sgml2x - Easily formats SGML/XML documents using DSSSL style-sheets

sgml2x [options] [sgmlfile | xmlfile ] ...

docclass-2-targetformat [options] [sgmlfile | xmlfile ] ...

sgml2x allows to easily format a SGML or XML document using DSSSL style-sheets, and provides the following features:

•

Multiple possible style-sheets per document class

•

Easy specification of style-sheets using aliases, with support for parameter inheritance

•

Easy integration of new style-sheets by adding a simple new definition file in a configuration directory

•

The caller can specify a PATH-like list of configuration directories, defaulting to a system-wide, a per-user, and a per-project configuration directories

•

Automatic selection of a default style-sheet to be used, based on assigned priorities

•

Pass arbitrary options to jade(1)

The document-class used to look for the style-sheets, and the output format, is for now only derived from the name with which the program is called, so you will want to call this program through symbolic links like docbook-2-pdf.

sgml2x is a implemented as a shell wrapper around jade(1) (or, preferably, openjade(1), although we use the generic name jade throughout this documentation), jadetex(1) and other tools.

If there is no jadetex.cfg file near the document, a default one is copied, that enables production of PDF bookmarks.

-c|--catalog catalog

Use the specified SGML catalog instead of the system default.

-C|--confdirs dir-list

Use (whitespace-separated) list of configuration directories. This option is cumulative, i.e. you can use several -C options and the lists will be concatenated.

The list elements should be ordered from the most generic configuration (e.g. system-wide) to the most specific (e.g. project-wide).

If any directory is provided through this option, the default directory list will be ignored.

-D|--dssslproc dsssl-processor

Use dsssl-processor to apply the style-sheet, instead of the default one. This processor has to support jade-like options, such as -V.

When this option is not present, the first found in the dssslproc files from confdirs is taken. See "Files" for more details.

-h|--help

Display an help message and exit.

-j|--jade dsssl-processor

Obsolete synonym for --dssslproc.

--jadetexfilter perl-filter

Post-process the jadetex output using a perl filter.

This can be useful to force pagebreaks at some specific places to overcome stylesheet problems, or to force hyphenations where TeX does not have enough patterns, or do any other clever transformation you'd think about.

See the examples/command-lines file for possible uses.

-n|--no-act

Print commands instead of running them. Useful to learn about lower-level tools, and for debugging the command-line.

-o|--openjade

This option is obsolete. openjade is now the default when available. Use --dssslproc or a dssslproc configuration file to force a specific processor.

This option used to use openjade(1) as a DSSSL processor instead of jade(1).

-O|--jadeopts jade-options

Additional options to pass to jade(1). This option is cumulative, you can specify several of them, the provided options will be concatenated.

-q|--quiet

Set verbosity to quiet

-r|--remarks

Render the content of document remarks in the document (remark elements in DocBook 4, comment elements in DocBook 3), making the produced output an internal-use-only document, printing a bold warning on the cover.

This is a docclass- and style-sheet-specific feature, and not all style-sheets will use this.

-s|--style style

Select an output style to override the (eventually document-derived) default.

Styles currently available for a specific document class and for each output format are dependent on the contents of the configuration directories, and can be displayed with the --help option.

Note that it is good practice to specify this option in a build procedure, so that you get reproducible results regardless of the available style-sheets.

-v|--verbose

Increase verbosity. This option can be specified multiple times.

--verbosity N

Set verbosity to N. The levels of verbosity are defined as follows:

quiet

Only print errors

default

Only print errors and warnings

verbose

Also print notices

trace

Also print significant commands as they are run (as --no-act does).

debug

Also print debugging messages

-V|--version

Print the program version and exit.

sgml2x uses a configuration directory tree instead of a configuration file, so that it is easy for other packages to plug in with a low risk of breaking an existing setup.

Styles hierarchies are located in directories named styles in each configuration directory. Old versions of this program used to put those hierarchies directly in the configuration directories.

A configuration directory contains one directory for each known document class, named with a document class nickname (e.g. docbook). Those docclass directories contain one sub-directory for each class of output-format (currently, only html and print are supported).

Currently, implementation issues enforce a limitation on nicknames for document classes and style-sheets: they can only contain alphanumeric and underscore characters. This limitation may be dropped in a future release, but that's not going to happen before this script gets rewritten in another language.

Each of those directory contain one file per available style. The names of these files may only contain alphanumeric characters, and are used as nicknames for the styles. This file contains lines with a key: value pattern, with the following keys being currently supported:

Id

The public identifier for the style-sheet

Desc

A short description of the styles, to be displayed in the help message

"pdfOverride,

rtfOverride, mifOverride" 10 A dsssl symbol from the print style-sheet to be set to #t (or a symbol=value pair, suitable as argument to jade's -V option), to be used for the given print format.

Only one symbol per override line is allowed. To define values for several symbols, use several lines.

Inherits

The nickname of a style-sheet this one inherits from, to avoid needless duplication of style definitions.

Currently, this only causes inheritance of the *Override parameters.

Priority

An positive integer to help selecting the default style when one cannot be derived from the document. Higher values get higher chance of being taken as default. Take care of using low priorities for hyper-specialized styles for a generic document-type, so that it does not get used by error.

For example, the current recommended policy for the DocBook style-sheets derived from Norman Walsh's is as follows (and may change if experience proves it to be inadequate).

10

The base style-sheets, which usually must be customized.

0

Any style-sheet that was written for an hyper-specialized purpose (e.g. marketing product sheet).

1000

A default style for all documents produced by an organization. Usually a light customization, featuring layout preferences, the organization's logo, or such things.

10-100

Miscellaneous generic customizations of the base style-sheets.

When you write an improved version of a style-sheet with priority n, you usually want to select a higher priority.

/etc/sgml/sgml2x/

~/.sgml2x/

./sgml2x/

The default configuration directories, in which the configuration files are searched for. See documentation for --confdirs for more details.

confdir/style/

The hierarchy that defines usable styles. See "Configuration" for more details.

confdir/dssslproc

A file containing an ordered list of dsssl processors to look for, separated by newlines and/or whitespace. Lines starting with a # character are treated as comments. Common values include openjade and jade.

DSSSL processors specified here should accept the -V and -D jade-compatible command-line options.

The configuration directories are looked for starting with the most specific one, so that, with the default confdirs, the project settings may override user settings, which in turn may override system settings.

The special value false can be used to stop the search and prevent looking into more generic directories. If for example a project must use the openjade-1.4devel command and no other, it can specify openjade-1.4devel false in its dssslproc file.

When using openjade-1.4devel as DSSSL processor, you'll see a complaint about the top-level flow-object generated by doctype.dsl, and automatic determination of the document-type will fail. This error is otherwise harmless. Ideas of how to deal with this, or confirmation that openjade-1.4devel is too strict, will be appreciated :)

Planned features for future releases include:

•

Integration of an index generator

•

Integration of a pretty-printing engine for code examples

•

Specification of transformations to be chained

•

Declaration of subset docclasses to allow the use with any docclass of the style-sheets that apply to its superset docclasses.

•

Work in a temporary location so as not to pollute the working directory with temporary files. This is not as easy as it sounds, because it breaks a document refers to image files using relative paths. That may be seen as a jade bug, however.

Browse the full TODO list and send us more ideas !

Copyright © 2001-2003 Alcove & Yann Dirson.

sgml2x is licensed under the GNU General Public License, version 2.

This documentation is licensed under the GNU Free Documentation License, version 1.

sgml2x is part of the AlcoveBook project (link to URL http://www.alcove-labs.org/en/software/alcovebook/) . Please use the AlcoveBook mailing lists (link to URL https://savannah.gnu.org/mail/?group_id=533) to get in touch with developers and users.

The list of bugs and feature requests is available through a Web interface (link to URL https://savannah.gnu.org/support/?group_id=533) . Please use it to submit problems and ideas.

openjade(1), jade(1), jadetex(1), collateindex.pl(1).