implement ONIGENC_CASE_FOLD_ASCII_ONLY

rivy-fix · Nov 28, 2020 · c9599e0 · c9599e0
1 parent b3cc984
commit c9599e0
Show file tree

Hide file tree

Showing 19 changed files with 208 additions and 101 deletions.
diff --git a/src/cp1251.c b/src/cp1251.c
@@ -105,12 +105,16 @@ static const unsigned short EncCP1251_CtypeTable[256] = {
 };
 
 static int
-cp1251_mbc_case_fold(OnigCaseFoldType flag ARG_UNUSED,
+cp1251_mbc_case_fold(OnigCaseFoldType flag,
              const UChar** pp, const UChar* end ARG_UNUSED, UChar* lower)
 {
   const UChar* p = *pp;
 
-  *lower = ENC_CP1251_TO_LOWER_CASE(*p);
+  if (CASE_FOLD_IS_NOT_ASCII_ONLY(flag) || ONIGENC_IS_ASCII_CODE(*p))
+    *lower = ENC_CP1251_TO_LOWER_CASE(*p);
+  else
+    *lower = *p;
+
   (*pp)++;
   return 1;
 }

diff --git a/src/iso8859_1.c b/src/iso8859_1.c
@@ -114,7 +114,7 @@ apply_all_case_fold(OnigCaseFoldType flag,
 }
 
 static int
-get_case_fold_codes_by_str(OnigCaseFoldType flag ARG_UNUSED,
+get_case_fold_codes_by_str(OnigCaseFoldType flag,
                            const OnigUChar* p, const OnigUChar* end,
                            OnigCaseFoldCodeItem items[])
 {
@@ -123,7 +123,8 @@ get_case_fold_codes_by_str(OnigCaseFoldType flag ARG_UNUSED,
 
   if (0x41 <= *p && *p <= 0x5a) {
     if (*p == LARGE_S && end > p + 1
-        && (*(p+1) == LARGE_S || *(p+1) == SMALL_S)) { /* SS */
+        && (*(p+1) == LARGE_S || *(p+1) == SMALL_S)
+        && CASE_FOLD_IS_NOT_ASCII_ONLY(flag)) { /* SS */
     ss_combination:
       items[0].byte_len = 2;
       items[0].code_len = 1;
@@ -152,7 +153,8 @@ get_case_fold_codes_by_str(OnigCaseFoldType flag ARG_UNUSED,
   }
   else if (0x61 <= *p && *p <= 0x7a) {
     if (*p == SMALL_S && end > p + 1
-        && (*(p+1) == SMALL_S || *(p+1) == LARGE_S)) { /* ss */
+        && (*(p+1) == SMALL_S || *(p+1) == LARGE_S)
+        && CASE_FOLD_IS_NOT_ASCII_ONLY(flag)) { /* ss */
       goto ss_combination;
     }
 
@@ -161,56 +163,58 @@ get_case_fold_codes_by_str(OnigCaseFoldType flag ARG_UNUSED,
     items[0].code[0] = (OnigCodePoint )(*p - 0x20);
     return 1;
   }
-  else if (0xc0 <= *p && *p <= 0xcf) {
-    items[0].byte_len = 1;
-    items[0].code_len = 1;
-    items[0].code[0] = (OnigCodePoint )(*p + 0x20);
-    return 1;
-  }
-  else if (0xd0 <= *p && *p <= 0xdf) {
-    if (*p == 0xdf) {
+  else if (CASE_FOLD_IS_NOT_ASCII_ONLY(flag)) {
+    if (0xc0 <= *p && *p <= 0xcf) {
       items[0].byte_len = 1;
-      items[0].code_len = 2;
-      items[0].code[0] = (OnigCodePoint )'s';
-      items[0].code[1] = (OnigCodePoint )'s';
+      items[0].code_len = 1;
+      items[0].code[0] = (OnigCodePoint )(*p + 0x20);
+      return 1;
+    }
+    else if (0xd0 <= *p && *p <= 0xdf) {
+      if (*p == 0xdf) {
+        items[0].byte_len = 1;
+        items[0].code_len = 2;
+        items[0].code[0] = (OnigCodePoint )'s';
+        items[0].code[1] = (OnigCodePoint )'s';
 
-      items[1].byte_len = 1;
-      items[1].code_len = 2;
-      items[1].code[0] = (OnigCodePoint )'S';
-      items[1].code[1] = (OnigCodePoint )'S';
+        items[1].byte_len = 1;
+        items[1].code_len = 2;
+        items[1].code[0] = (OnigCodePoint )'S';
+        items[1].code[1] = (OnigCodePoint )'S';
 
-      items[2].byte_len = 1;
-      items[2].code_len = 2;
-      items[2].code[0] = (OnigCodePoint )'s';
-      items[2].code[1] = (OnigCodePoint )'S';
+        items[2].byte_len = 1;
+        items[2].code_len = 2;
+        items[2].code[0] = (OnigCodePoint )'s';
+        items[2].code[1] = (OnigCodePoint )'S';
 
-      items[3].byte_len = 1;
-      items[3].code_len = 2;
-      items[3].code[0] = (OnigCodePoint )'S';
-      items[3].code[1] = (OnigCodePoint )'s';
+        items[3].byte_len = 1;
+        items[3].code_len = 2;
+        items[3].code[0] = (OnigCodePoint )'S';
+        items[3].code[1] = (OnigCodePoint )'s';
 
-      return 4;
-    }
-    else if (*p != 0xd7) {
-      items[0].byte_len = 1;
-      items[0].code_len = 1;
-      items[0].code[0] = (OnigCodePoint )(*p + 0x20);
-      return 1;
+        return 4;
+      }
+      else if (*p != 0xd7) {
+        items[0].byte_len = 1;
+        items[0].code_len = 1;
+        items[0].code[0] = (OnigCodePoint )(*p + 0x20);
+        return 1;
+      }
     }
-  }
-  else if (0xe0 <= *p && *p <= 0xef) {
-    items[0].byte_len = 1;
-    items[0].code_len = 1;
-    items[0].code[0] = (OnigCodePoint )(*p - 0x20);
-    return 1;
-  }
-  else if (0xf0 <= *p && *p <= 0xfe) {
-    if (*p != 0xf7) {
+    else if (0xe0 <= *p && *p <= 0xef) {
       items[0].byte_len = 1;
       items[0].code_len = 1;
       items[0].code[0] = (OnigCodePoint )(*p - 0x20);
       return 1;
     }
+    else if (0xf0 <= *p && *p <= 0xfe) {
+      if (*p != 0xf7) {
+        items[0].byte_len = 1;
+        items[0].code_len = 1;
+        items[0].code[0] = (OnigCodePoint )(*p - 0x20);
+        return 1;
+      }
+    }
   }
 
   return 0;
@@ -229,7 +233,11 @@ mbc_case_fold(OnigCaseFoldType flag, const UChar** pp,
     return 2;
   }
 
-  *lower = ONIGENC_ISO_8859_1_TO_LOWER_CASE(*p);
+  if (CASE_FOLD_IS_NOT_ASCII_ONLY(flag) || ONIGENC_IS_ASCII_CODE(*p))
+    *lower = ONIGENC_ISO_8859_1_TO_LOWER_CASE(*p);
+  else
+    *lower = *p;
+
   (*pp)++;
   return 1;
 }

diff --git a/src/iso8859_10.c b/src/iso8859_10.c
@@ -116,7 +116,11 @@ mbc_case_fold(OnigCaseFoldType flag,
     return 2;
   }
 
-  *lower = ENC_ISO_8859_10_TO_LOWER_CASE(*p);
+  if (CASE_FOLD_IS_NOT_ASCII_ONLY(flag) || ONIGENC_IS_ASCII_CODE(*p))
+    *lower = ENC_ISO_8859_10_TO_LOWER_CASE(*p);
+  else
+    *lower = *p;
+
   (*pp)++;
   return 1;
 }

diff --git a/src/iso8859_13.c b/src/iso8859_13.c
@@ -116,7 +116,11 @@ mbc_case_fold(OnigCaseFoldType flag,
     return 2;
   }
 
-  *lower = ENC_ISO_8859_13_TO_LOWER_CASE(*p);
+  if (CASE_FOLD_IS_NOT_ASCII_ONLY(flag) || ONIGENC_IS_ASCII_CODE(*p))
+    *lower = ENC_ISO_8859_13_TO_LOWER_CASE(*p);
+  else
+    *lower = *p;
+
   (*pp)++;
   return 1;
 }

diff --git a/src/iso8859_14.c b/src/iso8859_14.c
@@ -116,7 +116,11 @@ mbc_case_fold(OnigCaseFoldType flag,
     return 2;
   }
 
-  *lower = ENC_ISO_8859_14_TO_LOWER_CASE(*p);
+  if (CASE_FOLD_IS_NOT_ASCII_ONLY(flag) || ONIGENC_IS_ASCII_CODE(*p))
+    *lower = ENC_ISO_8859_14_TO_LOWER_CASE(*p);
+  else
+    *lower = *p;
+
   (*pp)++;
   return 1; /* return byte length of converted char to lower */
 }

diff --git a/src/iso8859_15.c b/src/iso8859_15.c
@@ -116,7 +116,11 @@ mbc_case_fold(OnigCaseFoldType flag,
     return 2;
   }
 
-  *lower = ENC_ISO_8859_15_TO_LOWER_CASE(*p);
+  if (CASE_FOLD_IS_NOT_ASCII_ONLY(flag) || ONIGENC_IS_ASCII_CODE(*p))
+    *lower = ENC_ISO_8859_15_TO_LOWER_CASE(*p);
+  else
+    *lower = *p;
+
   (*pp)++;
   return 1; /* return byte length of converted char to lower */
 }

diff --git a/src/iso8859_16.c b/src/iso8859_16.c
@@ -116,7 +116,11 @@ mbc_case_fold(OnigCaseFoldType flag,
     return 2;
   }
 
-  *lower = ENC_ISO_8859_16_TO_LOWER_CASE(*p);
+  if (CASE_FOLD_IS_NOT_ASCII_ONLY(flag) || ONIGENC_IS_ASCII_CODE(*p))
+    *lower = ENC_ISO_8859_16_TO_LOWER_CASE(*p);
+  else
+    *lower = *p;
+
   (*pp)++;
   return 1; /* return byte length of converted char to lower */
 }

diff --git a/src/iso8859_2.c b/src/iso8859_2.c
@@ -116,7 +116,11 @@ mbc_case_fold(OnigCaseFoldType flag,
     return 2;
   }
 
-  *lower = ENC_ISO_8859_2_TO_LOWER_CASE(*p);
+  if (CASE_FOLD_IS_NOT_ASCII_ONLY(flag) || ONIGENC_IS_ASCII_CODE(*p))
+    *lower = ENC_ISO_8859_2_TO_LOWER_CASE(*p);
+  else
+    *lower = *p;
+
   (*pp)++;
   return 1; /* return byte length of converted char to lower */
 }

diff --git a/src/iso8859_3.c b/src/iso8859_3.c
@@ -116,7 +116,11 @@ mbc_case_fold(OnigCaseFoldType flag, const UChar** pp,
     return 2;
   }
 
-  *lower = ENC_ISO_8859_3_TO_LOWER_CASE(*p);
+  if (CASE_FOLD_IS_NOT_ASCII_ONLY(flag) || ONIGENC_IS_ASCII_CODE(*p))
+    *lower = ENC_ISO_8859_3_TO_LOWER_CASE(*p);
+  else
+    *lower = *p;
+
   (*pp)++;
   return 1;
 }

diff --git a/src/iso8859_4.c b/src/iso8859_4.c
@@ -116,7 +116,11 @@ mbc_case_fold(OnigCaseFoldType flag,
     return 2;
   }
 
-  *lower = ENC_ISO_8859_4_TO_LOWER_CASE(*p);
+  if (CASE_FOLD_IS_NOT_ASCII_ONLY(flag) || ONIGENC_IS_ASCII_CODE(*p))
+    *lower = ENC_ISO_8859_4_TO_LOWER_CASE(*p);
+  else
+    *lower = *p;
+
   (*pp)++;
   return 1; /* return byte length of converted char to lower */
 }

diff --git a/src/iso8859_5.c b/src/iso8859_5.c
@@ -104,12 +104,16 @@ static const unsigned short EncISO_8859_5_CtypeTable[256] = {
 };
 
 static int
-mbc_case_fold(OnigCaseFoldType flag ARG_UNUSED,
+mbc_case_fold(OnigCaseFoldType flag,
               const UChar** pp, const UChar* end ARG_UNUSED, UChar* lower)
 {
   const UChar* p = *pp;
 
-  *lower = ENC_ISO_8859_5_TO_LOWER_CASE(*p);
+  if (CASE_FOLD_IS_NOT_ASCII_ONLY(flag) || ONIGENC_IS_ASCII_CODE(*p))
+    *lower = ENC_ISO_8859_5_TO_LOWER_CASE(*p);
+  else
+    *lower = *p;
+
   (*pp)++;
   return 1;
 }

diff --git a/src/iso8859_7.c b/src/iso8859_7.c
@@ -104,12 +104,16 @@ static const unsigned short EncISO_8859_7_CtypeTable[256] = {
 };
 
 static int
-mbc_case_fold(OnigCaseFoldType flag ARG_UNUSED,
+mbc_case_fold(OnigCaseFoldType flag,
               const UChar** pp, const UChar* end ARG_UNUSED, UChar* lower)
 {
   const UChar* p = *pp;
 
-  *lower = ENC_ISO_8859_7_TO_LOWER_CASE(*p);
+  if (CASE_FOLD_IS_NOT_ASCII_ONLY(flag) || ONIGENC_IS_ASCII_CODE(*p))
+    *lower = ENC_ISO_8859_7_TO_LOWER_CASE(*p);
+  else
+    *lower = *p;
+
   (*pp)++;
   return 1;
 }

diff --git a/src/iso8859_9.c b/src/iso8859_9.c
@@ -116,7 +116,11 @@ mbc_case_fold(OnigCaseFoldType flag,
     return 2;
   }
 
-  *lower = ENC_ISO_8859_9_TO_LOWER_CASE(*p);
+  if (CASE_FOLD_IS_NOT_ASCII_ONLY(flag) || ONIGENC_IS_ASCII_CODE(*p))
+    *lower = ENC_ISO_8859_9_TO_LOWER_CASE(*p);
+  else
+    *lower = *p;
+
   (*pp)++;
   return 1;
 }

diff --git a/src/koi8.c b/src/koi8.c
@@ -105,12 +105,16 @@ static const unsigned short EncKOI8_CtypeTable[256] = {
 
 
 static int
-koi8_mbc_case_fold(OnigCaseFoldType flag ARG_UNUSED,
+koi8_mbc_case_fold(OnigCaseFoldType flag,
                    const UChar** pp, const UChar* end ARG_UNUSED, UChar* lower)
 {
   const UChar* p = *pp;
 
-  *lower = ENC_KOI8_TO_LOWER_CASE(*p);
+  if (CASE_FOLD_IS_NOT_ASCII_ONLY(flag) || ONIGENC_IS_ASCII_CODE(*p))
+    *lower = ENC_KOI8_TO_LOWER_CASE(*p);
+  else
+    *lower = *p;
+
   (*pp)++;
   return 1;
 }

diff --git a/src/koi8_r.c b/src/koi8_r.c
@@ -109,7 +109,11 @@ koi8_r_mbc_case_fold(OnigCaseFoldType flag ARG_UNUSED,
 {
   const UChar* p = *pp;
 
-  *lower = ENC_KOI8_R_TO_LOWER_CASE(*p);
+  if (CASE_FOLD_IS_NOT_ASCII_ONLY(flag) || ONIGENC_IS_ASCII_CODE(*p))
+    *lower = ENC_KOI8_R_TO_LOWER_CASE(*p);
+  else
+    *lower = *p;
+
   (*pp)++;
   return 1;
 }

diff --git a/src/regcomp.c b/src/regcomp.c
@@ -771,7 +771,8 @@ node_char_len1(Node* node, regex_t* reg, MinMaxCharLen* ci, ScanEnv* env,
       StrNode* sn = STR_(node);
       UChar *s = sn->s;
 
-      if (NODE_IS_REAL_IGNORECASE(node)) {
+      if (NODE_IS_REAL_IGNORECASE(node) &&
+          CASE_FOLD_IS_NOT_ASCII_ONLY(env->case_fold_flag)) {
         /* Such a case is possible.
            ex. /(?i)(?<=\1)(a)/
            Backref node refer to capture group, but it doesn't tune yet.