1589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch// Copyright 2011 the V8 project authors. All rights reserved.
2589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch// Redistribution and use in source and binary forms, with or without
3589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch// modification, are permitted provided that the following conditions are
4589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch// met:
5589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch//
6589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch//     * Redistributions of source code must retain the above copyright
7589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch//       notice, this list of conditions and the following disclaimer.
8589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch//     * Redistributions in binary form must reproduce the above
9589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch//       copyright notice, this list of conditions and the following
10589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch//       disclaimer in the documentation and/or other materials provided
11589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch//       with the distribution.
12589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch//     * Neither the name of Google Inc. nor the names of its
13589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch//       contributors may be used to endorse or promote products derived
14589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch//       from this software without specific prior written permission.
15589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch//
16589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
17589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch// "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
18589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch// LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
19589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch// A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
20589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch// OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
21589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch// SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
22589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch// LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
23589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch// DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
24589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch// THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
25589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch// (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
26589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch// OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
27589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch
28589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch#ifndef V8_SCANNER_CHARACTER_STREAMS_H_
29589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch#define V8_SCANNER_CHARACTER_STREAMS_H_
30589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch
31589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch#include "scanner.h"
32589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch
33589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdochnamespace v8 {
34589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdochnamespace internal {
35589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch
36589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch// A buffered character stream based on a random access character
37589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch// source (ReadBlock can be called with pos_ pointing to any position,
38589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch// even positions before the current).
393ef787dbeca8a5fb1086949cda830dccee07bfbdBen Murdochclass BufferedUtf16CharacterStream: public Utf16CharacterStream {
40589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch public:
413ef787dbeca8a5fb1086949cda830dccee07bfbdBen Murdoch  BufferedUtf16CharacterStream();
423ef787dbeca8a5fb1086949cda830dccee07bfbdBen Murdoch  virtual ~BufferedUtf16CharacterStream();
43589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch
44589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch  virtual void PushBack(uc32 character);
45589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch
46589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch protected:
47589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch  static const unsigned kBufferSize = 512;
48589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch  static const unsigned kPushBackStepSize = 16;
49589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch
50589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch  virtual unsigned SlowSeekForward(unsigned delta);
51589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch  virtual bool ReadBlock();
52589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch  virtual void SlowPushBack(uc16 character);
53589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch
54589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch  virtual unsigned BufferSeekForward(unsigned delta) = 0;
55589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch  virtual unsigned FillBuffer(unsigned position, unsigned length) = 0;
56589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch
57589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch  const uc16* pushback_limit_;
58589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch  uc16 buffer_[kBufferSize];
59589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch};
60589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch
61589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch
62589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch// Generic string stream.
633ef787dbeca8a5fb1086949cda830dccee07bfbdBen Murdochclass GenericStringUtf16CharacterStream: public BufferedUtf16CharacterStream {
64589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch public:
653ef787dbeca8a5fb1086949cda830dccee07bfbdBen Murdoch  GenericStringUtf16CharacterStream(Handle<String> data,
663ef787dbeca8a5fb1086949cda830dccee07bfbdBen Murdoch                                    unsigned start_position,
673ef787dbeca8a5fb1086949cda830dccee07bfbdBen Murdoch                                    unsigned end_position);
683ef787dbeca8a5fb1086949cda830dccee07bfbdBen Murdoch  virtual ~GenericStringUtf16CharacterStream();
69589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch
70589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch protected:
71589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch  virtual unsigned BufferSeekForward(unsigned delta);
72589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch  virtual unsigned FillBuffer(unsigned position, unsigned length);
73589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch
74589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch  Handle<String> string_;
75589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch  unsigned start_position_;
76589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch  unsigned length_;
77589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch};
78589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch
79589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch
803ef787dbeca8a5fb1086949cda830dccee07bfbdBen Murdoch// Utf16 stream based on a literal UTF-8 string.
813ef787dbeca8a5fb1086949cda830dccee07bfbdBen Murdochclass Utf8ToUtf16CharacterStream: public BufferedUtf16CharacterStream {
82589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch public:
833ef787dbeca8a5fb1086949cda830dccee07bfbdBen Murdoch  Utf8ToUtf16CharacterStream(const byte* data, unsigned length);
843ef787dbeca8a5fb1086949cda830dccee07bfbdBen Murdoch  virtual ~Utf8ToUtf16CharacterStream();
85589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch
86589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch protected:
87589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch  virtual unsigned BufferSeekForward(unsigned delta);
88589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch  virtual unsigned FillBuffer(unsigned char_position, unsigned length);
89589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch  void SetRawPosition(unsigned char_position);
90589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch
91589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch  const byte* raw_data_;
92589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch  unsigned raw_data_length_;  // Measured in bytes, not characters.
93589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch  unsigned raw_data_pos_;
94589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch  // The character position of the character at raw_data[raw_data_pos_].
95589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch  // Not necessarily the same as pos_.
96589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch  unsigned raw_character_position_;
97589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch};
98589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch
99589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch
100589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch// UTF16 buffer to read characters from an external string.
1013ef787dbeca8a5fb1086949cda830dccee07bfbdBen Murdochclass ExternalTwoByteStringUtf16CharacterStream: public Utf16CharacterStream {
102589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch public:
1033ef787dbeca8a5fb1086949cda830dccee07bfbdBen Murdoch  ExternalTwoByteStringUtf16CharacterStream(Handle<ExternalTwoByteString> data,
1043ef787dbeca8a5fb1086949cda830dccee07bfbdBen Murdoch                                            int start_position,
1053ef787dbeca8a5fb1086949cda830dccee07bfbdBen Murdoch                                            int end_position);
1063ef787dbeca8a5fb1086949cda830dccee07bfbdBen Murdoch  virtual ~ExternalTwoByteStringUtf16CharacterStream();
107589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch
108589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch  virtual void PushBack(uc32 character) {
109589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch    ASSERT(buffer_cursor_ > raw_data_);
110589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch    buffer_cursor_--;
111589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch    pos_--;
112589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch  }
113589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch
114589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch protected:
115589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch  virtual unsigned SlowSeekForward(unsigned delta) {
116589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch    // Fast case always handles seeking.
117589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch    return 0;
118589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch  }
119589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch  virtual bool ReadBlock() {
120589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch    // Entire string is read at start.
121589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch    return false;
122589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch  }
123589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch  Handle<ExternalTwoByteString> source_;
124589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch  const uc16* raw_data_;  // Pointer to the actual array of characters.
125589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch};
126589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch
127589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch} }  // namespace v8::internal
128589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch
129589d6979ff2ef66fca2d8fa51404c369ca5e9250Ben Murdoch#endif  // V8_SCANNER_CHARACTER_STREAMS_H_
130