duke@435: /*
iklam@4851:  * Copyright (c) 1997, 2013, Oracle and/or its affiliates. All rights reserved.
duke@435:  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
duke@435:  *
duke@435:  * This code is free software; you can redistribute it and/or modify it
duke@435:  * under the terms of the GNU General Public License version 2 only, as
duke@435:  * published by the Free Software Foundation.
duke@435:  *
duke@435:  * This code is distributed in the hope that it will be useful, but WITHOUT
duke@435:  * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
duke@435:  * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
duke@435:  * version 2 for more details (a copy is included in the LICENSE file that
duke@435:  * accompanied this code).
duke@435:  *
duke@435:  * You should have received a copy of the GNU General Public License version
duke@435:  * 2 along with this work; if not, write to the Free Software Foundation,
duke@435:  * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
duke@435:  *
trims@1907:  * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA
trims@1907:  * or visit www.oracle.com if you need additional information or have any
trims@1907:  * questions.
duke@435:  *
duke@435:  */
duke@435: 
stefank@2314: #include "precompiled.hpp"
stefank@2314: #include "utilities/utf8.hpp"
duke@435: 
duke@435: // Assume the utf8 string is in legal form and has been
duke@435: // checked in the class file parser/format checker.
duke@435: char* UTF8::next(const char* str, jchar* value) {
duke@435:   unsigned const char *ptr = (const unsigned char *)str;
duke@435:   unsigned char ch, ch2, ch3;
duke@435:   int length = -1;              /* bad length */
duke@435:   jchar result;
duke@435:   switch ((ch = ptr[0]) >> 4) {
duke@435:     default:
duke@435:     result = ch;
duke@435:     length = 1;
duke@435:     break;
duke@435: 
duke@435:   case 0x8: case 0x9: case 0xA: case 0xB: case 0xF:
duke@435:     /* Shouldn't happen. */
duke@435:     break;
duke@435: 
duke@435:   case 0xC: case 0xD:
duke@435:     /* 110xxxxx  10xxxxxx */
duke@435:     if (((ch2 = ptr[1]) & 0xC0) == 0x80) {
duke@435:       unsigned char high_five = ch & 0x1F;
duke@435:       unsigned char low_six = ch2 & 0x3F;
duke@435:       result = (high_five << 6) + low_six;
duke@435:       length = 2;
duke@435:       break;
duke@435:     }
duke@435:     break;
duke@435: 
duke@435:   case 0xE:
duke@435:     /* 1110xxxx 10xxxxxx 10xxxxxx */
duke@435:     if (((ch2 = ptr[1]) & 0xC0) == 0x80) {
duke@435:       if (((ch3 = ptr[2]) & 0xC0) == 0x80) {
duke@435:         unsigned char high_four = ch & 0x0f;
duke@435:         unsigned char mid_six = ch2 & 0x3f;
duke@435:         unsigned char low_six = ch3 & 0x3f;
duke@435:         result = (((high_four << 6) + mid_six) << 6) + low_six;
duke@435:         length = 3;
duke@435:       }
duke@435:     }
duke@435:     break;
duke@435:   } /* end of switch */
duke@435: 
duke@435:   if (length <= 0) {
duke@435:     *value = ptr[0];    /* default bad result; */
duke@435:     return (char*)(ptr + 1); // make progress somehow
duke@435:   }
duke@435: 
duke@435:   *value = result;
duke@435: 
duke@435:   // The assert is correct but the .class file is wrong
duke@435:   // assert(UNICODE::utf8_size(result) == length, "checking reverse computation");
duke@435:   return (char *)(ptr + length);
duke@435: }
duke@435: 
duke@435: char* UTF8::next_character(const char* str, jint* value) {
duke@435:   unsigned const char *ptr = (const unsigned char *)str;
duke@435:   /* See if it's legal supplementary character:
duke@435:      11101101 1010xxxx 10xxxxxx 11101101 1011xxxx 10xxxxxx */
duke@435:   if (is_supplementary_character(ptr)) {
duke@435:     *value = get_supplementary_character(ptr);
duke@435:     return (char *)(ptr + 6);
duke@435:   }
duke@435:   jchar result;
duke@435:   char* next_ch = next(str, &result);
duke@435:   *value = result;
duke@435:   return next_ch;
duke@435: }
duke@435: 
duke@435: // Count bytes of the form 10xxxxxx and deduct this count
duke@435: // from the total byte count.  The utf8 string must be in
duke@435: // legal form which has been verified in the format checker.
duke@435: int UTF8::unicode_length(const char* str, int len) {
duke@435:   int num_chars = len;
duke@435:   for (int i = 0; i < len; i++) {
duke@435:     if ((str[i] & 0xC0) == 0x80) {
duke@435:       --num_chars;
duke@435:     }
duke@435:   }
duke@435:   return num_chars;
duke@435: }
duke@435: 
duke@435: // Count bytes of the utf8 string except those in form
duke@435: // 10xxxxxx which only appear in multibyte characters.
duke@435: // The utf8 string must be in legal form and has been
duke@435: // verified in the format checker.
duke@435: int UTF8::unicode_length(const char* str) {
duke@435:   int num_chars = 0;
duke@435:   for (const char* p = str; *p; p++) {
duke@435:     if (((*p) & 0xC0) != 0x80) {
duke@435:       num_chars++;
duke@435:     }
duke@435:   }
duke@435:   return num_chars;
duke@435: }
duke@435: 
duke@435: // Writes a jchar a utf8 and returns the end
duke@435: static u_char* utf8_write(u_char* base, jchar ch) {
duke@435:   if ((ch != 0) && (ch <=0x7f)) {
duke@435:     base[0] = (u_char) ch;
duke@435:     return base + 1;
duke@435:   }
duke@435: 
duke@435:   if (ch <= 0x7FF) {
duke@435:     /* 11 bits or less. */
duke@435:     unsigned char high_five = ch >> 6;
duke@435:     unsigned char low_six = ch & 0x3F;
duke@435:     base[0] = high_five | 0xC0; /* 110xxxxx */
duke@435:     base[1] = low_six | 0x80;   /* 10xxxxxx */
duke@435:     return base + 2;
duke@435:   }
duke@435:   /* possibly full 16 bits. */
duke@435:   char high_four = ch >> 12;
duke@435:   char mid_six = (ch >> 6) & 0x3F;
duke@435:   char low_six = ch & 0x3f;
duke@435:   base[0] = high_four | 0xE0; /* 1110xxxx */
duke@435:   base[1] = mid_six | 0x80;   /* 10xxxxxx */
duke@435:   base[2] = low_six | 0x80;   /* 10xxxxxx */
duke@435:   return base + 3;
duke@435: }
duke@435: 
duke@435: void UTF8::convert_to_unicode(const char* utf8_str, jchar* unicode_str, int unicode_length) {
duke@435:   unsigned char ch;
minqi@4267:   const char *ptr = utf8_str;
duke@435:   int index = 0;
duke@435: 
duke@435:   /* ASCII case loop optimization */
duke@435:   for (; index < unicode_length; index++) {
duke@435:     if((ch = ptr[0]) > 0x7F) { break; }
duke@435:     unicode_str[index] = ch;
duke@435:     ptr = (const char *)(ptr + 1);
duke@435:   }
duke@435: 
duke@435:   for (; index < unicode_length; index++) {
duke@435:     ptr = UTF8::next(ptr, &unicode_str[index]);
duke@435:   }
duke@435: }
duke@435: 
minqi@4267: // returns the quoted ascii length of a 0-terminated utf8 string
minqi@4267: int UTF8::quoted_ascii_length(const char* utf8_str, int utf8_length) {
minqi@4267:   const char *ptr = utf8_str;
minqi@4267:   const char* end = ptr + utf8_length;
minqi@4267:   int result = 0;
minqi@4267:   while (ptr < end) {
minqi@4267:     jchar c;
minqi@4267:     ptr = UTF8::next(ptr, &c);
minqi@4267:     if (c >= 32 && c < 127) {
minqi@4267:       result++;
minqi@4267:     } else {
minqi@4267:       result += 6;
minqi@4267:     }
minqi@4267:   }
minqi@4267:   return result;
minqi@4267: }
minqi@4267: 
minqi@4267: // converts a utf8 string to quoted ascii
iklam@4851: void UTF8::as_quoted_ascii(const char* utf8_str, int utf8_length, char* buf, int buflen) {
minqi@4267:   const char *ptr = utf8_str;
iklam@4851:   const char *utf8_end = ptr + utf8_length;
minqi@4267:   char* p = buf;
minqi@4267:   char* end = buf + buflen;
iklam@4851:   while (ptr < utf8_end) {
minqi@4267:     jchar c;
minqi@4267:     ptr = UTF8::next(ptr, &c);
minqi@4267:     if (c >= 32 && c < 127) {
minqi@4267:       if (p + 1 >= end) break;      // string is truncated
minqi@4267:       *p++ = (char)c;
minqi@4267:     } else {
minqi@4267:       if (p + 6 >= end) break;      // string is truncated
minqi@4267:       sprintf(p, "\\u%04x", c);
minqi@4267:       p += 6;
minqi@4267:     }
minqi@4267:   }
iklam@4851:   assert(p < end, "sanity");
minqi@4267:   *p = '\0';
minqi@4267: }
minqi@4267: 
minqi@4267: 
minqi@4267: const char* UTF8::from_quoted_ascii(const char* quoted_ascii_str) {
minqi@4267:   const char *ptr = quoted_ascii_str;
minqi@4267:   char* result = NULL;
minqi@4267:   while (*ptr != '\0') {
minqi@4267:     char c = *ptr;
minqi@4267:     if (c < 32 || c >= 127) break;
minqi@4267:   }
minqi@4267:   if (*ptr == '\0') {
minqi@4267:     // nothing to do so return original string
minqi@4267:     return quoted_ascii_str;
minqi@4267:   }
minqi@4267:   // everything up to this point was ok.
minqi@4267:   int length = ptr - quoted_ascii_str;
minqi@4267:   char* buffer = NULL;
minqi@4267:   for (int round = 0; round < 2; round++) {
minqi@4267:     while (*ptr != '\0') {
minqi@4267:       if (*ptr != '\\') {
minqi@4267:         if (buffer != NULL) {
minqi@4267:           buffer[length] = *ptr;
minqi@4267:         }
minqi@4267:         length++;
minqi@4267:       } else {
minqi@4267:         switch (ptr[1]) {
minqi@4267:           case 'u': {
minqi@4267:             ptr += 2;
minqi@4267:             jchar value=0;
minqi@4267:             for (int i=0; i<4; i++) {
minqi@4267:               char c = *ptr++;
minqi@4267:               switch (c) {
minqi@4267:                 case '0': case '1': case '2': case '3': case '4':
minqi@4267:                 case '5': case '6': case '7': case '8': case '9':
minqi@4267:                   value = (value << 4) + c - '0';
minqi@4267:                   break;
minqi@4267:                 case 'a': case 'b': case 'c':
minqi@4267:                 case 'd': case 'e': case 'f':
minqi@4267:                   value = (value << 4) + 10 + c - 'a';
minqi@4267:                   break;
minqi@4267:                 case 'A': case 'B': case 'C':
minqi@4267:                 case 'D': case 'E': case 'F':
minqi@4267:                   value = (value << 4) + 10 + c - 'A';
minqi@4267:                   break;
minqi@4267:                 default:
minqi@4267:                   ShouldNotReachHere();
minqi@4267:               }
minqi@4267:             }
minqi@4267:             if (buffer == NULL) {
minqi@4267:               char utf8_buffer[4];
minqi@4267:               char* next = (char*)utf8_write((u_char*)utf8_buffer, value);
minqi@4267:               length += next - utf8_buffer;
minqi@4267:             } else {
minqi@4267:               char* next = (char*)utf8_write((u_char*)&buffer[length], value);
minqi@4267:               length += next - &buffer[length];
minqi@4267:             }
minqi@4267:             break;
minqi@4267:           }
minqi@4267:           case 't': if (buffer != NULL) buffer[length] = '\t'; ptr += 2; length++; break;
minqi@4267:           case 'n': if (buffer != NULL) buffer[length] = '\n'; ptr += 2; length++; break;
minqi@4267:           case 'r': if (buffer != NULL) buffer[length] = '\r'; ptr += 2; length++; break;
minqi@4267:           case 'f': if (buffer != NULL) buffer[length] = '\f'; ptr += 2; length++; break;
minqi@4267:           default:
minqi@4267:             ShouldNotReachHere();
minqi@4267:         }
minqi@4267:       }
minqi@4267:     }
minqi@4267:     if (round == 0) {
minqi@4267:       buffer = NEW_RESOURCE_ARRAY(char, length + 1);
minqi@4267:       ptr = quoted_ascii_str;
minqi@4267:     } else {
minqi@4267:       buffer[length] = '\0';
minqi@4267:     }
minqi@4267:   }
minqi@4267:   return buffer;
minqi@4267: }
minqi@4267: 
minqi@4267: 
duke@435: // Returns NULL if 'c' it not found. This only works as long
duke@435: // as 'c' is an ASCII character
coleenp@2497: const jbyte* UTF8::strrchr(const jbyte* base, int length, jbyte c) {
duke@435:   assert(length >= 0, "sanity check");
duke@435:   assert(c >= 0, "does not work for non-ASCII characters");
duke@435:   // Skip backwards in string until 'c' is found or end is reached
duke@435:   while(--length >= 0 && base[length] != c);
duke@435:   return (length < 0) ? NULL : &base[length];
duke@435: }
duke@435: 
coleenp@2497: bool UTF8::equal(const jbyte* base1, int length1, const jbyte* base2, int length2) {
duke@435:   // Length must be the same
duke@435:   if (length1 != length2) return false;
duke@435:   for (int i = 0; i < length1; i++) {
duke@435:     if (base1[i] != base2[i]) return false;
duke@435:   }
duke@435:   return true;
duke@435: }
duke@435: 
duke@435: bool UTF8::is_supplementary_character(const unsigned char* str) {
duke@435:   return ((str[0] & 0xFF) == 0xED) && ((str[1] & 0xF0) == 0xA0) && ((str[2] & 0xC0) == 0x80)
duke@435:       && ((str[3] & 0xFF) == 0xED) && ((str[4] & 0xF0) == 0xB0) && ((str[5] & 0xC0) == 0x80);
duke@435: }
duke@435: 
duke@435: jint UTF8::get_supplementary_character(const unsigned char* str) {
duke@435:   return 0x10000 + ((str[1] & 0x0f) << 16) + ((str[2] & 0x3f) << 10)
duke@435:                  + ((str[4] & 0x0f) << 6)  + (str[5] & 0x3f);
duke@435: }
duke@435: 
duke@435: 
duke@435: //-------------------------------------------------------------------------------------
duke@435: 
duke@435: 
duke@435: int UNICODE::utf8_size(jchar c) {
duke@435:   if ((0x0001 <= c) && (c <= 0x007F)) return 1;
duke@435:   if (c <= 0x07FF) return 2;
duke@435:   return 3;
duke@435: }
duke@435: 
duke@435: int UNICODE::utf8_length(jchar* base, int length) {
duke@435:   int result = 0;
duke@435:   for (int index = 0; index < length; index++) {
duke@435:     jchar c = base[index];
duke@435:     if ((0x0001 <= c) && (c <= 0x007F)) result += 1;
duke@435:     else if (c <= 0x07FF) result += 2;
duke@435:     else result += 3;
duke@435:   }
duke@435:   return result;
duke@435: }
duke@435: 
duke@435: char* UNICODE::as_utf8(jchar* base, int length) {
duke@435:   int utf8_len = utf8_length(base, length);
duke@435:   u_char* result = NEW_RESOURCE_ARRAY(u_char, utf8_len + 1);
duke@435:   u_char* p = result;
duke@435:   for (int index = 0; index < length; index++) {
duke@435:     p = utf8_write(p, base[index]);
duke@435:   }
duke@435:   *p = '\0';
duke@435:   assert(p == &result[utf8_len], "length prediction must be correct");
duke@435:   return (char*) result;
duke@435: }
duke@435: 
duke@435: char* UNICODE::as_utf8(jchar* base, int length, char* buf, int buflen) {
duke@435:   u_char* p = (u_char*)buf;
duke@435:   u_char* end = (u_char*)buf + buflen;
duke@435:   for (int index = 0; index < length; index++) {
duke@435:     jchar c = base[index];
duke@435:     if (p + utf8_size(c) >= end) break;      // string is truncated
duke@435:     p = utf8_write(p, base[index]);
duke@435:   }
duke@435:   *p = '\0';
duke@435:   return buf;
duke@435: }
duke@435: 
duke@435: void UNICODE::convert_to_utf8(const jchar* base, int length, char* utf8_buffer) {
duke@435:   for(int index = 0; index < length; index++) {
duke@435:     utf8_buffer = (char*)utf8_write((u_char*)utf8_buffer, base[index]);
duke@435:   }
duke@435:   *utf8_buffer = '\0';
duke@435: }
minqi@4267: 
minqi@4267: // returns the quoted ascii length of a unicode string
minqi@4267: int UNICODE::quoted_ascii_length(jchar* base, int length) {
minqi@4267:   int result = 0;
minqi@4267:   for (int i = 0; i < length; i++) {
minqi@4267:     jchar c = base[i];
minqi@4267:     if (c >= 32 && c < 127) {
minqi@4267:       result++;
minqi@4267:     } else {
minqi@4267:       result += 6;
minqi@4267:     }
minqi@4267:   }
minqi@4267:   return result;
minqi@4267: }
minqi@4267: 
minqi@4267: // converts a utf8 string to quoted ascii
minqi@4267: void UNICODE::as_quoted_ascii(const jchar* base, int length, char* buf, int buflen) {
minqi@4267:   char* p = buf;
minqi@4267:   char* end = buf + buflen;
minqi@4267:   for (int index = 0; index < length; index++) {
minqi@4267:     jchar c = base[index];
minqi@4267:     if (c >= 32 && c < 127) {
minqi@4267:       if (p + 1 >= end) break;      // string is truncated
minqi@4267:       *p++ = (char)c;
minqi@4267:     } else {
minqi@4267:       if (p + 6 >= end) break;      // string is truncated
minqi@4267:       sprintf(p, "\\u%04x", c);
minqi@4267:       p += 6;
minqi@4267:     }
minqi@4267:   }
minqi@4267:   *p = '\0';
minqi@4267: }