threads / patch / 7975

patchSmall correction in reading of commit headers

Subject: [PATCH] Small correction in reading of commit headers

## tl;dr

One message between May 4, 2007 and May 4, 2007. Diffs are folded; open one to read it.

replies: 0people: 1as markdown or json

Alex Riesen· May 4, 2007, 21:51 UTC · lore

Check if a line of the header has enough characters to possibly contain the requested prefix.

Signed-off-by: Alex Riesen <raa.lkml@gmail.com>
---

Noticed by debugging git-log crash in broken repos with missing packs. It happened to be up in backtrace. At the moment the optimization is actually a slow down: no line in the header is shorter than the only key this function is ever asked for: encoding. But, in case the function ever gets asked for something longer, the old condition is just wrong: it does not take into account the amount of data in the line when comparing with key.

 commit.c |    4 +++-
 1 files changed, 3 insertions(+), 1 deletions(-)
Show changes to commit.c +3 −1
diff --git a/commit.c b/commit.c
index aa7059c..d01833d 100644
--- a/commit.c
+++ b/commit.c
@@ -638,7 +638,9 @@ static char *get_header(const struct commit *commit, const char *key)
 			next = NULL;
 		} else
 			next = eol + 1;
-		if (!strncmp(line, key, key_len) && line[key_len] == ' ') {
+		if (eol - line > key_len &&
+		    !strncmp(line, key, key_len) &&
+		    line[key_len] == ' ') {
 			int len = eol - line - key_len;
 			char *ret = xmalloc(len);
 			memcpy(ret, line + key_len + 1, len - 1);
-- 
1.5.2.rc1.21.g80e79

← back to recent threads