git/list[1] front-page[2] threads[3] people[4] search[5] about
 

[PATCH 05/12] load_subtree(): separate logic for internal vs. terminal entries

From
Michael Haggerty <mhagger@alum.mit.edu>
Date
Aug 26, 2017, 08:28 UTC
Message-ID
<b76c9b2596b4385ee40ea4da7f56190cf6d225f2.1503734566.git.mhagger@alum.mit.edu>
In-Reply-To
<cover.1503734566.git.mhagger@alum.mit.edu>
There are only two legitimate notes path components:
* A hexadecimal string that fills the rest of the SHA-1
* A two-digit hexadecimal string that constitutes another internal
  node.

So handle those two cases at the top level, and reject others as non-notes without trying to parse them. The logic separation also simplifies upcoming changes.

This prevents us from leaking memory for a leaf_node in the case of wrong-sized paths. There are still memory leaks in this code; they will be fixed in upcoming commits.

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
---
 notes.c | 52 +++++++++++++++++++++++++++++++---------------------
 1 file changed, 31 insertions(+), 21 deletions(-)
diff --git a/notes.c b/notes.c
index 62ab3f4ce3..768902055e 100644
--- a/notes.c
+++ b/notes.c
@@ -433,30 +433,40 @@ static void load_subtree(struct notes_tree *t, struct leaf_node *subtree,
 	while (tree_entry(&desc, &entry)) {
 		unsigned char type;
 		struct leaf_node *l;
-		int len, path_len = strlen(entry.path);
+		int path_len = strlen(entry.path);
+
+		if (path_len == 2 * (GIT_SHA1_RAWSZ - prefix_len)) {
+			/* This is potentially the remainder of the SHA-1 */
+			if (get_oid_hex_segment(entry.path, path_len,
+						object_oid.hash + prefix_len,
+						GIT_SHA1_RAWSZ - prefix_len) < 0)
+				goto handle_non_note; /* entry.path is not a SHA1 */
+
+			type = PTR_TYPE_NOTE;
+			l = (struct leaf_node *)
+				xcalloc(1, sizeof(struct leaf_node));
+			oidcpy(&l->key_oid, &object_oid);
+			oidcpy(&l->val_oid, entry.oid);
+		} else if (path_len == 2) {
+			/* This is potentially an internal node */
+			if (get_oid_hex_segment(entry.path, 2,
+						object_oid.hash + prefix_len,
+						GIT_SHA1_RAWSZ - prefix_len) < 0)
+				goto handle_non_note; /* entry.path is not a SHA1 */
 
-		len = get_oid_hex_segment(entry.path, path_len,
-				object_oid.hash + prefix_len, GIT_SHA1_RAWSZ - prefix_len);
-		if (len < 0)
-			goto handle_non_note; /* entry.path is not a SHA1 */
-		len += prefix_len;
-
-		/*
-		 * If object SHA1 is complete (len == 20), assume note object
-		 * If object SHA1 is incomplete (len < 20), and current
-		 * component consists of 2 hex chars, assume note subtree
-		 */
-		type = PTR_TYPE_NOTE;
-		l = (struct leaf_node *)
-			xcalloc(1, sizeof(struct leaf_node));
-		oidcpy(&l->key_oid, &object_oid);
-		oidcpy(&l->val_oid, entry.oid);
-		if (len < GIT_SHA1_RAWSZ) {
-			if (!S_ISDIR(entry.mode) || path_len != 2)
-				goto handle_non_note; /* not subtree */
-			l->key_oid.hash[KEY_INDEX] = (unsigned char) len;
 			type = PTR_TYPE_SUBTREE;
+			l = (struct leaf_node *)
+				xcalloc(1, sizeof(struct leaf_node));
+			oidcpy(&l->key_oid, &object_oid);
+			oidcpy(&l->val_oid, entry.oid);
+			if (!S_ISDIR(entry.mode))
+				goto handle_non_note; /* not subtree */
+			l->key_oid.hash[KEY_INDEX] = (unsigned char) (prefix_len + 1);
+		} else {
+			/* This can't be part of a note */
+			goto handle_non_note;
 		}
+
 		if (note_tree_insert(t, node, n, l, type,
 				     combine_notes_concatenate))
 			die("Failed to load %s %s into notes tree "
-- 
2.11.0
Previous: Michael HaggertyNext: Michael Haggerty
Message 11 of 23 in “Clean up notes-related code around `load_subtree()`”
  1. 00/12 Clean up notes-related code around `load_subtree()`Michael Haggerty, Aug 26, 2017
  2. 01/12 notes: make GET_NIBBLE macro more robustMichael Haggerty, Aug 26, 2017
  3. 03/12 load_subtree(): reduce the scope of some local variablesMichael Haggerty, Aug 26, 2017
  4. 04/12 load_subtree(): fix incorrect commentMichael Haggerty, Aug 26, 2017
  5. 06/12 load_subtree(): check earlier whether an internal node is a tree entryMichael Haggerty, Aug 26, 2017
  6. 08/12 get_oid_hex_segment(): return 0 on successMichael Haggerty, Aug 26, 2017
  7. 10/12 get_oid_hex_segment(): don't pad the rest of `oid`Michael Haggerty, Aug 26, 2017
  8. 11/12 hex_to_bytes(): simpler replacement for `get_oid_hex_segment()`Michael Haggerty, Aug 26, 2017
  9. 12/12 load_subtree(): declare some variables to be `size_t`Michael Haggerty, Aug 26, 2017
  10. 07/12 load_subtree(): only consider blobs to be potential notesMichael Haggerty, Aug 26, 2017
  11. 05/12 load_subtree(): separate logic for internal vs. terminal entriesMichael Haggerty, Aug 26, 2017
  12. 09/12 load_subtree(): combine some common codeMichael Haggerty, Aug 26, 2017
  13. 02/12 load_subtree(): remove unnecessary conditionalMichael Haggerty, Aug 26, 2017
  14. Junio C HamanoAug 26, 2017
  15. Michael HaggertyAug 27, 2017
  16. Michael HaggertyAug 28, 2017
  17. Junio C HamanoSep 1, 2017
  18. Johan HerlandAug 26, 2017
  19. Jeff KingSep 9, 2017
  20. Michael HaggertySep 10, 2017
  21. Jeff KingSep 10, 2017
  22. Michael HaggertySep 12, 2017
  23. Lars SchneiderSep 12, 2017

Read the whole thread, see it on lore, or plain text.

$ cat FOOTERMessages come from the public archive at lore.kernel.org/git, fetched every hour. The front page is chosen and written each morning by an AI editor and can be wrong; the threads themselves are the record. About and API. For agents: an MCP server at https://gitlist.dev/mcp, and any thread, story or person page as Markdown by adding .md to its URL (or sending Accept: text/markdown). Details in /llms.txt.