dom/xslt/xpath/txExprLexer.cpp@97036ab72558
dom/xslt/xpath/txExprLexer.cpp
Tue, 06 Jan 2015 21:39:09 +0100
- author
- Michael Schloh von Bennewitz <michael@schloh.com>
- date
- Tue, 06 Jan 2015 21:39:09 +0100
- branch
- TOR_BUG_9701
- changeset 8
- 97036ab72558
- permissions
- -rw-r--r--
Conditionally force memory storage according to privacy.thirdparty.isolate;
This solves Tor bug #9701, complying with disk avoidance documented in
https://www.torproject.org/projects/torbrowser/design/#disk-avoidance.
1 /* -*- Mode: C++; tab-width: 2; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
2 /* This Source Code Form is subject to the terms of the Mozilla Public
3 * License, v. 2.0. If a copy of the MPL was not distributed with this
4 * file, You can obtain one at http://mozilla.org/MPL/2.0/. */
6 /**
7 * Lexical analyzer for XPath expressions
8 */
10 #include "txExprLexer.h"
11 #include "nsGkAtoms.h"
12 #include "nsString.h"
13 #include "nsError.h"
14 #include "txXMLUtils.h"
16 /**
17 * Creates a new ExprLexer
18 */
19 txExprLexer::txExprLexer()
20 : mCurrentItem(nullptr),
21 mFirstItem(nullptr),
22 mLastItem(nullptr),
23 mTokenCount(0)
24 {
25 }
27 /**
28 * Destroys this instance of an txExprLexer
29 */
30 txExprLexer::~txExprLexer()
31 {
32 //-- delete tokens
33 Token* tok = mFirstItem;
34 while (tok) {
35 Token* temp = tok->mNext;
36 delete tok;
37 tok = temp;
38 }
39 mCurrentItem = nullptr;
40 }
42 Token*
43 txExprLexer::nextToken()
44 {
45 if (!mCurrentItem) {
46 NS_NOTREACHED("nextToken called on uninitialized lexer");
47 return nullptr;
48 }
50 if (mCurrentItem->mType == Token::END) {
51 // Do not progress beyond the end token
52 return mCurrentItem;
53 }
55 Token* token = mCurrentItem;
56 mCurrentItem = mCurrentItem->mNext;
57 return token;
58 }
60 void
61 txExprLexer::addToken(Token* aToken)
62 {
63 if (mLastItem) {
64 mLastItem->mNext = aToken;
65 }
66 if (!mFirstItem) {
67 mFirstItem = aToken;
68 mCurrentItem = aToken;
69 }
70 mLastItem = aToken;
71 ++mTokenCount;
72 }
74 /**
75 * Returns true if the following Token should be an operator.
76 * This is a helper for the first bullet of [XPath 3.7]
77 * Lexical Structure
78 */
79 bool
80 txExprLexer::nextIsOperatorToken(Token* aToken)
81 {
82 if (!aToken || aToken->mType == Token::NULL_TOKEN) {
83 return false;
84 }
85 /* This relies on the tokens having the right order in txExprLexer.h */
86 return aToken->mType < Token::COMMA ||
87 aToken->mType > Token::UNION_OP;
89 }
91 /**
92 * Parses the given string into a sequence of Tokens
93 */
94 nsresult
95 txExprLexer::parse(const nsASingleFragmentString& aPattern)
96 {
97 iterator start, end;
98 start = aPattern.BeginReading(mPosition);
99 aPattern.EndReading(end);
101 //-- initialize previous token, this will automatically get
102 //-- deleted when it goes out of scope
103 Token nullToken(nullptr, nullptr, Token::NULL_TOKEN);
105 Token::Type defType;
106 Token* newToken = nullptr;
107 Token* prevToken = &nullToken;
108 bool isToken;
110 while (mPosition < end) {
112 defType = Token::CNAME;
113 isToken = true;
115 if (*mPosition == DOLLAR_SIGN) {
116 if (++mPosition == end || !XMLUtils::isLetter(*mPosition)) {
117 return NS_ERROR_XPATH_INVALID_VAR_NAME;
118 }
119 defType = Token::VAR_REFERENCE;
120 }
121 // just reuse the QName parsing, which will use defType
122 // the token to construct
124 if (XMLUtils::isLetter(*mPosition)) {
125 // NCName, can get QName or OperatorName;
126 // FunctionName, NodeName, and AxisSpecifier may want whitespace,
127 // and are dealt with below
128 start = mPosition;
129 while (++mPosition < end && XMLUtils::isNCNameChar(*mPosition)) {
130 /* just go */
131 }
132 if (mPosition < end && *mPosition == COLON) {
133 // try QName or wildcard, might need to step back for axis
134 if (++mPosition == end) {
135 return NS_ERROR_XPATH_UNEXPECTED_END;
136 }
137 if (XMLUtils::isLetter(*mPosition)) {
138 while (++mPosition < end && XMLUtils::isNCNameChar(*mPosition)) {
139 /* just go */
140 }
141 }
142 else if (*mPosition == '*' && defType != Token::VAR_REFERENCE) {
143 // eat wildcard for NameTest, bail for var ref at COLON
144 ++mPosition;
145 }
146 else {
147 --mPosition; // step back
148 }
149 }
150 if (nextIsOperatorToken(prevToken)) {
151 nsDependentSubstring op(Substring(start, mPosition));
152 if (nsGkAtoms::_and->Equals(op)) {
153 defType = Token::AND_OP;
154 }
155 else if (nsGkAtoms::_or->Equals(op)) {
156 defType = Token::OR_OP;
157 }
158 else if (nsGkAtoms::mod->Equals(op)) {
159 defType = Token::MODULUS_OP;
160 }
161 else if (nsGkAtoms::div->Equals(op)) {
162 defType = Token::DIVIDE_OP;
163 }
164 else {
165 // XXX QUESTION: spec is not too precise
166 // badops is sure an error, but is bad:ops, too? We say yes!
167 return NS_ERROR_XPATH_OPERATOR_EXPECTED;
168 }
169 }
170 newToken = new Token(start, mPosition, defType);
171 }
172 else if (isXPathDigit(*mPosition)) {
173 start = mPosition;
174 while (++mPosition < end && isXPathDigit(*mPosition)) {
175 /* just go */
176 }
177 if (mPosition < end && *mPosition == '.') {
178 while (++mPosition < end && isXPathDigit(*mPosition)) {
179 /* just go */
180 }
181 }
182 newToken = new Token(start, mPosition, Token::NUMBER);
183 }
184 else {
185 switch (*mPosition) {
186 //-- ignore whitespace
187 case SPACE:
188 case TX_TAB:
189 case TX_CR:
190 case TX_LF:
191 ++mPosition;
192 isToken = false;
193 break;
194 case S_QUOTE :
195 case D_QUOTE :
196 start = mPosition;
197 while (++mPosition < end && *mPosition != *start) {
198 // eat literal
199 }
200 if (mPosition == end) {
201 mPosition = start;
202 return NS_ERROR_XPATH_UNCLOSED_LITERAL;
203 }
204 newToken = new Token(start + 1, mPosition, Token::LITERAL);
205 ++mPosition;
206 break;
207 case PERIOD:
208 // period can be .., .(DIGITS)+ or ., check next
209 if (++mPosition == end) {
210 newToken = new Token(mPosition - 1, Token::SELF_NODE);
211 }
212 else if (isXPathDigit(*mPosition)) {
213 start = mPosition - 1;
214 while (++mPosition < end && isXPathDigit(*mPosition)) {
215 /* just go */
216 }
217 newToken = new Token(start, mPosition, Token::NUMBER);
218 }
219 else if (*mPosition == PERIOD) {
220 ++mPosition;
221 newToken = new Token(mPosition - 2, mPosition, Token::PARENT_NODE);
222 }
223 else {
224 newToken = new Token(mPosition - 1, Token::SELF_NODE);
225 }
226 break;
227 case COLON: // QNames are dealt above, must be axis ident
228 if (++mPosition >= end || *mPosition != COLON ||
229 prevToken->mType != Token::CNAME) {
230 return NS_ERROR_XPATH_BAD_COLON;
231 }
232 prevToken->mType = Token::AXIS_IDENTIFIER;
233 ++mPosition;
234 isToken = false;
235 break;
236 case FORWARD_SLASH :
237 if (++mPosition < end && *mPosition == FORWARD_SLASH) {
238 ++mPosition;
239 newToken = new Token(mPosition - 2, mPosition, Token::ANCESTOR_OP);
240 }
241 else {
242 newToken = new Token(mPosition - 1, Token::PARENT_OP);
243 }
244 break;
245 case BANG : // can only be !=
246 if (++mPosition < end && *mPosition == EQUAL) {
247 ++mPosition;
248 newToken = new Token(mPosition - 2, mPosition, Token::NOT_EQUAL_OP);
249 break;
250 }
251 // Error ! is not not()
252 return NS_ERROR_XPATH_BAD_BANG;
253 case EQUAL:
254 newToken = new Token(mPosition, Token::EQUAL_OP);
255 ++mPosition;
256 break;
257 case L_ANGLE:
258 if (++mPosition == end) {
259 return NS_ERROR_XPATH_UNEXPECTED_END;
260 }
261 if (*mPosition == EQUAL) {
262 ++mPosition;
263 newToken = new Token(mPosition - 2, mPosition,
264 Token::LESS_OR_EQUAL_OP);
265 }
266 else {
267 newToken = new Token(mPosition - 1, Token::LESS_THAN_OP);
268 }
269 break;
270 case R_ANGLE:
271 if (++mPosition == end) {
272 return NS_ERROR_XPATH_UNEXPECTED_END;
273 }
274 if (*mPosition == EQUAL) {
275 ++mPosition;
276 newToken = new Token(mPosition - 2, mPosition,
277 Token::GREATER_OR_EQUAL_OP);
278 }
279 else {
280 newToken = new Token(mPosition - 1, Token::GREATER_THAN_OP);
281 }
282 break;
283 case HYPHEN :
284 newToken = new Token(mPosition, Token::SUBTRACTION_OP);
285 ++mPosition;
286 break;
287 case ASTERIX:
288 if (nextIsOperatorToken(prevToken)) {
289 newToken = new Token(mPosition, Token::MULTIPLY_OP);
290 }
291 else {
292 newToken = new Token(mPosition, Token::CNAME);
293 }
294 ++mPosition;
295 break;
296 case L_PAREN:
297 if (prevToken->mType == Token::CNAME) {
298 const nsDependentSubstring& val = prevToken->Value();
299 if (val.EqualsLiteral("comment")) {
300 prevToken->mType = Token::COMMENT_AND_PAREN;
301 }
302 else if (val.EqualsLiteral("node")) {
303 prevToken->mType = Token::NODE_AND_PAREN;
304 }
305 else if (val.EqualsLiteral("processing-instruction")) {
306 prevToken->mType = Token::PROC_INST_AND_PAREN;
307 }
308 else if (val.EqualsLiteral("text")) {
309 prevToken->mType = Token::TEXT_AND_PAREN;
310 }
311 else {
312 prevToken->mType = Token::FUNCTION_NAME_AND_PAREN;
313 }
314 isToken = false;
315 }
316 else {
317 newToken = new Token(mPosition, Token::L_PAREN);
318 }
319 ++mPosition;
320 break;
321 case R_PAREN:
322 newToken = new Token(mPosition, Token::R_PAREN);
323 ++mPosition;
324 break;
325 case L_BRACKET:
326 newToken = new Token(mPosition, Token::L_BRACKET);
327 ++mPosition;
328 break;
329 case R_BRACKET:
330 newToken = new Token(mPosition, Token::R_BRACKET);
331 ++mPosition;
332 break;
333 case COMMA:
334 newToken = new Token(mPosition, Token::COMMA);
335 ++mPosition;
336 break;
337 case AT_SIGN :
338 newToken = new Token(mPosition, Token::AT_SIGN);
339 ++mPosition;
340 break;
341 case PLUS:
342 newToken = new Token(mPosition, Token::ADDITION_OP);
343 ++mPosition;
344 break;
345 case VERT_BAR:
346 newToken = new Token(mPosition, Token::UNION_OP);
347 ++mPosition;
348 break;
349 default:
350 // Error, don't grok character :-(
351 return NS_ERROR_XPATH_ILLEGAL_CHAR;
352 }
353 }
354 if (isToken) {
355 NS_ENSURE_TRUE(newToken, NS_ERROR_OUT_OF_MEMORY);
356 NS_ENSURE_TRUE(newToken != mLastItem, NS_ERROR_FAILURE);
357 prevToken = newToken;
358 addToken(newToken);
359 }
360 }
362 // add a endToken to the list
363 newToken = new Token(end, end, Token::END);
364 if (!newToken) {
365 return NS_ERROR_OUT_OF_MEMORY;
366 }
367 addToken(newToken);
369 return NS_OK;
370 }
