Problem solved in recent versions are: 9 June 97 version Strfits was rewritten so that fits files with ascii file extensions buried in their rootnames are now converted to ascii even if they do not contain the flag that would mark them as tables. Several new functions were written: chk_adcname and chk_tabname in check.x, which check to see if a file is convertable to an ascii table by checking its extension, and gen_fname in genfname.x, which generates the output filename from the name encoded in the fits file. Change_name, which creates the filename for diagnostic prints and prt_tabinfo, which renames the table from its temporary name, both in fits_read.x, were also significantly modified. Smaller changes were made to tab_read_header and t_rfits. 14 May 97 version o Force parameter added A new task parameter, force, was added to strfits. There are certain fits files which strfits cannot convert, because the fits standard is more general than either the geis or tables standard and some fits files have no representation as either type of file. Generally these are fits files with extensions that were not originally converted from images or tables. Strfits tries to detect thes files and leave them as fits files, but sometimes it guesses wrong and leaves a file unconverted which should have been converted. If this task parameter is set to yes, strfits will try to convert all files into either an image or table. But it may not succeed, and may crash in the attempt. o Table format string changed Strfits was modifying the print format by sticking a negative sign in front of the format for string columns. A recent change to the tables library has it set print formats for string columns so that they start with a negative sign. As a result, print formats were winding up with two negative signs in front. I deleted the code in strfits which added a negative sign, so now the print format only gets one ngative sign. 11 May 97 version o Puzzling strfits error message Strfits would crash with a mysterious error message when the output file name was specified without an extension and intype was set to fits. The problem was that strfits cannot convert fits files to fits files. The code now checks the value of imtype when there is no filename and prints an error message if imtype is not imh or hhh. The only code change was two lines added to see_extn in t_rfits.x. o Strfits header info fixed The output from strfits was not printing the number of columns in an ascii file correctly (it should always be one.) This is because the variable ncols was left undefined in fits_read.x when the file was an ascii file. I modified the code to set ncols to one. o Argument list mismatch in strfits One of the fixes in the change to increase the maximum length of a string parameter in stwfits was to eliminate the length argument from wft_encodec in fits_params.x. This change was made for all the calling sequences in stwfits; however, strfits calls the same procedure and I did not make the changes there. This led to a segmentation violation in some cases. I have now corrected the calling sequences in strfits as well. Fortunately,the changes to the calling sequence were limited to rft_create_gpb in rft_subs.x. o Extension length increase I discovered that fitscopy, strfits, and r49fits all assumed that the extension length (SZ_EXTN) was only three characters. So I increased the length to 31 characters in all these programs to match stwfits and changed the call of strncmp to strcmp, just as I previously did in stwfits. o Strfits now handles fits file with extensions The former version of strfits would often crash when asked to convert fits files with complex sets of extensions. This is because fits files with extensions often have no equivalent geis image or sdas table. This presents a problem when reading files from tape. To resolve the problem, strfits now looks to see if a fits file has extensions and only converts files with extensions of the were originally derived from multigroup images or sdas tables. Otherwise, the fits file is copied as is.