t2html(1) Simple text to HTML converter. Relies on text indentation rules.

SYNOPSIS


t2html [options] file.txt > file.html

DESCRIPTION

Convert pure text files into nice looking, possibly framed, HTML pages. An example of conversion:

  1. Plain text source code
  http://pm-doc.git.sourceforge.net/git/gitweb.cgi?p=pm-doc/pm-doc;a=blob_plain;f=doc/index.txt;hb=HEAD
  2. reusult of conversion with custom --css-file option:
  http://pm-doc.sourceforge.net/pm-tips.html
  http://pm-doc.sourceforge.net/pm-tips.css
  3. An Emacs mode tinytf.el for writing the text files (optional)
  https://savannah.nongnu.org/projects/emacs-tiny-tools

Requirements for the input ascii files

The file must be written in Technical Format, whose layout is described in the this manual. Basicly the idea is simple and there are only two heading levels: one at column 0 and the other at column 4 (halfway between the tab width). Standard text starts at column 8 (the position after pressed tab-key).

The idea of technical format is that each column represents different rendering layout in the generated HTML. There is no special markup needed in the text file, so you can use the text version as a master copy of a FAQ etc. Bullets, numbered lists, word emphasis and quotation etc. can expressed in natural way.

HTML description

The generated HTML includes embedded Cascading Style Sheet 2 (CSS2) and a small piece of Java code. The CSS2 is used to colorize the page loyout and to define suitable printing font sizes. The generated HTML also takes an approach to support XHTML. See page http://www.w3.org/TR/xhtml1/#guidelines where the backward compatibility recommendations are outlined:

    Legal HTML          XHTML requires
    <P>                 <p> ..</p>
    <BR>                <br></br>
    <HR>                <hr></hr>

XHTML does not support fragment identifiers #foo, with the "name" element, but uses "id" instead. For backward compatibility both elements are defined:

    < ..name="tag">     Is now <.. name="tag" id="tag">

NOTE: This program was never designed to be used for XHTML and the strict XHTML validity is not to be expected.

Motivation

The easiest format to write large documents, like FAQs, is text. A text file offers WysiWyg editing and it can be turned easily into HTML format. Text files are easily maintained and there is no requirements for special text editors. Any text editor like notepad, vi, Emacs can be used to maintain the documents.

Text files are also the only sensible format if documents are kept under version control like RCS, CVS, SVN, Arch, Perforce, ClearCase. They can be asily compared with diff and patches can be easily received and sent to them.

To help maintining large documents, there is also available an Emacs minor mode, package called tinytf.el, which offers text fontification with colors, Indentation control, bullet filling, heading renumbering, word markup, syntax highlighting etc. See project http://freshmeat.net/projects/emacs-tiny-tools

OPTIONS

Html: Header and Footer options

--as-is
Any extra HTML formatting or text manipulation is suppressed. Text is preserved as it appears in file. Use this option if you plan to deliver or and print the text as seen.

    o  If file contains "Table of Contents" it is not removed
    o  Table of Content block is not created (it usually would)
--author -a STR
Author of document e.g. --author ``John Doe''
--disclaimer-file FILE
The text that appears at the footer is read from this file. If not given the default copyright text is added. Options "--quiet" and "--simple" suppress disclaimers.
--document FILE
Name of the document or filename. You could list all alternative URLs to the document with this option.
--email -e EMAIL
The contact address of the author of the document. Must be pure email address with no ``<'' and ``>'' characters included. Eg. --email [email protected]

    --email "<[email protected]>"     WRONG
    --email "[email protected]"       right
--simple -s
Print minimum footer only: contact, email and date. Use "--quiet" to completely discard footer.
--t2html-tags
Allow processing embedded #T2HTML-<tag> directives inside file. See full explanation by reading topic "EMBEDDED DIRECTIVES INSIDE TEXT". By default, you do not need to to supply this option - it is ``on'' by default.

To disregard embedded directives in text file, supply ``no'' option: --not2html-tags.

--title STR -t STR
The title text that appears in top frame of browser.
--url URL

Location of the HTML file. When --document gave the name, this gives the location. This information is printed at the Footer.

Html: Navigation urls

--base URL
URL location of the HTML file in the destination site where it will be put available. This option is needed only if the document is hosted on a FTP server (rare, but possible). A FTP server based document cannot use Table Of Contents links (fragment #tag identifiers) unless HTML tag BASE is also defined.

The argument can be full URL to the document:

    --base ftp://ftp.example.com/file.html
    --base ftp://ftp.example.com/
--button-heading-top
Add additional [toc] navigation button to the end of each heading. This may be useful in long non-framed HTML files.
--button-top URL
Buttons are placed at the top of document in order: [previous][top][next] and --button-* options define the URLs.

If URL is string none then no button is inserted. This may be handy if the buttons are defined by a separate program. And example using Perl:

    #!/usr/bin/perl
    my $top   = "index.html";             # set defaults
    my $prev  = "none";
    my $next  = "none";
    # ... somewhere $prev or $next may get set, or then not
    qx(t2html --button-top "$top" --button-prev "$prev" --button-next "$next" ...);
    # End of sample program
--button-prev URL
URL to go to previous document or string none.
--button-next URL
URL to go to next document or string none.
--reference tag=value
You can add any custom references (tags) inside text and get them expand to any value. This option can be given multiple times and every occurrance of TAG is replaced with VALUE. E.g. when given following options:

    --reference "#HOME-URL=http://www.example.com/dir"
    --reference "#ARCHIVE-URL=http://www.example.com/dir/dir2"

When referenced in text, the generated HTML includes expanded expanded to values. An example text:

        The homepage is #HOME-URL/page.html and the mirrot page it at
        #ARCHIVE-URL/page.html where you can find the latest version.
-R, --reference-separator STRING
See above. String that is used to split the TAG and VALUE. Default is equal sign ``=''.
-T, --toc-url-print
Display URLs (contructed from headings) that build up the Table of Contents (NAME AHREF tags) in a document. The list is outputted to stderr, so that it can be separated:

    % t2html --toc-url-print tmp.txt > file.html 2> toc-list.txt

Where would you need this? If you want to know the fragment identifies for your file, you need the list of names.

  http://www.example.com/myfile.html#fragment-identifier

Html: Controlling CSS generation (HTML tables)

--css-code-bg
This option affects how the code section (column 12) is rendered. Normally the section is surrounded with a <pre>..</pre> codes, but with this options, something more fancier is used. The code is wrapped inside a <table>...</table> and the background color is set to a shade of gray.
--css-code-note "REGEXP"
Option --css-code-bg is required to activate this option. A special word defined using regexp (defualt is 'Note:') will mark code sections specially. The "first word" is matched against the supplied Perl regexp.

The supplied regexp must not, repeat, must not, include any matching group operators. This simply means, that grouping parenthesis like "(one|two|three)" are not allowed. You must use the Perl non-grouping ones like "(?:one|two|three)". Please refer to perl manual page [perlre] if this short introduction did not give enough rope.

With this options, instead of rendering column 12 text with <pre>..</pre>, the text appears just like regular text, but with a twist. The background color of the text has been changed to darker grey to visually stand out form the text.

An example will clarify. Suppose that you passed options --css-code-bg and --css-code-note='(?:Notice|Note):', which instructed to treat the first paragraphs at column 12 differently. Like this:

    This is the regular text that appears somewhere at column 8.
    It may contain several lines of text in this paragraph.
        Notice: Here is the special section, at column 12,
        and the first word in this paragraph is 'Notice:'.
        Only that makes this paragraph at column 12 special.
    Now, we have some code to show to the user:
        for ( i = 0; i++; i < 10 )
        {
            //  Doing something in this loop
        }

One note, text written with initial special word, like "Notice:", must all fit in one full pragraph. Any other paragraphs that follow, are rendered as code sections. Like here:

    This is the regular text that appears somewhere
    It may contain several lines of text in this paragraph
        Notice: Here is the special section, at column 12,
        and the first word in this paragraph is 'Notice:'
        which makes it special
        Hoewver, this paragraph IS NOT rendered specially
        any more. Only the first paragraph above.
        for ( i = 0; i++; i < 10 )
        {
            //  Doing something in this loop
        }

As if this were not enough, there are some special table control directives that let you control the <table>..</table> which is put around the code section at column 12. Here are few examples:

    Here is example 1
        #t2html::td:bgcolor=#F7F7DE
        for ( i = 0; i++; i < 10 )
        {
            //  Doing something in this loop
        }
    Here is example 2
        #t2html::td:bgcolor=#F7F7DE:tableborder:1
        for ( i = 0; i++; i < 10 )
        {
            //  Doing something in this loop
        }
    Here is example 3
        #t2html::td:bgcolor="#FFFFFF":tableclass:dashed
        for ( i = 0; i++; i < 10 )
        {
            //  Doing something in this loop
        }
    Here is example 4
        #t2html::td:bgcolor="#FFFFFF":table:border=1_width=94%_border=0_cellpadding="10"_cellspacing="0"
        for ( i = 0; i++; i < 10 )
        {
            //  Doing something in this loop
        }

Looks cryptic? Cannot help that and in order for you to completely understand what these directives do, you need to undertand what elements can be added to the <table> and <td> tokens. Refer to HTML specification for available attributes. Here is briefing what you can do:

The start command is:

    #t2html::
            |
            After this comes attribute pairs in form key:value
            and multiple ones as key1:value1:key2:value2 ...

The "key:value" pairs can be:

    td:ATTRIBUTES
       |
       This is converted into <td attributes>
    table:ATTRIBUTES
          |
          This is converted into <table attributes>

There can be no spaces in the ATTRIBUTES, because the "First-word" must be one contiguous word. An underscore can be used in place of space:

    table:border=1_width=94%
          |
          Interpreted as <table border="1" width="94%">

It is also possible to change the default CLASS style with word "tableclass". In order the CLASS to be useful, its CSS definitions must be either in the default configuration or supplied from a external file. See option --script-file.

    tableclass:name
               |
               Interpreted as <table class="name">

For example, there are couple of default styles that can be used:

    1) Here is CLASS "dashed" example
        #t2html::tableclass:dashed
            for ( i = 0; i++; i < 10 )
            {
                //  Doing something in this loop
            }
    2) Here is CLASS "solid" example:
        #t2html::tableclass:solid
            for ( i = 0; i++; i < 10 )
            {
                //  Doing something in this loop
            }

You can change any individual value of the default table definition which is:

    <table  class="shade-note">

To change e.g. only value cellpadding, you would say:

     #t2html::table:tablecellpadding:2

If you are unsure what all of these were about, simply run program with --test-page and look at the source and generated HTML files. That should offer more rope to experiment with.

--css-file FILE
Include <LINK ...> which refers to external CSS style definition source. This option is ignored if --script-file option has been given, because that option imports whole content inside HEAD tag. This option can appear multiple times and the external CSS files are added in listed order.
--css-font-type CSS-DEFINITION
Set the BODY element's font defintion to CSS-DEFINITION. The default value used is the regular typeset used in newspapers and books:

    --css-font-type='font-family: "Times New Roman", serif;'
--css-font-size CSS-DEFINITION
Set the body element's font size to CSS-DEFINITION. The default font size is expressed in points:

    --css-font-size="font-size: 12pt;"

Html: Controlling the body of document

--delete REGEXP
Delete lines matching perl REGEXP. This is useful if you use some document tool that uses navigation tags in the text file that you do not want to show up in generated HTML.
--delete-email-headers
Delete email headers at the beginning of file, until first empty line that starts the body. If you keep your document ready for Usenet news posting, they may contain headers and body:

    From: ...
    Newsgroups: ...
    X-Sender-Info:
    Summary:
    BODY-OF-TEXT
--nodelete-default
Use this option to suppress default text deletion (which is on).

Emacs "folding.el" package and vi can be used with any text or programming language to place sections of text between tags {{{ and }}}. You can open or close such folds. This allows keeping big documents in order and manageable quite easily. For Emacs support, see. ftp://ftp.csd.uu.se/pub/users/andersl/beta/

The default value deletes these markers and special comments "#_comment" which make it possible to cinlude your own notes which are not included in the generated output.

  {{{ Security section
  #_comment Make sure you revise this section to
  #_comment the next release
  The seecurity is an important issue in everyday administration...
  More text ...
  }}}
--html-body STR
Additional attributes to add to HTML tag <BODY>. You could e.g. define language of the text with --html-body LANG=en which would generate HTML tag <BODY LANG=``en''> See section ``SEE ALSO'' for ISO 639.
--html-column-beg="SPEC HTML-SPEC"
The default interpretation of columns 1,2,3 5,6,7,8,9,10,11,12 can be changed with beg and end swithes. Columns 0,4 can't be changed because they are reserved for headings. Here are some samples:

    --html-column-beg="7quote <em class='quote7'>"
    --html-column-end="7quote </em>"
    --html-column-beg="10    <pre> class='column10'"
    --html-column-end="10    </pre>"
    --html-column-beg="quote <span class='word'>"
    --html-column-end="quote </span>"

Note: You can only give specifications up till column 12. If text is beyound column 12, it is interpreted like it were at column 12.

In addition to column number, the SPEC can also be one of the following strings

    Spec    equivalent word markup
    ------------------------------
    quote   `'
    bold    _
    emp     *
    small   +
    big     =
    ref     []   like: [Michael] referred to [rfc822]
    Other available Specs
    ------------------------------
    7quote      When column 7 starts with double quote.

For style sheet values for each color, refer to class attribute and use --script-file option to import definitions. Usually /usr/lib/X11/rgb.txt lists possible color values and the HTML standard at http://www.w3.org/ defines following standard named colors:

    Black       #000000  Maroon  #800000
    Green       #008000  Navy    #000080
    Silver      #C0C0C0  Red     #FF0000
    Lime        #00FF00  Blue    #0000FF
    Gray        #808080  Purple  #800080
    Olive       #808000  Teal    #008080
    White       #FFFFFF  Fuchsia #FF00FF
    Yellow      #FFFF00  Aqua    #00FFFF
--html-column-end="COL HTML-SPEC"
See --html-column-beg
--html-font SIZE
Define FONT SIZE. It might be useful to set bigger font size for presentations.
-F, --html-frame [FRAME-PARAMS]
If given, then three separate HTML files are generated. The left frame will contain TOC and right frame contains rest of the text. The FRAME-PARAMS can be any valid parameters for HTML tag FRAMESET. The default is "cols="25%,75%"".

Using this implies --out option automatically, because three files cannot be printed to stdout.

    file.html
    --> file.html       The Frame file, point browser here
        file-toc.html   Left frame (navigation)
        file-body.html  Right frame (content)
--language ID
Use language ID, a two character ISO identifier like ``en'' for English during the generation of HTML. This only affects the text that is shown to end-user, like text ``Table Of contents''. The default setting is ``en''. See section ``SEE ALSO'' for standards ISO 639 and ISO 3166 for proper codes.

The selected langauge changes propgram's internal arrays in two ways: 1) Instead of default ``Table of ocntents'' heading the national langaugage equivalent will be used 2) The text ``Pic'' below embedded sequentially numbered pictures will use natinal equivalent.

If your languagae is not supported, please send the phrase for ``Table of contents'' and word ``Pic'' in your langauge to the maintainer.

--script-file FILE
Include java code that must be complete <script...></script> from FILE. The code is put inside <head> of each HTML.

The --script-file is a general way to import anything into the HEAD element. Eg. If you want to keep separate style definitions for all, you could only import a pointer to a style sheet. See 14.3.2 Specifying external style sheets in HTML 4.0 standard.

--meta-keywords STR
Meta keywords. Used by search engines. Separate kwywords like ``AA, BB, CC'' with commas. Refer to HTML 4.01 specification and topic ``7.4.4 Meta data'' and see http://www.htmlhelp.com/reference/wilbur/ and

    --meta-keywords "AA,BB,CC"
--meta-description STR
Meta description. Include description string, max 1000 characters. This is used by search engines. Refer to HTML 4.01 specification and topic ``7.4.4 Meta data''
--name-uniq
First 1-4 words from the heading are used for the HTML name tags. However, it is possible that two same headings start with exactly the same 1-4 words. In those cases you have to turn on this option. It will use counter 00 - 999 instead of words from headings to construct HTML name references.

Please use this option only in emergencies, because referring to jump block name via

    httpI://example.com/doc.html#header_name

is more convenient than using obscure reference

    httpI://example.com/doc.html#11

In addition, each time you add a new heading the number changes, whereas the symbolic name picked from heading stays as long as you do not change the heading. Think about welfare of your netizens who bookmark you pages. Try to make headings to not have same subjects and you do not need this option.

Document maintenance and batch job commands

-A, --auto-detect
Convert file only if tag "#T2HTML-" is found from file. This option is handy if you run a batch command to convert all files to HTML, but only if they look like HTML base files:

    find . -name "*.txt" -type f \
         -exec t2html --auto-detect --verbose --out {} \;

The command searches all *.txt files under current directory and feeds them to conversion program. The --auto-detect only converts files which include "#T2HTML-" directives. Other text files are not converted.

--link-check -l
Check all http and ftp links. This option is supposed to be run standalone Option --quiet has special meaning when used with link check.

With this option you can regularly validate your document and remove dead links or update moved links. Problematic links are outputted to stderr. This link check feature is available only if you have the LWP web library installed. Program will check if you have it at runtime.

Links that are big, e.g. which match tar.gz .zip ... or that run programs (links with ? character) are ignored because the GET request used in checking would return whole content of the link and it would. be too expensive.

A suggestion: When you put binary links to your documents, add them with space:

    http://example.com/dir/dir/ filename.tar.gz

Then the program does check the http addresses. Users may not be able to get the file at one click, checker can validate at least the directory. If you are not the owner of the link, it is also possible that the file has moved of new version name has appeared.

-L, --link-check-single
Print condensed output in grep -n like manner FILE:LINE:MESSAGE

This option concatenates the url response text to single line, so that you can view the messages in one line. You can use programming tools (like Emacs M-x compile) that can parse standard grep syntax to jump to locations in your document to correct the links later.

-o, --out
write generated HTML to file that is derived from the input filename.

    --out --print /dir/file            --> /dir/file.html
    --out --print /dir/file.txt        --> /dir/file.html
    --out --print /dir/file.this.txt   --> /dir/file.this.html
--link-cache CACHE_FILE
When links are checked periodically, it would be quite a rigorous to check every link every time that has already succeeded. In order to save link checking time, the ``ok'' links can be cached into separate file. Next time you check the links, the cache is opened and only links found that were not in the cache are checked. This should dramatically improve long searches. Consider this example, where every text file is checked recursively.

    $ t2html --link-check-single \
      --quiet --link-cache ~tmp/link.cache \
      `find . -name "*.txt" -type f`
-O, --out-dir DIR
Like --out, but chop the directory part and write output files to DIR. The following would generate the HTML file to current directory:

    --out-dir .

If you have automated tool that fills in the directory, you can use word none to ignore this option. The following is a no-op, it will not generate output to directory ``none'':

    --out-dir none
-p, --print
Print filename to stdout after HTML processing. Normally program prints no file names, only the generated HTML.

    % t2html --out --print page.txt
    --> page.html
-P, --print-url
Print filename in URL format. This is useful if you want to check the layout immediately with your browser.

    % t2html --out --print-url page.txt | xargs lynx
    --> file: /users/foo/txt/page.html
--split REGEXP
Split document into smaller pieces when REGEXP matches. Split commands are standalone, meaning, that it starts and quits. No HTML conversion for the file is engaged.

If REGEXP is found from the line, it is a start point of a split. E.g. to split according to toplevel headings, which have no numbering, you would use:

    --split '^[A-Z]'

A sequential numbers, 3 digits, are added to the generated partials:

    filename.txt-NNN

The split feature is handy if you want to generate slides from each heading: First split the document, then convert each part to HTML and finally print each part (page) separately to printer.

-S1, --split1
This is shorthand of --split command. Define regexp to split on toplevel heading.
-S2, --split2
This is shorthand of --split command. Define regexp to split on second level heading.
-SN, --split-named-files
Additional directive for split commands. If you split e.g. by headings using --split1, it would be more informative to generate filenames according to first few words from the heading name. Suppose the heading names where split occur were:

    Program guidelines
    Conclusion

Then the generated partial filenames would be as follows.

    FILENAME-program_guidelines
    FILENAME-conclusion
-X, --xhtml
Render using strict XHTML. This means using <hr/>, <br/> and paragraphs use <p>..</p>.

"Note: this option is experimental. See BUGS"

Miscellaneous options

--debug LEVEL
Turn on debug with positive LEVEL number. Zero means no debug.
--help -h
Print help screen. Terminates program.
--help-css
Print default CSS used. Terminates program. You can copy and modify this output and instruct to use your own with --css-file=FILE. You can also embed the option to files with "#T2HTML-OPTION" directive.
--help-html
Print help in HTML format. Terminates program.
--help-man
Print help page in Unix manual page format. You want to feed this output to nroff -man in order to read it. Terminates program.
--test-page
Print the test page: HTML and example text file that demonstrates the capabilities.
--time
Print to stderr time spent used for handling the file.
-v, --verbose [LEVEL]
Print verbose messages.
-q, --quiet
Print no footer at all. This option has different meaning if --link-check option is turned on: print only errorneous links.
V, --version
Print program version information.

FORMAT DESCRIPTION

Program converts text files to HTML. The basic idea is to rely on indentation level, and the layout used is called 'Technical format' (TF) where only minimal conventions are used to mark italic, bold etc. text. The Basic principles can be demonstrated below. Notice the column poisiton ruler at the top:

 --//-- decription start
 123456789 123456789 123456789 123456789 123456789 column numbers
 Heading 1 starts with a big letter at leftmost column 1
  The column positions 1,2,3 are currently undefined and may not
  format correctly. Do ot place text at columns 1,2 or 3.
     Heading level 2 starts at half-tab column 4 with a big letter
      Normal but colored text at columns 5
       Normal but colored text at columns 6
        Heading 3 can be considered at position TAB minus 1, column 7.
        "Special <em> text at column 7 starts with double quote"
         Standard text starts at column 8, you can *emphatize* text or
         make it _strong_ and write =SmallText= or +BigText+ show
         variable name `ThisIsAlsoVariable'. You can `_*nest*_' `the'
         markup. more txt in this paragraph txt txt txt txt txt txt
         txt txt txt txt txt txt txt txt txt txt txt txt txt txt txt
         txt txt txt txt txt txt txt txt txt txt txt txt txt txt txt
         txt txt
          Strong text at column 9
           Column 10 is reserved for quotations
           Column 10 is reserved for quotations
           Column 10 is reserved for quotations
           Column 10 is reserved for quotations
          Strong text at column 11
           Column 12 and further is reserved for code examples
           Column 12 and further is reserved for code examples
           All text here are surrounded by <pre> HTML codes
           This CODE column in affected by the --css-code* options.
     Heading 2 at column 4 again
        If you want something like Heading level 3, use column 7 (bold)
         Column 8. Standard tab position. txt txt txt txt txt txt txt
         txt txt txt txt txt txt txt txt txt txt txt txt txt txt txt
         txt txt txt txt txt txt txt txt txt txt txt txt txt txt
         [1998-09-10 Mr. Foo said]:
           cited text cited text cited text cited text cited
           text cited text cited text cited text cited text
           cited text cited text cited text cited text cited
           text cited text
         *   Bullet at column 8. Notice 3 spaces after (*), so
             text starts at half-tab forward at column 12.
         *   Bullet. txt txt txt txt txt txt txt txt txt txt txt txt
         *   Bullet. txt txt txt txt txt txt txt txt txt txt txt txt
             ,txt txt txt txt
             Notice that previous paragraph ends to P-comma
             code, it tells this paragraph to continue in
             bullet mode, otherwise this text at column 12
             would be intepreted as code section surrpoundedn
             by <pre> HTML codes.
         .   This is ordered list.
         .   This is ordered list.
         .   This is ordered list.
         .This line starts wirg dot and is displayed in line by itself.
         .This line starts wirg dot and is displayed in line by itself.
         !! This adds an <hr> HTML code, text in line is marked with
         !! <strong> <em>
         Make this email address clickable <[email protected]> Do not
         make this email address clickable [email protected], because it
         is only an example and not a real address. Notice that the
         last one was not surrounded by <>. Common login names like
         foo, bar, quux, or internet site 'example' are ignored
         automatically.
         Also do not make < [email protected]> because there is extra
         white space. This may be more convenient way to disable email
         addresses temporarily.
 Heading1 again at colum 0
     Subheading at colum 4
         And regular text, column 8 txt txt txt txt txt txt txt txt txt
         txt txt txt txt txt txt txt txt txt txt txt txt txt txt txt txt
         txt txt txt txt txt txt txt txt txt txt txt
 --//-- decription end

That is it, there is the whole layout described. More formally the rules of text formatting are secribed below.

USED HEADINGS

  • There are only two heading levels in this style. Heading columns are 0 and 4 and the heading must start with big letter or number
  • at column 4, if the text starts with small letter, that line is interpreted as <strong>
  • A HTML <hr> mark is added just before printing heading at level 1.
  • The headings are gathered, the TOC is built and inserted to the beginning of HTML page. The HTML <name> references used in TOC are the first 4 sequential words from the headings. Make sure your headings are uniquely named, otherwise there will be same NAME references in the generated HTML. Spaces are converted into underscore when joining the words. If you can not write unique headings by four words, then you must use --name-uniq switch

TEXT PLACEMENT RULES

General

The basic rules for positioning text in certain columns:
  • Text at column 0 is undefined if it does not start with big letter or number to indicate Heading level 1.
  • Text between colums 1-3 is marked with <em>
  • Column 4 is reserved for heading level 2
  • Text between colums 5-7 is marked with <strong>
  • Text at column 7 is <em> if the first character is double quote.
  • Column 10 is reserved for <em> text. If you want to quote someone or to add reference text, place the text in this column.
  • Text at colums 9,11 are marked with <strong>

Column 8 for text and special codes

  • Column 8 is reserved for normal text
  • At the start of text, at colum 8, there can be DOT-code or COMMA-code.

Column 12 is special

  • Column 12 is treated specially: block is started with <pre> and lines are marked as <samp></samp>. When the last text at column 12 is found, the block is closed with </pre> Note follwing example

        txt txt txt         ;evenly placed block, fine, do it like this
        txt txt
        txt txt txt txt     ;Can not terminate the /pre, because last
        txt txt txt txt     ;column is not at 12
            txt txt txt txt
        txt txt txt txt
        txt txt txt txt
            txt txt txt txt
        ;; Finalizing comment, now the text is evenly placed
    

Additional tokens for use at column 8

  • If there is "."(dot) at the beginning of a line and immediately non-whitespace, then <br> code is added to the end of line.

        .This line will have a <BR> HTML tag at the end.
        While these two line are joined together
        by the browser, depending on the frame width.
    
  • If there is ","(comma) then the <p> code is not inserted if the previous line is empty. If you use both "."(dot) and ","(comma), they must be in order dot-comma. The ","(comma) works differently if it is used in bullet

    A <p> is always added if there is separation of paragraphs, but when you are writing a bullet, there is a problem, because a bullet exist only as long as text is kept together

        *   This is a bullet and it has all text ketp together
            even if there is another line in the bullet.
    

    But to write bullets tat spread multiple paragraphs, you must instruct that those are to kept together and the text in next paragraph is not <sample> while it is placed at column 12

        *   This is a bullet and it has all text ketp together
            ,even if there is another line in the bullet.
            This is new paragrah to the previous bullet and this is
            not a text sample. See continued COMMA-code above.
        *   This is new bullet
            // and this is code sample after bullet
            if ( $flag ) { ..do something.. }
    

Special text markings

italic, bold, code, small, big tokens
    _this_      is intepreted as <strong class='word'>this</strong>
    *this*      is intepreted as <em class='word'>this</em>
    `this'      is intepreted as <sample class='word'>this</sample> `

Exra modifiers that can be mixed with the above. Usually if you want bigger font, CAPITALIZE THE WORDS.

    =this=      is intepreted as <span class="word-small">this</span>
    +this+      is intepreted as <span class="word-big">this</span>
    [this]      is intepreted as <span class="word-ref">this</span>
superscripting
    word[this]  is intepreted as superscript. You can use like
                this[1], multiple[(2)] and almost any[(ab)] and
                imaginable[IV superscritps] as long as the left
                bracket is attached to the word.
subscripting
    12[[10]]    is representation of value 12 un base 10.
                This is intepreted as subscript. You can use like
                this[[1]], multiple[[(2)]] and almost any[[(ab)]] and
                imaginable[[IV superscritps]] as long as *two* left
                brackets are attached to the word.
embedding standard HTML tokens
Stanadard special HTML entities can be added inside text in a normal way, either using sybolic names or the hash code. Here are exmples:

    &times; &lt; &gt; &le; &ge; &ne; &radic; &minus;
    &alpha; &beta; &gamma; &divide;
    &laquo; &raquo; &lsaquo; &rsaquo; - &ndash; &mdash;
    &asymp; &equiv; &sum; &fnof; &infin;
    &deg; &plusmn;
    &trade; &copy; &reg;
    &euro; &pound; &yen;
embedding PURE HTML into text
This feature is highly experimental. It is possible to embed pure HTML inside text in occasions, where e.g. some special formatting is needed. The isea is simple: you write HTML as usual but double every '<' and '>' characters, like:

    <<p>>

The other rule is that all PURE HTML must be kept together. There must be no line breaks between pure HTML lines. This is incorrect:

    <<table>
        <<tr>>one
        <<tr>>two
    <</table>>

The pure HTML must be written without extra newlines:

    <<table>
        <<tr>>one
        <<tr>>two
    <</table>>

This ``doubling'' affects normal text writing rules as well. If you write documents, where you describe Unix styled HERE-documents, you MUST NOT put the tokens next to each other:

        bash$ cat<<EOF              # DON'T, this will confuse parser.
        one
        EOF

You must write the above code example using spaces to prevent ``<<'' from interpreting as PURE HTML:

        bash$ cat << EOF            # RIGHT, add spaces
        one
        EOF
drawing a short separator
A !! (two exclamation marks) at text column (position 8) causes adding immediate <hr> code. any text after !! in the same line is written with <strong> <em> and inserted just after <hr> code, therefore the word formatting commands have no effect in this line.

Http and email marking control

Lists and bullets

  • The bulletin table is contructed if there is ``o'' or ``*'' at column 8 and 3 spaces after it, so that text starts at column 12. Bulleted lines are advised to be kept together; no spaces between bullet blocks.
  • The ordered list is started with ``.'', a dot, and written like bullet where text starts at column 12.

Line breaks

  • All line breaks are visible in your document, do not use more than one line break to separate paragraphs.
  • Very important is that there is only one line break after headings.

EMBEDDED DIRECTIVES INSIDE TEXT

Command line options
You can cancel obeying all embedded directives by supplying option --not2html-tags.

You can include these lines anywhere in the document and their content is included in HTML output. Each directive line must fit in one line and it cannot be broken to separate lines.

    #T2HTML-TITLE            <as passed option --title>
    #T2HTML-EMAIL            <as passed option --email>
    #T2HTML-AUTHOR           <as passed option --author>
    #T2HTML-DOC              <as passed option --doc>
    #T2HTML-METAKEYWORDS     <as passed option --meta-keywords>
    #T2HTML-METADESCRIPTION  <as passed option --meta-description>

You can pass command line options embedded in the file. Like if you wanted the CODE section (column 12) to be coloured with shade of gray, you could add:

    #T2HTML-OPTION  --css-code-bg

Or you could request turning on particular options. Notice that each line is exactly as you have passed the argument in command line. Imagine surrounding double quoted around lines that are arguments to the associated options.

    #T2HTML-OPTION  --as-is
    #T2HTML-OPTION  --quiet
    #T2HTML-OPTION  --language
    #T2HTML-OPTION  en
    #T2HTML-OPTION  --css-font-type
    #T2HTML-OPTION  Trebuchet MS
    #T2HTML-OPTION --css-code-bg
    #T2HTML-OPTION --css-code-note
    #T2HTML-OPTION (?:Note|Notice|Warning):

You can also embed your own comments to the text. These are stripped away:

    #T2HTML-COMMENT  You comment here
    #T2HTML-COMMENT  You another comment here
Embedding files
#INCLUDE- command

This is used to include the content into current current position. The URL can be a filename reference, where every $VAR is subtituted from the environment variables. The tilde(~) expansion is not supported. The included filename is operating system supported path location.

A prefix "raw:" disables any normal formatting. The file content is included as is.

The URL can also be a HTTP reference to a remote location, whose content is included at the point. In case of remote content or when filename ends to extension ".html" or ".html", the content is stripped in order to make the inclusion of the content possible. In picture below, only the lines within the BODY, marked with !!, are included:

    <html>
      <head>
        ...
      </head>
      <body>
        this text                 !!
        and more of this          !!
      </body>
    </html>

Examples:

    #INCLUDE-$HOME/lib/html/picture1.html
    #INCLUDE-http://www.example.com/code.html
    #INCLUDE-raw:example/code.html
Embedding pictures
#PIC command is used to include pictures into the text

    #PIC picture.png#Caption Text#Picture HTML attributes#align#
          (1)        (2)          (3)                     (4)
    1.  The NAME or URL address of the picturere. Like image/this.png
    2.  The Text that appears below picture
    3.  Additional attributes that are attached inside <img> tag.
        For <img width="200" height="200">, the line would
        read:
        #PIC some.png#Caption Text#width=200 length=200##
    4.  The position of image: "left" (default), "center", "right"

Note: The "Caption Text" will also become the ALT text of the image which is used in case the browser is not capable of showing pictures. You can suppress the ALT text with option --no-picture-alt.

Fragment identifiers for named tags
#REF command is used for refering to HTML <name> tag inside current document. The whole command must be placed on one single line and cannot be broken to multiple lines. An example:

    #REF #how_to_profile;(Note: profiling);
          (1)            (2)
    1.  The NAME HTML tag reference in current document, a single word.
        This can also be a full URL link.
        You can get NAME list by enabling --toc-url-print option.
    2.  The clickable text is delimited by ; characters.
Referring to external documents.
"#URL" tag can be used to embed URLs inline, so that the full link is not visible. Only the shown text is used to jump to URL. This directive cannot be broken to separate lines,

     #URL<FULL-HTTP-LINK> <embedded inline text>
         |               |
         |               whitespace allowed here
         Must be kept together

Like if written:

     See search engine #URL<http://www.google.com> <Google>

TABLE OF CONTENT HEADING

If there is heading 1, which is named exactly ``Table of Contents'', then all text up to next heading are discarded from the generated HTML file. This is done because program generates its own TOC. It is supposed that you use some text formatting program to generate the toc for you in .txt file and you do not maintain it manually. For example Emacs package tinytf.el can be used.

TROUBLESHOOTING

Generated HTML document did not look what I intended

The most common mistake is that there are extra newlines in the document. Keeep one empty line between headings and text, keep one empty line between paragraphs, keep one empty line between body text and bullet. Make it your mantra: one one one ...

Next, you may have put text at wrong column position. Remember that the regular text is at column 8.

If generated HTML suddendly starts using only one font, eg <pre>, then you have forgot to close the block. Make it read even, like this:

    Code block
        Code block
        Code block
    ;;  Add empty comment here to "close" the code example at column 12

Headings start with a big letter or number, likein ``Heading'', not ``heading''. Double check the spelling.

EXAMPLES

To print the test page and show all the possibilities:

    t2html --test-page

To make simple HTML page without any meta information:

    t2html --title "Html Page Title" --author "Mr. Foo" \
           --simple --out --print file.txt

If you have periodic post in email format, use --delete-email-headers to ignore the header text:

    t2html --out --print --delete-email-headers page.txt

To make page fast

    t2html --html-frame --out --print page.txt

To convert page from a text document, including meta tags, buttons, colors and frames. Pay attention to switch --html-body which defines document language.

    t2html                                              \
    --print                                             \
    --out                                               \
    --author    "Mr. foo"                               \
    --email     "[email protected]"                       \
    --title     "This is manual page of page BAR"       \
    --html-body LANG=en                                 \
    --button-prev  previous.html                        \
    --button-top   index.html                           \
    --buttion-next next.html                            \
    --document  http://example.com/dir/this-page.html   \
    --url       manual.html                             \
    --css-code-bg                                       \
    --css-code-note '(?:Note|Notice|Warning):'          \
    --html-frame                                        \
    --disclaimer-file   $HOME/txt/my-html-footer.txt    \
    --meta-keywords    "language-en,manual,program"     \
    --meta-description "Bar program to do this that and more of those" \
    manual.txt

To check links and print status of all links in par with the http error message (most verbose):

    t2html --link-check file.txt | tee link-error.log

To print only problematic links:

    t2html --link-check --quiet file.txt | tee link-error.log

To print terse output in egep -n like manner: line number, link and error code:

    t2html --link-check-single --quiet file.txt | tee link-error.log

To check links from multiple pages and cache good links to separate file, use --link-cache option. The next link check will run much faster because cached valid links will not be fetched again. At regular intervals delete the link cache file to force complete check.

    t2html --link-check-single \
           --link-cache $HOME/tmp/link.cache \
           --quiet file.txt

To split large document into pieces, and convert each piece to HTML:

    t2html --split1 --split-name file.txt | t2html --simple --out

ENVIRONMENT

EMAIL
If environment variable EMAIL is defined, it is used in footer for contact address. Option --email overrides environment setting.
LANG
The default language setting for switch "--language" Make sure the first two characters contains the language definition, like in: LANG=en.iso88591

Related programs

Jan Ka.rrman <[email protected]> has written Perl html2ps which was 2004-11-11 available at http://www.tdb.uu.se/~jan/html2ps.html

HTML validator is at http://validator.w3.org/

iMATIX created htmlpp which is available at http://www.imatix.com/

Emacs minor mode to write documents based on TF layout is available. See package tinytf.el in project http://freshmeat.net/projects/emacs-tiny-tools

Standards

RFC 1766 contains list of langauge codes at http://www.rfc.net/

Latest HTML/XHTML and CSS specifications are at http://www.w3c.org/

ISO standards

639 Code for the representation of the names of languages http://www.oasis-open.org/cover/iso639a.html

3166 Standard Country Codes http://www.niso.org/3166.html and http://www.netstrider.com/tutorials/HTMLRef/standards/

BUGS

The implementation was originally designed to work linewise, so it is unfortunately impossible to add or modify any existing feature to look for items that span more than one line.

As the options --xhtml was much later added, it may not produce completely syntactically valid markup.

SCRIPT CATEGORIES

CPAN/Administrative html

PREREQUISITES

No additional CPAN modules needed for text to HTML conversion.

COREQUISITES

If link check feature is used to to validate URL links, then following modules are needed from CPAN "use LWP::UserAgent" "HTML::FormatText" and "HTML::Parse"

If you module "HTML::LinkExtractor" is available, it is used instead of included link extracting algorithm.

AVAILABILITY

Homepage is at http://freshmeat.net/projects/perl-text2html

AUTHOR

This program is free software; you can redistribute and/or modify program under the terms of GNU General Public license either version 2 of the License, or (at your option) any later version.

This documentation may be distributed subject to the terms and conditions set forth in GNU General Public License v2 or later; or, at your option, distributed under the terms of GNU Free Documentation License version 1.2 or later (GNU FDL).

POD ERRORS

Hey! The above document had some coding errors, which are explained below:
Around line 2663:
Non-ASCII character seen before =encoding in 'Ka.rrman'. Assuming ISO8859-1