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

[PATCH 4/5] log: handle integer overflow in timestamps

From
Jeff King <peff@peff.net>
Date
Feb 24, 2014, 07:46 UTC
Message-ID
<20140224074637.GD9969@sigill.intra.peff.net>
In-Reply-To
<20140224073348.GA20221@sigill.intra.peff.net>

If an ident line has a ridiculous date value like (2^64)+1, we currently just pass ULONG_MAX along to the date code, which can produce nonsensical dates.

On systems with a signed long time_t (e.g., 64-bit glibc systems), this actually doesn't end up too bad. The ULONG_MAX is converted to -1, we apply the timezone field to that, and the result ends up somewhere between Dec 31, 1969 and Jan 1, 1970.

However, there is still a few good reasons to detect the overflow explicitly:

  1. On systems where "unsigned long" is smaller than
     time_t, we get a nonsensical date in the future.
  2. Even where it would produce "Dec 31, 1969", it's easier
     to recognize "midnight Jan 1" as a consistent sentinel
     value for "we could not parse this".
  3.  Values which do not overflow strtoul but do overflow a
      signed time_t produce nonsensical values in the past.
      For example, on a 64-bit system with a signed long
      time_t, a timestamp of 18446744073000000000 produces a
      date in 1947.

We also recognize overflow in the timezone field, which could produce nonsensical results. In this case we show the parsed date, but in UTC.

Signed-off-by: Jeff King <peff@peff.net>
---
A note on these tests. They are designed for 64-bit systems, but should
run fine on 32-bit systems (they are both just overflow, and the second
test is not doing anything novel that the first is not).

However, the second test relies on finding a value that fits into an "unsigned long" but does not fit into a time_t. For systems with a 64-bit signed time_t, that's fine. But if a system has an unsigned 64-bit time_t, the test will fail (it will actually produce some value 300 billion years from now).

I'm inclined to include it as-is, as I do not know of any such system (and it would be kind of lame, since it could not represent dates before 1970). If somebody comes up with such a system, we can allow either output (we could do it preemptively, but somebody would have to calculate the exact date/time billions of years in the future; we would be just as likely to disagree with the system's gmtime about something silly like leapseconds).

 pretty.c               | 10 ++++++++--
 t/t4212-log-corrupt.sh | 16 ++++++++++++++++
 2 files changed, 24 insertions(+), 2 deletions(-)
diff --git a/pretty.c b/pretty.c
index 87db08b..3b811ed 100644
--- a/pretty.c
+++ b/pretty.c
@@ -401,8 +401,14 @@ static const char *show_ident_date(const struct ident_split *ident,
 
 	if (ident->date_begin && ident->date_end)
 		date = strtoul(ident->date_begin, NULL, 10);
-	if (ident->tz_begin && ident->tz_end)
-		tz = strtol(ident->tz_begin, NULL, 10);
+	if (date_overflows(date))
+		date = 0;
+	else {
+		if (ident->tz_begin && ident->tz_end)
+			tz = strtol(ident->tz_begin, NULL, 10);
+		if (tz == LONG_MAX || tz == LONG_MIN)
+			tz = 0;
+	}
 	return show_date(date, tz, mode);
 }
 
diff --git a/t/t4212-log-corrupt.sh b/t/t4212-log-corrupt.sh
index 83de981..ba25a2e 100755
--- a/t/t4212-log-corrupt.sh
+++ b/t/t4212-log-corrupt.sh
@@ -65,4 +65,20 @@ test_expect_success 'unparsable dates produce sentinel value (%ad)' '
 	test_cmp expect actual
 '
 
+# date is 2^64 + 1
+test_expect_success 'date parser recognizes integer overflow' '
+	commit=$(munge_author_date HEAD 18446744073709551617) &&
+	echo "Thu Jan 1 00:00:00 1970 +0000" >expect &&
+	git log -1 --format=%ad $commit >actual &&
+	test_cmp expect actual
+'
+
+# date is 2^64 - 2
+test_expect_success 'date parser recognizes time_t overflow' '
+	commit=$(munge_author_date HEAD 18446744073709551614) &&
+	echo "Thu Jan 1 00:00:00 1970 +0000" >expect &&
+	git log -1 --format=%ad $commit >actual &&
+	test_cmp expect actual
+'
+
 test_done
-- 
1.8.5.2.500.g8060133
Previous: Jeff KingNext: Junio C Hamano
Message 5 of 48 in “handle bogus commit dates”
  1. 0/5 handle bogus commit datesJeff King, Feb 24, 2014
  2. 1/5 t4212: test bogus timestamps with git-logJeff King, Feb 24, 2014
  3. 2/5 fsck: report integer overflow in author timestampsJeff King, Feb 24, 2014
  4. 3/5 date: check date overflow against time_tJeff King, Feb 24, 2014
  5. 4/5 log: handle integer overflow in timestampsJeff King, Feb 24, 2014
  6. Junio C HamanoFeb 24, 2014
  7. Jeff KingFeb 24, 2014
  8. Junio C HamanoFeb 24, 2014
  9. Jeff KingFeb 24, 2014
  10. Junio C HamanoFeb 24, 2014
  11. 5/5 log: do not segfault on gmtime errorsJeff King, Feb 24, 2014
  12. René ScharfeMar 22, 2014
  13. Jeff KingMar 24, 2014
  14. René ScharfeMar 24, 2014
  15. Jeff KingMar 24, 2014
  16. Charles BaileyMar 26, 2014
  17. Jeff KingMar 26, 2014
  18. t4212: handle systems with post-apocalyptic gmtimeJeff King, Mar 26, 2014
  19. Junio C HamanoMar 26, 2014
  20. Jeff KingMar 26, 2014
  21. Jeff KingMar 26, 2014
  22. Jeff KingMar 26, 2014
  23. Charles BaileyMar 26, 2014
  24. Jeff KingMar 26, 2014
  25. Charles BaileyMar 26, 2014
  26. Charles BaileyMar 26, 2014
  27. Jeff KingMar 26, 2014
  28. Charles BaileyMar 26, 2014
  29. Jeff KingMar 27, 2014
  30. Junio C HamanoMar 28, 2014
  31. Jeff KingMar 28, 2014
  32. Junio C HamanoMar 28, 2014
  33. Jeff KingMar 28, 2014
  34. Junio C HamanoMar 28, 2014
  35. Jeff KingApr 1, 2014
  36. 1/2 date: recognize bogus FreeBSD gmtime outputJeff King, Apr 1, 2014
  37. René ScharfeApr 1, 2014
  38. Junio C HamanoApr 1, 2014
  39. René ScharfeApr 1, 2014
  40. Jeff KingApr 1, 2014
  41. 2/2 t4212: loosen far-in-future test for AIXJeff King, Apr 1, 2014
  42. work around unreliable gmtime errors on AIXJeff King, Apr 1, 2014
  43. Junio C HamanoApr 1, 2014
  44. Jeff KingApr 1, 2014
  45. Junio C HamanoMar 26, 2014
  46. Jeff KingMar 26, 2014
  47. Junio C HamanoMar 26, 2014
  48. Jeff KingMar 26, 2014

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.