.help revisions May92 noao.rv .nf rvrvcor.x Modified so UTMIDDLE now properly accepts a DATE-OBS-like timestamp. Also modified so the date is properly incremented in the case the midpoint observation crosses midnight (7/20/10) rvrvcor.x An earlier change used stridx() but was passing in a singly-quoted character, this caused a runtime error when checking string values on Sun systems. (9/26/08) ======== V2.14.1 ======== rvrvcor.x Changed the rv_parse_timed() procedure to permit DATE-OBS strings. This is used to parse the time from UT/UTMIDDLE keywords, the change allows these to be in DATE-OBS format rather than restrict them to being sexigesimal strings. (7/25/08) ======== V2.14 ======== filtpars.x Fixed an error in the filtpars code where missing braces meant that a command such as ":cutoff" would not work when filtering is enabled (4/6/07) rvfgauss.x rvfuncs.x Merged in old changes to include an update to do the Gaussian fitting using double precision. (3/15/07) rvwrite.x Increased the HJD output to 5 decimals (8/15/06) ======== V2.12.2a ======== doc/fxcor.hlp Added an example for setting a zero heliocentric correction for the template image. (7/1/04) ======= V2.12.2 ======= doc/fxcor.hlp Added an example about saving the CCF which explains how to save a linearized CCF for large spectra. Also removed the obsolete 'time requirements' section. (9/15/03) wrtccf.x Fixed a bug in the text CCF output so it writes the velocity computed specifically for each point rather than the approximation as done before. (9/15/03) wrtccf.x Added CTYPE1='velocity' and CUNIT1='km/s' keywords to the CCF output image so it will display in velocity units (8/19/03) rvrebin.x In very rare cases the code could overrun a stack array and cause a segvio. This array wasn't actually being used in the routine and so the code was removed (8/15/03) rvidlines/idvhelio.x The previous change would trash the year, month, and day when parsing the UT. This is fixed. (2/6/03, Valdes) ======= V2.12.1 ======= ===== V2.12 ===== doc/rvidlines.hlp Expanded on what keywords from KEYWPARS are used and how. (2/12/02, Valdes) rvidlines/idvhelio.x The specified UT keyword is now used rather than the time given in the date keyword. To use the time in the date keyword, the UT keyword can be set to be the same as the date keyword and the UT will be parsed out of the string. So the UT keyword can be either a time value or a date/time string value. This is consistent with changes in ASTUTIL. (2/12/02, Valdes) splitplot.x Fixed a bug in moving the correlation X data to the plotting array which would sometimes cause the ccf plot to not be drawn correctly. (10/18/00) rvsumplot.x Move the 'fit did not converge' message to a more readable location on the plot. (10/18/00) fftutil.x Fixed a typo that was causing the fftmode plots to be computed using the entire spectrum instead of taking into account the sample regions specified. (08/22/00) fftmode.x prepspec.x splitplot.x rvfftcorr.x Additional fixes to the filter overlay plotting code. Also revealed a problem in the display of the FFT plots. (08/22/00) fftmode.x Fixed a bug in the filter overlay plotting (8/17/00) rvwrite.x Changed the output of the HJD to use modulo 10000 rather than constants such as 2450000. (3/30/00) coloncmds.x doc/fxcor.hlp Changed the behavior of the :apnum command so that it won't reset the sample regions. This was originally done to protect against bogus sample regions in echelle data where moving through orders would invalidate the regions, but harms multispec users who must reset the regions manually. Updated help page (3/23/00) ======= V2.11.3 ======= rvrvcor.x Fixed a bug where the DATE-OBS in the for CCYY-MM-DD wasn't checking for an INDEFD UT causing an overflow (12/23/99) rvidlines/idmap.x The SFREE added earlier was in the wrong place causing 1D images to abort with a salloc underflow. (10/12/99, Valdes) ======= V2.11.2 ======= mkpkg rvidlines/mkpkg Fixed missing/extra file dependencies (9/20/99) aplist.x rvidlines/idmap.x Fixed some problems picked up with SPPLINT. (7/28/99) rv.cl rv.hd x_rv.x rvcorrect.par - t_rvcorrect.x - doc/rvcorrect.x - The RVCORRECT task was a duplicate of the ASTUTIL package. To avoid code confusion sources were removed from the RV package and the task is now defined to be the ASTUTIL version. (5/27/99) Revisions rvpackage.spc keywpars.par rvflags.h keywpars.x Files were reviewed for Y2K compliance, no changes necessary. (5/27/99) keywpars.hlp Documentation updated about use of dtm_decode. (5/27/99) rvwparam.x The timestamp string written for the output log header was changed to use a 4-digit year. Does not affect formatting of output tables. (5/27/99) rvrvcor.x doc/fxcor.hlp Changed to use dtm_decode to parse date-obs field. If a time is included in the date string it has precedence over UT when computing observation midpoint and not using UTMIDDLE keyword. (5/27/99) Time from date-obs is used (if available) as the beginning time of the obs- ervation if there is no UT keyword present and we're not using the UTMIDDLE keyword to get the midpoint time. Documentation updated. (5/27/99) rvidlines/idvhelio.x rvidlines/rvidlines.hlp Modified to use dtm_decode to decode both the old and new FITS date string. If a time is included in the date string it has precedence over UT. (5/19/99, Valdes) rvwrite.x Fixed a bug in the output of the HJD to screen and txt (5/4/99) fftutil.x Increased max allowable exponent from 15 to 31 (5/3/99) rvvfit.x Fixed a bug in the output of deblended velocities (4/22/99) deblend.x Fixed a prompt error for the 't' keystroke. (4/8/99) t_rvcorrect.x Added code the catch a bad DATE-OBS keyword and print an informative error (2/21/99) fftmode.x Fixed a bug in the one-plot options. (2/21/99) rv.hd Modified to remove the 'rv' definition which was interfering with selecting the package help (1/27/99) rvcorrect.par Removed the KEYPARS pset from the parameter file. The pset is still available to the task, but it's presence interferes with the task when used in CL mode since the 'ut' parameter is no longer queried and the pset value is used, resulting in an illegal number error when getting the value (4/22/98) rvgetim.x When pixcorr was set the data were still being rebinned in the SMW code, fixed this. (4/2/98) splitplot.x Fixed a bug in the plotting of object spectrum in the summary plot (4/2/98) rvcursor.x calls to rv_do_save() had an unused 'type' arg which conflicted with usage in coloncmds.x. Since this was unused it was removed (3/27/98) rvrebin.x rv_getim() was being used as a subroutine, should be an int function. (3/27/98) deblend.x Fixed a type conversion problems causing an overflow when the computed velocities are INDEF (2/17/98) rvidlines/ididentify.x There was a parenthesis placement error in an id_zshift call. (2/5/98, Valdes) rvwrite.x Fixed HJD output as an offset from 2450000 for more recent dates. (1/5/98) +====== V2.11.1 +====== rvdrawfit.x Fixed a typo causing an incorrect array allocation (8/25/97) rvplot.x fftmode.x rvsample.x Removed the checks to see if the obj/temp image names are the same to decide if we're making a single or split plot. With only a single plot it's not possible to set different sample regions when autocorrelating. (7/24/97) fftmode.x Deleted unused pointers from fft_plot() routine (7/24/97) splitplot.x rvplot.x rvsample.x specmode.x Minor fix to plotting of sample regions (7/24/97) fftmode.x splitplot.x Fixed a fwe minor plotting bugs for power spectra (7/7/97) rvfgauss.x Changed check_converge() to rv_check_converge() to avoid name conflict with FITS kernel. (5/20/97) rvidlines/mkpkg rvidlines/idlinelist.x rvidlines/idlog.x rvidlines/t_reidentify.x rvidlines/idvelocity.x rvidlines/idgraph.x rvidlines/ididentify.x rvidlines/reidentify.x rvidlines/idrms.x Modified to work in input spectrum units. (3/13/97, Valdes) rvgetim.x Input spectra which are dispersion corrected are converted to Angstrom units on input. Thus the spectra can be in other recognized units. (3/13/97, Valdes) rvstrings.x rvwrite.x The nam_tempcode procedure had an unnecessary pointer arg. (3/6/97) rvwrite.x Added an extra space for the HJD output. Since the JD rolled over 245000 the number of decimal places decreased by 1, this is now restored (10/21/96) contin.x The normalized reference spectrum was being copied from the object spectrum when the image names were the same as a form of optimization. This could fail in cases where the 'n' key was used to move through a list to get the next object an the object spectrum was previously rebinned at a different dispersion causing artificial shifts in the data. The code was modified to always compute a normalization to avoid this (9/12/96) rvcorrect.par doc/rvcorrect.hlp Forgot to add the keypars pset for this task (8/26/96) rvlinefit.x Removed an extraneous sfree call causing a salloc underflow if a certain error condition happens. (8/13/96) specmode.x Fixed a bug where the spectrum mode could get confused on future plots on specmode. (8/8/96) rvfgauss.x A couple of *_old varables were declared but never used. (7/17/96) rvrvcor.x Fixed a bug in the rv correction code where the default kpno observatory was always being used in cases where the image did not have an OBSERVAT keyword, instead of using the obs supplied by the task parameter (6/26/96) rvgetim.x Commented out the check for a max number of points since this no longer appears to be needed. (5/14/96) rvfgauss.x Fixed a rare bug that can cause a floating overflow when a fit converges to a nonsensical large gaussian. Added a trap for a very large sigma in a computed fit. (7/26/95) rvplot.x Fixed an old bug in which the code to draw a single plot in spectrum mode was actually recursive. This would result in an infinite loop if the obj/temp were the same image and the 'i' key was used. Wrote code for a single plot for both the input and normalized single plots (3/10/95) ======= V2.10.4 ======= rvidlines/idvhelio.x Changed error action to a single warning to allow for the case of missing header keywords. (10/20/94, Valdes) rvbatch.x Fixed a bug where the RV_TEMPNUM wasn't being set before the rv_data_check() procedure was called. If the data were rebinned this corrupted the RV_TEMPVEL() pointer. (10/13/94) ======= V2.10.3 ======= rvbatch.x Fixed a bug where the image wasn't being update in batch mode (7/18/94) rvsample.x Fixed a bug erasing the sample regions with 'u' (5/19/94) splitplot.x Fixed a bug affecting filter overlay of FFT plots. (4/24/94) rvidlines/mkpkg The mkpkg was not complete in it's dependencies. Did a mkmlist to insure all dependencies are checked. (4/23/94, Valdes) t_fxcor.x The imtclose() calls when shutting down would fail if the file lists had changed during the task. Changed this to close the structure pointers instead. (2/22/94) rvgetim.x Added a check on the dispersion units. (2/18/94, Valdes) rvfftcorr.x The "prepared spectrum" plot wasn't showing the spectrum with the effects of the filtering. Filtering is normally done after the FFT but it made sense that the user should see the centered, apodized and filtered spectrum being correlated. (2/15/94) mkpkg asttools/ - Deleted the duplicate asttools source and made use of the noao package library. (8/25/93) mkpkg rv.cl rv.hd x_rv.x rv.men rvidlines/* + rvidlines.par + rvreidlines.par + doc/rvidlines.hlp + doc/rvreidlines.hlp + Added the new tasks RVIDLINES and RVREIDLINES for determining velocities from spectral lines. This task is similar to the IDENTIFY task for determining dispersion solutions. (8/24/93, Valdes) rvinit.x zzdebug.x rvfgauss.x rvdebug.par The type of the rvdebug.other parameter was changed but I forgot to update the files that were still treating this as a boolean. Also removed and option for using an old version of the lorentzian fitting. (8/12/93) rvcorrel.x Changed the computation of the antisymmetric noise function. The full discussion is in the RV mail logs but basically the previous algorithm was computing the differences incorrectly about the peak, which is obvious for the case of a shift of zero. The new algorithm is correct (even though the endpoints look funny). For consistancy the old agorithm is still available by setting rvdebug.other to a positive value. Setting rvdebug.other to -1 gets you the old alg- orithm with a bug fix to it. (8/7/93) rvanplot.x Added 'sigma' to the text output. (8/7/93) rvvfit.x Fixed a typo in the definition of c[3] as the Sigma of the Gaussian fit when it's really the Sigma^2 of the fit. It correctly used in the computation of width elsewhere in the code (7/9/93) mkpkg t_fxcor.x rvgetim.x rvrebin.x rvimutil.x coloncmds.x dispcor.x - dispcor.com - Changed the way the data rebinning to get a common dispersion is handled. Previously the V2.9 DISPCOR code was used but this ran into problems when trying to rebin a log spectrum to a different dispersion with many more points. What happens now is that the original image is read but then interpolated to the requested disp- ersion rather than the log equivalent of what's in the header. This This has the advantage that no spectrum is interpolated more than once prior to correlation. (6/29/93) rvplot.x Fixed a typo in the specmode.'i' plot annotation (6/21/93) numrep.x rvbatch.x rvrvcor.x rvwrite.x rvfparab.x rvimutil.x rvwparam.x rvcolon .x rvcursor.x coloncmds.x readtlist.x rvdrawfit.x rvfitfunc.x rvstrings.x splitplot.x Cleaned up "foo set but not used" messages reported by SPPLINT (5/13/93) mkpkg specmode.x rvlinefit.x + Implemented a new specmode keystroke command, 'v', which is basically an rip-off of the SPLOT 'k' keystroke that fits a gaussian to a line, asks for a standard wavelength and then computes a velocity for that single line. For the moment this will be an undocumented command since it will be part of other enhancements to specmode in the next release. (5/12/93) rvgetim.x The X coordinate array wasn't being initialized with the proper coords (5/11/93) rvrebin.x The rebinned spectrum was being normalized using the entire spec- trum as the sample. Any changes made interactively or samples selected were ignored. This didn't make sense so I commented this out so the fit is done using the same samples it would in any other case. (5/10/93) rvwrite.x Changed the FWHM output to .txt files so it prints velocity if present, otherwise it's the pixel width. The "km/s" units label was removed from the column heading. (5/7/93) rvrebin.x Fixed a type declaration problem causing compilation errors under AUX. (5/6/93) rvwrite.x The "Rebined WPC" field was being printed as some bogus value for pixel correlation. Changed to print INDEF. (5/5/93) rvcont.h contin.x continpars.x Set it up so that a :markrej command stays in effect until reset by another command. Requested by Daryl. (5/5/93) specmode.x doc/fxcor.hlp lib/scr/specmode.key Added a new 'b' keystroke command to set the sample region for both spectra simultaneously. (5/5/93) rvwrite.x In the case where redshifts are being printed an INDEF value was being converted to a redshift before printing. Added a bunch of if (IS_INDEF())s so that an INDEF redshift would be printed. (5/5/93) rvrvcor.x Reset the RV_PRINTZ flag in the case where missing header info prevented a full velocity computation. VREL was still being con- verted but it was always printed as a velocity, never a redshift. (5/5/93) rvsample.x Fixed a small problem w/ erasing sample regions. (4/30/93) rvbatch.x Fixed a bug in which the .txt file header was being written before the final WPC was computed. This meant that the WPC and vel-per-pixel values were wrong for batch runs if either spectrum had to be re- binned prior to the correlation. (4/29/93) rvrvcor.x The MJD was being computed incorrectly. As defined in the Astronomical Almanac it's the JD - 240 0000.5. (4/16/93) rvvfit.x Fixed a bug in the output of MJD to .log files. (4/14/93) rvfilter.x Fixed a bug in the square filter code. (4/7/93) doc/fxcor.hlp Added notes for the new package params. (3/31/93) keywpars.x rvcomdef.h rvkeywords.h keywpars.par doc/fxcor.hlp doc/keywpars.hlp Removed the W0 and WPC fields from KEYWPARS. Also made a note in the help pages about how the UT of the heliocentric correction is found. (3/29/93) mkpkg x_rv.x rvcorrect.par + rvcorrect.com + t_rvcorrect.x + Made a duplicate of the RVCORRECT task for the RV package. This task is identical to the one in ASTUTIL but it makes use of the KEYWPARS header keyword translation pset. (3/23/93) wrtccf.x rvpackage.h Fixed a bug in writing out the correlation function to an image or text file. Also removed an unused RV_AXIS macro (3/15/93) deblend.x splitplot.x Fixed some INDEF type conversions that were causing problems on the A/UX port. Conversion of INDEF is illegal but has only just now been caught because INDEFR == INDEFD on other systems (3/15/93) rvrebin.x rvdatacheck.x rvfftcorr.x An integer zero point coordinate offset was computed and applied in rv_fftcorr but the fractional part was ignored. This would produce a final velocity error of up to a pixel in those cases where the wavelength origins were not the same between the object and template spectra. This error shows up in tests with DOPCOR which introduces just such a zero point shift. o Moved the setting of w0 and npts from force_rebin to force_which. Force_which is now the main point for checking and resetting compatibility between the object and template spectral WCS. The key part of the fix was to adjust w0 so that it is an integer number of pixels shift from the w0 of the target spectrum. The integer part of the shift is then handled by rvfftcorr.x o The checking of the WCS in rv_data_check was moved to force_rebin (which then calls force_which). Rv_data_check now simply calls force_rebin every time. o Added a nint in rv_fftcorr to make sure that no funny roundoff occurs in setting integer pixel shift zero point offset. ishift = nint ((RV_OW0(rv) - RV_GLOB_W1(rv)) / RV_OWPC(rv)) (3/1/93, Valdes) mkpkg rv.par rvgetim.x aplist.x rvmwcs.h - rvmwcs.x - getdisp.x - doc/fxcor.hlp 1. Added dispaxis and nsum to the package parameters. These are used with 2D and 3D spectra which are now supported. 2. The spectrum header information and data are obtained through the smw/shdr routines from ONEDSPEC. This supports linear, log-linear, non-linear, equispec, multispec, 1D, 2D, and 3D spectral formats. (2/25/93, Valdes) coloncmds.x deblend.x fftmode.x rv.par rvanplot.x rvcolon.x rvcomdef.h rvdrawfit.x rvfitfunc.x rvflags.h rvinit.x rvpackage.h rvplot.x rvsample.x rvstrings.x rvsumplot.x splitplot.x Added support for color overlay vectors and text (12/14/93) rvplot.x deblend.x rvsample.x rvdrawfit.x rvfitfunc.x splitplot.x Changed the gseti() calls to use symbolic names when changing line and polymarker types. (12/11/92) rvplot.x fftmode.x rvanplot.x rvsumplot.x splitplot.x Removed the bold font definitions from gtext() calls. (12/7/92) rvrvcor.x Renamed procedure rv_correct() to rv_corr() to avoid possible name clashes at a later date. (10/13/92) rv0$rv0.par Added an "observatory" package parameter to allow the parameter indirection used by the RVCORRECT task. Until now that task couldn't be run. (10/13/92) rvrebin.x Had to clear a stack array since grabage was being used in the second correlation. I really need to go through all of the code and fix this once and for all (9/1/92) deblend.x rvfitfunc.x rvrvcor.x Cleaned up some INDEF expressions so there is no type conversion going on. This can caused problems on the VXUX systems (9/1/92) rvwrite.x Cleaned up some INDEF equality tests to us IS_INDEF() macro (9/1/92) rvcorrel.x Changed the decared dimension of the data array from 'npts' to 'ARB'. (8/26/92) rvfgauss.x rvfparab.x Added some debugging output to print weight info (8/26/92) aplists.x getdisp.x rvgetim.x rvrebin.x prepspec.x Made a simpler debug test macro for readability (8/26/92) prepspec.x The sample region masking could fail if one of the spectra were smaller than the other. The masking code assumed the whole global wavelength array was being passed in, but in fact only the data array was being passed. Also fixed a typo when initializing the array, too many points were being moved. (8/25/92) rvgetim.x The values of RV_X1 and RV_X2 were confused w/ log and linear values for data w/ lambda dispersion. (8/24/92) splitplot.x Removed the sample regions markers from the prepared spectrum plot (8/24/92) rvfilter.x There was a small "glitch" in the computed Welch or Hanning filters caused by an extra filter value being computed. (8/24/92) fftutil.x Changed some procedure arguments named 'data' to 'v' since this causes a compiler error on the RS/6000. Original complaint was for fft_fixwrap() but as also present in fft_cosbell(). (8/13/92) rvcursor.x rvfgauss.x Cleaned up some missing sfree() calls before error returns. (8/9/92) t_fxcor.x rvsample.x Fixed a bug in the code that parses the sample parameters. The loop could possibly skip past the EOS in the string, and if there is garbage in the buffer (as could happen when the task in the process cache) it could lead to incorrectly counting the number of ranges and cause a memory fault. I fixed the parser bug and cleared the array before it was filled. Reported by Oegerle. (8/7/92) rvsample.x Fixed a typo in which RV_ERANGE was not being freed (typo was RV_SRANGE in mfree). Found w/ SPPTOOLS. (8/5/92) rvcursor.x coloncmds.x The procedure cmd_write() was declared as an int procedure but not returning anything. Caused a compiler error on DSUX Dec fortran. Also changed rv_query_save() in rvcursor.x since it was returning something but it's return value was never used. (8/5/92) rvrvcor.x help/keywpars.hlp Fixed a typo in the explanation of the date field for DATE-OBS and made the error comment when it's parsed incorrectly more clear (7/31/92) rvmwcs.x Updated with yet another "old-format" special case fix (7/29/92) rvrvcor.x Fixed a memory leak caused by not unmapping images when there was an error return from rv_correct(). The task would grow in memory continuously until it finally ran out and died. (7/28/92) ---------------- V2.10.1 Released rvmwcs.x Updated with latest changes to onedspec$smw.x. (5/29/92) aplists.x Fixed a bug in the ordering of if-clauses that prevented a check against an aperture list trying to be applied to onedspec images. (5/29/92) rvinit.x t_fxcor.x Removed all of the stubs for reading observatory parameters from the image header keyword OBSERVAT. The observatory database tools now used will check the image header automatically to accomplish the same thing. This code was also prone to bugs since it would be possible to pass the wrong pointer the the obsopen procedure. (5/29/92) rvinit.x Fixed a segvio when the obervatory parameter set to "image" (5/26/92) rvmwcs.x Fixed a bug caused by the MWCS changes in which an image with no dispersion info a all could not be read. (5/7/92) aplists.x Fixed a bug wherein the apertures would not be found since the verify loop looked at all the rows in an image even though that may be greater than the number of apertures requested. (5/4/92) ------------------------------- Package frozen for 2.10 release (5/2/92) rvdatacheck.x Added a check to see if, when the dispersion for the two spectra are the same, the difference between starting wavelengths is the same. If not, then we must rebin so that the difference is an integral of the WPC, otherwise we could end up with up to a +/- one pixel error. (4/27/92) rvsinc.x Fixed a bug where the sinc fit would get into an infinite loop if the points in the fit lay above the half power point of the peak. The fix was to return fwhm=INDEF in this case. (4/22/92) rvfitfunc.x The sinc fit wasn't being redrawn if no fwhm was found. (4/22/92) splitplot.x Fixed a bug with contin=none and the 't' keystroke in fft mode not labelling the original template correctly. (4/20/92) fftmode.x splitplot.x Fixed a bug where the filter overlay wasn't being done correctly (4/20/92) rvgetim.x Fixed a bug with the template being a onedspec and having a different aperture number than a onedspec object. The template wasn't being rebinned. (4/16/92) fftmode.x Fixed a bug with the 'i' key not working on single plots (3/24/92) rvmwcs.x Made change to recognize data with APFORMAT='onedspec'. (2/21/92, FV) rvrvcor.x Modified the routine rv_correct() so it no longer uses the obsimcheck() procedure which was removed from xtools$obsdb.x. Now uses obsimopen() (2/8/92) rvbatch.x doc/fxcor.hlp Changed the wincenter parameter so that it takes into account a template velocity if present so the window will be centered on the expected velocity of the peak and not the relative velocity. If no template vel is present, the peak is centered on the relative velocity as before. Requested by oegerle@stsci.edu (1/16/92) rvwrite.x doc/fxcor.hlp Changed the format of the template code string since template names greater than 40 chars would trip a bug in sprintf and cause the write to fail with a "No write permission of file (StringFile)". For image names greater than 30 chars only the rightmost part of the string is printed. (1/14/92) rv.par rvinit.x rvflags.h dispcor.x rvrebin.x rvpackage.h doc/fxcor.hlp Made the rebinning interpolator a package parameter after it was found that the sinc interpolator had disadvantages for certain types of data. The default is once again poly5. (1/10/92) rvcorrel.x Added debug output to error calculations. (1/10/92) rvsinc.x 1) The value of the background wasn't being picked up when changed 2) MAXITER exceeded message wasn't going through error printing code. (1/10/92) coloncmds.x Redo the correlation after setting a new template velocity so we pick it up in the results. (1/10/92) rvfgauss.x 1) Silenced the convergence checking messages. 2) Cleared arrays before input to NLFIT. Was part of tracking a bug found in nlacpts() but should be done anyway. (1/9/92) rvfitfunc.x Fixed a real/double arg type error found w/ spplint. (1/6/92) rvpackage.h Removed any constraint from MAXTEMPS and set it to algorithm max of 702. Need to change this is template caching if re-implemented. (1/6/92) coloncmds.x Fixed the :tnum command to work with the new template naming scheme. Also fixed a small bug already there. (1/6/92) contin.x getdisp.x rvgetim.x rvcomdef.h continpars.x continpars.par doc/continpars.hlp Implemented a new 'replace' parameter for CONTINPARS to allow the rejected points to be replaced by the fit points. Useful for re- moving emmission lines or cosmic ray events. Requested by Oegerle. (1/3/92) rvwrite.x rvpackage.h rvstrings.x readtlist.x Increased the number of MAXTEMPS to 200. The algorithm codes the templates as A,...Z,AA,AB... so the max can be increased to almost 700 by changing the declaration in rvpackage.h. Requested by Oegerle & Hill (1/2/92) mkpkg rvinit.x rvmwcs.x + rvmwcs.h + getdisp.x aplists.x rvpackage.h Added support for new ONEDSPEC MWCS dispersion headers. Nonlinear cannot be rebinned and an error will be returned. Other log and linear formats can be read. This change should be backward compatible (12/31/91) rvfgauss.x rvfparab.x Changed calling sequence to nlerrors() to agree with the math$nlfit library calling sequence. Was causing an error on f68881 machines. (10/11/91) ----------------------------------------------------------------------- Package was installed in V2.10 noao package. 093091 ----------------------------------------------------------------------- rvpackage.h Updated package date. 091091 dispcor.x V210: Changed default interpolant to be II_SINC. 091091 rvwrite.x rvwriteln.x - V210: Consolidated procedures into one logical file. 091091 rvfgauss.x rvfparab.x V210: Changed NLFIT calls to be library version of single precision. 091091 t_fxcor.x rv0$rv0.cl rv0$rv0.hd rv0$rv0.par observatory.x - observatory.par - V210: Moved the obs_get_pars() procedure into inline code in rv_clpars(). Removed Observatory pset from package and use the NOAO observatory task instead. 091091 rvsinc.x numrep.x Moved the brent() procedure to rvsinc.x since it explicitly calls the sinc interpolator and ther routines in numrep.x may be provided by an xtools file in V2.10. 090491 deblend.x Removed Marquardt fitting code duplicated in xtools. 090491 rvparam.x Reset default maxwidth to 21 as in the par file. 090391 rvfitfunc.x Fixed a bug that was not getting the correct points for a 'y' fit. Also fixed a bug that was changing the width parameter and so in a list of stars it was possible to change the params being used without the user knowing it. The symptom was run- ning a list of stars, and then randomly picking out individual correlations without being able to reproduce the results. 090391 complex.x rvfilter.x Fixed a bug causing filter plots to be scaled incorrectly. 090291 rvbatch.x Made an error message printable only from interactive mode. 082991 rvfgauss.x rvfitfunc.x Added some checks for a proper convergence to avoid a FOE if the fit ended because of max iterations or with bogus answers. Fixes a bug reported by DSprayberry at UA. 082991 rvfgauss.x Fixed a type mismatch with the width of the fit region causing a FDZ in the weight fitting calc. Only part of another problem. 082991 coloncmds.x Fixed a typo in the output of dispersion info. 082091 fftutil.c Power spectra were computed with log(), not log10() as is used throughout the rest of the package. Fixed. 082091 fftmode.x - Fixed another segvio as below, but for the 'g' key. - Fixed a bug in fft_inverse which was reporting the wrong frequency with the 'i' keystroke. 071291 rvsinc.x Fixed a bug causing the ccf height to be computed wrongly. 071291 fftmode.x Fixed a segvio caused when pixcorr+ and continuum="none". 071091 coloncmds.x Fixed a bug caused by pixcorr+ trying to convert INDEF to an int when doing a ":wincenter" or ":window" call. 071091 rvwriteln.x Fixed a bug causing the fwhm to be output in pixels for a short .txt file. 070291 doc/fxcor.hlp Added a note that the "original" spectra are splotted on a linear wavelength scale and a slope may be noticed because of rebinning effects. 070291 rvcomdef.h rvcolon.x coloncmds.x doc/fxcor.hlp ../lib/scr/fxcor.key Implemented a ":comment" command to add a comment to the output logs. Requested by Daryl. 070191 rvfuncs.x rvfgauss.x rvfitfunc.x Changed the functional form of the Lorentzian to make it a bit more stable. Previously it was the form of the normalized function but that didn't apply. 062891 prepspec.x Fixed bug with specifying decimal samples on a pixel corr. 063091 asttools/asthjd.x asttools/asttimes.x Changed a constant in the JD calculation that fixes a 0.5 day bug on AOS systems. Also fixed some confused parameters to the procedure ast_hjd(). 062591 rvrvcor.x Added a check for the pixcorr parameter to abort a Vh calculation if it's not needed. 062591 fftutil.x Changed explicit argument dimensions to ARB. This was causing a "variable dimension error" on VMS systems. 061991 rvsample.x rvutil.x Fixed a problem with sample region for pixel-only correlations. 061891 coloncmds.x Fixed some confusion with the ":filter" commands. 061791 rvfitfunc.x Changed height calculation so it's the estimate of the correlation value and not the height of the function. 061791 rvdrawfit.x Made the function erasing a bit smarter so that the FWHM indicator is erased the same way it was drawn. On xterms, a line was being drawn across the whole screen because it didn't support graphics erase. 061491 fftmode.x Fixed two typos: One causing the 'f' keystroke to act like the 'p' keystroke, and another causing the fft plots to be incorrectly calculated. 061491 rvsample.x specmode.x Added some more functionality to the sample regions code, including a new 'u' keystroke that will automatically de-select a sample region. Also, sample may now be merged or expanded by using the 's' keystroke. 061491 prepspec.x Fixed a bug causing the sample regions to accidently be apodized for the filter plots. 061391 t_fxcor.x rvcursor.x coloncmds.x Added some checks so that the filter parameter cannot be changed (specifically, turned on) before filter values have been set. Previously the task would get in a confused state because the data check prior to the correlation would fail and there was no recovery. 061191 t_fxcor.x splitplot.x rvsample.[xh] Added code so that samples specified as pixels will still be plotted if dispersion info is present and the user hasn't requested a pixel-only correlation. 061191 splitplot.x Fixed a samples ranges plotting bug on summary plots. 061191 ------------------------------ Package updated on local systems for testing. 060591 wxcor.h rvwriteln.x Added the rebin parameter to the output codes. 060591 readtlist.x rvimutil..x Optimized a bit to remove extra rv_getim() calls that were only used to get template velocities. 060491 rvfitfunc.x rvfftcorr.x Fixed a memory bug being tripped up when images with differing sizes are passed in the list. Also made sure correlation array is big enough after the rebinning, previously is could have been off by a few elements. 060491 rvflags.h fxcor.par t_fxcor.x rvcolon.x rvparam.x rvcomdef.h rvpackage.h rvstrings.x coloncmds.x doc/fxcor.hlp rv0$lib/scr/fxcor.key Implemented a new "rebin" parameter to give user control over which spectrum is rebinned. 060391 rvwriteln.x doc/fxcor.hlp Added output of rebinned dispersion to logs when it differs from previous so user can be aware of rebinning changes. Since data are always rebinned to the lowest dispersion, log entries should be minimized although interpolation errors may increase. The help page was also modified to alert useras to this fact. 060291 contin.x rvplot.x fftmode.x rvrebin.x specmode.x readtlist.x Fixed usages of RV_(R)NPTS macros so the are consistent with the spectra used. Since relaxing the constraint that the rebinned npts match we need to make sure both object and templates are allowed to be arbitrary lengths, and that code know what the lengths should be. 060291 rvsinc.x numrep.x rvflags.h rvfgauss.x rvfparab.x rv0$rv0.par rvpackage.h Made fitting tolerances and max iterations a package param. 053091 rv0$rv0.par rvpackage.h rvrvcor.x Made the z_threshold a package parameter so that users have control over the threshold at which output is printed as redshift z values rather than velocities. 052391 rvinit.x rvplot.x rvutil.x fftutil.x fxcor.par rvcolon.x rvflags.h rvparam.x rvrvcor.x rvwrite.x t_fxcor.x prepspec.x rvcomdef.h rvfilter.x rvfparab.x rvsample.h + rvsample.x + specmode.x rvpackage.h coloncmds.x rvfftcorr.x rvsumplot.x splitplot.x rvdatacheck.x Implemented multiple and independent sample regions. This involved the addition of a new structure containing the ranges information. This structure can be accessed directly by referencing the struct pointer, or the structure for either object or template can be accessed with the main rv struct pointer by using macros. Work routines are passed only the sample struct pointer and calling routines figure out which spectra the sample applies to and pass in the appropriate pointer. 052191-052391 rvutil.x rvcursor.x specmode.x Fixed a GIO wcs bug causing endpoint marking keystrokes to get confused. 052091 fxcor.par doc/fxcor.hlp Changed default maxwidth to 21. 051891 rvsinc.x + rvvfit.x rvplot.x rvflags.h fxcor.par rvdrawfit.x rvstrings.x rvfitfunc.x doc/fxcor.hlp Began implementation of a sinc interpolator and Brent peak finding algorithm. 051791 rvbatch.x Fixed bug that was turning off object continuum subtraction. 051791 rvutil.x Fixed a small bug with using the 's' to specify a sample. Range was being checked against the log of the wavelength causing the range to be e.g. 3.09-5400. 051691 rvgetim.x rvrebin.x splitplot.x Fixed rebinning bug causing spectra to somtimes be truncated if the difference in dispersion was great enough. That is, the number of data points was being fixed rather than being allowed to vary so that the same wavelength region was covered. In the process I also fixed a small plotting bug for spectrum plots. 051691 rvsumplot.x rvfitfunc.x Fixed drawing of FWHM indicator for parabolic fits. This was reported as a bug but in fact the indicator was never drawn. Since we want the FWHM to be computed from the coefficients and not an empirical calculation, this is the way it is drawn although it looks like it is underestimated. 051491 rvcolon.x specmode.x rvcomdef.h coloncmds.x rv0$lib/scr/fxcor.key rv0$src/doc/fxcor.hlp Changed the ":wpc" command to be ":disp" so it prints out all dispersion information. 051491 rvcolon.x rvcomdef.h coloncmds.x rv0$lib/scr/fxcor.key rv0$src/doc/fxcor.hlp Implemented a ":printz" command to toggle redshift output. 051491 rvbatch.x Added more bounds checking to window computation. If a window was set too large or the center was outside the ccf bounds, a segvio would occur. 051491 aplists.x Fixed bug with onedspec images returning an ERR apnum. 051491 wxcor.h rvflags.h rvpackage.h rvwriteln.x Implemented output of redshift z values when relative velocity exceeds a threshold (set at 0.2). (User request) 051491 rvvfit.x rvbatch.x rvparam.x rvwrite.x t_fxcor.x rverrmsg.x coloncmds.x rvwriteln.x rvstrings.x Implemented the "verbose" parameter as an enumerated string to suppress metacode or log files (User request). 051391 deblend.x dispcor.x getdisp.x rvparam.x rvrebin.x continpars.x Cleanup up some more type inconsistencies found by SPPLINT. 051391 rvgetim.x Fixed typo causing template aperture to be incorrectly printed in logs if only 1-D. 051291 rvimutil.x Fixed bug so that aperture lists are updated for each new image. If aperture="*" and the apertures in the image varies, the original last was always being used causing an "Aperture out of range" message. 051291 rvdrawfit.x Changed line type of fitted function to be dashed so it's dis- tinquished from the ccf (user request). 051091 rvcorrel.x Removed an extra factor of 2 from the computation of the anti- symmetric noise component in rv_antisym(). This was causing an incorrect value of the T&D R to be computed. 050191 rvkeywords.h Changed size of the 'struct' to be 16. This clears up a memory bug that has been present in the system since dirt but has only recently made my life miserable. 051091 fxcor.par Changed boolean parameters declared as quoted "yes"/"no" to be true boolean values yes/no. 050991 wrtccf.x rvplot.x aplists.x rvparam.x fftmode.x rvparam.x zzdebug.x rvimutil.x specmode.x filtpars.x coloncmds.x rvdrawfit.x continpars.x Fixed some argument type mismatches and number of arguments. This may have been the cause of the long-time "Salloc Undeflow" error. All of these were caught using the public domain "f2c" program in combination with UNIX "lint" on the resulting C code. 050291 rvinit.x Fixed typo indentation which was making the logic look different than it actually was. 042991 rvgetim.x Returned from rv_ap2line() after determining format was ONEDSPEC rather than continue to parse APNUM keywords which may be different values. 042991 rvbatch.x Fixed bug causing template velocities to be confused in batch mode with a template list. 042691 wxcor.h Straightened out alignment of HGHT and VFWHM columns. 042591 rvdrawfit.x Fixed bug causing fitting function and ccf to not be drawn in batch mode. 042591 rvpackage.h rvdrawfit.x rvcursor.x Added code to avoid the 'erasing' of a possible residual plot when refitting a peak. On xterm windows this shows up as a draw and further messes up the plot. 042191 rvrvcor.x Removed calculation and reset of "verror" from the procedure. The actual error from the T&D computations is done in the fitting routines always called prior to the rv_rvcorrect() procedure and this was just trashing a correct error. By deleting the reset of verror = INDEF, we retain a velocity error estimate for the relative vel when a full heliocentric correction can't be computed. 042291 fftmode.x specmode.x rvxcomdef.h coloncmds.x Added output for ":wpc" command to print rebinned wpc. 042191 splitplot.x Fixed typo preventing points in summary plot ccf being plotted. 042191 /arc/ftp/iraf.old/rv0.tar.Z Fixed RV_VREL bug in archive since support for high redshift work is lacking and it seemed like an important fix (a 'no-no'). 4/17 22:23 doc/continpars.hlp Added a note on where to look for syntax of "c_sample". 040491 doc/fxcor.hlp Changed wording of rebinning order to reflect implementation. 040491 rvrvcor.x Changed argument to rv_shift2vel() so shift is real and not double. This was causing incorrect VREL values to be printed. 040491 ------------------------------------------------------------------------------- Package frozen and archived for Beta Release. 040191 ------------------------------------------------------------------------------- fftmode.x rvcolon.x rvfilter.x specmode.x coloncmds.x Cleaned up some missing sfree() calls. 032091 contin.x rvbatch.x rvgetim.x rvrebin.x rvcursor.x coloncmds.x readtlist.x Changed calls to do_continuum() so they pass int flags rather than a character in order to avoid confusion and problems with chars/int getting confused. 031991 rvbatch.x Fixed bug in window params when center is specified. 031991 rvbatch.x Fixed duplicate continuum subtraction being done in batch mode. 031591 rvplot.x Fixed min/max compiler bug under VMS. 031491 fftmode.x specmode.x Fixed bug with the stridx() function having an int argument when a char is requested. Since the integer*4 keystroke gets passed into a integer*2 argument, the info in the high two bytes is lost. 031491 rv0$lib/scr/fftmode.key Added to 'o' and 't' commands to menu. 031491 rvbatch.x coloncmds.x contin.x Removed 'b' option to do_continuum(), and removed salloc declarations in continuum() since there was no sfree and they weren't being used anyway. 031391 rvdatacheck.x splitplot.x Fixed mixed-type arguments to min/max causing problems on the VMS 5.4 compilers. 031291 rvutil.x rvflags.h Removed disp() function. Also fixed a bug in the rv_shiftspec() procedure causing bad approximations for large redshifts in the 'd' keystroke in specmode. 031291 rvvfit.x Expanded output on .log file to 24 chars for image names. Added note to docs saying image names may be truncated. 031191 getdisp.x Saved data format for later use in output log. 031191 specmode.x Added 'd' keystroke to not be remembered from SPECMODE. Fixed bug with 'q' key being remembered. 031191 rvplot.x rvutil.x rvvfit.x wrtccf.x rvbatch.x rvrvcor.x rvsumplot.x coloncmds.x rvdrawfit.x rvwriteln.x splitplot.x Made the calculation of the relative velocity uniform throughout the code. The relativistic (1+z) form is now used since things like plot scales and so on were slightly off at high redshift. Talked w/ Silva baout this and he agrees. 030891 splitplot.x Changed calculation of plot velocity scaling so it agrees w/ cursor readout at high redshifts. This was previously done as shift*deltav which is only an approximation. 030791 ------------------------------------------------------------------------ Updated noao systems with latest version prior to final release. 030791 ------------------------------------------------------------------------ rv0$rv0.cl Expanded min_lenuserarea to 40000 to avoid problems encountered by John Hill with large 2-D images. Code taken from nessie$nessie.cl script. 030491 wxcor.h rvwrite.x rvwparam.x rvwriteln.x Added height to output and reorganized fields. 030291 rvwrite.x specmode.x coloncmds.x rvwriteln.x Changed output of sample regions so it's only written to logs when the results are written. 030191 coloncmds.x Fixed bug with ":output" not working. 030191 rvcursor.x splitplot.x coloncmds.x doc/fxcor.hlp rv0$lib/scr/fxcor.key Implemented ymin/ymax commands to set ccf plot scaling. 030191 rvinit.x Fixed typo initializing RV_WINR. Added y1,y2 inits. 030191 rvfitfunc.x Added checks to the center1d fit to see if it returns an INDEF shift. A user reported floating overflows, which was tracked to the threshold level being set too high for his data. 022891 rvplot.x Fixed bug with 'z' not working on pixel-correlation plots. 022291 rv0$lib/scr/fxcor.key doc/fxcor.hlp rvcursor.x, splitplot.x Changed 'c' keystroke to be a cursor readout to help solve the problem of users trying to use 'C' to get cursor position. Also added note in the help page that 'C' is not recommended because of confused GIO w/ multiple WCSs. Renamed 'c' to 'm'. 022291 titles.x Removed fplot_title() and pplot_title() - not used. 022291 specmode.x Fixed bug causing 's' to be recognized as a 'sample' command when specmode was entered the 1st time and the last key was a sample sel- ection. Leaving specmode and re-entering with 's' from ccf mode caused odd behavior. Only occured with continuum=none. 021591 rv0$src/doc/fxcor.hlp, rv0$lib/scr/fxcor.key rvpackage.h, rvcomdef.h t_fxcor.x, getdisp.x, aplists.x, coloncmds.x Added a new parameter "pixcorr" which controls whether or not the dispersion info in the header is used to rebin the data. If set to 'y', then data are not rebinned and no velocities may be computed. 021591 coloncmds.x Fixed typo causing ":contin" to no take new value. 021591 splitplot.x Fixed bug causing the top FFT plot to be incorrectly scaled. 021391 rvvfit.x Fixed output of vel error when no dispersion info present. 020891 rvdatacheck.x Added check for dispersion info when calculating deltav. 020891 getdisp.x Added code to get CD_1 if CDELT1 not present, and made sure that w0/wpc are not uninitialized. 020891 rvranges.x Fixed bug that prevented range strings from being properly decoded. The parsing loop wasn't being executed and the rangecount flag was never reset from ALL_SPECTRUM. 020891 rvranges.x rvdatacheck.x Added more checks for rangecount and dc-flag values so regions would be interpreted inthe right units. 020891 specmode.x Fixed duplicate greactivate() call for ":show". 020891 ------------------------------------------------------------------------------- February 1, 1991 Test Version Released ------------------------------------------------------------------------------- .endhelp