SYNOPSIS

use Text::VimColor;
my $syntax = Text::VimColor->new(
file => $0,
filetype => 'perl',
);
print $syntax->html;
print $syntax->xml;

DESCRIPTION

This module tries to markup text files according to their syntax. It can be used to produce web pages with pretty-printed colourful source code samples. It can produce output in the following formats:

HTML

Valid XHTML 1.0, with the exact colouring and style left to a CSS stylesheet

XML

Pieces of text are marked with XML elements in a simple vocabulary, which can be converted to other formats, for example, using XSLT

Perl array

A simple Perl data structure, so that Perl code can be used to turn it into whatever is needed

This module works by running the Vim text editor and getting it to apply its excellent syntax highlighting (aka 'font-locking') to an input file, and mark pieces of text according to whether it thinks they are comments, keywords, strings, etc. The Perl code then reads back this markup and converts it to the desired output format.

This is an object-oriented module. To use it, create an object with the "new" function (as shown above in the SYNOPSIS) and then call methods to get the markup out.

METHODS

new(options)

Returns a syntax highlighting object. Pass it a hash of options.

The following options are recognised:

file

The file to syntax highlight. Can be either a filename or an open file handle.

Note that using a filename might allow Vim to guess the file type from its name if none is specified explicitly.

If the file isn't specified while creating the object, it can be given later in a call to the "syntax_mark_file" method (see below), allowing a single Text::VimColor object to be used with multiple input files.

string

Use this to pass a string to be used as the input. This is an alternative to the "file" option. A reference to a string will also work.

The "syntax_mark_string" method (see below) is another way to use a string as input.

filetype

Specify the type of file Vim should expect, in case Vim's automatic detection by filename or contents doesn't get it right. This is particularly important when providing the file as a string of file handle, since Vim won't be able to use the file extension to guess the file type.

The filetypes recognised by Vim are short strings like 'perl' or 'lisp'. They are the names of files in the 'syntax' directory in the Vim distribution.

This option, whether or not it is passed to "new()", can be overridden when calling "syntax_mark_file" and "syntax_mark_string", so you can use the same object to process multiple files of different types.

html_full_page

By default the "html()" output method returns a fragment of HTML, not a full file. To make useful output this must be wrapped in a "<pre>" element and a stylesheet must be included from somewhere. Setting the "html_full_page" option will instead make the "html()" method return a complete stand-alone XHTML file.

Note that while this is useful for testing, most of the time you'll want to put the syntax highlighted source code in a page with some other content, in which case the default output of the "html()" method is more appropriate.

html_inline_stylesheet

Turned on by default, but has no effect unless "html_full_page" is also enabled.

This causes the CSS stylesheet defining the colours to be used to render the markup to be be included in the HTML output, in a "<style>" element. Turn it off to instead use a "<link>" to reference an external stylesheet (recommended if putting more than one page on the web).

html_stylesheet

Ignored unless "html_full_page" and "html_inline_stylesheet" are both enabled.

This can be set to a stylesheet to include inline in the HTML output (the actual CSS, not the filename of it).

html_stylesheet_file

Ignored unless "html_full_page" and "html_inline_stylesheet" are both enabled.

This can be the filename of a stylesheet to copy into the HTML output, or a file handle to read one from. If neither this nor "html_stylesheet" are given, the supplied stylesheet light.css will be used instead.

html_stylesheet_url

Ignored unless "html_full_page" is enabled and "html_inline_stylesheet" is disabled.

This can be used to supply the URL (relative or absolute) or the stylesheet to be referenced from the HTML "<link>" element in the header. If this isn't given it will default to using a "file:" URL to reference the supplied light.css stylesheet, which is only really useful for testing.

xml_root_element

By default this is true. If set to a false value, XML output will not be wrapped in a root element called <syn:syntax>, but will be otherwise the same. This could allow XML output for several files to be concatenated, but to make it valid XML a root element must be added. Disabling this option will also remove the binding of the namespace prefix "syn:", so an "xmlns:syn" attribute would have to be added elsewhere.

vim_command

The name of the executable which will be run to invoke Vim. The default is "vim".

vim_options

A reference to an array of options to pass to Vim. The default options are:

   qw( -RXZ -i NONE -u NONE -N )

vim_let

A reference to a hash of options to set in Vim before the syntax file is loaded. Each of these is set using the ":let" command to the value specified. No escaping is done on the values, they are executed exactly as specified.

Values in this hash override some default options. Use a value of "undef" to prevent a default option from being set at all. The defaults are as follows:

   (
      perl_include_pod => 1,     # Recognize POD inside Perl code
      'b:is_bash' => 1,          # Allow Bash syntax in shell scripts
   )

These settings can be modified later with the "vim_let()" method.

vim_let(name => value, ...)

Change the options that are set with the Vim "let" command when Vim is run. See "new()" for details.

syntax_mark_file(file, options...)

Mark up the specified file. Subsequent calls to the output methods will then return the markup. It is not necessary to call this if a "file" or "string" option was passed to "new()".

Returns the object it was called on, so an output method can be called on it directly:

   my $syntax = Text::VimColor->new(
      vim_command => '/usr/local/bin/special-vim',
   );
   foreach (@files) {
      print $syntax->syntax_mark_file($_)->html;
   }

You can override the filetype set in new() by passing in a "filetype" option, like so:

   $syntax->syntax_mark_file($filename, filetype => 'perl');

This option will only affect the syntax colouring for that one call, not for any subsequent ones on the same object.

syntax_mark_string(string, options...)

Does the same as "syntax_mark_file" (see above) but uses a string as input. string can also be a reference to a string. Returns the object it was called on. Supports the "filetype" option just as "syntax_mark_file" does.

html()

Return XHTML markup based on the Vim syntax colouring of the input file.

Unless the "html_full_page" option is set, this will only return a fragment of HTML, which can then be incorporated into a full page. The fragment will be valid as either HTML and XHTML.

The only markup used for the actual text will be "<span>" elements wrapped round appropriate pieces of text. Each one will have a "class" attribute set to a name which can be tied to a foreground and background color in a stylesheet. The class names used will have the prefix "syn", for example "synComment". For the full list see the section HIGHLIGHTING TYPES below.

xml()

Returns markup in a simple XML vocabulary. Unless the "xml_root_element" option is turned off (it's on by default) this will produce a complete XML document, with all the markup inside a "<syntax>" element.

This XML output can be transformed into other formats, either using programs which read it with an XML parser, or using XSLT. See the text-vimcolor(1) program for an example of how XSLT can be used with XSL-FO to turn this into PDF.

The markup will consist of mixed content with elements wrapping pieces of text which Vim recognized as being of a particular type. The names of the elements used are the ones listed in the HIGHLIGHTING TYPES section below.

The "<syntax>" element will declare the namespace for all the elements prodeced, which will be "http://ns.laxan.com/text-vimcolor/1". It will also have an attribute called "filename", which will be set to the value returned by the "input_filename" method, if that returns something other than undef.

The XML namespace is also available as $Text::VimColor::NAMESPACE_ID.

marked()

This output function returns the marked-up text in the format which the module stores it in internally. The data looks like this:

   use Data::Dumper;
   print Dumper($syntax->marked);
   $VAR1 = [
      [ 'Statement', 'my' ],
      [ '', ' ' ],
      [ 'Identifier', '$syntax' ],
      [ '', ' = ' ],
       ...
   ];

The "marked()" method returns a reference to an array. Each item in the array is itself a reference to an array of two items: the first is one of the names listed in the HIGHLIGHTING TYPES section below (or the empty string if none apply), and the second is the actual piece of text.

input_filename()

Returns the filename of the input file, or undef if a filename wasn't specified.

HIGHLIGHTING TYPES

The following list gives the names of highlighting types which will be set for pieces of text. For HTML output, these will appear as CSS class names, except that they will all have the prefix "syn" added. For XML output, these will be the names of elements which will all be in the namespace "http://ns.laxan.com/text-vimcolor/1".

Here is the complete list:

• Comment
• Constant
• Identifier
• Statement
• PreProc
• Type
• Special
• Underlined
• Error
• Todo

RELATED MODULES

These modules allow Text::VimColor to be used more easily in particular environments:

Apache::VimColor

Kwiki::VimMode

Template-Plugin-VimColor

AUTHOR

Geoff Richards <[email protected]>

The Vim script mark.vim is a crufted version of 2html.vim by Bram Moolenaar <[email protected]> and David Ne\v{c}as (Yeti) <[email protected]>.

COPYRIGHT

Copyright 2002-2006, Geoff Richards.

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