From c28f4a47e9a9395781530de5839e426e4aa2b3e0 Mon Sep 17 00:00:00 2001
From: Konstantin Osipov <kostja@tarantool.org>
Date: Tue, 5 Feb 2019 14:23:20 +0300
Subject: [PATCH] json.c: add a comment
Add a comment explaining the logic behind intermediate lookups
in json_tree_lookup_path() function.
---
src/lib/json/json.c | 11 +++++++++++
1 file changed, 11 insertions(+)
diff --git a/src/lib/json/json.c b/src/lib/json/json.c
index 010a61d621..1b1a3ec2c0 100644
--- a/src/lib/json/json.c
+++ b/src/lib/json/json.c
@@ -565,6 +565,17 @@ json_tree_lookup_path(struct json_tree *tree, struct json_token *root,
json_lexer_create(&lexer, path, path_len, index_base);
while ((rc = json_lexer_next_token(&lexer, &token)) == 0 &&
token.type != JSON_TOKEN_END && ret != NULL) {
+ /*
+ * We could skip intermediate lookups by computing
+ * a rolling hash of all tokens produced by the
+ * lexer. But then we would still have to traverse
+ * the path back to ensure it is equal to the
+ * found to the found one. For that we would have
+ * to keep the stack of lexer tokens somewhere.
+ * Given the complications of the alternative,
+ * intermediate lookups don't seem to be so big of
+ * a problem.
+ */
ret = json_tree_lookup(tree, ret, &token);
}
if (rc != 0 || token.type != JSON_TOKEN_END)
--
GitLab