PURPOSE:   (one line only)
  General purpose routine for converting between sexigesimal and decimal.
  This routine encapsulates all of the parsing and conversion routines for
    handling all conversions between sexigeismal strings and double
    precision forms of angular coordinates.  Floating point is allowed on
    input but its precision is not really good enough for the taks so any
    decimal output will always be double precision.  Scalar and vector
    inputs are permitted.  A string on input is always an indication that
    the input is sexigesimal and the information is to be converted to
    decimal format.  If the input is floating point or double then the
    conversion is to a sexigeismal string.  There are a number of specialized
    conversion routines in the library but they all call this basic routine
    setting all the required options.

  A word about the parser: When reading strings, the information is parsed
    character by character.  This routine ignores leading blanks.  But, a
    blank is considered to be a terminating character once the parser has
    seen a non-blank.  Legal field separation characters are ':' and ','.
    Any and all fields can be floating point though decimal fields only
    really make sense if that is the last field provided.

  On output you can modify the separator character to anything you like but
    this routine will not be able to read arbitrary seperators.

  Enough options are provided to permit meaningful range checking on the
    input values for different types of angular quantities.  The quantities
    envisioned for this routine include Right Ascension [0-24h), Declination
    [-90,90], Hour angle [12W,12E), Longitude [0,360) to name a few.

  input - scalar or vector.  String          ==> sexigesimal to decimal
                             Double or float ==> decimal to sexigesimal
  minvalue - scalar, minimum allowed value for coordinate in decimal form.
  maxvalue - scalar, maximum allowed value for coordinates in decimal form.
  signchar - two-element string giving the characters to be used to indicate
               the sign of the number on reading or writing.
                 signchar[0] = Character for positive numbers
                 signchar[1] = Character for negative numbers
  wrap     - Flag, if set indicates the coordinates wrap across from max
                to min.  This further indicates the range is inclusive
                of minvalue and exclusive of maxvalue.  If not set, the
                range is inclusive of minvalue and maxvalue and an
                out-of-range is an error.


  HOURS    - Flag, if set, indicates the sexigesimal string is in units
                of hours.  The default is degrees.
  DEGREES  - Flag, if set, indicates the decimal value is in degrees.
                The default is radians.

  Keywords are grouped by the input data type since that dictates the
    operation that ensues:

  Double or float input:

    PLACES - Resolution of output string (default=0)
             = -8     Pure decimal output               XX.xxxxx
             = -7     nearest top level unit            XX
             = -6     nearest 30 minutes.               XX:(00,30)
             = -5     nearest 15 minutes.               XX:(00,15,30,45)
             = -4     nearest 10 minutes.               XX:M0
             = -3     nearest 5 minutes.                XX:M(0,5)
             = -2     nearest minute.                   XX:MM
             = -1     nearest ten seconds.              XX:MM:S0
             =  0     nearest second.                   XX:MM:SS
             =  1     nearest tenth of a second.        XX:MM:SS.s
             =  2     nearest hundredth of a second.
             =  3     nearest thousandth of a second.   XX:MM:SS.sss
             =  4     nearest thousandth of a second.   XX:MM:SS.ssss

    SEPCHAR - Scalar string of the separator character to use between
                fields on output.  Default=':'.  Note that if you set
                this to something other than ':', ',' or ' ' the output cannot
                be parsed by this routine as input.

  output - length matches input.  Double if input is string, string if input
              is double or float.
  CARRY  - Used on decimal to sexigesimal string conversion only.  This is
              a flag that will incidate if the number rolled over the cut
              in the coordinate (if a cut exists).
  ERROR  - Flag (length matches input).  Set if there was a conversion error.

  Right ascension, radians in decimal form, hours in sexigesimal
  Right ascension, degrees in decimal form, hours in sexigesimal
  Declination, radians in decimal form, degrees in sexigesimal
  Hour angle, radians in decimal form, hours in sexigesimal

  Written by Marc W. Buie, Southwest Research Institute, 2009/04/24
    This is created from a large suite of independent but very similar
    routines with the intent of removing all duplicate code and having
    the independent routines call this master routine.
  2010/04/18, MWB, modified internal loop variable to handle long input