From: Joel Rosdahl <joel@rosdahl.net>
Date: Sun, 13 Jan 2019 20:13:09 +0000 (+0100)
Subject: Check for word boundaries when scanning for __DATE__/__TIME__
X-Git-Tag: v3.6~7
X-Git-Url: http://git.ipfire.org/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=8602e1b19fd8f47943c8fb8cd3d795b3d8ea52c9;p=thirdparty%2Fccache.git

Check for word boundaries when scanning for __DATE__/__TIME__

As suggested in #347.
---

diff --git a/doc/NEWS.adoc b/doc/NEWS.adoc
index 9e129df02..07e7de661 100644
--- a/doc/NEWS.adoc
+++ b/doc/NEWS.adoc
@@ -45,6 +45,10 @@ Changes
 - The GCC variables “DEPENDENCIES_OUTPUT” and “SUNPRO_DEPENDENCIES” are now
   supported correctly.
 
+- The algorithm that scans for `__DATE_` and `__TIME__` tokens in the hashed
+  source code now doesn't produce false positives for tokens where `__DATE__`
+  or `__TIME__` is a substring.
+
 
 ccache 3.5.1
 ------------
diff --git a/src/hashutil.c b/src/hashutil.c
index 4ed10d774..789bf437b 100644
--- a/src/hashutil.c
+++ b/src/hashutil.c
@@ -1,4 +1,4 @@
-// Copyright (C) 2009-2018 Joel Rosdahl
+// Copyright (C) 2009-2019 Joel Rosdahl
 //
 // This program is free software; you can redistribute it and/or modify it
 // under the terms of the GNU General Public License as published by the Free
@@ -62,18 +62,18 @@ check_for_temporal_macros(const char *str, size_t len)
 		// Check whether the substring ending at str[i] has the form "__...E__". On
 		// the assumption that 'E' is less common in source than '_', we check
 		// str[i-2] first.
-		if (str[i - 2] == 'E' &&
-		    str[i - 0] == '_' &&
-		    str[i - 7] == '_' &&
-		    str[i - 1] == '_' &&
-		    str[i - 6] == '_') {
+		if (str[i - 2] == 'E'
+		    && str[i - 0] == '_'
+		    && str[i - 7] == '_'
+		    && str[i - 1] == '_'
+		    && str[i - 6] == '_'
+		    && (i < 8 || (str[i - 8] != '_' && !isalnum(str[i - 8])))
+		    && (i + 1 >= len || (str[i + 1] != '_' && !isalnum(str[i + 1])))) {
 			// Check the remaining characters to see if the substring is "__DATE__"
 			// or "__TIME__".
-			if (str[i - 5] == 'D' && str[i - 4] == 'A' &&
-			    str[i - 3] == 'T') {
+			if (str[i - 5] == 'D' && str[i - 4] == 'A' && str[i - 3] == 'T') {
 				result |= HASH_SOURCE_CODE_FOUND_DATE;
-			} else if (str[i - 5] == 'T' && str[i - 4] == 'I' &&
-			           str[i - 3] == 'M') {
+			} else if (str[i - 5] == 'T' && str[i - 4] == 'I' && str[i - 3] == 'M') {
 				result |= HASH_SOURCE_CODE_FOUND_TIME;
 			}
 		}
diff --git a/unittest/test_hashutil.c b/unittest/test_hashutil.c
index 2b2dbc9b1..d05769e5f 100644
--- a/unittest/test_hashutil.c
+++ b/unittest/test_hashutil.c
@@ -1,4 +1,4 @@
-// Copyright (C) 2010-2018 Joel Rosdahl
+// Copyright (C) 2010-2019 Joel Rosdahl
 //
 // This program is free software; you can redistribute it and/or modify it
 // under the terms of the GNU General Public License as published by the Free
@@ -149,6 +149,13 @@ TEST(check_for_temporal_macros)
 		"#define ab __DATE__";
 
 	const char no_temporal[] =
+		"#define ab a__DATE__\n"
+		"#define ab  __DATE__a\n"
+		"#define ab A__DATE__\n"
+		"#define ab  __DATE__A\n"
+		"#define ab 0__DATE__\n"
+		"#define ab  __DATE__0\n"
+		"#define ab _ _DATE__\n"
 		"#define ab _ _DATE__\n"
 		"#define ab __ DATE__\n"
 		"#define ab __D ATE__\n"