Skip to content
/ nuweb Public

Nuweb is a literate programming tool like Knuth's WEB only simpler.

Notifications You must be signed in to change notification settings

ZoBoRf/nuweb

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

97 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Nuweb

This is not the "official" nuweb repository.

Here I make available some small enhancements I made.

The official web site is http://nuweb.sourceforge.net/.

Nuweb is a literate programming tool like Knuth's WEB only simpler. More on literate programming you can find on Wikipedia.

A nuweb file contains program source code interleaved with documentation.

Knuth describes the system as follows:

A WEB user writes a program that serves as the source language for two different system routines.

                TeX
         [TeX] -----> [DVI]
  WEAVE /
       /
  [WEB]
       \
 TANGLE \
         [PAS] -----> [REL]
               PASCAL

One line of processing is called weaving the web; it produces a document that describes the program clearly and that facilitates program maintenance. The other line of processing is called tangling the web; it produces a machine-executable program. The program and its documentation are both generated from the same source, so they are consistent with each other.

When nuweb is given nuweb a file, it writes the program file(s), and also produces LaTeX source for typeset documentation, i.e. nuweb tangles and weaves at the same time.

Prerequisites

Prerequisites for building and using nuweb:

  • GNU C Compiler gcc
    • On Windows I use MinGW, the Minimalist GNU for Windows.
  • A TeX distribution

Bootstrap nuweb

Use already tangled sources to build the initial nuweb.exe.

$ cd nuweb
$ make nuweb
cc -g   -c -o main.o main.c
cc -g   -c -o pass1.o pass1.c
cc -g   -c -o latex.o latex.c
cc -g   -c -o html.o html.c
cc -g   -c -o output.o output.c
cc -g   -c -o input.o input.c
cc -g   -c -o scraps.o scraps.c
cc -g   -c -o names.o names.c
cc -g   -c -o arena.o arena.c
cc -g   -c -o global.o global.c
cc -o nuweb main.o pass1.o latex.o html.o output.o input.o scraps.o names.o arena.o global.o

Rebuild from nuweb.w

The current directory and TeX must be in your PATH:

$ export PATH=.:$PATH
$ export PATH=/C/Program\ Files\ \(x86\)/MiKTeX\ 2.9/miktex/bin:$PATH

Now use nuweb.exe to rebuild itself from it's nuweb.w source.

$ make
make nuweb.tex
make[1]: Entering directory '.../nuweb'
nuweb -r nuweb.w
nuweb: you'll need to rerun nuweb after running latex
nuweb -r nuweb.w
nuweb: you'll need to rerun nuweb after running latex
make[1]: Leaving directory '.../nuweb'
make nuweb
make[1]: Entering directory '.../nuweb'
cc -g   -c -o main.o main.c
cc -g   -c -o pass1.o pass1.c
cc -g   -c -o latex.o latex.c
cc -g   -c -o html.o html.c
cc -g   -c -o output.o output.c
cc -g   -c -o input.o input.c
cc -g   -c -o scraps.o scraps.c
cc -g   -c -o names.o names.c
cc -g   -c -o arena.o arena.c
cc -g   -c -o global.o global.c
cc -o nuweb main.o pass1.o latex.o html.o output.o input.o scraps.o names.o arena.o global.o
make[1]: Leaving directory '.../nuweb'

Check the Build

Optionally test the build

$ make check
...
Testing test/00/t0001a.sh
/c/storage/project/nuweb.github/nuweb/./nuweb: you'll need to rerun nuweb after running latex
/c/storage/project/nuweb.github/nuweb/./nuweb: <Test with parameter> never referenced.
Testing test/00/t0002a.sh
...
Output written on test.dvi (1 page, 1752 bytes).
Transcript written on test.log.
35 done
0 failed

Build the weaved literate program PDF

$ cd nuweb
$ make nuweb.pdf

If the references in the PDF file are not right (i.e. are shown as [?]), repeat the following steps up to three times (TeX generates different intermediate files, which it includes in later runs):

$ touch nuweb.w
$ make nuweb.tex
$ make nuweb.pdf

Build the nuweb documentation PDF

Make the nuweb docs nuwebdoc.pdf:

$ make nuwebdoc.pdf
sed -e '/^\\ifshowcode$/,/^\\fi$/d' nuweb.tex > nuwebdoc.tex
pdflatex nuwebdoc
This is pdfTeX, Version 3.14159265-2.6-1.40.21 (MiKTeX 2.9.7400)
...
...
mtt10.pfb>
Output written on nuwebdoc.pdf (12 pages, 300734 bytes).
Transcript written on nuwebdoc.log.

RTFM

Read the fine manuals.

How to Read a Weaved Nuweb Document

Keith Harwood, one of the developers and users of nuweb suggests in his z-vimes project's weaved documentation (file z.pdf), how to read a weaved nuweb document:

This document is written using the literate programming style. The program is presented in a form which can, in principle, be read as a continuous narrative. Within the narrative the program code is presented as fragments. Each fragment is given a title which describes the action of the fragment. The bodies of the fragments contain code in the source language interspersed with references to other fragments. This allows us to describe the essential structure of the program, but defer details until later.

Each fragment is given an index indicating the page on which it occurs and, if there are more than one fragment on a page, a letter indicating which one. These indexes form cross-references between the definition and use of fragments. These cross-references are implemented as hyperlinks. The same cross-references also appear as comments in the code.

(If you are viewing this document with a pdf reader the coloured section names and fragment cross-references are hyperlinks. Click on them to switch attention to where they reference. If you are reading it on paper you have to turn the pages yourself. I apologise for that.)

Changes Made

Changes made so far are documented in changes/changes.md.