hxtoc - insert a table of contents in an HTML file


hxtoc [ -x ] [ -l low ] [ -h high ] [ -t ] [ -d ] [ -c class ] [ -f ] [ file-or-URL ]


The hxtoc command reads an HTML file, inserts missing ID attributes in all H1 to H6 elements between the levels -l and -h (unless the option -d is in effect, see below). Unless the option -t is given, it also inserts A elements with NAME attributes, because old browsers do not recognize ID attributes as target anchors. The output is written to stdout.

If there is a comment of the form


or a pair of comments


then the comment, or the pair with everything in between, will be replaced by a table of contents, consisting of a list (UL) of links to all headers in the document.

The text of headers is copied to this table of contents, including any inline markup, except that ID attributes, DFN tags and SPAN tags with a CLASS of "index" are omitted (but the elements’ content is copied).

The copied text can optionally be "flattened" first, see option -f.

If a header has a CLASS attribute with as value (or one of its values) the keyword "no-toc", then that header will not appear in the table of contents.


The following options are supported:


Use XML conventions: empty elements are written with a slash at the end: <IMG />

-l low

Sets the lowest numbered header to appear in the table of content. Default is 1 (i.e., H1).

-h high

Sets the highest numbered header to appear in the table of content. Default is unlimited.


Normally, hxtoc adds both ID attributes and empty A elements with a NAME attribute and CLASS="bctarget", so that older browsers that do no understand ID will still find the target. With this option, the A elements will not be generated.

-c class

The generated UL elements in the table of contents will have a CLASS attribute with the value class. The default is "toc".


Tries to use sectioning elements as targets in the table of contents instead of H1 to H6. A sectioning elements is a DIV, SECTION, ARTICLE, ASIDE or NAV element that contains at least one heading element (H1 to H6) or HGROUP. The sectioning element will be given an ID if it doesn’t have one yet. With this option, the level of any H1 to H6 that is the first heading of a sectioning element is not determined by its name, but by the nesting depth of the sectioning elements. (Any H1 to H6 that are not the first heading of a sectioning element still have their level implied by their name.)


Flatten the text of the table of contents. Without -f, the contents of header elements are copied to the table of contents almost unchanged, i.e., including any child elements and their attributes (except for ID attributes, DFN elements and certain SPAN elements, as explained above). With -f, the contents are flattened instead: All child elements are removed and only their contents are copied to the table of contents. Additionally elements with an ALT attribute, such as IMG, are replaced by the contents of the ALT attribute. Exception: BDO tags are copied unchanged and elements with a DIR attribute are replaced by a SPAN with that DIR attribute. (BDO and DIR may occur in languages written right-to-left.)


The following operand is supported:

The name or URL of an HTML file. If absent, standard input is read instead.


The following exit values are returned:


Successful completion.

> 0

An error occurred in the parsing of the HTML file. hxtoc will try to correct the error and produce output anyway.


asc2xml(1), hxnormalize(1), hxnum(1), xml2asc(1)


The error recovery for incorrect HTML is primitive.