Function:     HERINP

Purpose:      Decodes a string of characters into reals, integers,
              logicals, characters or bytes.

Category:     MATH, UTILITY, USER-INTERFACE

File:         herinp.c

Author:       K.G. Begeman

Use:          CALL HERINP( EXPR ,    Input   CHARACTER*(*)
                           TYPE ,    Input   CHARACTER*(*)
                           LENT ,    Input   INTEGER
                           OUTV ,    Output  ARRAY of BYTES
                           NOUT ,    Input   INTEGER
                           NRET ,    Output  INTEGER
                           IERH )    Output  INTEGER

              EXPR  Character string containing the expressions to be
                    evaluated.
              TYPE  Character denoting type of conversion wanted. See
                    the notes for a list of possible types.
              LENT  Length of type in bytes. See the notes for a list
                    of possibilities.
              OUTV  Array of bytes containing the decoded values.
              NOUT  Maximum number of values of type TYPE to be
                    returned.
              NRET  Number of values actual decoded.
              IERH  Error return code. A list of error codes is given
                    in the notes.

Notes:        1. The syntax of the input for this routine is described
                 in document INPUT.DC2
              2. List of possible types and allowed lengths:
                 TYPE         LENT     type
                  'F'         4,8      reals
                  'I'          2,4      integers
                  'L'         1,2,4    logicals
                  'A'         any      plain bytes
                  'C'         any      bytes with separator recognition
              3. List of error returns:
                 IERH  explanation
                   0   no error
                 -11   bad call
                 -12   unknown function
                 -13   syntax error
                 -14   illegal character
                 -15   wrong repeat argument
                 -16   wrong number of arguments
                 -17   arithmetic error
                 -18   not enough internal memory
                 -19   conversion error
                 -20   unequal list length
                 -21   empty list
                 -22   nested lists
                 -23   output buffer overflow (not for 'A' type of input)
                 -24   floating overflow/underflow in conversion

Updates:      Dec 16, 1987: KGB, Type C implemented .
              Jan 25, 1988: KGB, Document created.
              Feb 16, 1988: KGB, In GENLIB.
              Mar 23, 1988: KGB, Added argument NCHR.
              Apr 11, 1988: KGB, Disabled check of buffer overflow
                                 for type 'A'
              Apr 14, 1988: KGB, Error if no value after a ','
              May  5, 1988: KGB, Small but important change in
                          document
              Oct 13, 1988: KGB, Implemented ANSI F77 calls to DCDCHAR,
                                 DCDINT, DCDLOG, DCDREAL, DCDDBLE
              Nov 23, 1988: KGB, Handles now BLANKS
              Dec  5, 1988: KGB, Bug in UNIX version removed
              Mar  1, 1989: KGB, Bug in decoding logicals removed
              Jul 27, 1989: KGB, FtoC interface implemented.
              Jul 22, 1994: JPT, Call Hermes for decoding numbers.