#!/bin/sh
# html_cfg.sh 0.0.3, (c) 2008 by Patrick Boylan
# not to be confused with cfg2html.pl .
#
# You can modify, redistribute, and pretty much do whatever you want
# with this script under the terms of the GNU General Public License,
# Version 2, or the new one if you prefer. If you didn't get a copy of
# the GNU GPL with this program, look in /usr/share/common-licenses/ on
# Debian-based systems, or http://gnu.org/licenses/gpl-2.0.txt . If
# the URL should fail for some reason, they have a search engine.
#
# Again, this shouldn't be confused with Vlad Harchev's cfg2html.pl ,
# which produces much nicer HTML anyway. My goals are different. I
# wanted a set of hyperlinked config files which are easily navigated
# from within Lynx -- so, via the e)dit command, Lynx becomes its own
# best configuration editor. You should be able to find the settings
# you want to change, and change them more easily this way.
#
# With that in mind, I wanted the results to look as similar as
# possible in a web browser or text editor. The plaintext view might
# not be pretty, but I've tried to keep it legible.
# CHANGES:
# 0.0.3 * I *think* it's a POSIX-compliant /bin/sh script now.
# * DOCTYPE declaration was omitted, because I wasn't sure
# what DTD to put it under. WDG validator validates it
# now [except for those Lynx-specific TAB elements].
# * Uses separate for loops to generate the HTML pages and
# main index. Adds some unnecessary reptition, but it
# should be easier to read.
# * Fixed some of those multi-line echoes, so they won't
# screw up the indentation [purely cosmetic; also for
# legibility].
#
# 0.0.2 * Tried doing more in memory, less with temp-files, but
# that didn't seem to speed things up any.
# * Rearranged the `case "$foo" in' thing so the most likely
# possiblities are tested first. Went from a five minute
# job to like, a minute on my old 233 mHz Compaq Armada.
# * Went back to mostly temp-files.
# * Commented the hell out of it.
#
# 0.0.1 * You mean, it actually did something this time? Whoah...
nom="$(basename "$0")"
tmp="lynx_cfg_tmp"
sub="lynx_cfg.d"
idx="lynx_cfg.html"
phile="lynx_cfg"
intro="html"
hr="\n
"
dtd='
'
how2 () {
echo "
Usage: $nom /path/to/lynx.cfg [/path/to/output_dir/]
If no output directory is specified, the current directory is used.
Either path may be relative or absolute, but the output directory
must already exist.
Also note that if you wish to use $idx as your lynx.cfg ,
the output directory *must* be its final destination. Moving your
files around after running the script will invalidate all \`INCLUDE:'
directives inside $idx , because these use absolute paths.
If you only want a browsable HTML version of lynx.cfg to look at,
this is not a problem; the HTML links are relative. But if this is
what you're trying to do, I recommend using Vlad Harchev's excellent
cfg2html.pl instead.
"
}
if [ "$1" = "" ] || [ "$1" = "-h" ] || [ "$1" = "-help" ] || [ "$1" = "--help" ]
then
how2
exit 0
fi
if [ -f "$1" ]
then
lynx_cfg="$(cat "$1")"
if [ "$lynx_cfg" = "" ]
then
echo "That Didn't Work."
how2
exit 1
fi
else
echo "\`$1' is not a file."
how2
exit 1
fi
# Some half-hearted idiot-proofing.
if [ "$2" = "" ]
then
src="$0"
dest="`pwd`"
elif [ -d "$2" ]
then
if [ "$(echo "$0" | grep '^\./')" != "" ]
then
src="`pwd`/$0"
elif [ "$(echo "$0" | grep '^/')" != "" ]
then
src="$0"
elif [ -x "`pwd`/$0" ]
then
src="`pwd`/$0"
# Having `.' in your $PATH is a BAD idea,
# but it could happen.
fi
cd "$2"
dest="`pwd`"
else
echo "\`$2' is not a directory."
how2
exit 1
fi
# Still half-hearted.
echo "Loaded \`$1', sorting lines..."
# Declaring a few useful functions...
scrubHeader () {
# Strip the leading `.h1' and `.h2' from input.
echo "$1" | cut --delimiter=' ' --fields='2-222'
}
scrubFileName () {
# Convert the current line into a sensible filename
# [eg, all lowercase, no spaces, no funky puctuation].
echo "$1" | tr '_' ' ' | tr --delete '[:punct:]' | \
tr -s ' ' '_' | tr '[:upper:]' '[:lower:]'
}
scrubLine () {
# Use an HTML escape sequence if absolutely necessary.
scrubbedLine="$1"
while [ "$(echo "$scrubbedLine" | grep '<' -)" != "" ]
do
scrubbedLine="$(echo "$scrubbedLine" | sed 's/</\</')"
# Any way to make sed do its own recursion?
# This is a bit silly.
done
echo "$scrubbedLine"
}
num=0
tmp2="$tmp"
while [ -e "$tmp2" ]
do
tmp2="${tmp}.$num"
num=$(($num + 1))
# Just making sure the temp directory *does_not*
# already exist, because it will be deleted later.
done
tmp="$tmp2"
mkdir "$tmp"
unset $num
unset $tmp2
echo "Using \`$dest/$tmp' for temp-files..."
if [ ! -d "$sub" ]
then
mkdir "$sub"
fi
IFS='
'
for line in $lynx_cfg
do
case "$line" in
\#*\<*) datOut="$datOut$(scrubLine "$line")\n"
;;
\#*) datOut="$datOut$line\n"
;;
.*) case "$line" in
.h2\ *) line="$(scrubHeader "$line")"
if [ "$datOut" != "" ]
then
if [ "$intro" = "intro" ]
then
echo "<PRE>\n$datOut</PRE>" \
>> "$tmp/${phile}.$intro"
else
echo "<PRE>\n$datOut\n</PRE>$hr" \
>> "$tmp/${phile}.$intro"
fi
datOut=''
#
# Flush the output buffer before changing
# the subject.
fi
intro="main"
aname="$(scrubFileName "$line")"
echo "<H2><A HREF=\"../$idx#$phile\" NAME=\"$aname\"\n\n\n >$line</A></H2>" \
>> "$tmp/${phile}.$intro"
# H2 header for this section.
echo " <LI><A HREF=\"$sub/${phile}.html#$aname\">$line</A></LI>" \
>> "$tmp/${phile}.lst"
# Sub-headings list for the main index.
echo " <LI><A HREF=\"#$aname\">$line</A></LI>" \
>> "$tmp/${phile}.ls2"
# Sub-headings list for each file [ToC].
;;
.h1\ *) line="$(scrubHeader "$line")"
lsH1="$lsH1$line\n"
#
# Store .h1 headers for future reference.
if [ "$datOut" != "" ]
then
if [ "$intro" = "intro" ]
then
echo "<PRE>\n$datOut</PRE>" \
>> "$tmp/${phile}.$intro"
else
echo "<PRE>\n$datOut\n</PRE>$hr" \
>> "$tmp/${phile}.$intro"
fi
datOut=''
#
# Flush the output buffer before changing
# the subject.
fi
intro="intro"
#
# Any text which occurs after the .h1 line, but
# before any .h2 lines, will be an introduction
# to this section.
phile="$(scrubFileName "$line")"
#
# Further output under this .h1 header will go into
# `${phile}.html'.
;;
.ex) datOut="$datOut#\n# Example:\n"
;;
.ex\ *) datOut="$datOut#\n# Examples:\n"
;;
*) datOut="$datOut#\n"
# Missed .nf and .fi, but with no formatting
# to begin with, there's no need to turn it
# off.
# No need to turn it back on again, either.
;;
esac
;;
*\<*) datOut="$datOut<!>$(scrubLine "$line") <!--\n"
datOut="$datOut# The line above is escaped for HTML viewers.\n"
datOut="$datOut# Active configuration item follows:\n"
datOut="$datOut$line\n"
datOut="$datOut# -->\n"
;;
*) datOut="$datOut$line\n"
;;
esac
done
unset $lynx_cfg
if [ "$datOut" != "" ]
then
echo "<PRE>\n$datOut\n</PRE>$hr" \
>> "$tmp/${phile}.$intro"
unset $datOut
#
# Don't forget to flush.
fi
# Okay, so... Now that we've dissected lynx.cfg ,
# let's try to put it back together again.
#
# Won't this be fun...
echo "Split \`$1' into temp-files. Re-assembling..."
lsH1="$(echo "$lsH1" | sort -u)"
for line in $lsH1
do
phile="$(scrubFileName "$line")"
{ echo "$dtd$line"
echo ""
echo "$line
"
} > "$sub/${phile}.html"
#
# Start a new file, based on .h1 directives in lynx.cfg .
if [ -f "$tmp/${phile}.intro" ]
then
cat "$tmp/${phile}.intro" \
>> "$sub/${phile}.html"
fi
if [ -f "$tmp/${phile}.ls2" ]
then
{ echo
echo "Sub-headings within this document:
"
echo ""
echo "$(cat "$tmp/${phile}.ls2")"
echo "
$hr"
} >> "$sub/${phile}.html"
fi
# Those .h2 headings go here.
if [ -f "$tmp/${phile}.main" ]
then
cat "$tmp/${phile}.main" \
>> "$sub/${phile}.html"
fi
# Dump the main text into your new file [after ToC].
done
echo "Generating index, \`$idx'..."
echo "${dtd}lynx.cfg file.
lynx.cfg file.
" > "$idx"
cat "$tmp/$idx" >> "$idx"
#
# Opening lines and intro for lynx_cfg.html .
echo "\nIndex by category:
\n" > "$tmp/$idx"
#
# Note that we're still using the temp-file for the detailed
# index [no ToC in main file yet, and that should go first].
# Also clears the temp-file, because whatever was in it has
# already been catted into the final copy.
inc=''
toc=''
for line in $lsH1
do
phile="$(scrubFileName "$line")"
echo "" \
>> "$tmp/$idx"
#
# Add a new header for this section of lynx_cfg.html
# [the temp-file, not the real one yet].
if [ -f "$tmp/${phile}.intro" ]
then
cat "$tmp/${phile}.intro" \
>> "$tmp/$idx"
fi
# Re-use the introduction for this section.
if [ -f "$tmp/${phile}.lst" ]
then
{ echo ""
echo "$(cat "$tmp/${phile}.lst")"
echo "
$hr"
} >> "$tmp/$idx"
fi
# Those .h2 sub-headings go here, too.
inc="${inc}INCLUDE:$dest/$sub/${phile}.html\n"
#
# Add this file to the list of INCLUDE directives.
toc="${toc}
| $line |
$line | \n
"
#
# Add this header to the table of contents [literally a table,
# in this case].
done
echo "" \
>> "$idx"
unset $inc
#
# Dump your INCLUDE directives into lynx_cfg.html [final, not temp].
echo "
Primary configuration headings:
| In this document: |
In its own file: |
$toc
$hr" \
>> "$idx"
unset $toc
#
# Now the table of contents.
cat "$tmp/$idx" >> "$idx"
# And the detailed index [stored in the temp-file].
echo "Removing temp-files, $dest/$tmp ..."
rm -rf "$dest/$tmp"
echo "Add closing to finish files..."
prevF=''
thisF=''
for nextT in $lsH1
do
nextF="$(scrubFileName "$nextT").html"
if [ "$thisF" = "" ]
then
{ echo
echo ""
echo " First Item: $nextT"
echo "
"
echo " HTML generated on"
echo "
$(date)"
echo "
by $nom -"
echo " For Free."
echo "
"
echo "