DXR is a code search and navigation tool aimed at making sense of large projects. It supports full-text and regex searches as well as structural queries.

Implementation

Mercurial (fddffdeab170)

VCS Links

Line Code
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332
/* -*- Mode: C++; tab-width: 8; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
/* vim: set ts=8 sts=2 et sw=2 tw=80: */
/* This Source Code Form is subject to the terms of the Mozilla Public
 * License, v. 2.0. If a copy of the MPL was not distributed with this
 * file, You can obtain one at http://mozilla.org/MPL/2.0/. */

#ifndef Tokenizer_h__
#define Tokenizer_h__

#include "nsString.h"
#include "mozilla/CheckedInt.h"

namespace mozilla {

/**
 * This is a simple implementation of a lexical analyzer or maybe better
 * called a tokenizer.  It doesn't allow any user dictionaries or
 * user define token types.
 *
 * It is limited only to ASCII input for now. UTF-8 or any other input
 * encoding must yet be implemented.
 */
class Tokenizer {
public:
  /**
   * The analyzer works with elements in the input cut to a sequence of token
   * where each token has an elementary type
   */
  enum TokenType {
    TOKEN_UNKNOWN,
    TOKEN_ERROR,
    TOKEN_INTEGER,
    TOKEN_WORD,
    TOKEN_CHAR,
    TOKEN_WS,
    TOKEN_EOL,
    TOKEN_EOF
  };

  /**
   * Class holding the type and the value of a token.  It can be manually created
   * to allow checks against it via methods of Tokenizer or are results of some of
   * the Tokenizer's methods.
   */
  class Token {
    TokenType mType;
    nsDependentCSubstring mWord;
    char mChar;
    uint64_t mInteger;

    // If this token is a result of the parsing process, this member is referencing
    // a sub-string in the input buffer.  If this is externally created Token this
    // member is left an empty string.
    nsDependentCSubstring mFragment;

    friend class Tokenizer;
    void AssignFragment(nsACString::const_char_iterator begin,
                        nsACString::const_char_iterator end);

  public:
    Token() : mType(TOKEN_UNKNOWN), mChar(0), mInteger(0) {}
    Token(const Token& aOther);
    Token& operator=(const Token& aOther);

    // Static constructors of tokens by type and value
    static Token Word(const nsACString& aWord);
    static Token Char(const char aChar);
    static Token Number(const uint64_t aNumber);
    static Token Whitespace();
    static Token NewLine();
    static Token EndOfFile();
    static Token Error();

    // Compares the two tokens, type must be identical and value
    // of one of the tokens must be 'any' or equal.
    bool Equals(const Token& aOther) const;

    TokenType Type() const { return mType; }
    char AsChar() const;
    nsDependentCSubstring AsString() const;
    uint64_t AsInteger() const;

    nsDependentCSubstring Fragment() const { return mFragment; }
  };

public:
  /**
   * @param aSource
   *    The string to parse.
   *    IMPORTANT NOTE: Tokenizer doesn't ensure the input string buffer lifetime.
   *    It's up to the consumer to make sure the string's buffer outlives the Tokenizer!
   * @param aWhitespaces
   *    If non-null Tokenizer will use this custom set of whitespaces for CheckWhite()
   *    and SkipWhites() calls.
   *    By default the list consists of space and tab.
   * @param aAdditionalWordChars
   *    If non-null it will be added to the list of characters that consist a word.
   *    This is useful when you want to accept e.g. '-' in HTTP headers.
   *    By default a word character is consider any character for which upper case
   *    is different from lower case.
   *
   * If there is an overlap between aWhitespaces and aAdditionalWordChars, the check for
   * word characters is made first.
   */
  explicit Tokenizer(const nsACString& aSource,
                     const char* aWhitespaces = nullptr,
                     const char* aAdditionalWordChars = nullptr);
  explicit Tokenizer(const char* aSource,
                     const char* aWhitespaces = nullptr,
                     const char* aAdditionalWordChars = nullptr);

  /**
   * When there is still anything to read from the input, tokenize it, store the token type
   * and value to aToken result and shift the cursor past this just parsed token.  Each call
   * to Next() reads another token from the input and shifts the cursor.
   * Returns false if we have passed the end of the input.
   */
  MOZ_WARN_UNUSED_RESULT
  bool Next(Token& aToken);

  /**
   * Parse the token on the input read cursor position, check its type is equal to aTokenType
   * and if so, put it into aResult, shift the cursor and return true.  Otherwise, leave
   * the input read cursor position intact and return false.
   */
  MOZ_WARN_UNUSED_RESULT
  bool Check(const TokenType aTokenType, Token& aResult);
  /**
   * Same as above method, just compares both token type and token value passed in aToken.
   * When both the type and the value equals, shift the cursor and return true.  Otherwise
   * return false.
   */
  MOZ_WARN_UNUSED_RESULT
  bool Check(const Token& aToken);

  /**
   * Return false iff the last Check*() call has returned false or when we've read past
   * the end of the input string.
   */
  MOZ_WARN_UNUSED_RESULT
  bool HasFailed() const;

  /**
   * SkipWhites method (below) may also skip new line characters automatically.
   */
  enum WhiteSkipping {
    /**
     * SkipWhites will only skip what is defined as a white space (default).
     */
    DONT_INCLUDE_NEW_LINE = 0,
    /**
     * SkipWhites will skip definited white spaces as well as new lines
     * automatically.
     */
    INCLUDE_NEW_LINE = 1
  };

  /**
   * Skips any occurence of whitespaces specified in mWhitespaces member,
   * optionally skip also new lines.
   */
  void SkipWhites(WhiteSkipping aIncludeNewLines = DONT_INCLUDE_NEW_LINE);

  // These are mostly shortcuts for the Check() methods above.

  /**
   * Check whitespace character is present.
   */
  MOZ_WARN_UNUSED_RESULT
  bool CheckWhite() { return Check(Token::Whitespace()); }
  /**
   * Check there is a single character on the read cursor position.  If so, shift the read
   * cursor position and return true.  Otherwise false.
   */
  MOZ_WARN_UNUSED_RESULT
  bool CheckChar(const char aChar) { return Check(Token::Char(aChar)); }
  /**
   * This is a customizable version of CheckChar.  aClassifier is a function called with
   * value of the character on the current input read position.  If this user function
   * returns true, read cursor is shifted and true returned.  Otherwise false.
   * The user classifiction function is not called when we are at or past the end and
   * false is immediately returned.
   */
  MOZ_WARN_UNUSED_RESULT
  bool CheckChar(bool (*aClassifier)(const char aChar));
  /**
   * Check for a whole expected word.
   */
  MOZ_WARN_UNUSED_RESULT
  bool CheckWord(const nsACString& aWord) { return Check(Token::Word(aWord)); }
  /**
   * Shortcut for literal const word check with compile time length calculation.
   */
  template <uint32_t N>
  MOZ_WARN_UNUSED_RESULT
  bool CheckWord(const char (&aWord)[N]) { return Check(Token::Word(nsDependentCString(aWord, N - 1))); }
  /**
   * Checks \r, \n or \r\n.
   */
  MOZ_WARN_UNUSED_RESULT
  bool CheckEOL() { return Check(Token::NewLine()); }
  /**
   * Checks we are at the end of the input string reading.  If so, shift past the end
   * and returns true.  Otherwise does nothing and returns false.
   */
  MOZ_WARN_UNUSED_RESULT
  bool CheckEOF() { return Check(Token::EndOfFile()); }

  /**
   * These are shortcuts to obtain the value immediately when the token type matches.
   */
  bool ReadChar(char* aValue);
  bool ReadChar(bool (*aClassifier)(const char aChar), char* aValue);
  bool ReadWord(nsACString& aValue);
  bool ReadWord(nsDependentCSubstring& aValue);

  /**
   * This is an integer read helper.  It returns false and doesn't move the read
   * cursor when any of the following happens:
   *  - the token at the read cursor is not an integer
   *  - the final number doesn't fit the T type
   * Otherwise true is returned, aValue is filled with the integral number
   * and the cursor is moved forward.
   */
  template <typename T>
  bool ReadInteger(T* aValue)
  {
    MOZ_RELEASE_ASSERT(aValue);

    nsACString::const_char_iterator rollback = mRollback;
    nsACString::const_char_iterator cursor = mCursor;
    Token t;
    if (!Check(TOKEN_INTEGER, t)) {
      return false;
    }

    mozilla::CheckedInt<T> checked(t.AsInteger());
    if (!checked.isValid()) {
      // Move to a state as if Check() call has failed
      mRollback = rollback;
      mCursor = cursor;
      mHasFailed = true;
      return false;
    }

    *aValue = checked.value();
    return true;
  }

  /**
   * Returns the read cursor position back as it was before the last call of any parsing
   * method of Tokenizer (Next, Check*, Skip*) so that the last operation can be repeated.
   * Rollback cannot be used multiple times, it only reverts the last successfull parse
   * operation.  It also cannot be used before any parsing operation has been called
   * on the Tokenizer.
   */
  void Rollback();

  /**
   * Record() and Claim() are collecting the input as it is being parsed to obtain
   * a substring between particular syntax bounderies defined by any recursive
   * descent parser or simple parser the Tokenizer is used to read the input for.
   * Inlucsion of a token that has just been parsed can be controlled using an arguemnt.
   */
  enum ClaimInclusion {
    /**
     * Include resulting (or passed) token of the last lexical analyzer operation in the result.
     */
    INCLUDE_LAST,
    /**
     * Do not include it.
     */
    EXCLUDE_LAST
  };

  /**
   * Start the process of recording.  Based on aInclude value the begining of the recorded
   * sub-string is at the current position (EXCLUDE_LAST) or at the position before the last
   * parsed token (INCLUDE_LAST).
   */
  void Record(ClaimInclusion aInclude = EXCLUDE_LAST);
  /**
   * Claim result of the record started with Record() call before.  Depending on aInclude
   * the ending of the sub-string result includes or excludes the last parsed or checked
   * token.
   */
  void Claim(nsACString& aResult, ClaimInclusion aInclude = EXCLUDE_LAST);
  void Claim(nsDependentCSubstring& aResult, ClaimInclusion aInclude = EXCLUDE_LAST);

protected:
  // false if we have already read the EOF token.
  bool HasInput() const;
  // Main parsing function, it doesn't shift the read cursor, just returns the next
  // token position.
  nsACString::const_char_iterator Parse(Token& aToken) const;
  // Is read cursor at the end?
  bool IsEnd(const nsACString::const_char_iterator& caret) const;
  // Is read cursor on a character that is a word start?
  bool IsWordFirst(const char aInput) const;
  // Is read cursor on a character that is an in-word letter?
  bool IsWord(const char aInput) const;
  // Is read cursor on a character that is a valid number?
  // TODO - support multiple radix
  bool IsNumber(const char aInput) const;

  // true iff we have already read the EOF token
  bool mPastEof;
  // true iff the last Check*() call has returned false, reverts to true on Rollback() call
  bool mHasFailed;

  // Customizable list of whitespaces
  const char* mWhitespaces;
  // Additinal custom word characters
  const char* mAdditionalWordChars;

  // All these point to the original buffer passed to the Tokenizer
  nsACString::const_char_iterator mRecord; // Position where the recorded sub-string for Claim() is
  nsACString::const_char_iterator mRollback; // Position of the previous token start
  nsACString::const_char_iterator mCursor; // Position of the current (actually next to read) token start
  nsACString::const_char_iterator mEnd; // End of the input position

private:
  Tokenizer() = delete;
  Tokenizer(const Tokenizer&) = delete;
  Tokenizer(Tokenizer&&) = delete;
  Tokenizer(const Tokenizer&&) = delete;
  Tokenizer &operator=(const Tokenizer&) = delete;
};

} // mozilla

#endif // Tokenizer_h__