Provided by: libfortran-format-perl_0.90-1_all bug


       Fortran::Format - Read and write data according to a standard Fortran 77 FORMAT


           use Fortran::Format;

           my $f = Fortran::Format->new("2('N: ',I4,2X)");
           print $f->write(1 .. 10);
           # prints the following:
           # N:    1  N:    2
           # N:    3  N:    4
           # N:    5  N:    6
           # N:    7  N:    8
           # N:    9  N:   10

           # if you don't want to save the format object,
           # just chain the calls:
           Fortran::Format->new("2('N: ',I4,2X)")->write(1 .. 10);


       This is a Perl implementation of the Fortran 77 formatted input/output facility.  One
       possible use is for producing input files for old Fortran programs, making sure that their
       column-oriented records are rigorously correct. Fortran formats may also have some
       advantages over "printf" in some cases: it is very easy to output an array, reusing the
       format as needed; and the syntax for repeated columns is more concise. Unlike "printf",
       for good or ill, Fortran-formatted fields never exceed their desired width. For example,

           printf "%3d", 12345;                            # prints "12345"
           print Fortran::Format->new("I3")->write(12345); # prints "***"

       This implementation was written in pure Perl, with portability and correctness in mind. It
       implements the full ANSI standard for Fortran 77 Formats (or at least it should). It was
       not written with speed in mind, so if you need to process millions of records it may not
       be what you need.


       What follows is a very brief summary of Fortran formats. For a rigorous description, see
       the ANSI standard. A format consists of a list of "edit descriptors" or sublists of edit
       descriptors. Edit descriptors are separated by commas, but the comma may be omitted if
       there's no ambiguity. Spaces and case are ignored, except within strings, so 'i 1 2' is
       the same as 'I12'.

   Repeatable edit descriptors
       The following edit descriptors may be repeated if they are preceded by a number; for
       example, '3I4' is the same as 'I4,I4,I4' or 'I4I4I4' or 'I4,2I4'. Lists can be nested by
       using parentheses, so '2(I2I3)' is the same as 'I2I3I2I3'. Most descriptors include a
       width w. If the width is larger than needed, the output is right-justified. If the width
       is not large enough, the entire field is filled with asterisks.

           An integer with width w, and optionally a minimum number of digits m (adding zeroes on
           the left if needed).

           An fixed precision floating-point number with width w, and d digits after the decimal

           A number in exponential notation with width w, d digits after the decimal point, and
           optionally e digits after the exponent.

           For values between 0.1 and 10^d, format like F. For values outside that range, format
           like E.

       Fw  Treat the variable as Boolean and output either T or F in a field of width w.

       Aw  Insert a string variable. If the width is not specified, it outputs the entire string.
           If the width is smaller than the string, the string is truncated (instead of filling
           with asterisks).

   Non-repeatable edit descriptors
       Most of the following descriptors don't output anything but act as control strings. "Non-
       repeatable" descriptors can be repeated only by including them in a repeated list within

           Insert string as is. Quotes may be escaped by doubling them; for example, 'Joe''s'
           produces Joe's.

           Insert The next n characters after the H as is.

       TRc Move to position c of the current record (T), or c characters to the left (TL), or c
           characters to the right (TR).

       nX  Move n characters to the right.

       /   Move to the begining of the next record (the next line).

       :   Stop producing output immediately if there are no more variables left to format.

       SS  Control whether the plus sign is included for positive numbers. Include it for SP, do
           not include it for SS, and use the default (do not include) for S.

       kP  Scaling factor for output in exponential notation. By default, a number such as 1.23
           would be written as 0.123E+01. When a scaling factor k is given, the decimal point is
           shifted k positions to the left and the exponent is decreased by k orders of
           magnitude. With 1P the output would be 1.23E+00.


               my $format = Fortran::Format->new($format_string);

           Create a new format object. The string is parsed and compiled when the object is
           constructed. Croaks if there is a syntax error.

               my $format_string = $format->format;

           Returns the format string used by the object.

               $output = $format->write(@data);

           Formats the data. This is equivalent to the Fortran "WRITE" statement, except that it
           just returns the formatted string. It does not write directly to a file. Data items
           may be either scalar or array references (which can be nested).

           For matrices (multidimensional arrays), the contents are formatted in column-major
           order, same as in Fortran. For example,

               my $a = [[1,2],[3,4]];

           will print

               1   3   2   4



           will print

               1   3
               2   4

           This is effectively equivalent to transposing the matrix before printing it in the
           row-major order that would be expected by most non-Fortran programmers. This kludge is
           necessary to ensure that the output can be read properly by a Fortran program.

           Note: this is incompatible with Fortran::Format 0.5x, which simply flattened the
           nested arrays, producing the output in row-major order. Also note that the behavior is
           undefined if the nested array is not rectangular. For example, [[1],[2,3]] will give
           strange results.

               my (@results) = $format->read($fh, @input_list);

           Read data from the filehandle $fh using the format ($fh can also be a string instead
           of a filehandle). The input list is a list of array sizes: 1 for simple scalars, n for
           simple arrays, and an array reference of dimensions (such as [3,3]) for
           multidimensional arrays. For example,

               my ($i, $matrix, $j) = $format->read($fh, 1, [3,3], 2)

           will read one scalar, followed by a 3x3 matrix, followed by an array with size two.
           Note: this method should be called in list context!

           The input list is needed because Fortran formats are reused automatically for
           subsequent lines until all the variables are read.

           Matrices are read in column-major order. See "write" for details.

           When reading, it is also possible to specify the length of the resulting string
           variables by appending "Alength". For example,

               my $s = $format->read($fh, '1A40')

           will read the data into a 40-character long string variable (this is regardless of the
           field width specified in the format string itself). The string will be padded with
           trailing spaces if needed to ensure that it is exactly 40 characters long. This
           attempts to emulate Fortran's peculiar string length semantics. It is needed if you
           want to read a string, write it back, and be sure that you get the exact same output
           that you would get with Fortran.

           For example,

               my $in = 'hello world';
               my $a5  = Fortran::Format->new('A5');
               my $a20 = Fortran::Format->new('A20');

               my ($s) = $a5->read($in, '1A10');

               print $a20->write($s);
               # prints "          hello     "

           Notice that 1) $s was padded with five space, to a length of ten characters; 2) the
           output is right-justified to a total width of 20 characters.

           Now, if we do this instead:

               my ($s) = $a5->read($in, '1A3');
               print $a20->write($s);
               # prints "                 llo"

           Five character are read from the left of the string ("hello"), but only the rightmost
           three are copied to the 3-character-long variable ("llo").




       The Fortran format specification:


       Ivan Tubert-Brohman <>


       Copyright (c) 2005 Ivan Tubert-Brohman. All rights reserved. This program is free
       software; you can redistribute it and/or modify it under the same terms as Perl itself.