From 45dcb3de900b77583f4e9daa663004c55fad4794 Mon Sep 17 00:00:00 2001
From: Zoltan Herczeg <hzmester@freemail.hu>
Date: Wed, 22 Nov 2023 10:22:59 +0000
Subject: [PATCH] Fix \X matching in 32 bit mode without UTF in JIT

---
 src/pcre2_jit_compile.c  | 6 +++---
 testdata/testinput12     | 4 ++++
 testdata/testoutput12-16 | 9 +++++++++
 testdata/testoutput12-32 | 5 +++++
 4 files changed, 21 insertions(+), 3 deletions(-)

diff --git a/src/pcre2_jit_compile.c b/src/pcre2_jit_compile.c
index 510c392..8d64e1c 100644
--- a/src/pcre2_jit_compile.c
+++ b/src/pcre2_jit_compile.c
@@ -8718,7 +8718,7 @@ c = *cc++;
 
 #if PCRE2_CODE_UNIT_WIDTH == 32
 if (c >= 0x110000)
-  return NULL;
+  return cc;
 #endif /* PCRE2_CODE_UNIT_WIDTH == 32 */
 lgb = UCD_GRAPHBREAK(c);
 
@@ -8958,7 +8958,7 @@ switch(type)
 #else
   sljit_emit_icall(compiler, SLJIT_CALL, SLJIT_ARGS2(W, W, W), SLJIT_IMM,
     common->invalid_utf ? SLJIT_FUNC_ADDR(do_extuni_utf_invalid) : SLJIT_FUNC_ADDR(do_extuni_no_utf));
-  if (!common->utf || common->invalid_utf)
+  if (common->invalid_utf)
     add_jump(compiler, backtracks, CMP(SLJIT_EQUAL, SLJIT_RETURN_REG, 0, SLJIT_IMM, 0));
 #endif
 
@@ -12044,7 +12044,7 @@ switch(opcode)
     }
 
 #if defined SUPPORT_UNICODE && PCRE2_CODE_UNIT_WIDTH != 32
-  if (common->utf)
+  if (type == OP_EXTUNI || common->utf)
     {
     OP1(SLJIT_MOV, tmp_base, tmp_offset, STR_PTR, 0);
     detect_partial_match(common, &no_match);
diff --git a/testdata/testinput12 b/testdata/testinput12
index 5a2d8d2..a6678bb 100644
--- a/testdata/testinput12
+++ b/testdata/testinput12
@@ -569,4 +569,8 @@
 /\x{802a0000}*/
     \x{802a0000}\x{802a0000}
 
+# UTF matching without UTF, check invalid UTF characters
+/\X++/
+    a\x{110000}\x{ffffffff}
+
 # End of testinput12
diff --git a/testdata/testoutput12-16 b/testdata/testoutput12-16
index 9ac403e..f3b40a3 100644
--- a/testdata/testoutput12-16
+++ b/testdata/testoutput12-16
@@ -1814,4 +1814,13 @@ No match
 Failed: error 134 at offset 11: character code point value in \x{} or \o{} is too large
     \x{802a0000}\x{802a0000}
 
+# UTF matching without UTF, check invalid UTF characters
+/\X++/
+    a\x{110000}\x{ffffffff}
+** Character \x{110000} is greater than 0xffff and UTF-16 mode is not enabled.
+** Truncation will probably give the wrong result.
+** Character \x{ffffffff} is greater than 0xffff and UTF-16 mode is not enabled.
+** Truncation will probably give the wrong result.
+ 0: a\x00\x{ffff}
+
 # End of testinput12
diff --git a/testdata/testoutput12-32 b/testdata/testoutput12-32
index 9396305..dd42f86 100644
--- a/testdata/testoutput12-32
+++ b/testdata/testoutput12-32
@@ -1812,4 +1812,9 @@ No match
     \x{802a0000}\x{802a0000}
  0: \x{802a0000}\x{802a0000}
 
+# UTF matching without UTF, check invalid UTF characters
+/\X++/
+    a\x{110000}\x{ffffffff}
+ 0: a\x{110000}\x{ffffffff}
+
 # End of testinput12
-- 
2.33.0