dtplite was written with the following three use cases in mind.

[1]

Validation of a single document, i.e. checking that it was written in valid doctools format. This mode can also be used to get a preliminary version of the formatted output for a single document, for display in a browser, nroff, etc., allowing proofreading of the formatting.

[2]

Generation of the formatted documentation for a single package, i.e. all the manpages, plus a table of contents and an index of keywords.

[3]

An extension of the previous mode of operation, a method for the easy generation of one documentation tree for several packages. Especially unified table of contents and keyword index.

Beyond the above we also want to make use of the customization features provided by the HTML formatter. It is not the only format the application should be able to generate, but we anticipiate it to be the most commonly used, and it is one of the few which do provide customization hooks.

We allow the caller to specify a header string, footer string, a stylesheet, and data for a bar of navigation links at the top of the generated document. While all can be set as long as the formatting engine provides an appropriate engine parameter (See section OPTIONS) the last two have internal processing which make them specific to HTML.

dtplite -o output ?options? format inputfile

This is the form for use case [1]. The options will be explained later, in section OPTIONS.

output path (in)

This argument specifies where to write the generated document. It can be the path to a file or directory, or -. The last value causes the application to write the generated documented to stdout. If the output does not exist then [file dirname $output] has to exist and must be a writable directory. The generated document will be written to a file in that directory, and the name of that file will be derived from the inputfile, the format, and the value given to option -ext (if present).

format (path|handle) (in)

This argument specifies the formatting engine to use when processing the input, and thus the format of the generated document. See section FORMATS for the possibilities recognized by the application.

inputfile path (in)

This argument specifies the path to the file to process. It has to exist, must be readable, and written in doctools format.

dtplite -o output ?options? format inputdirectory

This is the form for use case [2]. It differs from the form for use case [1] by having the input documents specified through a directory instead of a file. The other arguments are identical, except for output, which now has to be the path to an existing and writable directory. The input documents are all files in inputdirectory or any of its subdirectories which were recognized by fileutil::fileType as containing text in doctools format.

dtplite -merge -o output ?options? format inputdirectory

This is the form for use case [3]. The only difference to the form for use case [2] is the additional option -merge. Each such call will merge the generated documents coming from processing the input documents under inputdirectory or any of its subdirectories to the files under output. In this manner it is possible to incrementally build the unified documentation for any number of packages. Note that it is necessary to run through all the packages twice to get fully correct cross-references (for formats supporting them).

This section describes all the options available to the user of the application, with the exception of the options -o and -merge. These two were described already, in section COMMAND LINE.

-ext string

If the name of an output file has to be derived from the name of an input file it will use the name of the format as the extension by default. This option here will override this however, forcing it to use string as the file extension. This option is ignored if the name of the output file is fully specified through option -o. When used multiple times only the last definition is relevant.

-header file

This option can be used if and only if the selected format provides an engine parameter named "header". It takes the contents of the specified file and assign them to that parameter, for whatever use by the engine. The HTML engine will insert the text just after the tag <body>. If navigation buttons are present (see option -nav below), then the HTML generated for them is appended to the header data originating here before the final assignment to the parameter. When used multiple times only the last definition is relevant.

-footer file

Like -header, except that: Any navigation buttons are ignored, the corresponding required engine parameter is named "footer", and the data is inserted just before the tag </body>. When used multiple times only the last definition is relevant.

-style file

This option can be used if and only if the selected format provides an engine parameter named "meta". When specified it will generate a piece of HTML code declaring the file as the stylesheet for the generated document and assign that to the parameter. The HTML engine will insert this inot the document, just after the tag <head>. When processing an input directory the stylesheet file is copied into the output directory and the generated HTML will refer to the copy, to make the result more self-contained. When processing an input file we have no location to copy the stylesheet to and so just reference it as specified. When used multiple times only the last definition is relevant.

-nav label url

Use this option to specify a navigation button with label to display and the url to link to. This option can be used if and only if the selected format provides an engine parameter named "header". The HTML generated for this is appended to whatever data we got from option -header before it is inserted into the generated documents. When used multiple times all definitions are collected and a navigation bar is created, with the first definition shown at the left edge and the last definition to the right.

At first the format argument will be treated as a path to a tcl file containing the code for the requested formatting engine. The argument will be treated as the name of one of the predefined formats listed below if and only if the path does not exist.

Note a limitation: If treating the format as path to the tcl script implementing the engine was sucessful, then this script has to implement not only the engine API for doctools, i.e. doctools_api, but for doctoc_api and docidx_api as well. Otherwise the generation of a table of contents and of a keyword index will fail.

List of predefined formats, i.e. as provided by the package doctools:

nroff

The processor generates *roff output, the standard format for unix manpages.

html

The processor generates HTML output, for usage in and display by web browsers. This engine is currently the only one providing the various engine parameters required for the additional customaization of the output.

tmml

The processor generates TMML output, the Tcl Manpage Markup Language, a derivative of XML.

latex

The processor generates LaTeX output.

wiki

The processor generates Wiki markup as understood by wikit.

list

The processor extracts the information provided by manpage_begin. This format is used internally to extract the meta data from which both table of contents and keyword index are derived from.

null

The processor does not generate any output.

In this section we describe the directory structures generated by the application under output when processing all documents in an inputdirectory. In other words, this is only relevant to the use cases [2] and [3].

[2]

The following directory structure is created when processing a single set of input documents. The file extension used is for output in HTML, but that is not relevant to the structure and was just used to have proper file names.

    output/
        toc.html
        index.html
        files/
            path/to/FOO.html

The last line in the example shows the document generated for a file FOO located at

    inputdirectory/path/to/FOO

[3]

When merging many packages into a unified set of documents the generated directory structure is a bit deeper:

    output
        .toc
        .idx
        .xrf
        toc.html
        index.html
        FOO1/
            ...
        FOO2/
            toc.html
            files/
                path/to/BAR.html

Each of the directories FOO1, ... contains the documents generated for the package FOO1, ... and follows the structure shown for use case [2]. The only exception is that there is no per-package index. The files ".toc", ".idx", and ".xrf" contain the internal status of the whole output and will be read and updated by the next invokation. Their contents will not be documented. Remove these files when all packages wanted for the output have been processed, i.e. when the output is complete.