NAME

C::Include - Package to easy operate with binary data via describing they like C/C++ structs.

CLASSES AND PACKAGES

      C::Include         - Header file base parser class
      C::Include::Struct - Struct wraper class

USAGE

use C::Include;

POSSIBILITY

CLASS C::Include

This class is a header file parser

FUNCTIONS

new ( params )

        As params there should be pairs (-name, value). If the name of the
        param is absent - then it is accepted as 'source'.

        Current version support two params types: -src and -cache.

        -src param describe header file. May be \*FILEHANDLE or
        'path/to/file.h'.

        -cache say, that was cached the compiled header file. If the value
        of this parameter is absent and -src the parameter describes path to
        the file, the cache the file, will be created as 'Path/to/file.ph'.
        Pay attention that, while the value of this parameter is absent and
        -src the parameter is \*FILEHANDLE, the cache the file will not be
        created.

         Examples of usage:
   
            # Creating object from external header file
            my $include = new C::Include( -src, '/usr/src/include.h' );
            my $include = new C::Include( '/usr/src/include.h' );
            my $include = new C::Include( '/usr/src/include.h', -cache );
            my $include = new C::Include( '/usr/src/include.h', -cache, '/tmp/cache.ph' );

            # Create object from embended header file in Perl
            # script after __DATE__ directive
            my $include = new C::Include( \main::DATA );
            my $include = new C::Include( \main::DATA, -cache, '/tmp/cache.ph' );

            # Alternative constructor call 
            use C::Include qw(/usr/src/include.h -cache);

sizeof

Return length of type in bytes.

Examples of usage:

            my $size_of_int = $include->sizeof(int);
            my $size_of_struct = $include->sizeof(HEADER_STRUCT);

make_struct ( 'struct name' )

        Returns a copy of the object C::Include::Struct - wrapper of
        structure.

         Examples of usage:

            # Create wraper class for struct described in $include as 'HEADER_STRUCT'.
            my $header = $include->make_struct( 'HEADER_STRUCT' );

INC( [module_name] )

        Return REF to alternative constructed instance C::Include in module
        "module_name". If "module_name" param not defined, then - current
        call module name;

        Note: There should be only one created instance of C::Include for
        each calling unit module.

         Examples of usage:

            package My::Cfg;

            # Alternative constructor call 
            use C::Include qw(/usr/src/include.h -cache);

            # Create wraper class for struct described in /usr/src/include.h
            # as 'HEADER_STRUCT'.
            my $header = INC->make_struct( 'HEADER_STRUCT' );

         in other scripts may be called instance of C::Include created in package My::Cfg as

            use C::Include;
            my $header = INC('My::Cfg')->make_struct( 'HEADER_STRUCT' );

CLASS C::Include::Struct

This class is a struct wrapper.

FUNCTIONS

pack ( [\*FILEHANDLE] )

Pack struct fields to binary and, if need, save it to file.

        As param may be REF to FILEHANDLE. If FILEHANDLE exists then packed
        binary data will be saved to this file.

        Function return resulted binary data as scalar.

         Examples of usage:

            $header_struct->pack( \*OUT_FILE );

            my data =  $header_struct->pack();

unpack

            unpack( data )
            unpack( \*FILEHANDLE, [offset, [size]] )

        To unpack the binary informations, and to fill by them in the
        fields.

        If first param is scalar, then it is binary data to unpacking.
        Otherwise if first param is \*FILEHANDLE, then input binary data
        will be reader from this file. If 'size' param not defined then from
        file will be readed same counte of bytes as struct length. Param
        'offset' describe count of bytes to skip before reading binary data.
        Default offset equal 0;

         Examples of usage:

            $header_struct->unpack( $binary_data );
            $header_struct->unpack( \CFG );
            $header_struct->unpack( \CFG, 10, $header_struct->size );

size

Return length of struct in bytes.

Examples of usage:

            # 8)
            print 'Size of struct HEADER_STRUCT: ', $header_struct->size(), ' bytes';

EXAMPLE

use C::Include;

        # Create instance of object and cache it to 'data/database.ph'
        # If 'data/database.ph' exists then only read compiled data from they.
        my $include = new C::Include( 'data/database.h', -cache )
            or die "Can't read include: $!";
  
        my $cfg = $include->make_struct('CONFIG')
            or die "Can't create instance of CONFIG: $!";

        # Open binary file 
        open CFG, 'database.cfg';

        # Read from CFG length of struct wraped by $cfg and fill self fields
        $cfg->unpack( \*CFG );

        close CFG;

FAQ

INTERNAL REMARKS

  1. For all field definitions with type array of chars will be produced one string field:
        struct file{
          char path[__MAXPATH];   // first string field
          char name[12];          // second string field
        }

If you want to create array of bytes for definitions as array of chars, you must change type, for example, to 'byte':

        struct array{
          char path[__MAXPATH];   // first string field
          byte flags[12];         // second field as array of unsigned chars
        }

        or 

        #define BYTE char
        struct array{
          char path[__MAXPATH];   // first string field
          BYTE flags[12];         // second field as array of signed chars
        }

2. All properties of C::Include::Struct object be in element with key {''}. Do not delete and do not change them!:

delete $header_struct->{''};

3. You can describe in headers exclusive definitions like compiller sensivity

Example
        #ifdef __WATCOMC_
            // some definitions
        #endif

        // __PERL__ is predefined in package default
        // and processed correctly
        #ifdef __PERL__  
            #define skiptype neganull
        #endif

4. For optimizing memory usage and speed for pack/unpack machanizm:

in this sample code:

        1   typedef struct HEADER{
        2       word revision;
        3       byte reserved[1024];  // in wraper object will be allocated
        4   };                        // array with 1024 elements and take CPU
                                      // to process in every pack/unpack operation
      line 3. best replace to 

        char reserved[1024];      // will be created one string element
                                  // with length 1024
      or 

        #ifndef __PERL__
        #define nullbyte byte
        #endif
        nullbyte reserved[1024];  // this 1024 bytes will be skiped
                                  // in operations pack/unpack and
                                  // not allocated in wraper object
                                  // as element

5. Note, that the current version processes of the preprocessor commands only once, before parsing of the main code!

In this code:

        #define __MAXPATH 56
        struct file{
          char path[__MAXPATH];
          char name[12];
        }
        #define __MAXPATH 127

      Length of a field file.path will be, equal 127 bytes!

6. All defined values via '#define' and 'enum' are availablis:

        my $include = new C::Include( 'data/database.h', -cache )
            or die "Can't read include: $!";

        print $include->{__MAXPATH};

7. All copies of the predetermined structures also are availablis:

        struct FILE{
          char path[__MAXPATH];
          char name[12];
        }root, files[100];

        $\ = "\n";
        print $include->{root}{name};
        print $include->{files}[$_]{name} for 0..99;

8. For parsing progress indicating, need switch a flag 'TEST' to '1':

        use C::Include;
        $C::Include::TEST = 1;

        my $include = new C::Include( 'data/database.h', -cache )

AUTHOR

      Albert N. MICHEEV
      email: Albert@f80.n5049.z2.fidonet.org
      fido:  2:5049/80

COPYRIGHT

Copyright 2000, Albert N. MICHEEV (Albert@f80.n5049.z2.fidonet.org)

This library is free software; you can redistribute it and/or modify it under the same terms as Perl itself.

AVAILABILITY

The latest version of this library is likely to be available from:

http://www.perl.com/CPAN

SEE ALSO

Storable.