pgLab/SqlLexer.cpp

#include "SqlLexer.h"

SqlLexer::SqlLexer(const QString &block, LexerState currentstate)
	: m_block(block)
	, m_state(currentstate)
{}

QChar SqlLexer::nextChar()
{
	QChar result = QChar::Null;
	if (m_pos < m_block.size()) {
		result = m_block.at(m_pos++);
	}
	return result;
}

QChar SqlLexer::peekChar()
{
	QChar result = QChar::Null;
	if (m_pos < m_block.size()) {
		result = m_block.at(m_pos);
	}
	return result;
}

/**
 * @brief NextBasicToken
 * @param in
 * @param ofs
 * @param start
 * @param length
 * @return false when input seems invalid, it will return what it did recognize but something wasn't right, parser should try to recover
 */
bool SqlLexer::nextBasicToken(int &startpos, int &length, BasicTokenType &tokentype, QString &out)
{
	// Basically chops based on white space
	// it does also recognize comments and quoted strings/identifiers
	bool result = false;
	while (true) {
		startpos = m_pos;
		QChar c = nextChar();
		if (c.isSpace()) {
			// Just skip whitespace
		}
		else if (c == '-' && peekChar() == '-') { // two dashes, start of comment
			// Loop till end of line or end of block
			c = nextChar();
			for (;;) {
				c = peekChar();
				if (c != QChar::Null && c != '\n')
					nextChar();
				else
					break;
			}
			length = m_pos - startpos;
			tokentype = BasicTokenType::Comment;
			return true;
		}
		else if (c == '\'') {
			// Single quoted string so it's an SQL text literal
			while (true) {
				c = peekChar();
				if (c == QChar::Null || c == '\n') {
					// unexpected end, pretend nothings wrong
					length = m_pos - startpos;
					tokentype = BasicTokenType::QuotedString;
					return true;
				}
				else {
					nextChar();
					if (c == '\'') {
						// maybe end of string literal
						if (peekChar() == '\'') {
							// Nope, just double quote to escape quote
							nextChar(); // eat it
						}
						else {
							length = m_pos - startpos;
							tokentype = BasicTokenType::QuotedString;
							return true;
						}
					}
				}
			}
		}
		else if (c == '"') {
			// Double quoted identifier
			while (true) {
				c = peekChar();
				if (c == QChar::Null || c == '\n') {
					// unexpected end, pretend nothings wrong
					length = m_pos - startpos;
					tokentype = BasicTokenType::QuotedIdentifier;
					return true;
				}
				else {
					nextChar();
					if (c == '"') {
						// maybe end of string literal
						if (peekChar() == '"') {
							// Nope, just double quote to escape quote
							nextChar(); // eat it
						}
						else {
							length = m_pos - startpos;
							tokentype = BasicTokenType::QuotedIdentifier;
							return true;
						}
					}
				}
			}
		}
		else if (c == QChar::Null) {
			break;
		}
		else {
			// Undetermined symbol
			for (;;) {
				c = peekChar();
				if (c.isLetterOrNumber() || c == '_')
					nextChar();
				else
					break;
			}
			length = m_pos - startpos;
			tokentype = BasicTokenType::Symbol;
			QStringRef sr(&m_block, startpos, length);
			out = sr.toString();
			return true;
		}
	}
	return false;
}