diff options
Diffstat (limited to 'src/headerparser.h')
| -rw-r--r-- | src/headerparser.h | 400 |
1 files changed, 400 insertions, 0 deletions
diff --git a/src/headerparser.h b/src/headerparser.h new file mode 100644 index 0000000..eb0ea8e --- /dev/null +++ b/src/headerparser.h @@ -0,0 +1,400 @@ +// Copyright 2008 Google Inc. +// Author: Lincoln Smith +// +// Licensed under the Apache License, Version 2.0 (the "License"); +// you may not use this file except in compliance with the License. +// You may obtain a copy of the License at +// +// http://www.apache.org/licenses/LICENSE-2.0 +// +// Unless required by applicable law or agreed to in writing, software +// distributed under the License is distributed on an "AS IS" BASIS, +// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +// See the License for the specific language governing permissions and +// limitations under the License. + +#ifndef OPEN_VCDIFF_HEADERPARSER_H_ +#define OPEN_VCDIFF_HEADERPARSER_H_ + +#include <config.h> +#include <stddef.h> // NULL +#include <stdint.h> // int32_t, uint32_t +#include "checksum.h" // VCDChecksum +#include "vcdiff_defs.h" // VCDiffResult + +namespace open_vcdiff { + +// This class contains a contiguous memory buffer with start and end pointers, +// as well as a position pointer which shows how much of the buffer has been +// parsed and how much remains. +// +// Because no virtual destructor is defined for ParseableChunk, a pointer to +// a child class of ParseableChunk must be destroyed using its specific type, +// rather than as a ParseableChunk*. +class ParseableChunk { + public: + ParseableChunk(const char* data_start, size_t data_size) { + SetDataBuffer(data_start, data_size); + } + + const char* End() const { return end_; } + + // The number of bytes remaining to be parsed. This is not necessarily the + // same as the initial size of the buffer; it changes with each call to + // Advance(). + size_t UnparsedSize() const { + return end_ - position_; + } + + // The number of bytes that have already been parsed. + size_t ParsedSize() const { + return position_ - start_; + } + + bool Empty() const { return 0 == UnparsedSize(); } + + // The start of the data remaining to be parsed. + const char* UnparsedData() const { return position_; } + + // Returns a pointer to the start of the data remaining to be parsed. + const char** UnparsedDataAddr() { return &position_; } + + // Moves the parsing position forward by number_of_bytes. + void Advance(size_t number_of_bytes); + + // Jumps the parsing position to a new location. + void SetPosition(const char* position); + + // Jumps the parsing position to the end of the data chunk. + void Finish() { + position_ = end_; + } + + // Jumps the parsing position so that there are now number_of_bytes + // bytes left to parse. This number should be smaller than the size of data + // to be parsed before the function was called. + void FinishExcept(size_t number_of_bytes); + + void SetDataBuffer(const char* data_start, size_t data_size) { + start_ = data_start; + end_ = data_start + data_size; + position_ = start_; + } + + private: + const char* start_; + const char* end_; + + // The current parsing position within the data chunk. + // Must always respect start_ <= position_ <= end_. + const char* position_; + + // Making these private avoids implicit copy constructor & assignment operator + ParseableChunk(const ParseableChunk&); + void operator=(const ParseableChunk&); +}; + +// Represents one of the three sections in the delta window, as described in +// RFC section 4.3: +// * Data section for ADDs and RUNs +// * Instructions and sizes section +// * Addresses section for COPYs +// When using the interleaved format, data and addresses are pulled from the +// instructions and sizes section rather than being stored in separate sections. +// For that reason, this class allows one DeltaWindowSection to be based on +// another, such that the same position pointer is shared by both sections; +// i.e., UnparsedDataAddr() returns the same value for both objects. +// To achieve this end, one extra level of indirection (a pointer to a +// ParseableChunk object) is added. +class DeltaWindowSection { + public: + DeltaWindowSection() : parseable_chunk_(NULL), owned_(true) { } + + ~DeltaWindowSection() { + FreeChunk(); + } + + void Init(const char* data_start, size_t data_size) { + if (owned_ && parseable_chunk_) { + // Reuse the already-allocated ParseableChunk object. + parseable_chunk_->SetDataBuffer(data_start, data_size); + } else { + parseable_chunk_ = new ParseableChunk(data_start, data_size); + owned_ = true; + } + } + + void Init(DeltaWindowSection* original) { + FreeChunk(); + parseable_chunk_ = original->parseable_chunk_; + owned_ = false; + } + + void Invalidate() { FreeChunk(); } + + bool IsOwned() const { return owned_; } + + // The following functions just pass their arguments to the underlying + // ParseableChunk object. + + const char* End() const { + return parseable_chunk_->End(); + } + + size_t UnparsedSize() const { + return parseable_chunk_->UnparsedSize(); + } + + size_t ParsedSize() const { + return parseable_chunk_->ParsedSize(); + } + + bool Empty() const { + return parseable_chunk_->Empty(); + } + + const char* UnparsedData() const { + return parseable_chunk_->UnparsedData(); + } + + const char** UnparsedDataAddr() { + return parseable_chunk_->UnparsedDataAddr(); + } + + void Advance(size_t number_of_bytes) { + return parseable_chunk_->Advance(number_of_bytes); + } + private: + void FreeChunk() { + if (owned_) { + delete parseable_chunk_; + } + parseable_chunk_ = NULL; + } + + // Will be NULL until Init() has been called. If owned_ is true, this will + // point to a ParseableChunk object that has been allocated with "new" and + // must be deleted by this DeltaWindowSection object. If owned_ is false, + // this points at the parseable_chunk_ owned by a different DeltaWindowSection + // object. In this case, it is important to free the DeltaWindowSection which + // does not own the ParseableChunk before (or simultaneously to) freeing the + // DeltaWindowSection that owns it, or else deleted memory may be accessed. + ParseableChunk* parseable_chunk_; + bool owned_; + + // Making these private avoids implicit copy constructor & assignment operator + DeltaWindowSection(const DeltaWindowSection&); + void operator=(const DeltaWindowSection&); +}; + +// Used to parse the bytes and Varints that make up the delta file header +// or delta window header. +class VCDiffHeaderParser { + public: + // header_start should be the start of the header to be parsed; + // data_end is the position just after the last byte of available data + // (which may extend far past the end of the header.) + VCDiffHeaderParser(const char* header_start, const char* data_end); + + // One of these functions should be called for each element of the header. + // variable_description is a description of the value that we are attempting + // to parse, and will only be used to create descriptive error messages. + // If the function returns true, then the element was parsed successfully + // and its value has been placed in *value. If the function returns false, + // then *value is unchanged, and GetResult() can be called to return the + // reason that the element could not be parsed, which will be either + // RESULT_ERROR (an error occurred), or RESULT_END_OF_DATA (the limit data_end + // was reached before the end of the element to be parsed.) Once one of these + // functions has returned false, further calls to any of the Parse... + // functions will also return false without performing any additional actions. + // Typical usage is as follows: + // int32_t segment_length = 0; + // if (!header_parser.ParseInt32("segment length", &segment_length)) { + // return header_parser.GetResult(); + // } + // + // The following example takes advantage of the fact that calling a Parse... + // function after an error or end-of-data condition is legal and does nothing. + // It can thus parse more than one element in a row and check the status + // afterwards. If the first call to ParseInt32() fails, the second will have + // no effect: + // + // int32_t segment_length = 0, segment_position = 0; + // header_parser.ParseInt32("segment length", &segment_length)); + // header_parser.ParseInt32("segment position", &segment_position)); + // if (RESULT_SUCCESS != header_parser.GetResult()) { + // return header_parser.GetResult(); + // } + // + bool ParseByte(unsigned char* value); + bool ParseInt32(const char* variable_description, int32_t* value); + bool ParseUInt32(const char* variable_description, uint32_t* value); + bool ParseChecksum(const char* variable_description, VCDChecksum* value); + bool ParseSize(const char* variable_description, size_t* value); + + // Parses the first three elements of the delta window header: + // + // Win_Indicator - byte + // [Source segment size] - integer (VarintBE format) + // [Source segment position] - integer (VarintBE format) + // + // Returns true if the values were parsed successfully and the values were + // found to be acceptable. Returns false otherwise, in which case + // GetResult() can be called to return the reason that the two values + // could not be validated. This will be either RESULT_ERROR (an error + // occurred and was logged), or RESULT_END_OF_DATA (the limit data_end was + // reached before the end of the values to be parsed.) If return value is + // true, then *win_indicator, *source_segment_length, and + // *source_segment_position are populated with the parsed values. Otherwise, + // the values of these output arguments are undefined. + // + // dictionary_size: The size of the dictionary (source) file. Used to + // validate the limits of source_segment_length and + // source_segment_position if the source segment is taken from the + // dictionary (i.e., if the parsed *win_indicator equals VCD_SOURCE.) + // decoded_target_size: The size of the target data that has been decoded + // so far, including all target windows. Used to validate the limits of + // source_segment_length and source_segment_position if the source segment + // is taken from the target (i.e., if the parsed *win_indicator equals + // VCD_TARGET.) + // allow_vcd_target: If this argument is false, and the parsed *win_indicator + // is VCD_TARGET, then an error is produced; if true, VCD_TARGET is + // allowed. + // win_indicator (output): Points to a single unsigned char (not an array) + // that will receive the parsed value of Win_Indicator. + // source_segment_length (output): The parsed length of the source segment. + // source_segment_position (output): The parsed zero-based index in the + // source/target file from which the source segment is to be taken. + bool ParseWinIndicatorAndSourceSegment(size_t dictionary_size, + size_t decoded_target_size, + bool allow_vcd_target, + unsigned char* win_indicator, + size_t* source_segment_length, + size_t* source_segment_position); + + // Parses the following two elements of the delta window header: + // + // Length of the delta encoding - integer (VarintBE format) + // Size of the target window - integer (VarintBE format) + // + // Return conditions and values are the same as for + // ParseWinIndicatorAndSourceSegment(), above. + // + bool ParseWindowLengths(size_t* target_window_length); + + // May only be called after ParseWindowLengths() has returned RESULT_SUCCESS. + // Returns a pointer to the end of the delta window (which might not point to + // a valid memory location if there is insufficient input data.) + // + const char* EndOfDeltaWindow() const; + + // Parses the following element of the delta window header: + // + // Delta_Indicator - byte + // + // Because none of the bits in Delta_Indicator are used by this implementation + // of VCDIFF, this function does not have an output argument to return the + // value of that field. It may return RESULT_SUCCESS, RESULT_ERROR, or + // RESULT_END_OF_DATA as with the other Parse...() functions. + // + bool ParseDeltaIndicator(); + + // Parses the following 3 elements of the delta window header: + // + // Length of data for ADDs and RUNs - integer (VarintBE format) + // Length of instructions and sizes - integer (VarintBE format) + // Length of addresses for COPYs - integer (VarintBE format) + // + // If has_checksum is true, it also looks for the following element: + // + // Adler32 checksum - unsigned 32-bit integer (VarintBE format) + // + // Return conditions and values are the same as for + // ParseWinIndicatorAndSourceSegment(), above. + // + bool ParseSectionLengths(bool has_checksum, + size_t* add_and_run_data_length, + size_t* instructions_and_sizes_length, + size_t* addresses_length, + VCDChecksum* checksum); + + // If one of the Parse... functions returned false, this function + // can be used to find the result code (RESULT_ERROR or RESULT_END_OF_DATA) + // describing the reason for the most recent parse failure. If none of the + // Parse... functions has returned false, returns RESULT_SUCCESS. + VCDiffResult GetResult() const { + return return_code_; + } + + // The following functions just pass their arguments to the underlying + // ParseableChunk object. + + const char* End() const { + return parseable_chunk_.End(); + } + + size_t UnparsedSize() const { + return parseable_chunk_.UnparsedSize(); + } + + size_t ParsedSize() const { + return parseable_chunk_.ParsedSize(); + } + + const char* UnparsedData() const { + return parseable_chunk_.UnparsedData(); + } + + private: + // Parses two variable-length integers representing the source segment length + // and source segment position (== offset.) Checks whether the source segment + // length and position would cause it to exceed the size of the source file or + // target file. Returns true if the values were parsed successfully and the + // values were found to be acceptable. Returns false otherwise, in which case + // GetResult() can be called to return the reason that the two values could + // not be validated, which will be either RESULT_ERROR (an error occurred and + // was logged), or RESULT_END_OF_DATA (the limit data_end was reached before + // the end of the integers to be parsed.) + // from_size: The requested size of the source segment. + // from_boundary_name: A NULL-terminated string naming the end of the + // source or target file, used in error messages. + // from_name: A NULL-terminated string naming the source or target file, + // also used in error messages. + // source_segment_length (output): The parsed length of the source segment. + // source_segment_position (output): The parsed zero-based index in the + // source/target file from which the source segment is to be taken. + // + bool ParseSourceSegmentLengthAndPosition(size_t from_size, + const char* from_boundary_name, + const char* from_name, + size_t* source_segment_length, + size_t* source_segment_position); + + ParseableChunk parseable_chunk_; + + // Contains the result code of the last Parse...() operation that failed + // (RESULT_ERROR or RESULT_END_OF_DATA). If no Parse...() method has been + // called, or if all calls to Parse...() were successful, then this contains + // RESULT_SUCCESS. + VCDiffResult return_code_; + + // Will be zero until ParseWindowLengths() has been called. After + // ParseWindowLengths() has been called successfully, this contains the + // parsed length of the delta encoding. + size_t delta_encoding_length_; + + // Will be NULL until ParseWindowLengths() has been called. After + // ParseWindowLengths() has been called successfully, this points to the + // beginning of the section of the current window titled "The delta encoding" + // in the RFC, i.e., to the position just after the length of the delta + // encoding. + const char* delta_encoding_start_; + + // Making these private avoids implicit copy constructor & assignment operator + VCDiffHeaderParser(const VCDiffHeaderParser&); + void operator=(const VCDiffHeaderParser&); +}; + +} // namespace open_vcdiff + +#endif // OPEN_VCDIFF_HEADERPARSER_H_ |