//
// This is the lexer. Check the character at the current index, and put the found token in dtok and
// some raw date/time information in raw.
//
private static void Lex(
int dps, __DTString str, DateTimeToken dtok, DateTimeRawInfo raw, DateTimeResult result, ref DateTimeFormatInfo dtfi) {
int sep;
dtok.dtt = DTT_Unk; // Assume the token is unkown.
//
// Skip any white spaces.
//
if (!str.SkipWhiteSpaceComma()) {
//
// SkipWhiteSpaceComma() will return true when end of string is reached.
//
dtok.dtt = DTT_End;
return;
}
char ch = str.GetChar();
if (Char.IsLetter(ch))
{
//
// This is a letter.
//
int month, dayOfWeek, era, timeMark;
//
// Check if this is a beginning of a month name.
// And check if this is a day of week name.
//
if (raw.month == -1 && (month = GetMonthNumber(str, dtfi)) >= 1)
{
//
// This is a month name
//
switch(sep=GetSeparator(str, raw, dtfi))
{
case SEP_End:
dtok.dtt = DTT_MonthEnd;
break;
case SEP_Space:
dtok.dtt = DTT_MonthSpace;
break;
case SEP_Date:
dtok.dtt = DTT_MonthDatesep;
break;
default:
//Invalid separator after month name
throw new FormatException(Environment.GetResourceString("Format_BadDateTime"));
}
raw.month = month;
}
else if (raw.dayOfWeek == -1 && (dayOfWeek = GetDayOfWeekNumber(str, dtfi)) >= 0)
{
//
// This is a day of week name.
//
raw.dayOfWeek = dayOfWeek;
dtok.dtt = DTT_DayOfWeek;
//
// Discard the separator.
//
GetSeparator(str, raw, dtfi);
}
else if (GetTimeZoneName(str))
{
//
// This is a timezone designator
//
// NOTENOTE : for now, we only support "GMT" and "Z" (for Zulu time).
//
dtok.dtt = DTT_TimeZone;
result.timeZoneUsed = true;
result.timeZoneOffset = new TimeSpan(0);
} else if ((raw.era == -1) && ((era = GetEra(str, result, ref dtfi)) != -1)) {
raw.era = era;
dtok.dtt = DTT_Era;
} else if (raw.timeMark == -1 && (timeMark = GetTimeMark(str, dtfi)) != -1) {
raw.timeMark = timeMark;
GetSeparator(str, raw, dtfi);
} else {
//
// Not a month name, not a day of week name. Check if this is one of the
// known date words. This is used to deal case like Spanish cultures, which
// uses 'de' in their Date string.
//
//
if (!str.MatchWords(dtfi.DateWords)) {
throw new FormatException(
String.Format(Environment.GetResourceString("Format_UnknowDateTimeWord"), str.Index));
}
GetSeparator(str, raw, dtfi);
}
} else if (Char.IsDigit(ch)) {
if (raw.numCount == 3) {
throw new FormatException(Environment.GetResourceString("Format_BadDateTime"));
}
//
// This is a digit.
//
int number = ch - '0';
int digitCount = 1;
//
// Collect other digits.
//
while (str.GetNextDigit())
{
number = number * 10 + str.GetDigit();
digitCount++;
}
// If the previous parsing state is DS_T_NNt (like 12:01), and we got another number,
// so we will have a terminal state DS_TX_NNN (like 12:01:02).
// If the previous parsing state is DS_T_Nt (like 12:), and we got another number,
// so we will have a terminal state DS_TX_NN (like 12:01:02).
//
// Look ahead to see if the following character is a decimal point or timezone offset.
// This enables us to parse time in the forms of:
// "11:22:33.1234" or "11:22:33-08".
if (dps == DS_T_NNt || dps == DS_T_Nt) {
char nextCh;
if ((str.Index < str.len - 1)) {
nextCh = str.Value[str.Index];
switch (nextCh) {
case '.':
if (dps == DS_T_NNt) {
// Yes, advance to the next character.
str.Index++;
// Collect the second fraction.
raw.fraction = ParseFraction(str);
}
break;
case '+':
case '-':
if (result.timeZoneUsed) {
// Should not have two timezone offsets.
throw new FormatException(Environment.GetResourceString("Format_BadDateTime"));
}
result.timeZoneUsed = true;
result.timeZoneOffset = ParseTimeZone(str, nextCh);
break;
}
}
}
if (number >= 0)
{
dtok.num = number;
if (digitCount >= 3)
{
if (raw.year == -1)
{
raw.year = number;
//
// If we have number which has 3 or more digits (like "001" or "0001"),
// we assume this number is a year. Save the currnet raw.numCount in
// raw.year.
//
switch (sep = GetSeparator(str, raw, dtfi))
{
case SEP_End:
dtok.dtt = DTT_YearEnd;
break;
case SEP_Am:
case SEP_Pm:
case SEP_Space:
dtok.dtt = DTT_YearSpace;
break;
case SEP_Date:
dtok.dtt = DTT_YearDateSep;
break;
case SEP_YearSuff:
case SEP_MonthSuff:
case SEP_DaySuff:
dtok.dtt = DTT_NumDatesuff;
dtok.suffix = sep;
break;
case SEP_HourSuff:
case SEP_MinuteSuff:
case SEP_SecondSuff:
dtok.dtt = DTT_NumTimesuff;
dtok.suffix = sep;
break;
default:
// Invalid separator after number number.
throw new FormatException(Environment.GetResourceString("Format_BadDateTime"));
}
//
// Found the token already. Let's bail.
//
return;
}
throw new FormatException(Environment.GetResourceString("Format_BadDateTime"));
}
} else
{
//
// number is overflowed.
//
throw new FormatException(Environment.GetResourceString("Format_BadDateTime"));
}
switch (sep = GetSeparator(str, raw, dtfi))
{
//
// Note here we check if the numCount is less than three.
// When we have more than three numbers, it will be caught as error in the state machine.
//
case SEP_End:
dtok.dtt = DTT_NumEnd;
raw.num[raw.numCount++] = dtok.num;
break;
case SEP_Am:
case SEP_Pm:
dtok.dtt = DTT_NumAmpm;
raw.num[raw.numCount++] = dtok.num;
break;
case SEP_Space:
dtok.dtt = DTT_NumSpace;
raw.num[raw.numCount++] = dtok.num;
break;
case SEP_Date:
dtok.dtt = DTT_NumDatesep;
raw.num[raw.numCount++] = dtok.num;
break;
case SEP_Time:
if (!result.timeZoneUsed) {
dtok.dtt = DTT_NumTimesep;
raw.num[raw.numCount++] = dtok.num;
} else {
// If we already got timezone, there should be no
// time separator again.
throw new FormatException(Environment.GetResourceString("Format_BadDateTime"));
}
break;
case SEP_YearSuff:
dtok.num = dtfi.Calendar.ToFourDigitYear(number);
dtok.dtt = DTT_NumDatesuff;
dtok.suffix = sep;
break;
case SEP_MonthSuff:
case SEP_DaySuff:
dtok.dtt = DTT_NumDatesuff;
dtok.suffix = sep;
break;
case SEP_HourSuff:
case SEP_MinuteSuff:
case SEP_SecondSuff:
dtok.dtt = DTT_NumTimesuff;
dtok.suffix = sep;
break;
case SEP_LocalTimeMark:
dtok.dtt = DTT_NumLocalTimeMark;
raw.num[raw.numCount++] = dtok.num;
break;
default:
// Invalid separator after number number.
throw new FormatException(Environment.GetResourceString("Format_BadDateTime"));
}
}
else
{
//
// Not a letter, not a digit. Just ignore it.
//
str.Index++;
}
return;
}
//
// Starting at str.Index, check the type of the separator.
//
private static int GetSeparator(__DTString str, DateTimeRawInfo raw, DateTimeFormatInfo dtfi) {
int separator = SEP_Space; // Assume the separator is a space. And try to find a better one.
//
// Check if we found any white spaces.
//
if (!str.SkipWhiteSpaceComma()) {
//
// SkipWhiteSpaceComma() will return true when end of string is reached.
//
//
// Return the separator as SEP_End.
//
return (SEP_End);
}
if (Char.IsLetter(str.GetChar())) {
//
// This is a beginning of a word.
//
if (raw.timeMark == -1)
{
//
// Check if this is an AM time mark.
//
int timeMark;
if ((timeMark = GetTimeMark(str, dtfi)) != -1)
{
raw.timeMark = timeMark;;
return (timeMark == TM_AM ? SEP_Am: SEP_Pm);
}
}
if (MatchWord(str, LocalTimeMark, false)) {
separator = SEP_LocalTimeMark;
} else if (MatchWord(str, CJKYearSuff, false) || MatchWord(str, KoreanYearSuff, false)) {
separator = SEP_YearSuff;
}
else if (MatchWord(str, CJKMonthSuff, false) || MatchWord(str, KoreanMonthSuff, false))
{
separator = SEP_MonthSuff;
}
else if (MatchWord(str, CJKDaySuff, false) || MatchWord(str, KoreanDaySuff, false))
{
separator = SEP_DaySuff;
}
else if (MatchWord(str, CJKHourSuff, false) || MatchWord(str, ChineseHourSuff, false))
{
separator = SEP_HourSuff;
}
else if (MatchWord(str, CJKMinuteSuff, false))
{
separator = SEP_MinuteSuff;
}
else if (MatchWord(str, CJKSecondSuff, false))
{
separator = SEP_SecondSuff;
}
} else {
//
// Not a letter. Check if this is a date separator.
//
if ((MatchWord(str, dtfi.DateSeparator, false)) ||
(MatchWord(str, invariantInfo.DateSeparator, false)) ||
(MatchWord(str, alternativeDateSeparator, false)))
{
//
// NOTENOTE : alternativeDateSeparator is a special case because some cultures
// (e.g. the invariant culture) use "/". However, in RFC format, we use "-" as the
// date separator. Therefore, we should check for it.
//
separator = SEP_Date;
}
//
// Check if this is a time separator.
//
else if ((MatchWord(str, dtfi.TimeSeparator, false)) ||
(MatchWord(str, invariantInfo.TimeSeparator, false)))
{
separator = SEP_Time;
} else if (dtfi.CultureID == 0x041c) {
// Special case for sq-AL (0x041c)
// Its time pattern is "h:mm:ss.tt"
if (str.GetChar() == '.') {
if (raw.timeMark == -1)
{
//
// Check if this is an AM time mark.
//
int timeMark;
str.Index++;
if ((timeMark = GetTimeMark(str, dtfi)) != -1)
{
raw.timeMark = timeMark;;
return (timeMark == TM_AM ? SEP_Am: SEP_Pm);
}
str.Index--;
}
}
}
}
return (separator);
}