build

#!/usr/bin/env sh
# Generate a static HTML website from Markdown with pandoc(1).
set -eu

if [ $# -ne 0 ]; then
	echo usage: $0 1>&2
	echo Generate a static HTML website. 1>&2
	exit 1
fi

if ! command -v pandoc >/dev/null; then
	echo error: pandoc not found 1>&2
	exit 1
fi

# HTML Tidy executable name
if command -v tidy5 >/dev/null; then
	tidy="tidy5"
elif command -v tidy >/dev/null; then
	tidy="tidy"
else
	echo warning: HTML Tidy not found 1>&2
	tidy=""
fi

cd "$(dirname "$0")"
rootdir="$(pwd)"
outdir="$rootdir"/public

# Create scratch directory (deleted when script exits)
tmpdir="$(mktemp -d)"
trap 'rm -rf -- "$tmpdir"' EXIT

# if the output directory exists, temporarily save a copy of it as an
# optimization for reusing pages which haven't changed
olddir=
if [ -d "$outdir" ]; then
	olddir="$tmpdir"/old
	mv "$outdir" "$olddir"
fi

# delete and recreate the output directory so that it's a clean slate
rm -rf "$outdir"
mkdir -p "$outdir"
outdir="$(readlink -f "$outdir")"

# Everything in static/ is deployed as-is.  A good place for stylesheets,
# scripts, images, standalone HTML pages, etc.
if [ -d static ]; then
	cp -r static/* "$outdir"
fi

# is system using GNU coreutils stat(1)?
isgnustat() {
	set +e
	stat --version 2>/dev/null | grep GNU >/dev/null
	grep_status=$?
	set -e
	return $grep_status
}

# file modification time in seconds since Unix epoch
mtime() {
	$(isgnustat) && echo "$(stat -c'%Y' "$1")" || echo "$(stat -f'%m' "$1")"
}

# file modifications fractional seconds
mtime_frac() {
	$(isgnustat) && \
		echo "$(stat -c'%y' "$1" | sed 's/.*\.\([0-9]*\).*/\1/')" || \
		echo "$(stat -f'%Fm' "$1" | sed 's/[0-9]*\.//')"
}

# is first file older than second file?
# yes, this script is reinventing make(1)
older_than() {
	[ ! -f "$2" ] && return 1
	[ $(mtime "$1") -lt $(mtime "$2") ] && return 0
	[ $(mtime "$1") -gt $(mtime "$2") ] && return 1
	[ $(mtime_frac "$1") -lt $(mtime_frac "$2") ] && return 0
	return 1
}

# are all given files (after the first) older than the first given file?
all_older_than() {
	first="$1"
	shift
	for f in "$@"; do
		! $(older_than "$f"  "$first") && return 1
	done
	return 0
}

# Update href="./" and src="./" paths to replace the "./" with the relative
# path to the top-level site directory.
update_relative_path() {
	path_to_root="$1"
	html_input="$2"
	html_output="$3"
	cat "$html_input" |
		sed "s|\( href=\"\)\./|\1$path_to_root|" | \
		sed "s|\( src=\"\)\./|\1$path_to_root|" \
		>"$html_output"
}

cd content
find . -type f -name '*.md' | while IFS= read page; do
	page_dir="$(dirname "$page")"
	page_md_name="$(basename "$page")"
	page_html_name="$(echo "$page_md_name" | sed 's/\.md$/.html/')"
	page_html_relpath="$page_dir"/"$page_html_name"

	mkdir -p "$outdir"/"$page_dir"

	if [ ! -z "$olddir" ]; then
		# if the HTML file is newer than all the files used to build it...
		if all_older_than "$olddir"/"$page_html_relpath" "$page" "$rootdir"/layouts/*; then
			# reuse the previous HTML file
			echo "$page_md_name unchanged"
			mv "$olddir"/"$page_html_relpath" "$outdir"/"$page_html_relpath"
			continue
		fi
	fi

	echo "$page_md_name -> $page_html_name"

	# Update relative paths in the header/footer to be correct for the
	# location of this page.
	page_toroot="$(echo "$page_dir/" | sed 's|^./*||' | sed -r 's|[^/]+|..|g')"
	update_relative_path "$page_toroot" "$rootdir"/layouts/header.html "$tmpdir"/header.html
	update_relative_path "$page_toroot" "$rootdir"/layouts/footer.html "$tmpdir"/footer.html

	tmpfile="$tmpdir"/"$page_html_name"

	pandoc \
		--from=markdown \
		--to=html5 \
		--standalone \
		--preserve-tabs \
		--eol=lf \
		--toc \
		--shift-heading-level-by=1 \
		--highlight-style=monochrome \
		--css="$page_toroot"style.css \
		--include-before-body="$tmpdir"/header.html \
		--include-after-body="$tmpdir"/footer.html \
		"$page" --output="$tmpfile"

	# The Pandoc HTML template includes the html5shiv script, using Internet
	# Explorer conditional comments so that the script is only included for
	# IE8 and earlier.  Remove it.  Removing it here is easier than
	# maintaining a fork of the template.
	cat "$tmpfile" | \
		grep -v '<!--\[if .*IE .*\]>' | \
		grep -v '<script src=".*html5shiv.*"></script>' | \
		grep -v '<!\[endif\]-->' \
		> "$tmpfile.tmp"
	mv "$tmpfile.tmp" "$tmpfile"

	if [ "$tidy" != "" ]; then
		set +e
		"$tidy" \
			-quiet \
			-modify \
			-utf8 \
			-ashtml \
			-omit \
			--doctype html5 \
			--newline LF \
			--keep-tabs yes \
			--wrap 0 \
			"$tmpdir"/"$page_html_name"
		tidy_status=$?
		set -e
		if [ $tidy_status -eq 2 ]; then
			echo "error: $tidy exited with errors"
			exit 1
		fi
	fi

	mv "$tmpdir"/"$page_html_name" "$outdir"/"$page_html_relpath"
done