_KEYWORD, since KEY_LEN > LENGTH=LEN(STRING) (the string was harmless truncated). The CALL H_ADD_KEYWORD('TUNIT#' .... have been moved before the CALL SAX_DF_KEYWORDS to fix the problem SAXTACCUM 0.7 cures such a problem ----------------------------------------------------------------------- Bug # : 50 Date : 17 Sep 1996 Status : fixed with updated routines Module : routines SAX*RANGE in library FOTLIB Title : correction of syntax errors Routines SAX_ACC_RANGE and SAX_ACC_OTHER_RANGE did not fully declare all variables used in the calls required by the range checking code, and contained an erroneous statement IF (FOUND.EQ.FALSE) THEN [where FALSE is NOT a logical value !) instead of the correct IF (.NOT.FOUND) THEN. This is fixed in versions 0.10 and 0.9 respectively. Contextually a minor warning in SAX_ACC_OPEN_TLM (missing END statement) is fixed too. ----------------------------------------------------------------------- Bug # : 51 Date : 18 Sep 1996 Status : fixed with routine and program upgrades Module : several routines and programs Title : (NOT A BUG) porting to Alpha OSF The following routines and programs have been modified for the porting of the VOS to Alpha OSF : zc_stat 0.3 in VOS different types int vs long z_logintime 2.0 in VOS different layout of utmp z_op_sys 0.1 in VOS system='OSF' z_open 1.1 in VOS (comment) same code as for DEC z_initenv 1.4 in VOS use same code as DEC blksyscommon 1.2 in GENERAL added description for OSF syscommon.inc 1.2 added a new system importback 1.2 use same code as DEC ----------------------------------------------------------------------- Bug # : 52 Date : 18 Sep 1996 Status : fixed with routine and program upgrades Module : several routines and programs Title : syntax bugs & features in several modules During the porting to Alpha undeclared variables or other syntax errors and related bugs have been discovered in a number of modules. Where they did not affect the behaviour, the bug has been corrected without changing the version number. z_exit and z_break in VOS and programs xtaccum, xpaccum, wplot,iwindow, psserver and xwserver contained undeclared items createserver 0.4 fixes a bug due to a line too long fromfits and routines tetafi_xy and write_rmf_ebo contained strings split across two lines ----------------------------------------------------------------------- Bug # : 53 Date : 18 Sep 1996 Status : fixed with version 2.1 (OSF) Module : routine Z_LOGINTIME in library VOS (OSF) Title : cannot get login time of dxterm sessions The routine Z_LOGINTIME is used to get the login time (to be compared with the date of the saved environment file). To do this it looks in the utmp file for an entry for the current tty. Now it appears that dxterm windows which are not login shells are not recorded there, therefore the routine failed with an error 9999. This occurs only on OSF (Ultrix was fine, and so is OSF with dxterm -ls or xterm windows). This affected at least the behaviour of program OVERTRACE which relies on a "transient" environment variable passed across overlaid programs. A fix has been made so that an user is logged in on a tty not present in utmp, the login time on the console (:0) is assumed (this does not coincide with the dxterm creation time but should be harmless unless one creates a window, destroys it and recreates it). ----------------------------------------------------------------------- Bug # : 54 Date : 20 Sep 1996 Status : fixed with version 0.2 Module : routine Z_BREAK in library VOS (all Unix systems) Title : break is triggered immediately Theroutine Z_BREAK contained TWO errors, one of which was a syntax error (see previous entry on "syntax bugs & features in several modules", 18 Sep). This syntax error (ZC_BREAK not declared INTEGER) caused the routine to return correctly with FALSE the first time it was called. However correcting this error the routine now returns incorrectly TRUE as if control-C had been pressed, because of a missing statement. The result is that programs like SAX*ACCUM or TWINDOW terminate prematurely. ----------------------------------------------------------------------- Bug # : 55 Date : 18 Sep 1996 Status : fixed on 20 Sep 1996 with sytem patch Module : routine Z_BREAK in library VOS (OSF) Title : (OPERATING SYSTEM BUG) control-C break does not work The handling of interrupts (control-C i.e. SIGINT) does not work with some versions of the OSF Fortran Runtime Support Library. This appear to affects programs calling the Z_BREAK routine to trap a control-C (i.e. SAX*ACCUM and TWINDOW) but in fact any Fortran program is affected (simply control-C is ignored). If you have kit DFARTL361 installed the problem surely occurs, you have to upgrade to DFARTL365, avaialable from Digital as ECO kit F77RTLAUE01038 (e.g. in file://ftp.service.digital.com/pub/osf/v3.0/fortran_unix/3.8) ----------------------------------------------------------------------- Bug # : 56/70 Date : 18 Sep 1996 Status : *** PROVISIONAL FIX 24 Dec 96 Module : program MECSMACCUM Title : program gives error 101 on RMF (OGIP) (OSF) The program cannot write an RMF in OGIP format (ARF, and XAS format work) with an incongrous error 101 (illegal LU) on the FTCLOS call in WRITE_RMF. It appears that the RMF has been correctly written to disk anyhow when the error occurs. This occurs under OSF 3.2. The main program has been changed to ignore such spurious error. ----------------------------------------------------------------------- Bug # : 57 Date : 02 Oct 1996 Status : fixed with version 1.3 Module : routine BUILDPATH in library XASLIB Title : improved performance of BUILDPATH per DAWG-REP.7/92 The routine BUILDPATH builds the path where a file of a given class (DATA, FOT or PRINT) has to be located. In doing this it uses several environment variables (ROOTDIR, DATADIR|FOTDIR|PRINTDIR, TARGET, DATE INSTRUMENT and ORDER). If ORDER is used a complex multilevel path using all other variables may be constructed (e.g. /root/target/date/instr/data). The original requirement in DAWG-REP.7/92 was that one could have such multilevel structure for some classes (typically DATA) and single-level structure for other (e.g. typically FOT). However this was not implemented and, in the case ORDER was used, a multilevel structure was imposed on all classes. The original requirement is now satisfied in the case one wishes different levels in the hierarchy using two new variables (FOTORDER and PRINTORDER) to specify bypassing of the default level for the FOT and PRINT classes. e.g. if your arrangement produces an order of /root/target/date/instr/data and you do NOT want also /root/target/date/instr/fot, just specify XASSET FOTORDER C to obtain instead /root/fot. ----------------------------------------------------------------------- Bug # : 58 Date : 19 Nov 1996 Status : fixed with version 0.5 Module : program FOTFILE (step 2) Title : FOTFILE now supports selection in data type The question about which data "class" (to be chosen among ABCDEF) to be file, which is asked by pass 2 (FOTFILE_2), was not honoured. Version 0.4 (unofficial) by D.Dal Fiume introduced a parsing routine to support selection by data class. Version 0.5 is a minor variant (e.g. supports classes in upper or lower case). ----------------------------------------------------------------------- Bug # : 59 Date : 19 Nov 1996 Status : fixed with version 0.10 Module : routine SAX_ACC_PRELOAD in library FOTLIB Title : code simplification in SAX_ACC_PRELOAD The correction in buglist #48 has been simplified, namely for all packetcap entries which are ALWAYS 2-character long, one passes directly FIELD(:2) and not FIELD(:J), avoiding useless calls to TRUE_LENGTH to set J. ----------------------------------------------------------------------- Bug # : 60 Date : 22 Nov 1996 Status : fixed with version 0.2 Module : routine GET_OBS_CHAIN in library XASLIB Title : accumulator crashes on very long observation chain Accumulator programs crashed when dealing with chains composed of a very large number of observations. In fact when CONCATENATE was updated to allow long chains (XAS11.BUGLIST entry #38), it was forgotten to enlarge the string containing the OBSCHAIN variable from 80 to 200 characters also in GET_OBS_CHAIN. This alignment has now been done (all accumulators to be recompiled). ----------------------------------------------------------------------- Bug # : 61 Date : 02 Dec 1996 Status : fixed with new routine Module : routine ZC_MEMCPY in library VOS Title : (NOT A BUG) added new memory copy routine A new routine (in C directly callable from Fortran) is added to ALL versions of the VOS. The routine allows to copy memory areas (byte strings) from one location to another (even if not aligned). ----------------------------------------------------------------------- Bug # : 62 Date : 02 Dec 1996 Status : fixed with new routine Module : routine UDOUBLE in library GENERAL Title : (NOT A BUG) added new unsigned integer support routine A new routine (in C directly callable from Fortran) is added to allow to convert unsigned integer quantities (which cannot be handled in Fortran) into double precision quantities. ----------------------------------------------------------------------- Bug # : 63 Date : 02 Dec 1996 Status : fixed with new routine and packetcap changes Module : routine SAX_ACC_B3S4 in library FOTLIB Title : correct alignment in time of HK and ENG ratemeters The times of the ratemeters in the HKD files are defined by Alenia in a different way (end of the 2nd sample) with respect to the times of the SAME ratemeters in the ENG files, causing a shift in time profiles. To overcome this, the routine SAX_ACC_B3S4 now recognises a new field "sh" in the instrument PCF. This field must be set to -2 sec (sh#-2) for most ratemeters. This change is care of each instrument team. A revised MECS.PCF is issued contextually with this upgrade. ----------------------------------------------------------------------- Bug # : 64 Date : 02 Dec 1996 Status : fixed with program and routine upgrade Module : SAX*ACCUM and other programs and several routines Title : (NOT A BUG) support to time in double precision The format of the XAS time profiles has been changed from the previous use of time as an I*4 quantity with resolution of 2*-n sec, to time as a DOUBLE PRECISION quantity in seconds and fractions. This allows to avoid any recycle problem with observations longer than 32 Ksec. Contextually some unofficial upgrades, like computation of the exposure time in accumulation from science data, and support to s/c HK, attitude and ephemeris files by SAXHKACCUM, have been incorporated. At the same time also the provision for ground calibration data ( This involved changes in the definition of the following common blocks 1) HKCOMMON.INC 0.3 (new fields, or change of type to REAL*8) 2) TIMECOMMON.INC 1.3 (also for support to exposure time computation) 3) WCOMMON.INC 0.1 (time window boundaries now are REAL*8) 4) ACCUMCOMMON .INC 0.9 (new fields, or change of type to REAL*8) In addition to the routines added/modified as described in the three previous bug entries, the following routines have been added/modified : in XASLIB 5) ASKTIME 0.2 and ASKBIN 0.2 now support REAL*8 times 6) LOADWINDOW 0.1 and INIT_TIMEWINDOW 0.4 : time window boundaries now are REAL*8) 7) TIME_CONSTANT_SETUP 0.5 is a new initialization routine 8) WRITE_BIN 1.1 and READ_BIN 1.1 now support REAL*8 fields in XAS tabular files via the ZC_MEMCPY VOS call. 9) UPDATE_START_END 0.2 now supports REAL*8 times in FOTLIB 10) SAX_ACC_RANGE 0.11 and SAX_ACC_OTHER_RANGE 0.10 include a separate dialogue for times in double precision 11) SAX_ACC_HKRANGE 0.4 is a similar routine formerly within SAXHKACCUM 12) GET_START_END 0.3 has been simplified removing ground data support and adding double precision time handling 13) SAX_ACC_PRELOAD 0.11 initializes time recycle and handles indirect modes 14) SAX_ACC_BT_3 0.3 supports the new s/c HK and ASCII HK files, which are handled by the new routines SAX_ACC_B3S5 and B3S6 respectively 15) SAX_ACC_B3S3 0.5, B3S4 0.7, B1S1_Y 0.8 and B2S1_Y 0.5 support double precision times (and the latter two also exposure time computation) 16) SAX_ACC_OPEN_TLM 0.7 : removed ground data support 17) SAX_ACC_OPEN_SC_TLM 0.0 is a new routine to handle telemetry files which occur per OP (like s/c HK, attitude and ephemeris) 18) SAX_PCF_LOAD 0.3 and SAX_PKTCAP_LOAD 0.7 now search also the spacecraft related PCF and packetcap respectively The following programs have been modified, adding support for the new time profiles and removing support to old-style time profiles : 19) SAXHKACCUM 0.16 now handles not only instrument HKD and ENG ratemeters,but also s/c HK, attitude and ephemeris files (formerly unofficially handled by SAXAUXACCUM which is now merged with this program). 20) SAXTACCUM 0.7, SAXHACCUM 0.14, SAXPACCUM 0.8 and SAXGACCUM 0.8 support double precision time in selections 21) SAXTACCUM and SAXHKACCUM produce new-style time profiles (REAL*8 times). 22) IWINDOW 0.3 and MERGEWINDOW 0.4 support REAL*8 times. 23) TLIST 0.5 and TPLOT 0.6 support REAL*8 times 24) the demo program OLDTIMER is the only quick-and-dirty form of support to old-style time profiles (which are converted in the new format), which for the rest are declared obsolete. ----------------------------------------------------------------------- Bug # : 65 Date : 02 Dec 1996 Status : fixed with program and routine upgrade Module : call of routine INSTR_KEYWORDS in library FOTLIB Title : (NOT A BUG) changed argument syntax of INSTR_KEYWORDS The argument PRODUCT introduced into INSTR_KEYWORDS 0.2 was a one-letter code. For uniformity with the call to SAX_ACC_WHICH_DATA this is now changed to a verbose string like 'Image', 'Photon', 'HK' etc. Only the first letter of such string is significant. This involves an edit to all SAX*ACCUM programmes to expand the argument (this change is not numbered but included in the major restructuring for double precision times), it involves no change to INSTR_KEYWORDS itself while routine PDS_KEYWORDS 0.5 has been updated to use code 'HK' for housekeeping instead of code 'K'. ----------------------------------------------------------------------- Bug # : 66 Date : 05 Dec 1996 Status : fixed with version 0.9 of routine and new packetcaps Module : routine SAX_ACC_B1S1_Y in library FOTLIB Title : support to direct modes with non-maximal time resolution The new accumulation routines which use times in double precision need to take account of recycles under all possible conditions, and make use both of event times and packet start-end times. However for some direct modes where the ORIGINAL time resolution in the telemetry was not the maximal one (2**-16 s), the zero-padding included in the FOT time reformatting has to be taken duly in account when comparing packet (full resolution) and event (zero-padded) times. To do this it is necessary that the original resolution is known to the routine. This is done declaring it in the packetcap "or" field, which is supported by the revised routine. Contextually the MECS packetcap SAXMECSDIR.PACKETCAP including the new field is released. Other instrument packetcaps are to be verified yet. ----------------------------------------------------------------------- Bug # : 67 Date : 13 Dec 1996 Status : fixed with routine modification Module : routine INSTR_KEYWORDS in library FOTLIB routine PDS_KEYWORDS in library PDSLIB Title : definition of PRODUCT variable aligned to XAS use The variable PRODUCT in the INSTR_KEYWORDS and PDS_KEYWORDS routines was defines as CHARACTER*(*), while it was compared in an IF statement with a one-character constant. That misalignment produced a crashed when accumulation a PDS product of any kind. INSTR_KEYWORDS 0.4 and PDS_KEYWORDS 0.6 cures such a problem (in the latter also a comment has been modified to align with buglist entry #65 modifications) ----------------------------------------------------------------------- Bug # : missing in SDC version Date : 16 Dec 1996 Status : fixed with version 0.2 Module : routine INIT_CORRECT in library FOTLIB Title : always set "correctable" quantities in common The routine INIT_CORRECT used to set indices CORRINDX in common ACCUMCOMMON to point to the quantities (e.g. X,Y,PHA) liable to be used for instrument dependent corrections *only when* the corrections were enabled. Now some programs will need such information even if corrections are disabled, so they are set anyhow. ----------------------------------------------------------------------- Bug # : 68 Date : 18 Dec 1996 Status : fixed Module : routine SAX_ACC_B1S1_Y in library FOTLIB Title : wrong type definition of variable PREVIOUS_TIMEHI The variable PREVIOUS_TIMEHI was defined INTEGER instead of DOUBLE PRECISION. The correct type definition is inserted. ----------------------------------------------------------------------- Bug # : 69 Date : 18 Dec 1996 Status : fixed with version 0.15 Module : program SAXHACCUM Title : (NOT A BUG) changed definition of LOWER and UPPER boundaries The previous (preliminary) version of SAXHACCUM indicated LOWER and UPPER boundaries in PHA channels with the SAME value (e.g. 0,0 ; 1,1 etc.). It has to be noted that the final version is intended to have LOWER and UPPER boundaries in keV (currently done by additional programs). The definition has now been changed to have different values in LOWER and UPPER (e.g. 0,1 ; 1,2 etc. for zoom one : interpret them as "bin one has as lower bound the lower bound of PHA 0, and as upper bound the lower bound of PHA 1"). This is more consistent with other bin definitions and also allows nicer plotting in histogram or errorbar mode. ----------------------------------------------------------------------- Bug # : 70 Date : 30 Dec 1996 Status : see entry 56 Module : program MECSMACCUM Title : program gives error 101 on RMF (OGIP) (OSF) ----------------------------------------------------------------------- Bug # : 71 Date : 24 Dec 1996 Status : fixed with new version of routines Module : routines in FOTLIB and PDSLIB Title : (NOT A BUG) default is now FLIGHT ON, alas ! The following routines now default to "xasset FLIGHT ON" : CHECK_PACKET 0.1, SAX_OPEN_DIR 0.4, SAX_WHICH_DATA 0.7 in FOTLIB and SAX_FROOT_NAME 0.2 in PDSLIB. We remind that support to ground data was removed in SAX_ACC_OPEN_TLM in the jumbo patch, so the use of "FLIGHT OFF" is vestigial and won't work (for instrument groups desiring it) unless an older version is used. ----------------------------------------------------------------------- Bug # : 72 Date : 24 Dec 1996 Status : fixed with program upgrades Module : accumulation programs Title : last part of runstring ignored When a command like "ACCUMULATE xxxx" is invoked with all or almost all of its arguments on the runstring, it may happen that the last arguments are ignored, or even truncated. This is due to an insufficient buffer length (80 character) in the primary dispatcher ACCUMULATE, and to an insufficient number of arguments parsed (20) in the secondary dispatchers for spectra and images (other accumulations do not have secondaries). The size of BUFFER is increased to MAXRUN (as defined in RUNSTRING.INC) in ACCUMULATE 0.5 and XACCUMULATE 0.3. The number of arguments is increased to 20 in SAXSACCUM 0.3 and SAXIACCUM 0.3. This allows handling of the complete runstring. In principle all programs using Z_GET_COMMAND and Z_RUN are liable to this truncation, but the problem is in practice not encountered with the other dispatchers. ----------------------------------------------------------------------- Bug # : 73 Date : 18 Dec 1996 Status : fixed with new version of programs/routines Module : programs DISPLAY and TPLOT Title : (NOT A BUG) improved graphics DISPLAY 0.4 allows the new syntax "DISPLAY WINDOW file" (as well as "OVERTRACE WINDOW file") to invoke WPLOT to overlay time windows on the current time profile plot (it ALWAYS overlays). TPLOT 0.7 allows plotting of horizontal error bars for time profiles and spectra in histogram and error bar mode, and use the middle point of a bin in solid mode. It also allows to overtrace time profiles pertaining to different days. Routines PLOT_XXY_JOIN 0.1 and PLOT_XXY_HISTO 0.1 in XASGRAPH have been upgraded so that SPLOT and TPLOT now do not plot connecting line in gaps but skip (where a gap is defined when the end of a bin is not contiguous with the beginning of the next bin within a tolerance of 1 percent) ----------------------------------------------------------------------- Bug # : 74 Date : 18 Dec 1997 Status : fixed with version 0.6 of TLIST Module : program TLIST Title : (NOT A BUG) list times also as hh:mm:ss TLIST 0.6 allows an option to display double precision times after the jumbo patch also in the format hh:mm:ss.ff. This is enabled for a column named TIME by doing XASSET HMS YES ----------------------------------------------------------------------- Bug # : 75 Date : 18 Dec 1996 Status : fixed with new or updated programs Module : programs MECSGAINACCUM, SAX*ACCUM and several libraries Title : (NOT A BUG) MECS gain vs time correction The new program MECSGAINACCUM, invoked via ACCUMULATE GAIN, produces a XAS gain history (time profile), which may be used by all SAX*ACCUM accumulators to perform gain time dependent correction. For this the following new or modified files are necessary : Mx_CALSOURCE.COEFF (x=1,2,3) in $XASTOP/calib/sax/mecs contain reference positions and energies for the calibration sources In GENERAL there are new routines for the CURFIT package (CURFT, FDER, FCHI and MTINV) and a variant of EXTRP called EXTRPD which supports double precision arguments for extrapolation. In MECSLIB there is a new routine ME_GAIN_TIME, while there are updates for MECOMMON.INC 0.1, INIT_CORRECT_ME 0.3, MECSCORRECT 0.4 and MECS_KEYWORDS 0.1 ----------------------------------------------------------------------- Bug # : 76 Date : 24 Dec 1996 Status : fixed with new version of routines Module :program MECSMACCUM several MECSLIB routines and calibration files Title : (NOT A BUG) new version of MECS matrix Version 0.1 of MECSMACCUM, a new set of routines in MECSLIB and the associated calibration files in CALIB/SAX/MECS are released. They can be used to generate updated response matrices. ----------------------------------------------------------------------- Bug # : 77 Date : 28 Dec 1996 Status : fixed with new or upgraded routines Module : accumulation programs Title : (NOT A BUG) spatial region selection available for MECS The scientific data accumulation programs now support the selection of spatial regions defined in external files (XASSET SELECTFILE name) as XAS image masks (XASSET SELECTMETHOD MAP) or SAOimage-style region files (XASSET SELECTMETHOD REGION). This is achieved by adding or modifiying the following routines : MECOMMON.INC 0.2, INIT_CORRECT_ME 0.4, MECSCORRECT 0.5 and MECS_KEYWORDS 0.2 are modified for/in MECSLIB RADIANT.INC is a new general purpose include file. ME_INIT_MAP is a new initialization routine in MECSLIB. BIT_INIT_HANDLE is a provisional pseudo-bit-array interface in GENERAL. ----------------------------------------------------------------------- Bug # : 78 Date : 31 Dec 1996 Status : fixed Module : xasbuild script in config/Scripts Title : no export of the environment variable $system The environment variable $system is not exported. The scripts directly scheduled by xasbuild are affected if they do not redefine this variable inside themselves. Now the variable is correctly exported as soon as it is defined. RCS version: 1.1.7.3 ----------------------------------------------------------------------- Bug # : 79 Date : 31 Dec 1996 Status : fixed Module : xasbuild.sh script in config/Scripts Title : incomplete cleaning of fake dependency files If fake dependency files must be created for more than one 'bin' directory during a bootstrap run, the fake files are recreated only for the last directory scanned. This was due to a reblanking of a variable in the loop scanning sourcedir.names. The problem is fixed blanking the variable outside the loop, before its beginning. RCS version: 1.1.7.3 ----------------------------------------------------------------------- Bug # : 80 Date : 31 Dec 1996 Status : fixed Module : xasdepend script in config/Scripts Title : NOT A BUG - better management of ambiguous dependencies An ambiguous dependency is defined as a case in which a root file name $i without extension has multiple occurrences in the target source directory (e.g. $i=$XASTOP/vos/z_open corresponds both to $XASTOP/vos/z_open.f and to $XASTOP/vos/z_open_stream.f). In order to simplify the regeneration of the depend files, this ambiguity is automatically solved when $i.[f|c] exists. In this case, a dependency from this file is assigned to the executable whose load map is being scanned. Real ambiguities, when the name is truncated, are solved manually as usual. RCS version: 1.1.7.4 ----------------------------------------------------------------------- Bug # : 81 Date : 08 Jan 1997 Status : fixed with upgraded routines Module : routine MECS_ARF in MECSLIB Title : Lower threshold set on ARF file XSPEC on Alpha/OSF systems crashed if the values of effective areas are lower than ~10^-30. A hreshold has been set in the routine which writes the OGIP format effective area in order it to be qual to 0 when its calculated value is lower than 10^-8. MECS_ARF 1.3 cures such a problem ----------------------------------------------------------------------- Bug # : 82 Date : 15 Jan 1997 Status : fixed with version 0.5 Module : routine INIT_CORRECT_ME in MECSLIB Title : artifacts due to inconsistent format of gain coeff. file While the format of the gain files MiGAIN.COEFF has been changed for the matrix program MECSMACCUM, the routine INIT_CORRECT_ME was still reading it in the old format. As a result the energy in keV correspon- ding to a channel was computed in an awfully wrong way in MECSCORRECT, introducing all sorts of artifacts in images (and other accumulations?). The routine now reads the new format of gain files. ------------------------------------------------------------------------ Bug # : 83 Date : 20 Jan 1997 Status : fixed with version 0.4 Module : program IWINDOW Title : end of intensity windows incomplete by one time bin Program IWINDOW was terminating time windows using the time bin reference time (i.e. the start time) instead of the end time. Time windows were therefore too short or too long by one time bin. Now the correct time is used. Contextually a couple of prompts have been made less misleading. Now upper and lower values can be given in any order.