michael@0: // -*- mode: c++ -*-
michael@0: 
michael@0: // Copyright (c) 2010, Google Inc.
michael@0: // All rights reserved.
michael@0: //
michael@0: // Redistribution and use in source and binary forms, with or without
michael@0: // modification, are permitted provided that the following conditions are
michael@0: // met:
michael@0: //
michael@0: //     * Redistributions of source code must retain the above copyright
michael@0: // notice, this list of conditions and the following disclaimer.
michael@0: //     * Redistributions in binary form must reproduce the above
michael@0: // copyright notice, this list of conditions and the following disclaimer
michael@0: // in the documentation and/or other materials provided with the
michael@0: // distribution.
michael@0: //     * Neither the name of Google Inc. nor the names of its
michael@0: // contributors may be used to endorse or promote products derived from
michael@0: // this software without specific prior written permission.
michael@0: //
michael@0: // THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
michael@0: // "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
michael@0: // LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
michael@0: // A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
michael@0: // OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
michael@0: // SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
michael@0: // LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
michael@0: // DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
michael@0: // THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
michael@0: // (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
michael@0: // OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
michael@0: 
michael@0: // Original author: Jim Blandy <jimb@mozilla.com> <jimb@red-bean.com>
michael@0: 
michael@0: // byte_cursor.h: Classes for parsing values from a buffer of bytes.
michael@0: // The ByteCursor class provides a convenient interface for reading
michael@0: // fixed-size integers of arbitrary endianness, being thorough about
michael@0: // checking for buffer overruns.
michael@0: 
michael@0: #ifndef COMMON_BYTE_CURSOR_H_
michael@0: #define COMMON_BYTE_CURSOR_H_
michael@0: 
michael@0: #include <assert.h>
michael@0: #include <stdint.h>
michael@0: #include <stdlib.h>
michael@0: #include <string.h>
michael@0: #include <string>
michael@0: 
michael@0: #include "common/using_std_string.h"
michael@0: 
michael@0: namespace google_breakpad {
michael@0: 
michael@0: // A buffer holding a series of bytes.
michael@0: struct ByteBuffer {
michael@0:   ByteBuffer() : start(0), end(0) { }
michael@0:   ByteBuffer(const uint8_t *set_start, size_t set_size)
michael@0:       : start(set_start), end(set_start + set_size) { }
michael@0:   ~ByteBuffer() { };
michael@0: 
michael@0:   // Equality operators. Useful in unit tests, and when we're using
michael@0:   // ByteBuffers to refer to regions of a larger buffer.
michael@0:   bool operator==(const ByteBuffer &that) const {
michael@0:     return start == that.start && end == that.end;
michael@0:   }
michael@0:   bool operator!=(const ByteBuffer &that) const {
michael@0:     return start != that.start || end != that.end;
michael@0:   }
michael@0: 
michael@0:   // Not C++ style guide compliant, but this definitely belongs here.
michael@0:   size_t Size() const {
michael@0:     assert(start <= end);
michael@0:     return end - start;
michael@0:   }
michael@0: 
michael@0:   const uint8_t *start, *end;
michael@0: };
michael@0: 
michael@0: // A cursor pointing into a ByteBuffer that can parse numbers of various
michael@0: // widths and representations, strings, and data blocks, advancing through
michael@0: // the buffer as it goes. All ByteCursor operations check that accesses
michael@0: // haven't gone beyond the end of the enclosing ByteBuffer.
michael@0: class ByteCursor {
michael@0:  public:
michael@0:   // Create a cursor reading bytes from the start of BUFFER. By default, the
michael@0:   // cursor reads multi-byte values in little-endian form.
michael@0:   ByteCursor(const ByteBuffer *buffer, bool big_endian = false)
michael@0:       : buffer_(buffer), here_(buffer->start), 
michael@0:         big_endian_(big_endian), complete_(true) { }
michael@0: 
michael@0:   // Accessor and setter for this cursor's endianness flag.
michael@0:   bool big_endian() const { return big_endian_; }
michael@0:   void set_big_endian(bool big_endian) { big_endian_ = big_endian; }
michael@0: 
michael@0:   // Accessor and setter for this cursor's current position. The setter
michael@0:   // returns a reference to this cursor.
michael@0:   const uint8_t *here() const { return here_; }
michael@0:   ByteCursor &set_here(const uint8_t *here) {
michael@0:     assert(buffer_->start <= here && here <= buffer_->end);
michael@0:     here_ = here;
michael@0:     return *this;
michael@0:   }
michael@0: 
michael@0:   // Return the number of bytes available to read at the cursor.
michael@0:   size_t Available() const { return size_t(buffer_->end - here_); }
michael@0: 
michael@0:   // Return true if this cursor is at the end of its buffer.
michael@0:   bool AtEnd() const { return Available() == 0; }
michael@0: 
michael@0:   // When used as a boolean value this cursor converts to true if all
michael@0:   // prior reads have been completed, or false if we ran off the end
michael@0:   // of the buffer.
michael@0:   operator bool() const { return complete_; }
michael@0: 
michael@0:   // Read a SIZE-byte integer at this cursor, signed if IS_SIGNED is true,
michael@0:   // unsigned otherwise, using the cursor's established endianness, and set
michael@0:   // *RESULT to the number. If we read off the end of our buffer, clear
michael@0:   // this cursor's complete_ flag, and store a dummy value in *RESULT.
michael@0:   // Return a reference to this cursor.
michael@0:   template<typename T>
michael@0:   ByteCursor &Read(size_t size, bool is_signed, T *result) {
michael@0:     if (CheckAvailable(size)) {
michael@0:       T v = 0;
michael@0:       if (big_endian_) {
michael@0:         for (size_t i = 0; i < size; i++)
michael@0:           v = (v << 8) + here_[i];
michael@0:       } else {
michael@0:         // This loop condition looks weird, but size_t is unsigned, so
michael@0:         // decrementing i after it is zero yields the largest size_t value.
michael@0:         for (size_t i = size - 1; i < size; i--)
michael@0:           v = (v << 8) + here_[i];
michael@0:       }
michael@0:       if (is_signed && size < sizeof(T)) {
michael@0:         size_t sign_bit = (T)1 << (size * 8 - 1);
michael@0:         v = (v ^ sign_bit) - sign_bit;
michael@0:       }
michael@0:       here_ += size;
michael@0:       *result = v;
michael@0:     } else {
michael@0:       *result = (T) 0xdeadbeef;
michael@0:     }
michael@0:     return *this;
michael@0:   }
michael@0: 
michael@0:   // Read an integer, using the cursor's established endianness and
michael@0:   // *RESULT's size and signedness, and set *RESULT to the number. If we
michael@0:   // read off the end of our buffer, clear this cursor's complete_ flag.
michael@0:   // Return a reference to this cursor.
michael@0:   template<typename T>
michael@0:   ByteCursor &operator>>(T &result) {
michael@0:     bool T_is_signed = (T)-1 < 0;
michael@0:     return Read(sizeof(T), T_is_signed, &result); 
michael@0:   }
michael@0: 
michael@0:   // Copy the SIZE bytes at the cursor to BUFFER, and advance this
michael@0:   // cursor to the end of them. If we read off the end of our buffer,
michael@0:   // clear this cursor's complete_ flag, and set *POINTER to NULL.
michael@0:   // Return a reference to this cursor.
michael@0:   ByteCursor &Read(uint8_t *buffer, size_t size) {
michael@0:     if (CheckAvailable(size)) {
michael@0:       memcpy(buffer, here_, size);
michael@0:       here_ += size;
michael@0:     }
michael@0:     return *this;
michael@0:   }
michael@0: 
michael@0:   // Set STR to a copy of the '\0'-terminated string at the cursor. If the
michael@0:   // byte buffer does not contain a terminating zero, clear this cursor's
michael@0:   // complete_ flag, and set STR to the empty string. Return a reference to
michael@0:   // this cursor.
michael@0:   ByteCursor &CString(string *str) {
michael@0:     const uint8_t *end
michael@0:       = static_cast<const uint8_t *>(memchr(here_, '\0', Available()));
michael@0:     if (end) {
michael@0:       str->assign(reinterpret_cast<const char *>(here_), end - here_);
michael@0:       here_ = end + 1;
michael@0:     } else {
michael@0:       str->clear();
michael@0:       here_ = buffer_->end;
michael@0:       complete_ = false;
michael@0:     }
michael@0:     return *this;
michael@0:   }
michael@0: 
michael@0:   // Like CString(STR), but extract the string from a fixed-width buffer
michael@0:   // LIMIT bytes long, which may or may not contain a terminating '\0'
michael@0:   // byte. Specifically:
michael@0:   //
michael@0:   // - If there are not LIMIT bytes available at the cursor, clear the
michael@0:   //   cursor's complete_ flag and set STR to the empty string.
michael@0:   //
michael@0:   // - Otherwise, if the LIMIT bytes at the cursor contain any '\0'
michael@0:   //   characters, set *STR to a copy of the bytes before the first '\0',
michael@0:   //   and advance the cursor by LIMIT bytes.
michael@0:   //   
michael@0:   // - Otherwise, set *STR to a copy of those LIMIT bytes, and advance the
michael@0:   //   cursor by LIMIT bytes.
michael@0:   ByteCursor &CString(string *str, size_t limit) {
michael@0:     if (CheckAvailable(limit)) {
michael@0:       const uint8_t *end
michael@0:         = static_cast<const uint8_t *>(memchr(here_, '\0', limit));
michael@0:       if (end)
michael@0:         str->assign(reinterpret_cast<const char *>(here_), end - here_);
michael@0:       else
michael@0:         str->assign(reinterpret_cast<const char *>(here_), limit);
michael@0:       here_ += limit;
michael@0:     } else {
michael@0:       str->clear();
michael@0:     }
michael@0:     return *this;
michael@0:   }
michael@0: 
michael@0:   // Set *POINTER to point to the SIZE bytes at the cursor, and advance
michael@0:   // this cursor to the end of them. If SIZE is omitted, don't move the
michael@0:   // cursor. If we read off the end of our buffer, clear this cursor's
michael@0:   // complete_ flag, and set *POINTER to NULL. Return a reference to this
michael@0:   // cursor.
michael@0:   ByteCursor &PointTo(const uint8_t **pointer, size_t size = 0) {
michael@0:     if (CheckAvailable(size)) {
michael@0:       *pointer = here_;
michael@0:       here_ += size;
michael@0:     } else {
michael@0:       *pointer = NULL;
michael@0:     }
michael@0:     return *this;
michael@0:   }
michael@0: 
michael@0:   // Skip SIZE bytes at the cursor. If doing so would advance us off
michael@0:   // the end of our buffer, clear this cursor's complete_ flag, and
michael@0:   // set *POINTER to NULL. Return a reference to this cursor.
michael@0:   ByteCursor &Skip(size_t size) {
michael@0:     if (CheckAvailable(size))
michael@0:       here_ += size;
michael@0:     return *this;
michael@0:   }
michael@0: 
michael@0:  private:
michael@0:   // If there are at least SIZE bytes available to read from the buffer,
michael@0:   // return true. Otherwise, set here_ to the end of the buffer, set
michael@0:   // complete_ to false, and return false.
michael@0:   bool CheckAvailable(size_t size) {
michael@0:     if (Available() >= size) {
michael@0:       return true;
michael@0:     } else {
michael@0:       here_ = buffer_->end;
michael@0:       complete_ = false;
michael@0:       return false;
michael@0:     }
michael@0:   }
michael@0: 
michael@0:   // The buffer we're reading bytes from.
michael@0:   const ByteBuffer *buffer_;
michael@0: 
michael@0:   // The next byte within buffer_ that we'll read.
michael@0:   const uint8_t *here_;
michael@0: 
michael@0:   // True if we should read numbers in big-endian form; false if we
michael@0:   // should read in little-endian form.
michael@0:   bool big_endian_;
michael@0: 
michael@0:   // True if we've been able to read all we've been asked to.
michael@0:   bool complete_;
michael@0: };
michael@0: 
michael@0: }  // namespace google_breakpad
michael@0: 
michael@0: #endif  // COMMON_BYTE_CURSOR_H_