DORMANTno replies

[PATCH] fix reflog approxidate parsing bug

From: Jeff King <hidden>
Date: 2016-06-15 22:44:33
Subsystem: the rest · Maintainer: Linus Torvalds

In get_sha1_basic, we parse a string like

  HEAD@{10 seconds ago}:path/to/file

into its constituent ref, reflog date, and path components.
We never actually munge the string itself, but instead keep
offsets into the string with their associated lengths.

When we call approxidate on the contents inside braces,
however, we pass just a string without a length. This means
that approxidate could sometimes look past the closing brace
and (erroneously) interpret the rest of the string as part
of the date.

Signed-off-by: Jeff King <redacted>
---
Sorry, I don't have a concise way of writing a test-case. Geoff's
example shows the bug when parsing:

  HEAD{10 seconds ago}:data/node/node.newds1

but it works OK with:

  HEAD{10 seconds ago}:data/node

I didn't track down the exact conditions that cause approxidate to not
like the string, since it seems wrong to be sending it a non-terminated
string anyway.

 sha1_name.c |    7 +++++--
 1 files changed, 5 insertions(+), 2 deletions(-)
diff --git a/sha1_name.c b/sha1_name.c
index 491d2e7..b0b2167 100644
--- a/sha1_name.c
+++ b/sha1_name.c
@@ -351,8 +351,11 @@ static int get_sha1_basic(const char *str, int len, unsigned char *sha1)
 		}
 		if (0 <= nth)
 			at_time = 0;
-		else
-			at_time = approxidate(str + at + 2);
+		else {
+			char *tmp = xstrndup(str + at + 2, reflog_len);
+			at_time = approxidate(tmp);
+			free(tmp);
+		}
 		if (read_ref_at(real_ref, at_time, nth, sha1, NULL,
 				&co_time, &co_tz, &co_cnt)) {
 			if (at_time)
-- 
1.5.5.1.178.gc1eb9.dirty
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help