Attachment #6159 for bug #36162





		if (c1 == INVALID_CODEPOINT ||
		    c2 == INVALID_CODEPOINT) {
			/*
			 * Fall back to byte
			 * comparison. We must
			 * step back by the codepoint
			 * length we just incremented
			 * - otherwise we are not
			 * checking the bytes that
			 * failed the conversion.
			 */
			s1 -= size1;
			s2 -= size2;
			return strcasecmp(s1, s2);
		}

- 
in the face of bad conversions.
--
lib/util/charset/util_str.c | 29 +++++++++++++++++++++++++++--
1 file changed, 27 insertions(+), 2 deletions(-)





		if (c1 == INVALID_CODEPOINT ||
		    c2 == INVALID_CODEPOINT) {
			/*
			 * Fall back to byte
			 * comparison. We must
			 * step back by the codepoint
			 * length we just incremented
			 * by - otherwise we are not
			 * checking the bytes that
			 * failed the conversion.
			 */
			s1 -= size1;
			s2 -= size2;
			/*
			 * n was specified in characters,
			 * now we must convert it to bytes.
			 * As bytes are the smallest
			 * character unit, the following
			 * increment and strncasecmp is always
			 * safe.
			 *
			 * The source string was already known
			 * to be n characters long, so we are
			 * guaranteed to be able to look at the
			 * (n remaining + size1) bytes from the
			 * new (s1 - size1) position).
			 */
			n += size1;
			return strncasecmp(s1, s2, n);
		}

		if (toupper_m(c1) != toupper_m(c2)) {
- 
constantly crashes when filename contains non-ascii character
--
lib/util/charset/tests/charset.c | 12 ++++++++++++
1 file changed, 12 insertions(+)

Lines 50-61 static bool test_codepoint_cmpi(struct torture_context *tctx) Link Here

(-)a/lib/util/charset/tests/charset.c (-2 / +12 lines)
50		50
51	static bool test_strcasecmp_m(struct torture_context *tctx)	51	static bool test_strcasecmp_m(struct torture_context *tctx)
52	{	52	{
		53	/* file.{accented e} in iso8859-1 */
		54	const char file_iso8859_1[7] = { 0x66, 0x69, 0x6c, 0x65, 0x2d, 0xe9, 0 };
		55	/* file.{accented e} in utf8 */
		56	const char file_utf8[8] = { 0x66, 0x69, 0x6c, 0x65, 0x2d, 0xc3, 0xa9, 0 };
53	torture_assert(tctx, strcasecmp_m("foo", "bar") != 0, "different strings");	57	torture_assert(tctx, strcasecmp_m("foo", "bar") != 0, "different strings");
54	torture_assert(tctx, strcasecmp_m("foo", "foo") == 0, "same case strings");	58	torture_assert(tctx, strcasecmp_m("foo", "foo") == 0, "same case strings");
55	torture_assert(tctx, strcasecmp_m("foo", "Foo") == 0, "different case strings");	59	torture_assert(tctx, strcasecmp_m("foo", "Foo") == 0, "different case strings");
56	torture_assert(tctx, strcasecmp_m(NULL, "Foo") != 0, "one NULL");	60	torture_assert(tctx, strcasecmp_m(NULL, "Foo") != 0, "one NULL");
57	torture_assert(tctx, strcasecmp_m("foo", NULL) != 0, "other NULL");	61	torture_assert(tctx, strcasecmp_m("foo", NULL) != 0, "other NULL");
58	torture_assert(tctx, strcasecmp_m(NULL, NULL) == 0, "both NULL");	62	torture_assert(tctx, strcasecmp_m(NULL, NULL) == 0, "both NULL");
		63	torture_assert(tctx, strcasecmp_m(file_iso8859_1, file_utf8) != 0,
		64	"file.{accented e} should differ");
59	return true;	65	return true;
60	}	66	}
61		67
Lines 102-107 static bool test_string_replace_m(struct torture_context *tctx) Link Here
102		108
103	static bool test_strncasecmp_m(struct torture_context *tctx)	109	static bool test_strncasecmp_m(struct torture_context *tctx)
104	{	110	{
		111	/* file.{accented e} in iso8859-1 */
		112	const char file_iso8859_1[7] = { 0x66, 0x69, 0x6c, 0x65, 0x2d, 0xe9, 0 };
		113	/* file.{accented e} in utf8 */
		114	const char file_utf8[8] = { 0x66, 0x69, 0x6c, 0x65, 0x2d, 0xc3, 0xa9, 0 };
105	torture_assert(tctx, strncasecmp_m("foo", "bar", 3) != 0, "different strings");	115	torture_assert(tctx, strncasecmp_m("foo", "bar", 3) != 0, "different strings");
106	torture_assert(tctx, strncasecmp_m("foo", "foo", 3) == 0, "same case strings");	116	torture_assert(tctx, strncasecmp_m("foo", "foo", 3) == 0, "same case strings");
107	torture_assert(tctx, strncasecmp_m("foo", "Foo", 3) == 0, "different case strings");	117	torture_assert(tctx, strncasecmp_m("foo", "Foo", 3) == 0, "different case strings");
Lines 111-116 static bool test_strncasecmp_m(struct torture_context *tctx) Link Here
111	torture_assert(tctx, strncasecmp_m(NULL, "Foo", 3) != 0, "one NULL");	121	torture_assert(tctx, strncasecmp_m(NULL, "Foo", 3) != 0, "one NULL");
112	torture_assert(tctx, strncasecmp_m("foo", NULL, 3) != 0, "other NULL");	122	torture_assert(tctx, strncasecmp_m("foo", NULL, 3) != 0, "other NULL");
113	torture_assert(tctx, strncasecmp_m(NULL, NULL, 3) == 0, "both NULL");	123	torture_assert(tctx, strncasecmp_m(NULL, NULL, 3) == 0, "both NULL");
		124	torture_assert(tctx, strncasecmp_m(file_iso8859_1, file_utf8, 6) != 0,
		125	"file.{accented e} should differ");
114	return true;	126	return true;
115	}	127	}
116		128
117	-
118	--
119	lib/util/charset/util_str.c \| 48 ++++++++++++---------------------------------	129	lib/util/charset/util_str.c \| 48 ++++++++++++---------------------------------
120	1 file changed, 13 insertions(+), 35 deletions(-)	130	1 file changed, 13 insertions(+), 35 deletions(-)




		c1 = next_codepoint_handle(iconv_handle, s1, &size1);
		c2 = next_codepoint_handle(iconv_handle, s2, &size2);

		if (c1 == INVALID_CODEPOINT ||
		    c2 == INVALID_CODEPOINT) {
			return strcasecmp(s1, s2);
		}

		s1 += size1;
		s2 += size2;


			continue;
		}

		if (c1 == INVALID_CODEPOINT ||
		    c2 == INVALID_CODEPOINT) {
			/*
			 * Fall back to byte
			 * comparison. We must
			 * step back by the codepoint
			 * length we just incremented
			 * - otherwise we are not
			 * checking the bytes that
			 * failed the conversion.
			 */
			s1 -= size1;
			s2 -= size2;
			return strcasecmp(s1, s2);
		}

		if (toupper_m(c1) != toupper_m(c2)) {
			return c1 - c2;
		}

		c1 = next_codepoint_handle(iconv_handle, s1, &size1);
		c2 = next_codepoint_handle(iconv_handle, s2, &size2);

		s1 += size1;
		s2 += size2;

		if (c1 == c2) {
			continue;
		}

		if (c1 == INVALID_CODEPOINT ||
		    c2 == INVALID_CODEPOINT) {
			/*
			 * Fall back to byte
			 * comparison. We must
			 * step back by the codepoint
			 * length we just incremented
			 * by - otherwise we are not
			 * checking the bytes that
			 * failed the conversion.
			 */
			s1 -= size1;
			s2 -= size2;
			/*
			 * n was specified in characters,
			 * now we must convert it to bytes.
			 * As bytes are the smallest

			 * to be n characters long, so we are
			 * guaranteed to be able to look at the
			 * (n remaining + size1) bytes from the
			 * s1 position).
			 */
			n += size1;
			return strncasecmp(s1, s2, n);
		}

		s1 += size1;
		s2 += size2;

		if (c1 == c2) {
			continue;
		}

		if (toupper_m(c1) != toupper_m(c2)) {
			return c1 - c2;
		}
- 

Return to bug 36162

Lines 56-62 _PUBLIC_ int strcasecmp_m_handle(struct smb_iconv_handle *iconv_handle, Link Here

(-)a/lib/util/charset/util_str.c (-3 / +11 lines)
56		56
57	if (c1 == INVALID_CODEPOINT \|\|	57	if (c1 == INVALID_CODEPOINT \|\|
58	c2 == INVALID_CODEPOINT) {	58	c2 == INVALID_CODEPOINT) {
59	/* what else can we do?? */	59	/*
		60	* Fall back to byte
		61	* comparison. We must
		62	* step back by the codepoint
		63	* length we just incremented
		64	* - otherwise we are not
		65	* checking the bytes that
		66	* failed the conversion.
		67	*/
		68	s1 -= size1;
		69	s2 -= size2;
60	return strcasecmp(s1, s2);	70	return strcasecmp(s1, s2);
61	}	71	}
62		72
63	-
64	in the face of bad conversions.	73	in the face of bad conversions.
65	--
66	lib/util/charset/util_str.c \| 29 +++++++++++++++++++++++++++--	74	lib/util/charset/util_str.c \| 29 +++++++++++++++++++++++++++--
67	1 file changed, 27 insertions(+), 2 deletions(-)	75	1 file changed, 27 insertions(+), 2 deletions(-)

Lines 116-123 _PUBLIC_ int strncasecmp_m_handle(struct smb_iconv_handle *iconv_handle, Link Here

(-)a/lib/util/charset/util_str.c (-4 / +27 lines)
116		116
117	if (c1 == INVALID_CODEPOINT \|\|	117	if (c1 == INVALID_CODEPOINT \|\|
118	c2 == INVALID_CODEPOINT) {	118	c2 == INVALID_CODEPOINT) {
119	/* what else can we do?? */	119	/*
120	return strcasecmp(s1, s2);	120	* Fall back to byte
		121	* comparison. We must
		122	* step back by the codepoint
		123	* length we just incremented
		124	* by - otherwise we are not
		125	* checking the bytes that
		126	* failed the conversion.
		127	*/
		128	s1 -= size1;
		129	s2 -= size2;
		130	/*
		131	* n was specified in characters,
		132	* now we must convert it to bytes.
		133	* As bytes are the smallest
		134	* character unit, the following
		135	* increment and strncasecmp is always
		136	* safe.
		137	*
		138	* The source string was already known
		139	* to be n characters long, so we are
		140	* guaranteed to be able to look at the
		141	* (n remaining + size1) bytes from the
		142	* new (s1 - size1) position).
		143	*/
		144	n += size1;
		145	return strncasecmp(s1, s2, n);
121	}	146	}
122		147
123	if (toupper_m(c1) != toupper_m(c2)) {	148	if (toupper_m(c1) != toupper_m(c2)) {
124	-
125	constantly crashes when filename contains non-ascii character	149	constantly crashes when filename contains non-ascii character
126	--
127	lib/util/charset/tests/charset.c \| 12 ++++++++++++	150	lib/util/charset/tests/charset.c \| 12 ++++++++++++
128	1 file changed, 12 insertions(+)	151	1 file changed, 12 insertions(+)

Lines 47-52 _PUBLIC_ int strcasecmp_m_handle(struct smb_iconv_handle *iconv_handle, Link Here

(-)a/lib/util/charset/util_str.c (-36 / +13 lines)
47	c1 = next_codepoint_handle(iconv_handle, s1, &size1);	47	c1 = next_codepoint_handle(iconv_handle, s1, &size1);
48	c2 = next_codepoint_handle(iconv_handle, s2, &size2);	48	c2 = next_codepoint_handle(iconv_handle, s2, &size2);
49		49
		50	if (c1 == INVALID_CODEPOINT \|\|
		51	c2 == INVALID_CODEPOINT) {
		52	return strcasecmp(s1, s2);
		53	}
		54
50	s1 += size1;	55	s1 += size1;
51	s2 += size2;	56	s2 += size2;
52		57
Lines 54-75 _PUBLIC_ int strcasecmp_m_handle(struct smb_iconv_handle *iconv_handle, Link Here
54	continue;	59	continue;
55	}	60	}
56		61
57	if (c1 == INVALID_CODEPOINT \|\|
58	c2 == INVALID_CODEPOINT) {
59	/*
60	* Fall back to byte
61	* comparison. We must
62	* step back by the codepoint
63	* length we just incremented
64	* - otherwise we are not
65	* checking the bytes that
66	* failed the conversion.
67	*/
68	s1 -= size1;
69	s2 -= size2;
70	return strcasecmp(s1, s2);
71	}
72
73	if (toupper_m(c1) != toupper_m(c2)) {	62	if (toupper_m(c1) != toupper_m(c2)) {
74	return c1 - c2;	63	return c1 - c2;
75	}	64	}
Lines 107-133 _PUBLIC_ int strncasecmp_m_handle(struct smb_iconv_handle *iconv_handle, Link Here
107	c1 = next_codepoint_handle(iconv_handle, s1, &size1);	96	c1 = next_codepoint_handle(iconv_handle, s1, &size1);
108	c2 = next_codepoint_handle(iconv_handle, s2, &size2);	97	c2 = next_codepoint_handle(iconv_handle, s2, &size2);
109		98
110	s1 += size1;
111	s2 += size2;
112
113	if (c1 == c2) {
114	continue;
115	}
116
117	if (c1 == INVALID_CODEPOINT \|\|	99	if (c1 == INVALID_CODEPOINT \|\|
118	c2 == INVALID_CODEPOINT) {	100	c2 == INVALID_CODEPOINT) {
119	/*	101	/*
120	* Fall back to byte
121	* comparison. We must
122	* step back by the codepoint
123	* length we just incremented
124	* by - otherwise we are not
125	* checking the bytes that
126	* failed the conversion.
127	*/
128	s1 -= size1;
129	s2 -= size2;
130	/*
131	* n was specified in characters,	102	* n was specified in characters,
132	* now we must convert it to bytes.	103	* now we must convert it to bytes.
133	* As bytes are the smallest	104	* As bytes are the smallest
Lines 139-150 _PUBLIC_ int strncasecmp_m_handle(struct smb_iconv_handle *iconv_handle, Link Here
139	* to be n characters long, so we are	110	* to be n characters long, so we are
140	* guaranteed to be able to look at the	111	* guaranteed to be able to look at the
141	* (n remaining + size1) bytes from the	112	* (n remaining + size1) bytes from the
142	* new (s1 - size1) position).	113	* s1 position).
143	*/	114	*/
144	n += size1;	115	n += size1;
145	return strncasecmp(s1, s2, n);	116	return strncasecmp(s1, s2, n);
146	}	117	}
147		118
		119	s1 += size1;
		120	s2 += size2;
		121
		122	if (c1 == c2) {
		123	continue;
		124	}
		125
148	if (toupper_m(c1) != toupper_m(c2)) {	126	if (toupper_m(c1) != toupper_m(c2)) {
149	return c1 - c2;	127	return c1 - c2;
150	}	128	}
151	-