1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/dom/xslt/xpath/txExprLexer.cpp Wed Dec 31 06:09:35 2014 +0100
1.3 @@ -0,0 +1,370 @@
1.4 +/* -*- Mode: C++; tab-width: 2; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
1.5 +/* This Source Code Form is subject to the terms of the Mozilla Public
1.6 + * License, v. 2.0. If a copy of the MPL was not distributed with this
1.7 + * file, You can obtain one at http://mozilla.org/MPL/2.0/. */
1.8 +
1.9 +/**
1.10 + * Lexical analyzer for XPath expressions
1.11 + */
1.12 +
1.13 +#include "txExprLexer.h"
1.14 +#include "nsGkAtoms.h"
1.15 +#include "nsString.h"
1.16 +#include "nsError.h"
1.17 +#include "txXMLUtils.h"
1.18 +
1.19 +/**
1.20 + * Creates a new ExprLexer
1.21 + */
1.22 +txExprLexer::txExprLexer()
1.23 + : mCurrentItem(nullptr),
1.24 + mFirstItem(nullptr),
1.25 + mLastItem(nullptr),
1.26 + mTokenCount(0)
1.27 +{
1.28 +}
1.29 +
1.30 +/**
1.31 + * Destroys this instance of an txExprLexer
1.32 + */
1.33 +txExprLexer::~txExprLexer()
1.34 +{
1.35 + //-- delete tokens
1.36 + Token* tok = mFirstItem;
1.37 + while (tok) {
1.38 + Token* temp = tok->mNext;
1.39 + delete tok;
1.40 + tok = temp;
1.41 + }
1.42 + mCurrentItem = nullptr;
1.43 +}
1.44 +
1.45 +Token*
1.46 +txExprLexer::nextToken()
1.47 +{
1.48 + if (!mCurrentItem) {
1.49 + NS_NOTREACHED("nextToken called on uninitialized lexer");
1.50 + return nullptr;
1.51 + }
1.52 +
1.53 + if (mCurrentItem->mType == Token::END) {
1.54 + // Do not progress beyond the end token
1.55 + return mCurrentItem;
1.56 + }
1.57 +
1.58 + Token* token = mCurrentItem;
1.59 + mCurrentItem = mCurrentItem->mNext;
1.60 + return token;
1.61 +}
1.62 +
1.63 +void
1.64 +txExprLexer::addToken(Token* aToken)
1.65 +{
1.66 + if (mLastItem) {
1.67 + mLastItem->mNext = aToken;
1.68 + }
1.69 + if (!mFirstItem) {
1.70 + mFirstItem = aToken;
1.71 + mCurrentItem = aToken;
1.72 + }
1.73 + mLastItem = aToken;
1.74 + ++mTokenCount;
1.75 +}
1.76 +
1.77 +/**
1.78 + * Returns true if the following Token should be an operator.
1.79 + * This is a helper for the first bullet of [XPath 3.7]
1.80 + * Lexical Structure
1.81 + */
1.82 +bool
1.83 +txExprLexer::nextIsOperatorToken(Token* aToken)
1.84 +{
1.85 + if (!aToken || aToken->mType == Token::NULL_TOKEN) {
1.86 + return false;
1.87 + }
1.88 + /* This relies on the tokens having the right order in txExprLexer.h */
1.89 + return aToken->mType < Token::COMMA ||
1.90 + aToken->mType > Token::UNION_OP;
1.91 +
1.92 +}
1.93 +
1.94 +/**
1.95 + * Parses the given string into a sequence of Tokens
1.96 + */
1.97 +nsresult
1.98 +txExprLexer::parse(const nsASingleFragmentString& aPattern)
1.99 +{
1.100 + iterator start, end;
1.101 + start = aPattern.BeginReading(mPosition);
1.102 + aPattern.EndReading(end);
1.103 +
1.104 + //-- initialize previous token, this will automatically get
1.105 + //-- deleted when it goes out of scope
1.106 + Token nullToken(nullptr, nullptr, Token::NULL_TOKEN);
1.107 +
1.108 + Token::Type defType;
1.109 + Token* newToken = nullptr;
1.110 + Token* prevToken = &nullToken;
1.111 + bool isToken;
1.112 +
1.113 + while (mPosition < end) {
1.114 +
1.115 + defType = Token::CNAME;
1.116 + isToken = true;
1.117 +
1.118 + if (*mPosition == DOLLAR_SIGN) {
1.119 + if (++mPosition == end || !XMLUtils::isLetter(*mPosition)) {
1.120 + return NS_ERROR_XPATH_INVALID_VAR_NAME;
1.121 + }
1.122 + defType = Token::VAR_REFERENCE;
1.123 + }
1.124 + // just reuse the QName parsing, which will use defType
1.125 + // the token to construct
1.126 +
1.127 + if (XMLUtils::isLetter(*mPosition)) {
1.128 + // NCName, can get QName or OperatorName;
1.129 + // FunctionName, NodeName, and AxisSpecifier may want whitespace,
1.130 + // and are dealt with below
1.131 + start = mPosition;
1.132 + while (++mPosition < end && XMLUtils::isNCNameChar(*mPosition)) {
1.133 + /* just go */
1.134 + }
1.135 + if (mPosition < end && *mPosition == COLON) {
1.136 + // try QName or wildcard, might need to step back for axis
1.137 + if (++mPosition == end) {
1.138 + return NS_ERROR_XPATH_UNEXPECTED_END;
1.139 + }
1.140 + if (XMLUtils::isLetter(*mPosition)) {
1.141 + while (++mPosition < end && XMLUtils::isNCNameChar(*mPosition)) {
1.142 + /* just go */
1.143 + }
1.144 + }
1.145 + else if (*mPosition == '*' && defType != Token::VAR_REFERENCE) {
1.146 + // eat wildcard for NameTest, bail for var ref at COLON
1.147 + ++mPosition;
1.148 + }
1.149 + else {
1.150 + --mPosition; // step back
1.151 + }
1.152 + }
1.153 + if (nextIsOperatorToken(prevToken)) {
1.154 + nsDependentSubstring op(Substring(start, mPosition));
1.155 + if (nsGkAtoms::_and->Equals(op)) {
1.156 + defType = Token::AND_OP;
1.157 + }
1.158 + else if (nsGkAtoms::_or->Equals(op)) {
1.159 + defType = Token::OR_OP;
1.160 + }
1.161 + else if (nsGkAtoms::mod->Equals(op)) {
1.162 + defType = Token::MODULUS_OP;
1.163 + }
1.164 + else if (nsGkAtoms::div->Equals(op)) {
1.165 + defType = Token::DIVIDE_OP;
1.166 + }
1.167 + else {
1.168 + // XXX QUESTION: spec is not too precise
1.169 + // badops is sure an error, but is bad:ops, too? We say yes!
1.170 + return NS_ERROR_XPATH_OPERATOR_EXPECTED;
1.171 + }
1.172 + }
1.173 + newToken = new Token(start, mPosition, defType);
1.174 + }
1.175 + else if (isXPathDigit(*mPosition)) {
1.176 + start = mPosition;
1.177 + while (++mPosition < end && isXPathDigit(*mPosition)) {
1.178 + /* just go */
1.179 + }
1.180 + if (mPosition < end && *mPosition == '.') {
1.181 + while (++mPosition < end && isXPathDigit(*mPosition)) {
1.182 + /* just go */
1.183 + }
1.184 + }
1.185 + newToken = new Token(start, mPosition, Token::NUMBER);
1.186 + }
1.187 + else {
1.188 + switch (*mPosition) {
1.189 + //-- ignore whitespace
1.190 + case SPACE:
1.191 + case TX_TAB:
1.192 + case TX_CR:
1.193 + case TX_LF:
1.194 + ++mPosition;
1.195 + isToken = false;
1.196 + break;
1.197 + case S_QUOTE :
1.198 + case D_QUOTE :
1.199 + start = mPosition;
1.200 + while (++mPosition < end && *mPosition != *start) {
1.201 + // eat literal
1.202 + }
1.203 + if (mPosition == end) {
1.204 + mPosition = start;
1.205 + return NS_ERROR_XPATH_UNCLOSED_LITERAL;
1.206 + }
1.207 + newToken = new Token(start + 1, mPosition, Token::LITERAL);
1.208 + ++mPosition;
1.209 + break;
1.210 + case PERIOD:
1.211 + // period can be .., .(DIGITS)+ or ., check next
1.212 + if (++mPosition == end) {
1.213 + newToken = new Token(mPosition - 1, Token::SELF_NODE);
1.214 + }
1.215 + else if (isXPathDigit(*mPosition)) {
1.216 + start = mPosition - 1;
1.217 + while (++mPosition < end && isXPathDigit(*mPosition)) {
1.218 + /* just go */
1.219 + }
1.220 + newToken = new Token(start, mPosition, Token::NUMBER);
1.221 + }
1.222 + else if (*mPosition == PERIOD) {
1.223 + ++mPosition;
1.224 + newToken = new Token(mPosition - 2, mPosition, Token::PARENT_NODE);
1.225 + }
1.226 + else {
1.227 + newToken = new Token(mPosition - 1, Token::SELF_NODE);
1.228 + }
1.229 + break;
1.230 + case COLON: // QNames are dealt above, must be axis ident
1.231 + if (++mPosition >= end || *mPosition != COLON ||
1.232 + prevToken->mType != Token::CNAME) {
1.233 + return NS_ERROR_XPATH_BAD_COLON;
1.234 + }
1.235 + prevToken->mType = Token::AXIS_IDENTIFIER;
1.236 + ++mPosition;
1.237 + isToken = false;
1.238 + break;
1.239 + case FORWARD_SLASH :
1.240 + if (++mPosition < end && *mPosition == FORWARD_SLASH) {
1.241 + ++mPosition;
1.242 + newToken = new Token(mPosition - 2, mPosition, Token::ANCESTOR_OP);
1.243 + }
1.244 + else {
1.245 + newToken = new Token(mPosition - 1, Token::PARENT_OP);
1.246 + }
1.247 + break;
1.248 + case BANG : // can only be !=
1.249 + if (++mPosition < end && *mPosition == EQUAL) {
1.250 + ++mPosition;
1.251 + newToken = new Token(mPosition - 2, mPosition, Token::NOT_EQUAL_OP);
1.252 + break;
1.253 + }
1.254 + // Error ! is not not()
1.255 + return NS_ERROR_XPATH_BAD_BANG;
1.256 + case EQUAL:
1.257 + newToken = new Token(mPosition, Token::EQUAL_OP);
1.258 + ++mPosition;
1.259 + break;
1.260 + case L_ANGLE:
1.261 + if (++mPosition == end) {
1.262 + return NS_ERROR_XPATH_UNEXPECTED_END;
1.263 + }
1.264 + if (*mPosition == EQUAL) {
1.265 + ++mPosition;
1.266 + newToken = new Token(mPosition - 2, mPosition,
1.267 + Token::LESS_OR_EQUAL_OP);
1.268 + }
1.269 + else {
1.270 + newToken = new Token(mPosition - 1, Token::LESS_THAN_OP);
1.271 + }
1.272 + break;
1.273 + case R_ANGLE:
1.274 + if (++mPosition == end) {
1.275 + return NS_ERROR_XPATH_UNEXPECTED_END;
1.276 + }
1.277 + if (*mPosition == EQUAL) {
1.278 + ++mPosition;
1.279 + newToken = new Token(mPosition - 2, mPosition,
1.280 + Token::GREATER_OR_EQUAL_OP);
1.281 + }
1.282 + else {
1.283 + newToken = new Token(mPosition - 1, Token::GREATER_THAN_OP);
1.284 + }
1.285 + break;
1.286 + case HYPHEN :
1.287 + newToken = new Token(mPosition, Token::SUBTRACTION_OP);
1.288 + ++mPosition;
1.289 + break;
1.290 + case ASTERIX:
1.291 + if (nextIsOperatorToken(prevToken)) {
1.292 + newToken = new Token(mPosition, Token::MULTIPLY_OP);
1.293 + }
1.294 + else {
1.295 + newToken = new Token(mPosition, Token::CNAME);
1.296 + }
1.297 + ++mPosition;
1.298 + break;
1.299 + case L_PAREN:
1.300 + if (prevToken->mType == Token::CNAME) {
1.301 + const nsDependentSubstring& val = prevToken->Value();
1.302 + if (val.EqualsLiteral("comment")) {
1.303 + prevToken->mType = Token::COMMENT_AND_PAREN;
1.304 + }
1.305 + else if (val.EqualsLiteral("node")) {
1.306 + prevToken->mType = Token::NODE_AND_PAREN;
1.307 + }
1.308 + else if (val.EqualsLiteral("processing-instruction")) {
1.309 + prevToken->mType = Token::PROC_INST_AND_PAREN;
1.310 + }
1.311 + else if (val.EqualsLiteral("text")) {
1.312 + prevToken->mType = Token::TEXT_AND_PAREN;
1.313 + }
1.314 + else {
1.315 + prevToken->mType = Token::FUNCTION_NAME_AND_PAREN;
1.316 + }
1.317 + isToken = false;
1.318 + }
1.319 + else {
1.320 + newToken = new Token(mPosition, Token::L_PAREN);
1.321 + }
1.322 + ++mPosition;
1.323 + break;
1.324 + case R_PAREN:
1.325 + newToken = new Token(mPosition, Token::R_PAREN);
1.326 + ++mPosition;
1.327 + break;
1.328 + case L_BRACKET:
1.329 + newToken = new Token(mPosition, Token::L_BRACKET);
1.330 + ++mPosition;
1.331 + break;
1.332 + case R_BRACKET:
1.333 + newToken = new Token(mPosition, Token::R_BRACKET);
1.334 + ++mPosition;
1.335 + break;
1.336 + case COMMA:
1.337 + newToken = new Token(mPosition, Token::COMMA);
1.338 + ++mPosition;
1.339 + break;
1.340 + case AT_SIGN :
1.341 + newToken = new Token(mPosition, Token::AT_SIGN);
1.342 + ++mPosition;
1.343 + break;
1.344 + case PLUS:
1.345 + newToken = new Token(mPosition, Token::ADDITION_OP);
1.346 + ++mPosition;
1.347 + break;
1.348 + case VERT_BAR:
1.349 + newToken = new Token(mPosition, Token::UNION_OP);
1.350 + ++mPosition;
1.351 + break;
1.352 + default:
1.353 + // Error, don't grok character :-(
1.354 + return NS_ERROR_XPATH_ILLEGAL_CHAR;
1.355 + }
1.356 + }
1.357 + if (isToken) {
1.358 + NS_ENSURE_TRUE(newToken, NS_ERROR_OUT_OF_MEMORY);
1.359 + NS_ENSURE_TRUE(newToken != mLastItem, NS_ERROR_FAILURE);
1.360 + prevToken = newToken;
1.361 + addToken(newToken);
1.362 + }
1.363 + }
1.364 +
1.365 + // add a endToken to the list
1.366 + newToken = new Token(end, end, Token::END);
1.367 + if (!newToken) {
1.368 + return NS_ERROR_OUT_OF_MEMORY;
1.369 + }
1.370 + addToken(newToken);
1.371 +
1.372 + return NS_OK;
1.373 +}
