private static boolean number(Text t, DataType requiredType, List<Object> valueList) {
// posNumber
// : cardinality
// | '0b' BinaryDigits+ [zZ]?
// | '0x' HexDigit+ [zZ]?
// | '-'? Pint? ( '.' Digit+ ) ('e' Pint) [fF]
// | '-'? Pint [zZfF]?
// ;
final boolean result;
assert !t.isEof() : "Should have been checked by caller";
if (binary(t, valueList, requiredType) || hex(t, valueList, requiredType)) {
// ?todo I could allow negatives here
// ( binary | hex | '0' )
result = true;
} else {
// ( Int DecPart | Int | DecPart ) ( 'e' Int )? [zZfF]?
final int start = t.cursor();
if ((t.consumeInt() && (decPart(t) || true) && (exponent(t) || true))
|| (t.consume('-') || true) && decPart(t) && (exponent(t) || true)) {
final String string = t.getString(start);
final Object value = deriveType(string, requiredType, t, 10);
valueList.add(value);
result = true;
} else {
result = false;
}
}
return result;
}
private static boolean text(Text t, List<Object> valueList) {
final boolean result;
final int start = t.cursor();
if (t.consumeString()) {
// +/- 1s: drop the quotes
final String textValue = unescape(t, start + 1, t.cursor() - 1);
valueList.add(textValue);
result = true;
} else {
// ?todo Could easily detect unterminated string and throw here
result = false;
}
return result;
}
private static boolean decPart(Text t) {
final boolean result;
// '.' Digits+
final int save = t.cursor();
if (t.consume('.') && t.consumeAscii(Text.ASCII_0_9)) {
result = true;
} else {
t.setCursor(save);
result = false;
}
return result;
}
private static boolean exponent(Text t) {
final boolean result;
// 'e' Int
final int save = t.cursor();
if (t.consume('e') && t.consumeInt()) {
result = true;
} else {
t.setCursor(save);
result = false;
}
return result;
}
private static boolean identifier(Text t, List<Object> valueList) {
final boolean result;
final int start = t.cursor();
if (Identifier.consume(t)) {
final String string = t.getString(start);
valueList.add(new Identifier(string));
result = true;
} else {
result = false;
}
return result;
}
private static boolean time(Text t) {
final boolean result;
// HH ':' mm ( : ss )?
// Try for HH:mm...
int reset = t.cursor();
if (t.consumeAscii(Text.ASCII_0_9) && t.consume(':') && t.consumeAscii(Text.ASCII_0_9)) {
result = true;
reset = t.cursor();
// Try for :ss...
if (t.consume(':') && t.consumeAscii(Text.ASCII_0_9)) {
reset = t.cursor();
}
} else {
result = false;
}
t.setCursor(reset);
return result;
}
private static boolean date(Text t, List<Object> valueList) {
final boolean result;
final int start = t.cursor();
if (t.consume('@') && date(t)) {
final String string = t.getString(start + 1); // Jump the'@'
final LocalDate value = DateU.parseStandardDate(string);
valueList.add(value);
result = true;
} else {
result = false;
}
return result;
}
/**
* There's a bug in this as it requires seconds in the string whereas the time() function does
* not. It seems to be in the Java parser.
*
* @param tb Source to parse.
* @return A Temporal or null if none found.
*/
@Nullable
private static Temporal temporal(Text t) {
Temporal result;
final int save = t.cursor();
t.consume('@');
final int start = save + 1;
// yyyy/MM/dd HH:mm:ss
try {
if (date(t)) {
final int save2 = t.cursor();
t.ws();
if (time(t)) {
final String string = t.getString(start);
result = DateU.parseStandardDatetime(string);
} else {
t.setCursor(save2);
final String string = t.getString(start);
result = DateU.parseStandardDate(string);
}
} else if (time(t)) {
final String string = t.getString(start);
result = DateU.parseStandardTime(string);
} else {
result = null;
t.setCursor(save);
}
} catch (final DateTimeParseException e) {
result = null;
t.setCursor(save);
}
return result;
}
private static List<Object> elementList(Text t, DataType type) {
final List<Object> result = new ArrayList<>();
// elementList: ( WS? value ( WS? ',' WS? value )* WS? )?
if (t.ws() && value(t, type, result)) {
while (true) {
final int save = t.cursor();
if (t.ws() && t.consume(',') && t.ws() && value(t, type, result)) {
continue;
}
t.setCursor(save);
break;
}
}
return result;
}
private static boolean binary(Text t, List<Object> valueList, DataType requiredType) {
final boolean result;
// '0b' [01]+ [zZfF]?
final int save = t.cursor();
if (t.consume("0b") && t.consumeAscii(Text.ASCII_0_1)) {
final String string = t.getString(save + 2);
final Object value = deriveType(string, requiredType, t, 2);
valueList.add(value);
result = true;
} else {
t.setCursor(save);
result = false;
}
return result;
}
private static boolean date(Text t) {
final boolean result;
// yyyy '/' MM '/' dd
final int save = t.cursor();
if (t.consumeAscii(Text.ASCII_0_9)
&& t.consume('/')
&& t.consumeAscii(Text.ASCII_0_9)
&& t.consume('/')
&& t.consumeAscii(Text.ASCII_0_9)) {
result = true;
} else {
result = false;
t.setCursor(save);
}
return result;
}
/**
* In Oak it is possible to embed a tab character as an ASCII 8 or as a \t but internally we store
* this in ASCII. The same goes for newlines and other tables (see the Oak reference). Here we
* convert external format to internal format.
*
* @param source The text to process, may be null. E.g. "Hello\tworld" (with the quotes)
* @return Return an internal format string.
*/
@Nullable
public static String textToInternalFormat(@Nullable String source) throws ParsingException {
final String result;
assert source == null
|| source.length() >= 2
&& source.charAt(0) == '"'
&& source.charAt(source.length() - 1) == '"';
if (source == null) {
result = null;
} else {
final Text t = new Text();
final char[] ca = source.toCharArray();
final int length = ca.length - 1; // 1 because remove quotes
for (int i = 1; i < length; i++) { // 1 because remove quotes
final char c = ca[i];
if (c == '\\') {
if (i == length - 1) {
error("Invalid text " + source + ", escape at end of line");
} else {
i++;
final char next = ca[i];
if (next == 't') {
t.append('\t');
} else if (next == 'n') {
t.append('\n');
} else if (next == '"') {
t.append('"');
} else if (next == '\\') {
t.append('\\');
} else if (next == 'u') {
// Unicode, 1-4 hex characters...
i++;
final Text hex = new Text();
hex.append(source);
hex.setCursor(i);
final int start = i;
if (hex.consumeAscii(Text.ASCII_0_F)) {
final int end = start + Math.min(4, hex.cursor() - start);
final String string = hex.getString(start, end);
final char u = (char) Integer.parseInt(string, 16);
t.append(u);
i = end - 1;
} else {
error("Invalid text " + source + ", incorrect unicode");
}
} else {
error("Invalid text: \\" + next);
}
}
} else {
t.append(c);
}
}
result = t.toString();
}
return result;
}