Fix inifinite loop when a single byte newline is searched in JIT.

git-svn-id: svn://vcs.exim.org/pcre2/code/trunk@1258 6239d852-aaf2-0410-a92c-79f79f948069

3 files changed, 36 insertions, 14 deletions

diff --git a/ChangeLog b/ChangeLog
index 3cb3895..94ab843 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -9,6 +9,9 @@ compiler. This fixes https://bugs.exim.org/show_bug.cgi?id=2578. Patch for
 Makefile.am and configure.ac by H.J. Lu. Equivalent patch for CMakeLists.txt 
 invented by PH.
 
+2. Fix inifinite loop when a single byte newline is searched in JIT when
+invalid utf8 mode is enabled.
+
 
 Version 10.35 09-May-2020
 ---------------------------
diff --git a/src/pcre2_jit_compile.c b/src/pcre2_jit_compile.c
index 33ad7e6..4a3ddd8 100644
--- a/src/pcre2_jit_compile.c
+++ b/src/pcre2_jit_compile.c
@@ -4578,7 +4578,14 @@ if (common->nltype != NLTYPE_ANY)
   /* All newlines are ascii, just skip intermediate octets. */
   jump[0] = CMP(SLJIT_GREATER_EQUAL, STR_PTR, 0, STR_END, 0);
   loop = LABEL();
-  OP1(MOV_UCHAR, TMP2, 0, SLJIT_MEM1(STR_PTR), IN_UCHARS(0));
+  if (sljit_emit_mem(compiler, MOV_UCHAR | SLJIT_MEM_SUPP | SLJIT_MEM_POST, TMP2, SLJIT_MEM1(STR_PTR), IN_UCHARS(1)) == SLJIT_SUCCESS)
+    sljit_emit_mem(compiler, MOV_UCHAR | SLJIT_MEM_POST, TMP2, SLJIT_MEM1(STR_PTR), IN_UCHARS(1));
+  else
+    {
+    OP1(MOV_UCHAR, TMP2, 0, SLJIT_MEM1(STR_PTR), IN_UCHARS(0));
+    OP2(SLJIT_ADD, STR_PTR, 0, STR_PTR, 0, SLJIT_IMM, IN_UCHARS(1));
+    }
+
   OP2(SLJIT_AND, TMP2, 0, TMP2, 0, SLJIT_IMM, 0xc0);
   CMPTO(SLJIT_EQUAL, TMP2, 0, SLJIT_IMM, 0x80, loop);
   OP2(SLJIT_SUB, STR_PTR, 0, STR_PTR, 0, SLJIT_IMM, IN_UCHARS(1));
diff --git a/src/pcre2_jit_test.c b/src/pcre2_jit_test.c
index a29fffa..16dade7 100644
--- a/src/pcre2_jit_test.c
+++ b/src/pcre2_jit_test.c
@@ -1831,7 +1831,9 @@ struct invalid_utf8_regression_test_case {
 	const char *input;
 };
 
-static struct invalid_utf8_regression_test_case invalid_utf8_regression_test_cases[] = {
+static const char invalid_utf8_newline_cr;
+
+static const struct invalid_utf8_regression_test_case invalid_utf8_regression_test_cases[] = {
 	{ UDA, CI, 0, 0, 0, 0, 4, { ".", NULL }, "\xf4\x8f\xbf\xbf" },
 	{ UDA, CI, 0, 0, 0, 0, 4, { ".", NULL }, "\xf0\x90\x80\x80" },
 	{ UDA, CI, 0, 0, 0, -1, -1, { ".", NULL }, "\xf4\x90\x80\x80" },
@@ -1974,6 +1976,8 @@ static struct invalid_utf8_regression_test_case invalid_utf8_regression_test_cas
 	{ 0, PCRE2_JIT_COMPLETE, 0, 0, 1, -1, -1, { "\\X{2}", NULL }, "\r\n\n" },
 	{ 0, PCRE2_JIT_COMPLETE, 0, 0, 1, -1, -1, { "\\R{2}", NULL }, "\r\n\n" },
 
+	{ PCRE2_UTF | PCRE2_MULTILINE, CI, 0, 0, 0, -1, -1, { "^.a", &invalid_utf8_newline_cr }, "\xc3\xa7#a" },
+
 	{ 0, 0, 0, 0, 0, 0, 0, { NULL, NULL }, NULL }
 };
 
@@ -1981,7 +1985,7 @@ static struct invalid_utf8_regression_test_case invalid_utf8_regression_test_cas
 #undef CI
 #undef CPI
 
-static int run_invalid_utf8_test(struct invalid_utf8_regression_test_case *current,
+static int run_invalid_utf8_test(const struct invalid_utf8_regression_test_case *current,
 	int pattern_index, int i, pcre2_compile_context_8 *ccontext, pcre2_match_data_8 *mdata)
 {
 	pcre2_code_8 *code;
@@ -2034,7 +2038,7 @@ static int run_invalid_utf8_test(struct invalid_utf8_regression_test_case *curre
 
 static int invalid_utf8_regression_tests(void)
 {
-	struct invalid_utf8_regression_test_case *current;
+	const struct invalid_utf8_regression_test_case *current;
 	pcre2_compile_context_8 *ccontext;
 	pcre2_match_data_8 *mdata;
 	int total = 0, successful = 0;
@@ -2051,10 +2055,18 @@ static int invalid_utf8_regression_tests(void)
 		total++;
 
 		result = 1;
-		if (!run_invalid_utf8_test(current, total - 1, 0, ccontext, mdata))
-			result = 0;
-		if (!run_invalid_utf8_test(current, total - 1, 1, ccontext, mdata))
-			result = 0;
+		if (current->pattern[1] != &invalid_utf8_newline_cr)
+		{
+			if (!run_invalid_utf8_test(current, total - 1, 0, ccontext, mdata))
+				result = 0;
+			if (!run_invalid_utf8_test(current, total - 1, 1, ccontext, mdata))
+				result = 0;
+		} else {
+			pcre2_set_newline_8(ccontext, PCRE2_NEWLINE_CR);
+			if (!run_invalid_utf8_test(current, total - 1, 0, ccontext, mdata))
+				result = 0;
+			pcre2_set_newline_8(ccontext, PCRE2_NEWLINE_ANY);
+		}
 
 		if (result) {
 			successful++;
@@ -2128,7 +2140,7 @@ static PCRE2_UCHAR16 test16_10[] = { ' ', 0xdc00, 0xd800, 0x2028, '#', 0 };
 static PCRE2_UCHAR16 test16_11[] = { 0xdc00, 0xdc00, 0xd800, 0xdc00, 0xdc00, '#', 0xd800, 0xdc00, '#', 0 };
 static PCRE2_UCHAR16 test16_12[] = { '#', 0xd800, 0xdc00, 0xd800, '#', 0xd800, 0xdc00, 0xdc00, 0xdc00, '#', 0xd800, 0xdc00, '#', 0 };
 
-static struct invalid_utf16_regression_test_case invalid_utf16_regression_test_cases[] = {
+static const struct invalid_utf16_regression_test_case invalid_utf16_regression_test_cases[] = {
 	{ UDA, CI, 0, 0, 0, 0, 1, { allany16, NULL }, test16_1 },
 	{ UDA, CI, 1, 0, 0, 1, 2, { allany16, NULL }, test16_1 },
 	{ UDA, CI, 2, 0, 0, 2, 3, { allany16, NULL }, test16_1 },
@@ -2182,7 +2194,7 @@ static struct invalid_utf16_regression_test_case invalid_utf16_regression_test_c
 #undef CI
 #undef CPI
 
-static int run_invalid_utf16_test(struct invalid_utf16_regression_test_case *current,
+static int run_invalid_utf16_test(const struct invalid_utf16_regression_test_case *current,
 	int pattern_index, int i, pcre2_compile_context_16 *ccontext, pcre2_match_data_16 *mdata)
 {
 	pcre2_code_16 *code;
@@ -2242,7 +2254,7 @@ static int run_invalid_utf16_test(struct invalid_utf16_regression_test_case *cur
 
 static int invalid_utf16_regression_tests(void)
 {
-	struct invalid_utf16_regression_test_case *current;
+	const struct invalid_utf16_regression_test_case *current;
 	pcre2_compile_context_16 *ccontext;
 	pcre2_match_data_16 *mdata;
 	int total = 0, successful = 0;
@@ -2329,7 +2341,7 @@ static PCRE2_UCHAR32 test32_4[] = { '#', 0x10ffff, 0x110000, 0 };
 static PCRE2_UCHAR32 test32_5[] = { ' ', 0x2028, '#', 0 };
 static PCRE2_UCHAR32 test32_6[] = { ' ', 0x110000, 0x2028, '#', 0 };
 
-static struct invalid_utf32_regression_test_case invalid_utf32_regression_test_cases[] = {
+static const struct invalid_utf32_regression_test_case invalid_utf32_regression_test_cases[] = {
 	{ UDA, CI, 0, 0, 0, 0, 1, { allany32, NULL }, test32_1 },
 	{ UDA, CI, 2, 0, 0, -1, -1, { allany32, NULL }, test32_1 },
 	{ UDA, CI, 0, 0, 0, 0, 1, { allany32, NULL }, test32_2 },
@@ -2369,7 +2381,7 @@ static struct invalid_utf32_regression_test_case invalid_utf32_regression_test_c
 #undef CI
 #undef CPI
 
-static int run_invalid_utf32_test(struct invalid_utf32_regression_test_case *current,
+static int run_invalid_utf32_test(const struct invalid_utf32_regression_test_case *current,
 	int pattern_index, int i, pcre2_compile_context_32 *ccontext, pcre2_match_data_32 *mdata)
 {
 	pcre2_code_32 *code;
@@ -2429,7 +2441,7 @@ static int run_invalid_utf32_test(struct invalid_utf32_regression_test_case *cur
 
 static int invalid_utf32_regression_tests(void)
 {
-	struct invalid_utf32_regression_test_case *current;
+	const struct invalid_utf32_regression_test_case *current;
 	pcre2_compile_context_32 *ccontext;
 	pcre2_match_data_32 *mdata;
 	int total = 0, successful = 0;