#!/bin/bash
# Check html files using html-tidy.
#   This wrapper sets options suitable for posting to dreamwidth
#   and similar sites.  You might want a different set of defaults
#   for other use cases, although this is pretty generic except for
#   suppressing warnings for unescaped ampersands (which occur all
#   the time in query strings) and munging DW/LJ-specific-tags so
#   that they won't look like tags, because there doesn't seem to
#   be a reliable way of doing that in my version of tidy.

#   Ampersands are handled by escaping them wherever they occur, not
#   just in URLs -- that will of course fail if they were introducing
#   real entities, but we're only interested in the errors.

#   Since we need to shove the munged file through standard input,
#   the --gnu-emacs option doesn't do anything.  So we mung the
#   output to put them in file:line:column format

OPTIONS="--doctype auto --quote-ampersand yes \
	 --show-body-only yes --show-info no -q --gnu-emacs yes "

for f in $*; do
    sed -E -e 's/(<\/?)(user|lj|lj-cut|cut)/$1 $2/g' -e 's/&/&amp;/g' $1 \
	| tidy $OPTIONS 2>&1 >/dev/null \
	| sed   -e "s|^|${1}:|" -e "s/line //" -e 's/ column /:/' ;
done