#!/bin/sed -f
# Thu May 18 12:43:45 CEST 2000 by tilmann@bitterberg.de
#
# Description:
# Creates an index of links from a HTML file
# Does something similar like lynx -force_html -dump but
# leaves the document html (generate an index of links)
#
# Example: Input
#
# foo1 Click here foo2
#
#
# Output:
#
# foo1 [1] Click here foo2
#
[1] http://link.org
#
#
# NOTE:
# 1) Will break at links like foo2\n 1 2 3 4 .. 500
# using newline as separator to the 's' command and 'I' for casei
s
\(a *href\) *= *\("\([^"]\+\)"[^>]*>\)\([^\n]*\(\n\)\) \([^ ]*\)\(.*$\)
\1|||||=\2[\6] \4\7\5[\6] \3
I
#|----1----| |----------2---------||-------4------| |---6---||--7--|
# |---3----| |--5-|
# Field Contains:
# \1 a href
# \2 the link text up to the closing >
# \3 the link itself (http://foo.com)
# \4 the rest of the input line
# \5 a newline (\n)
# \6 the number we would like to use
# \7 everything up to the end of patternspace
#
# Now the line looks like:
# foo1 [1] foo2\n 2 3 4 .. 500\n[1] blah.html
t loop; # look if there is another link in that line
s/|||||//g; # delete marker
h; # save how many numbers are used
s/\n.*//; # "restore" the original line
x
s/[^\n]*\n//
x
}
}
# Just before the insert index
/<\/[Bb][Oo][Dd][Yy]>/{
x; # insert saved stuff
s/[^\n]*\n//; # delete unused numbers
s/^/