X-Git-Url: https://git.proxmox.com/?a=blobdiff_plain;f=MdeModulePkg%2FUniversal%2FRegularExpressionDxe%2FOniguruma%2Futf16_le.c;fp=MdeModulePkg%2FUniversal%2FRegularExpressionDxe%2FOniguruma%2Futf16_le.c;h=535713852f7d3a72a1c278b8118d3c7e9b4a9158;hb=b26691c47188ce255b8a4d920bf07ddf1431e2cd;hp=b139f83ff012ab7a739c1e226c35b6f6d301470f;hpb=ecc32c90ee4ad557205cb2725619a3cc2f45ebd0;p=mirror_edk2.git

diff --git a/MdeModulePkg/Universal/RegularExpressionDxe/Oniguruma/utf16_le.c b/MdeModulePkg/Universal/RegularExpressionDxe/Oniguruma/utf16_le.c
index b139f83ff0..535713852f 100644
--- a/MdeModulePkg/Universal/RegularExpressionDxe/Oniguruma/utf16_le.c
+++ b/MdeModulePkg/Universal/RegularExpressionDxe/Oniguruma/utf16_le.c
@@ -2,7 +2,7 @@
   utf16_le.c -  Oniguruma (regular expression library)
 **********************************************************************/
 /*-
- * Copyright (c) 2002-2018  K.Kosako  <sndgk393 AT ybb DOT ne DOT jp>
+ * Copyright (c) 2002-2019  K.Kosako  <sndgk393 AT ybb DOT ne DOT jp>
  * All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
@@ -95,7 +95,15 @@ static const int EncLen_UTF16[] = {
 static int
 utf16le_code_to_mbclen(OnigCodePoint code)
 {
-  return (code > 0xffff ? 4 : 2);
+  if (code > 0xffff) {
+    if (code > 0x10ffff)
+      return ONIGERR_INVALID_CODE_POINT_VALUE;
+    else
+      return 4;
+  }
+  else {
+    return 2;
+  }
 }
 
 static int
@@ -110,7 +118,16 @@ is_valid_mbc_string(const UChar* p, const UChar* end)
   const UChar* end1 = end - 1;
 
   while (p < end1) {
-    p += utf16le_mbc_enc_len(p);
+    int len = utf16le_mbc_enc_len(p);
+    if (len == 4) {
+      if (p + 3 < end && ! UTF16_IS_SURROGATE_SECOND(*(p + 3)))
+        return FALSE;
+    }
+    else
+      if (UTF16_IS_SURROGATE_SECOND(*(p + 1)))
+        return FALSE;
+
+    p += len;
   }
 
   if (p != end)
@@ -184,7 +201,7 @@ utf16le_code_to_mbc(OnigCodePoint code, UChar *buf)
 
 static int
 utf16le_mbc_case_fold(OnigCaseFoldType flag,
-		      const UChar** pp, const UChar* end, UChar* fold)
+                      const UChar** pp, const UChar* end, UChar* fold)
 {
   const UChar* p = *pp;
 
@@ -207,13 +224,13 @@ utf16le_mbc_case_fold(OnigCaseFoldType flag,
   }
   else
     return onigenc_unicode_mbc_case_fold(ONIG_ENCODING_UTF16_LE, flag, pp, end,
-					 fold);
+                                         fold);
 }
 
 #if 0
 static int
 utf16le_is_mbc_ambiguous(OnigCaseFoldType flag, const UChar** pp,
-			 const UChar* end)
+                         const UChar* end)
 {
   const UChar* p = *pp;
 
@@ -252,7 +269,8 @@ utf16le_left_adjust_char_head(const UChar* start, const UChar* s)
     s--;
   }
 
-  if (UTF16_IS_SURROGATE_SECOND(*(s+1)) && s > start + 1)
+  if (UTF16_IS_SURROGATE_SECOND(*(s+1)) && s > start + 1 &&
+      UTF16_IS_SURROGATE_FIRST(*(s-1)))
     s -= 2;
 
   return (UChar* )s;
@@ -263,7 +281,7 @@ utf16le_get_case_fold_codes_by_str(OnigCaseFoldType flag,
     const OnigUChar* p, const OnigUChar* end, OnigCaseFoldCodeItem items[])
 {
   return onigenc_unicode_get_case_fold_codes_by_str(ONIG_ENCODING_UTF16_LE,
-						    flag, p, end, items);
+                                                    flag, p, end, items);
 }
 
 OnigEncodingType OnigEncodingUTF16_LE = {
@@ -286,6 +304,6 @@ OnigEncodingType OnigEncodingUTF16_LE = {
   init,
   0, /* is_initialized */
   is_valid_mbc_string,
-  ENC_FLAG_UNICODE,
+  ENC_FLAG_UNICODE|ENC_FLAG_SKIP_OFFSET_1,
   0, 0
 };