unsigned int FAR PASCAL SJIS2UTF8(WORD KCode, int *byte, char *locale)
{
wchar_t wchar;
int ret;
unsigned int code = KCode;
unsigned int c, c1, c2, c3;
unsigned char *ptr, buf[3];
unsigned short cset;
int len = 0;
*byte = 2;
// CP932からUTF-16LEへ変換する
setlocale(LC_ALL, locale);
buf[0] = KCode >> 8;
if (buf[0] > 0) {
len++;
}
buf[len++] = KCode & 0xff;
ret = mbtowc(&wchar, buf, len);
if (ret <= 0) { // 変換失敗
cset = 0;
if (_stricmp(locale, DEFAULT_LOCALE) == 0) {
cset = ConvertUnicode(KCode, mapSJISToUnicode, sizeof(mapSJISToUnicode)/sizeof(mapSJISToUnicode[0]));
}
if (cset == 0) {
c = 0xfffd; // U+FFFD: Replacement Character
} else {
c = cset;
}
} else {
ptr = (unsigned char *)&wchar;
c = ((ptr[1] << 8) | ptr[0]);
}
// UTF-16LEからUTF-8へ変換する
if (0x00000000 <= c && c <= 0x0000007f) {
code = (c & 0xff);
*byte = 1;
} else if (0x00000080 <= c && c <= 0x000007ff) {
c1 = ((c >> 6) & 0x1f) | 0xc0;
c2 = (c & 0x3f) | 0x80;
code = (c1 << 8) | c2;
*byte = 2;
} else if (0x00000800 <= c && c <= 0x0000ffff) {
void ProcessNextChar(char c) {
if (c == '\n') {
state = jsBegin; // abandon current parse (if any) and start again
return;
}
switch (state) {
case jsBegin: // initial state, expecting '{'
if (c == '{') {
parser_watcher::StartReceivedMessage();
state = jsExpectId;
fieldVal.clear();
fieldId.clear();
arrayDepth = 0;
}
break;
case jsExpectId: // expecting a quoted ID
switch (c) {
case ' ': // ignore space
break;
case '"':
state = jsId;
break;
case '}': // empty object, or extra comma at end of field list
RemoveLastId();
if (fieldId.size() == 0) {
// TODO: should we report this to the monitor as an error?
parser_watcher::EndReceivedMessage();
state = jsBegin;
} else {
RemoveLastIdChar();
state = jsEndVal;
}
break;
default:
state = JsError();
break;
}
break;
case jsId: // expecting an identifier, or in the middle of one
switch (c) {
case '"':
state = jsHadId;
break;
default:
if (c < ' ') {
state = JsError();
} else if (c != ':' && c != '^') {
if (!fieldId.add(c)) {
state = JsError();
}
}
break;
}
break;
case jsHadId: // had a quoted identifier, expecting ':'
switch (c) {
case ':':
state = jsVal;
break;
case ' ':
break;
default:
state = JsError();
break;
}
break;
case jsVal: // had ':' or ':[', expecting value
switch (c) {
case ' ':
break;
case '"':
fieldVal.clear();
state = jsStringVal;
break;
case '[':
if (arrayDepth < MAX_ARRAY_NESTING && fieldId.add('^')) {
arrayIndices[arrayDepth] = 0; // start an array
++arrayDepth;
} else {
state = JsError();
}
break;
case ']':
if (InArray()) {
EndArray(); // empty array
state = jsEndVal;
} else {
state = JsError(); // ']' received without a matching '[' first
}
break;
case '-':
fieldVal.clear();
// TODO: an report error to the watcher
state = (fieldVal.add(c)) ? jsNegIntVal : JsError();
break;
case '{': // start of a nested object
// TODO: report an error to the watcher
state = (fieldId.add(':')) ? jsExpectId : JsError();
break;
default:
if (c >= '0' && c <= '9') {
fieldVal.clear();
fieldVal.add(c); // must succeed because we just cleared fieldVal
state = jsIntVal;
break;
} else {
state = JsError();
}
}
break;
case jsStringVal: // just had '"' and expecting a string value
switch (c) {
case '"':
ConvertUnicode();
ProcessField();
state = jsEndVal;
break;
case '\\':
state = jsStringEscape;
break;
default:
if (c < ' ') {
state = JsError();
} else {
fieldVal.add(c); // ignore any error so that long string parameters
// just get truncated
}
break;
}
break;
case jsStringEscape: // just had backslash in a string
if (!fieldVal.full()) {
switch (c) {
case '"':
case '\\':
case '/':
if (!fieldVal.add(c)) {
state = JsError();
}
break;
case 'n':
case 't':
if (!fieldVal.add(' ')) // replace newline and tab by space
{
// parser_watcher::ProcessError();
state = JsError();
}
break;
case 'b':
case 'f':
case 'r':
default:
break;
}
}
state = jsStringVal;
break;
case jsNegIntVal: // had '-' so expecting a integer value
// TODO: report an error to the watcher
state = (c >= '0' && c <= '9' && fieldVal.add(c)) ? jsIntVal : JsError();
break;
case jsIntVal: // receiving an integer value
switch (c) {
case '.':
state = (fieldVal.add(c)) ? jsFracVal : JsError();
break;
case ',':
ProcessField();
if (InArray()) {
++arrayIndices[arrayDepth - 1];
fieldVal.clear();
state = jsVal;
} else {
RemoveLastId();
state = jsExpectId;
}
break;
case ']':
if (InArray()) {
ProcessField();
++arrayIndices[arrayDepth - 1];
EndArray();
state = jsEndVal;
} else {
state = JsError();
}
break;
case '}':
if (InArray()) {
state = JsError();
} else {
ProcessField();
RemoveLastId();
if (fieldId.size() == 0) {
parser_watcher::EndReceivedMessage();
state = jsBegin;
} else {
RemoveLastIdChar();
state = jsEndVal;
}
}
break;
default:
if (!(c >= '0' && c <= '9' && fieldVal.add(c))) {
state = JsError();
}
break;
}
break;
case jsFracVal: // receiving a fractional value
switch (c) {
case ',':
ProcessField();
if (InArray()) {
++arrayIndices[arrayDepth - 1];
state = jsVal;
} else {
RemoveLastId();
state = jsExpectId;
}
break;
case ']':
if (InArray()) {
ProcessField();
++arrayIndices[arrayDepth - 1];
EndArray();
state = jsEndVal;
} else {
state = JsError();
}
break;
case '}':
if (InArray()) {
state = JsError();
} else {
ProcessField();
RemoveLastId();
if (fieldId.size() == 0) {
parser_watcher::EndReceivedMessage();
state = jsBegin;
} else {
RemoveLastIdChar();
state = jsEndVal;
}
}
break;
default:
if (!(c >= '0' && c <= '9' && fieldVal.add(c))) {
state = JsError();
}
break;
}
break;
case jsEndVal: // had the end of a string or array value, expecting comma
// or ] or }
switch (c) {
case ',':
if (InArray()) {
++arrayIndices[arrayDepth - 1];
fieldVal.clear();
state = jsVal;
} else {
RemoveLastId();
state = jsExpectId;
}
break;
case ']':
if (InArray()) {
++arrayIndices[arrayDepth - 1];
EndArray();
} else {
state = JsError();
}
break;
case '}':
if (InArray()) {
state = JsError();
} else {
RemoveLastId();
if (fieldId.size() == 0) {
parser_watcher::EndReceivedMessage();
state = jsBegin;
} else {
RemoveLastIdChar();
// state = jsEndVal; // not needed, state == jsEndVal already
}
}
break;
default:
break;
}
break;
case jsError:
// Ignore all characters. State will be reset to jsBegin at the start of
// this function when we receive a newline.
break;
}
}