Filters and parsers for use with noweb
[noweave|notangle|noweb] [-filter filter|-markup parser] ...
This man page documents the various filters and parsers that are supplied with noweb(1). A filter, which transforms the pipeline representation, is used with the -filter option. A parser, which reads a source file and emits the pipeline representation, is used with the -markup option. Not all filters and parsers are available at all installations.
noweave's -autodefs and -showautodefs options use these filters, which automatically find defined identifiers. The * stands for the name of a programming language. (Icon installation only.)
The btdefn filter uses Dave Hanson's convention to mark definitions. In code chunks, each defining instances of an identifier should be preceded by a single backtick character (ASCII 140). The btdefn filter removes these backticks from the code and converts them to definitions, thus: `definition. It, not @ %def, should be used with both notangle(1) and noweave(1), at least for programs written in C-like languages.
The disambiguate filter makes it possible to abbreviate chunk names using a trailing ellipsis (three periods), in the style of WEB and nuweb(1).
The docs2comments filter converts documentation chunks to comments in code chunks. Options include -all, which converts all documentation chunks into comments; -one converts only one documentation chunk preceding each code chunk. The -wk option formats comments k columns wide.
The elide filter removes from the stream any code chunk matching any of the arguments. Each argument is a pattern in which ? matches any single character and * matches any sequence of characters.
The emptydefn filter makes an empty definition (<<>>=) an abbreviation for a continuation of the previous definition.
The l2h filter converts LaTeX to HTML in documentation chunks. If the -show-unknowns options is given, unrecognized LaTeX control sequences appear in bold in the output. l2h uses the LaTeX-generated .toc and .bbl files to help create table of contents and bibliography. (Icon installation only.)
The pipedocs filter runs each documentation chunk through an invocation of command, which may be any Unix command. The Noweb library includes the command h2a, which uses lynx(1) to convert HTML to ASCII.
The xchunks filter pulls in code chunks from an external source. The files argument lists one or more noweb files, which are used only for their definitions of code chunks. Empty chunk definitions in the main document are replaced with the external definitions of the same name. This filter makes it possible to include the same code chunks in multiple documents, or even the same code chunks multiple times in the same document. (Icon installation only.)
A parser converts a source file to the noweb pipeline format, interpreting the ``markup'' in that file. By default, the tools use the standard noweb markup, but some installations may support nuweb markup as well. Parsers include
Converts standard noweb(1) markup to pipeline format.
Converts nuweb(1) markup to noweb pipeline format. It is not part of the standard installation, but it can be found in noweb's contrib/norman directory.
All filters and parsers are stored in /usr/lib/noweb
It is possible to tell l2h about new control sequences by putting special TeX comments in one's documentation chunks, but there's no documentation. Use the source, Luke.
numarkup loses information, e.g., about the proper treatment of tabs. It also doesn't compensate for notangle's newline-removal heuristic, which means results on nuweb files may differ subtly from those produced by nuweb.
Many filters are not available in installations that use Awk instead of Icon.
This man page is from noweb version 2.11b.
Norman Ramsey, Harvard University. Internet address [email protected].
Noweb home page at http://www.eecs.harvard.edu/~nr/noweb.