[PATCH v5 4/9] lib: introduce json_path_cmp routine

[Kirill Shcherbatov <kshcherbatov@tarantool.org>]

Introduced a new json_path_cmp routine as a part of JSON library
to compare JSON paths that may have different representation.

Need for #1012
---
 src/lib/json/json.c        | 28 ++++++++++++++++++++++++++++
 src/lib/json/json.h        | 11 +++++++++++
 test/unit/json_path.c      | 31 ++++++++++++++++++++++++++++++-
 test/unit/json_path.result | 12 +++++++++++-
 4 files changed, 80 insertions(+), 2 deletions(-)

diff --git a/src/lib/json/json.c b/src/lib/json/json.c
index 9198dca..20fbbba 100644
--- a/src/lib/json/json.c
+++ b/src/lib/json/json.c
@@ -507,3 +507,31 @@ json_tree_postorder_next(struct json_token *root, struct json_token *pos)
 	}
 	return pos->parent != root ? pos->parent : NULL;
 }
+
+int
+json_path_cmp(const char *a, uint32_t a_len, const char *b, uint32_t b_len)
+{
+	struct json_lexer lexer_a, lexer_b;
+	json_lexer_create(&lexer_a, a, a_len);
+	json_lexer_create(&lexer_b, b, b_len);
+	struct json_token token_a, token_b;
+	int rc_a, rc_b;
+	while ((rc_a = json_lexer_next_token(&lexer_a, &token_a)) == 0 &&
+		(rc_b = json_lexer_next_token(&lexer_b, &token_b)) == 0 &&
+		token_a.key.type != JSON_TOKEN_END &&
+		token_b.key.type != JSON_TOKEN_END) {
+		int rc = json_token_key_cmp(&token_a, &token_b);
+		if (rc != 0)
+			return rc;
+	}
+	/* Path "a" should be valid. */
+	assert(rc_a == 0);
+	if (rc_b != 0)
+		return rc_b;
+	/*
+	 * The parser stopped because the end of one of the paths
+	 * was reached. As JSON_TOKEN_END > JSON_TOKEN_{NUM, STR},
+	 * the path having more tokens has lower key.type value.
+	 */
+	return token_b.key.type - token_a.key.type;
+}
diff --git a/src/lib/json/json.h b/src/lib/json/json.h
index dd09f5a..7d46601 100644
--- a/src/lib/json/json.h
+++ b/src/lib/json/json.h
@@ -137,6 +137,17 @@ json_lexer_create(struct json_lexer *lexer, const char *src, int src_len)
 int
 json_lexer_next_token(struct json_lexer *lexer, struct json_token *token);
 
+/**
+ * Compare two JSON paths using Lexer class.
+ * - @a path must be valid
+ * - at the case of paths that have same token-sequence prefix,
+ *   the path having more tokens is assumed to be greater
+ * - when @b path contains an error, the path "a" is assumed to
+ *   be greater
+ */
+int
+json_path_cmp(const char *a, uint32_t a_len, const char *b, uint32_t b_len);
+
 /** Create a JSON tree object to manage data relations. */
 int
 json_tree_create(struct json_tree *tree);
diff --git a/test/unit/json_path.c b/test/unit/json_path.c
index f6b0472..35c2164 100644
--- a/test/unit/json_path.c
+++ b/test/unit/json_path.c
@@ -352,15 +352,44 @@ test_tree()
 	footer();
 }
 
+void
+test_path_cmp()
+{
+	const char *a = "Data[1][\"FIO\"].fname";
+	uint32_t a_len = strlen(a);
+	const struct path_and_errpos rc[] = {
+		{a, 0},
+		{"[\"Data\"][1].FIO[\"fname\"]", 0},
+		{"Data[[1][\"FIO\"].fname", 6},
+		{"Data[1]", 1},
+		{"Data[1][\"FIO\"].fname[1]", -2},
+		{"Data[1][\"Info\"].fname[1]", -1},
+	};
+	header();
+	plan(lengthof(rc));
+
+	for (size_t i = 0; i < lengthof(rc); ++i) {
+		const char *path = rc[i].path;
+		int errpos = rc[i].errpos;
+		int rc = json_path_cmp(a, a_len, path, strlen(path));
+		is(rc, errpos, "path cmp result \"%s\" with \"%s\": "
+		   "have %d, expected %d", a, path, rc, errpos);
+	}
+
+	check_plan();
+	footer();
+}
+
 int
 main()
 {
 	header();
-	plan(3);
+	plan(4);
 
 	test_basic();
 	test_errors();
 	test_tree();
+	test_path_cmp();
 
 	int rc = check_plan();
 	footer();
diff --git a/test/unit/json_path.result b/test/unit/json_path.result
index df68210..5c1de38 100644
--- a/test/unit/json_path.result
+++ b/test/unit/json_path.result
@@ -1,5 +1,5 @@
 	*** main ***
-1..3
+1..4
 	*** test_basic ***
     1..71
     ok 1 - parse <[0]>
@@ -138,4 +138,14 @@ ok 2 - subtests
     ok 35 - records iterated count 4 of 4
 ok 3 - subtests
 	*** test_tree: done ***
+	*** test_path_cmp ***
+    1..6
+    ok 1 - path cmp result "Data[1]["FIO"].fname" with "Data[1]["FIO"].fname": have 0, expected 0
+    ok 2 - path cmp result "Data[1]["FIO"].fname" with "["Data"][1].FIO["fname"]": have 0, expected 0
+    ok 3 - path cmp result "Data[1]["FIO"].fname" with "Data[[1]["FIO"].fname": have 6, expected 6
+    ok 4 - path cmp result "Data[1]["FIO"].fname" with "Data[1]": have 1, expected 1
+    ok 5 - path cmp result "Data[1]["FIO"].fname" with "Data[1]["FIO"].fname[1]": have -2, expected -2
+    ok 6 - path cmp result "Data[1]["FIO"].fname" with "Data[1]["Info"].fname[1]": have -1, expected -1
+ok 4 - subtests
+	*** test_path_cmp: done ***
 	*** main: done ***
-- 
2.7.4