void consume_multiline_comment(TokenizeContext& context)
{
int lookahead = 0;
// Keep track of the current depth, for nested blocks.
int depth = 0;
while (context.withinRange(lookahead)) {
if (context.next(lookahead) == '{' && context.next(lookahead + 1) == '-') {
// Found a comment opener, increase depth. Also advance lookahead so that
// we don't get confused by this: {-}
lookahead += 2;
depth++;
continue;
}
if (context.next(lookahead) == '-' && context.next(lookahead + 1) == '}') {
// Found a comment ender.
depth--;
lookahead += 2;
if (depth == 0)
break;
continue;
}
lookahead++;
}
context.consume(tok_Comment, lookahead);
}
bool try_to_consume_keyword(TokenizeContext& context, int keyword)
{
const char* str = get_token_text(keyword);
int str_len = (int) strlen(str);
// Check if every letter matches
for (int i=0; i < str_len; i++) {
if (context.next(i) != str[i])
return false;
}
// Check that this is really the end of the word
if (is_acceptable_inside_identifier(context.next(str_len)))
return false;
// Don't match as a keyword if the next character is (. This might be
// a bad idea.
if (context.next(str_len) == '(')
return false;
// Keyword matches, now consume it
context.consume(keyword, str_len);
return true;
}
void consume_string_literal(TokenizeContext &context)
{
int lookahead = 0;
// Consume starting quote, this can be ' or "
char quote_type = context.next();
lookahead++;
bool escapedNext = false;
while (context.withinRange(lookahead)) {
char c = context.next(lookahead);
if (c == quote_type && !escapedNext)
break;
if (c == '\\' && !escapedNext)
escapedNext = true;
else
escapedNext = false;
lookahead++;
}
// consume ending quote
lookahead++;
context.consume(tok_String, lookahead);
}
bool match_number(TokenizeContext &context)
{
int lookahead = 0;
if (context.next(lookahead) == '.')
lookahead++;
if (is_number(context.next(lookahead)))
return true;
return false;
}
void consume_whitespace(TokenizeContext &context)
{
int lookahead = 0;
while (is_whitespace(context.next(lookahead)))
lookahead++;
context.consume(TK_WHITESPACE, lookahead);
}
void consume_whitespace(TokenizeContext &context)
{
int lookahead = 0;
while (is_whitespace(context.next(lookahead)))
lookahead++;
context.consume(tok_Whitespace, lookahead);
}
void consume_identifier(TokenizeContext &context)
{
int lookahead = 0;
while (is_acceptable_inside_identifier(context.next(lookahead)))
lookahead++;
context.consume(tok_Identifier, lookahead);
}
void consume_triple_quoted_string_literal(TokenizeContext &context)
{
int lookahead = 0;
// Consume initial <<<
lookahead += 3;
while (context.withinRange(lookahead) &&
!(context.next(lookahead) == '>'
&& context.next(lookahead + 1) == '>'
&& context.next(lookahead + 2) == '>'))
lookahead++;
// Consume closing >>>
lookahead += 3;
context.consume(tok_String, lookahead);
}
void consume_comment(TokenizeContext& context)
{
int lookahead = 0;
while (context.withinRange(lookahead) && !is_newline(context.next(lookahead)))
lookahead++;
context.consume(tok_Comment, lookahead);
}
void consume_number(TokenizeContext &context)
{
int lookahead = 0;
bool dot_encountered = false;
// Possibly consume minus sign
if (context.next(lookahead) == '-') {
lookahead++;
}
while (true) {
if (is_number(context.next(lookahead))) {
lookahead++;
} else if (context.next(lookahead) == '.') {
// If we've already encountered a dot, finish and don't consume
// this one.
if (dot_encountered)
break;
// Special case: if this dot is followed by another dot, then it should
// be tokenized as TWO_DOTS, so don't consume it here.
if (context.next(lookahead+1) == '.')
break;
// Another special case, if the dot is followed by an identifier, then
// don't consume it here. It might be an object call.
if (is_identifier_first_letter(context.next(lookahead + 1)))
break;
// Otherwise, consume the dot
lookahead++;
dot_encountered = true;
}
else {
break;
}
}
if (dot_encountered)
context.consume(tok_Float, lookahead);
else
context.consume(tok_Integer, lookahead);
}
void consume_hex_number(TokenizeContext &context)
{
int lookahead = 0;
// consume the 0x part
lookahead += 2;
while (is_hexadecimal_digit(context.next(lookahead)))
lookahead++;
context.consume(tok_HexInteger, lookahead);
}
void consume_name(TokenizeContext &context)
{
int lookahead = 0;
// consume the leading :
lookahead++;
while (is_acceptable_inside_identifier(context.next(lookahead)))
lookahead++;
context.consume(TK_NAME, lookahead);
}
void consume_hex_number(TokenizeContext &context)
{
int lookahead = 0;
// consume the 0x part
lookahead += 2;
while (is_hexadecimal_digit(context.next(lookahead)))
lookahead++;
context.consume(TK_HEX_INTEGER, lookahead);
}
void consume_symbol(TokenizeContext &context)
{
int lookahead = 0;
// consume the leading :
lookahead++;
while (is_acceptable_inside_identifier(context.next(lookahead)))
lookahead++;
context.consume(tok_ColonString, lookahead);
}
void consume_color_literal(TokenizeContext &context)
{
int lookahead = 0;
// consume #
lookahead++;
while (is_hexadecimal_digit(context.next(lookahead)))
lookahead++;
int hex_digits = lookahead - 1;
// acceptable lengths are 3, 4, 6 or 8 characters (not including #)
if (hex_digits == 3 || hex_digits == 4 || hex_digits == 6 || hex_digits == 8)
context.consume(tok_Color, lookahead);
else
context.consume(tok_Unrecognized, lookahead);
}
void top_level_consume_token(TokenizeContext &context)
{
if (is_identifier_first_letter(context.next())) {
if (try_to_consume_keyword(context, TK_DEF)) return;
if (try_to_consume_keyword(context, TK_TYPE)) return;
if (try_to_consume_keyword(context, TK_BEGIN)) return;
if (try_to_consume_keyword(context, TK_END)) return;
if (try_to_consume_keyword(context, TK_IF)) return;
if (try_to_consume_keyword(context, TK_ELSE)) return;
if (try_to_consume_keyword(context, TK_ELIF)) return;
if (try_to_consume_keyword(context, TK_FOR)) return;
if (try_to_consume_keyword(context, TK_STATE)) return;
if (try_to_consume_keyword(context, TK_IN)) return;
if (try_to_consume_keyword(context, TK_TRUE)) return;
if (try_to_consume_keyword(context, TK_FALSE)) return;
// check 'do once' before 'do'
if (try_to_consume_keyword(context, TK_DO_ONCE)) return;
if (try_to_consume_keyword(context, TK_DO)) return;
if (try_to_consume_keyword(context, TK_NAMESPACE)) return;
if (try_to_consume_keyword(context, TK_INCLUDE)) return;
if (try_to_consume_keyword(context, TK_IMPORT)) return;
if (try_to_consume_keyword(context, TK_AND)) return;
if (try_to_consume_keyword(context, TK_OR)) return;
if (try_to_consume_keyword(context, TK_DISCARD)) return;
if (try_to_consume_keyword(context, TK_NULL)) return;
if (try_to_consume_keyword(context, TK_RETURN)) return;
if (try_to_consume_keyword(context, TK_BREAK)) return;
if (try_to_consume_keyword(context, TK_CONTINUE)) return;
if (try_to_consume_keyword(context, TK_SWITCH)) return;
if (try_to_consume_keyword(context, TK_CASE)) return;
if (try_to_consume_keyword(context, TK_WHILE)) return;
consume_identifier(context);
return;
}
if (is_whitespace(context.next())) {
consume_whitespace(context);
return;
}
if (context.next() == '0'
&& context.next(1) == 'x') {
consume_hex_number(context);
return;
}
if (match_number(context)) {
consume_number(context);
return;
}
// Check for specific characters
switch(context.next()) {
case '(':
context.consume(TK_LPAREN, 1);
return;
case ')':
context.consume(TK_RPAREN, 1);
return;
case '{':
if (context.next(1) == '-') {
consume_multiline_comment(context);
return;
}
context.consume(TK_LBRACE, 1);
return;
case '}':
context.consume(TK_RBRACE, 1);
return;
case '[':
context.consume(TK_LBRACKET, 1);
return;
case ']':
context.consume(TK_RBRACKET, 1);
return;
case ',':
context.consume(TK_COMMA, 1);
return;
case '@':
if (context.next(1) == '.') {
context.consume(TK_AT_DOT, 2);
} else {
context.consume(TK_AT_SIGN, 1);
}
return;
case '=':
if (context.next(1) == '=') {
context.consume(TK_DOUBLE_EQUALS, 2);
return;
}
context.consume(TK_EQUALS, 1);
return;
case '"':
case '\'':
consume_string_literal(context);
return;
case '\n':
context.consume(TK_NEWLINE, 1);
return;
case '.':
if (context.next(1) == '.') {
if (context.next(2) == '.') {
context.consume(TK_ELLIPSIS, 3);
} else {
context.consume(TK_TWO_DOTS, 2);
}
} else {
context.consume(TK_DOT, 1);
}
return;
case '?':
context.consume(TK_QUESTION, 1);
return;
case '*':
if (context.next(1) == '=') {
context.consume(TK_STAR_EQUALS, 2);
return;
}
context.consume(TK_STAR, 1);
return;
case '/':
if (context.next(1) == '=') {
context.consume(TK_SLASH_EQUALS, 2);
return;
}
if (context.next(1) == '/') {
context.consume(TK_DOUBLE_SLASH, 2);
return;
}
context.consume(TK_SLASH, 1);
return;
case '!':
if (context.next(1) == '=') {
context.consume(TK_NOT_EQUALS, 2);
return;
}
break;
case ':':
if (context.next(1) == '=') {
context.consume(TK_COLON_EQUALS, 2);
return;
}
else if (context.next(1) == ':') {
context.consume(TK_DOUBLE_COLON, 2);
return;
} else if (is_identifier_first_letter(context.next(1))) {
return consume_name(context);
}
context.consume(TK_COLON, 1);
return;
case '+':
if (context.next(1) == '=') {
context.consume(TK_PLUS_EQUALS, 2);
} else {
context.consume(TK_PLUS, 1);
}
return;
case '-':
if (context.next(1) == '>') {
context.consume(TK_RIGHT_ARROW, 2);
return;
}
if (context.next(1) == '-')
return consume_comment(context);
if (context.next(1) == '=') {
context.consume(TK_MINUS_EQUALS, 2);
return;
}
context.consume(TK_MINUS, 1);
return;
case '<':
if (context.next(1) == '<' && context.next(2) == '<') {
consume_triple_quoted_string_literal(context);
return;
}
if (context.next(1) == '=') {
context.consume(TK_LTHANEQ, 2);
return;
}
if (context.next(1) == '-') {
context.consume(TK_LEFT_ARROW, 2);
return;
}
context.consume(TK_LTHAN, 1);
return;
case '>':
if (context.next(1) == '=') {
context.consume(TK_GTHANEQ, 2);
return;
}
context.consume(TK_GTHAN, 1);
return;
case '%':
context.consume(TK_PERCENT, 1);
return;
case '|':
if (context.next(1) == '|') {
context.consume(TK_DOUBLE_VERTICAL_BAR, 2);
return;
}
break;
case '&':
if (context.next(1) == '&') {
context.consume(TK_DOUBLE_AMPERSAND, 2);
return;
}
context.consume(TK_AMPERSAND, 1);
return;
case ';':
context.consume(TK_SEMICOLON, 1);
return;
case '#':
consume_color_literal(context);
return;
}
// Fall through, consume the next letter as UNRECOGNIZED
context.consume(TK_UNRECOGNIZED, 1);
}
void top_level_consume_token(TokenizeContext &context)
{
if (is_identifier_first_letter(context.next())) {
if (context.next() <= 'm') {
// a through m
if (try_to_consume_keyword(context, tok_And)) return;
if (try_to_consume_keyword(context, tok_Break)) return;
if (try_to_consume_keyword(context, tok_Case)) return;
if (try_to_consume_keyword(context, tok_Continue)) return;
if (try_to_consume_keyword(context, tok_Def)) return;
if (try_to_consume_keyword(context, tok_Discard)) return;
if (try_to_consume_keyword(context, tok_Else)) return;
if (try_to_consume_keyword(context, tok_Elif)) return;
if (try_to_consume_keyword(context, tok_False)) return;
if (try_to_consume_keyword(context, tok_For)) return;
if (try_to_consume_keyword(context, tok_If)) return;
if (try_to_consume_keyword(context, tok_In)) return;
if (try_to_consume_keyword(context, tok_Import)) return;
if (try_to_consume_keyword(context, tok_Include)) return;
if (try_to_consume_keyword(context, tok_Let)) return;
} else {
// n through z
if (try_to_consume_keyword(context, tok_Namespace)) return;
if (try_to_consume_keyword(context, tok_Not)) return;
if (try_to_consume_keyword(context, tok_Nil)) return;
if (try_to_consume_keyword(context, tok_Or)) return;
if (try_to_consume_keyword(context, tok_Return)) return;
if (try_to_consume_keyword(context, tok_State)) return;
if (try_to_consume_keyword(context, tok_Struct)) return;
if (try_to_consume_keyword(context, tok_Switch)) return;
if (try_to_consume_keyword(context, tok_True)) return;
if (try_to_consume_keyword(context, tok_Require)) return;
if (try_to_consume_keyword(context, tok_RequireLocal)) return;
if (try_to_consume_keyword(context, tok_Package)) return;
if (try_to_consume_keyword(context, tok_Section)) return;
if (try_to_consume_keyword(context, tok_While)) return;
}
consume_identifier(context);
return;
}
if (is_whitespace(context.next())) {
consume_whitespace(context);
return;
}
if (context.next() == '0'
&& context.next(1) == 'x') {
consume_hex_number(context);
return;
}
if (match_number(context)) {
consume_number(context);
return;
}
// Check for specific characters
switch(context.next()) {
case '(':
context.consume(tok_LParen, 1);
return;
case ')':
context.consume(tok_RParen, 1);
return;
case '{':
if (context.next(1) == '-') {
consume_multiline_comment(context);
return;
}
context.consume(tok_LBrace, 1);
return;
case '}':
context.consume(tok_RBrace, 1);
return;
case '[':
context.consume(tok_LSquare, 1);
return;
case ']':
context.consume(tok_RSquare, 1);
return;
case ',':
context.consume(tok_Comma, 1);
return;
case '@':
context.consume(tok_At, 1);
return;
case '=':
if (context.next(1) == '=') {
context.consume(tok_DoubleEquals, 2);
return;
} else if (context.next(1) == '>') {
context.consume(tok_FatArrow, 2);
return;
}
context.consume(tok_Equals, 1);
return;
case '"':
case '\'':
consume_string_literal(context);
return;
case '\n':
context.consume(tok_Newline, 1);
return;
case '.':
if (context.next(1) == '.') {
if (context.next(2) == '.') {
context.consume(tok_Ellipsis, 3);
} else {
context.consume(tok_TwoDots, 2);
}
} else if (context.next(1) == '@') {
context.consume(tok_DotAt, 2);
} else {
context.consume(tok_Dot, 1);
}
return;
case '?':
context.consume(tok_Question, 1);
return;
case '*':
if (context.next(1) == '=') {
context.consume(tok_StarEquals, 2);
return;
}
if (context.next(1) == '*') {
context.consume(tok_DoubleStar, 2);
return;
}
context.consume(tok_Star, 1);
return;
case '/':
if (context.next(1) == '=') {
context.consume(tok_SlashEquals, 2);
return;
}
if (context.next(1) == '/') {
context.consume(tok_DoubleSlash, 2);
return;
}
context.consume(tok_Slash, 1);
return;
case '!':
if (context.next(1) == '=') {
context.consume(tok_NotEquals, 2);
return;
}
break;
case ':':
if (context.next(1) == '=') {
context.consume(tok_ColonEquals, 2);
return;
}
else if (context.next(1) == ':') {
context.consume(tok_DoubleColon, 2);
return;
} else if (is_acceptable_inside_identifier(context.next(1))) {
return consume_symbol(context);
}
context.consume(tok_Colon, 1);
return;
case '+':
if (context.next(1) == '=') {
context.consume(tok_PlusEquals, 2);
} else {
context.consume(tok_Plus, 1);
}
return;
case '-':
if (context.next(1) == '>') {
context.consume(tok_RightArrow, 2);
return;
}
if (context.next(1) == '-')
return consume_comment(context);
if (context.next(1) == '=') {
context.consume(tok_MinusEquals, 2);
return;
}
context.consume(tok_Minus, 1);
return;
case '<':
if (context.next(1) == '<' && context.next(2) == '<') {
consume_triple_quoted_string_literal(context);
return;
}
if (context.next(1) == '=') {
context.consume(tok_LThanEq, 2);
return;
}
if (context.next(1) == '-') {
context.consume(tok_LeftArrow, 2);
return;
}
context.consume(tok_LThan, 1);
return;
case '>':
if (context.next(1) == '=')
context.consume(tok_GThanEq, 2);
else
context.consume(tok_GThan, 1);
return;
case '%':
context.consume(tok_Percent, 1);
return;
case '|':
if (context.next(1) == '|')
context.consume(tok_DoubleVerticalBar, 2);
else
context.consume(tok_VerticalBar, 1);
return;
case '&':
if (context.next(1) == '&')
context.consume(tok_DoubleAmpersand, 2);
else
context.consume(tok_Ampersand, 1);
return;
case ';':
context.consume(tok_Semicolon, 1);
return;
case '#':
consume_color_literal(context);
return;
}
// Fall through, consume the next letter as UNRECOGNIZED
context.consume(tok_Unrecognized, 1);
}