From b2f474336ab19a33827ba062dd8d93d839b9c9f0 Mon Sep 17 00:00:00 2001
From: Camden Dixie O'Brien <cdo@wip.sh>
Date: Fri, 25 Oct 2024 18:40:38 +0100
Subject: [PATCH] Implement character class parsing

---
 lib/parser.c         | 52 ++++++++++++++++++++++++++++++++++++++++++++
 tests/parser_tests.c | 36 ++++++++++++++++++++++++++++++
 2 files changed, 88 insertions(+)

diff --git a/lib/parser.c b/lib/parser.c
index 94d2418..3a5acc9 100644
--- a/lib/parser.c
+++ b/lib/parser.c
@@ -8,6 +8,7 @@
 #include <stdbool.h>
 #include <stdlib.h>
 
+#define CLASS_START_CAPACITY 4
 #define SEQUENCE_START_CAPACITY 8
 
 static bool is_special(char c)
@@ -21,6 +22,9 @@ static bool is_special(char c)
 	case '*':
 	case '+':
 	case '?':
+	case '[':
+	case ']':
+	case '^':
 		return true;
 	default:
 		return false;
@@ -40,6 +44,48 @@ static int parse_literal(const char *input, int rem, char *out)
 	}
 }
 
+static int parse_class(const char *input, int rem, class_t *out)
+{
+	int result, used = 0;
+
+	if (used >= rem || '[' != input[used])
+		return -1;
+	++used;
+
+	if (used < rem && '^' == input[used]) {
+		out->negated = true;
+		++used;
+	}
+
+	out->count = 0;
+	out->capacity = CLASS_START_CAPACITY;
+	out->contents = malloc(out->capacity);
+	if (NULL == out->contents)
+		return -1;
+
+	while (used < rem) {
+		if (out->count >= out->capacity) {
+			out->capacity *= 2;
+			out->contents = realloc(out->contents, out->capacity);
+			if (NULL == out->contents)
+				return -1;
+		}
+
+		result = parse_literal(
+		    input + used, rem - used, &out->contents[out->count]);
+		if (result < 0)
+			break;
+		used += result;
+		++out->count;
+	}
+
+	if (used >= rem || ']' != input[used])
+		return -1;
+	++used;
+
+	return out->count > 0 ? used : -1;
+}
+
 static int parse_term(const char *input, int rem, term_t *out)
 {
 	int result, used = 0;
@@ -62,6 +108,12 @@ static int parse_term(const char *input, int rem, term_t *out)
 		if (')' != input[used])
 			return -1;
 		++used;
+	} else if ('[' == input[0]) {
+		result = parse_class(input + used, rem - used, &out->class);
+		if (result < 0)
+			return -1;
+		out->type = TERM_TYPE_CLASS;
+		used += result;
 	} else {
 		result = parse_literal(input + used, rem - used, &out->literal);
 		if (result < 0)
diff --git a/tests/parser_tests.c b/tests/parser_tests.c
index b914cc4..027f6b5 100644
--- a/tests/parser_tests.c
+++ b/tests/parser_tests.c
@@ -211,6 +211,40 @@ static void dot_question_mark_is_parsed_as_zero_or_one_wildcard(void)
 	regex_free_children(&r);
 }
 
+static void a_in_brackets_is_parsed_as_class_containing_only_a(void)
+{
+	regex_t r = { 0 };
+	const int result = PARSE_REGEX_STRING("[a]", &r);
+	ASSERT_NE(-1, result);
+
+	ASSERT_EQ(1, r.sequence.len);
+	ASSERT_EQ(QUANTIFIER_NONE, r.sequence.contents[0].quantifier);
+	ASSERT_EQ(TERM_TYPE_CLASS, r.sequence.contents[0].type);
+	ASSERT_FALSE(r.sequence.contents[0].class.negated);
+	ASSERT_EQ(1, r.sequence.contents[0].class.count);
+	ASSERT_NOT_NULL(r.sequence.contents[0].class.contents);
+	ASSERT_EQ('a', r.sequence.contents[0].class.contents[0]);
+
+	regex_free_children(&r);
+}
+
+static void caret_a_in_brackets_parses_as_negated_class(void)
+{
+	regex_t r = { 0 };
+	const int result = PARSE_REGEX_STRING("[^a]", &r);
+	ASSERT_NE(-1, result);
+
+	ASSERT_EQ(1, r.sequence.len);
+	ASSERT_EQ(QUANTIFIER_NONE, r.sequence.contents[0].quantifier);
+	ASSERT_EQ(TERM_TYPE_CLASS, r.sequence.contents[0].type);
+	ASSERT_TRUE(r.sequence.contents[0].class.negated);
+	ASSERT_EQ(1, r.sequence.contents[0].class.count);
+	ASSERT_NOT_NULL(r.sequence.contents[0].class.contents);
+	ASSERT_EQ('a', r.sequence.contents[0].class.contents[0]);
+
+	regex_free_children(&r);
+}
+
 int main(void)
 {
 	TESTING_BEGIN();
@@ -228,5 +262,7 @@ int main(void)
 	dot_star_is_parsed_as_zero_or_more_wildcard();
 	dot_plus_is_parsed_as_one_or_more_wildcard();
 	dot_question_mark_is_parsed_as_zero_or_one_wildcard();
+	a_in_brackets_is_parsed_as_class_containing_only_a();
+	caret_a_in_brackets_parses_as_negated_class();
 	return TESTING_END();
 }