From f6e55a026abf33867141896b1e227e791942c2a3 Mon Sep 17 00:00:00 2001
From: P. J. McDermott <pj@pehjota.net>
Date: Fri, 19 Feb 2016 04:01:32 -0500
Subject: [WIP] Add lexer/parser demo

---
(limited to 'parsing/parse.sh')

diff --git a/parsing/parse.sh b/parsing/parse.sh
new file mode 100644
index 0000000..2ac4215
--- /dev/null
+++ b/parsing/parse.sh
@@ -0,0 +1,133 @@
+HT="$(printf '\t.')";   HT="${HT%.}"
+LF="$(printf '\n.')";   LF="${LF%.}"
+RS="$(printf '\036.')"; RS="${RS%.}"
+US="$(printf '\037.')"; US="${US%.}"
+
+. ./tokens.sh
+. ./lexer.sh
+
+complete_command()
+{
+	if list; then
+		separator
+		return 0
+	fi
+	return 1
+}
+
+list()
+{
+	if and_or; then
+		while separator_op; do
+			if ! and_or; then
+				return 1
+			fi
+		done
+		return 0
+	fi
+	return 1
+}
+and_or()
+{
+	if pipeline; then
+		while accept T_AND_IF || accept T_OR_IF; do
+			if ! linebreak || ! pipeline; then
+				return 1
+			fi
+		done
+		return 0
+	fi
+	return 1
+}
+pipeline()
+{
+	accept T_BANG
+	if pipe_sequence; then
+		return 0
+	fi
+	return 1
+}
+
+pipe_sequence()
+{
+	if command; then
+		while accept T_PIPE; do
+			if ! linebreak || ! command; then
+				return 1
+			fi
+		done
+		return 0
+	fi
+	return 1
+}
+
+command()
+{
+	# XXX: Unfinished
+	accept T_WORD
+}
+
+
+
+newline_list()
+{
+	if accept T_NEWLINE; then
+		while accept T_NEWLINE; do
+			:
+		done
+		return 0
+	fi
+	return 1
+}
+linebreak()
+{
+	newline_list
+	return 0
+}
+
+separator_op()
+{
+	if accept T_AND || accept T_SEMI; then
+		return 0
+	fi
+	return 1
+}
+
+separator()
+{
+	if separator_op && linebreak; then
+		return 0
+	elif newline_list; then
+		return 0
+	fi
+	return 1
+}
+
+parse()
+{
+	local fn="${1}"
+	shift 1
+
+	init_lexer "${fn}"
+	while complete_command; do :; done
+	if :; then  # TODO: Test for EOF or errors
+		get_tokens
+		return 0
+	fi
+	return 1
+}
+
+if tokens="$(printf '%s\n' '"foo bar" && $baz || qux' '${quux%uux } quuux' | \
+		parse -)"; then
+	IFS="${RS}"
+	for t in ${tokens}; do
+		printf 'Token: %s\n' "$(tokname "${t}")"
+		case "${t%${US}*}" in T_WORD)
+			printf '       "%s"\n' "${t#T_WORD${US}}"
+			;;
+		esac
+	done
+	unset IFS
+else
+	echo FAIL
+fi
--
cgit v0.9.1