diff options
Diffstat (limited to 'rangeset.h')
| -rw-r--r-- | rangeset.h | 278 |
1 files changed, 278 insertions, 0 deletions
diff --git a/rangeset.h b/rangeset.h new file mode 100644 index 000000000..f224a08be --- /dev/null +++ b/rangeset.h @@ -0,0 +1,278 @@ +/* + * Copyright (C) 2017 The Android Open Source Project + * + * 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. + */ + +#pragma once + +#include <stddef.h> + +#include <string> +#include <utility> +#include <vector> + +#include <android-base/logging.h> +#include <android-base/parseint.h> +#include <android-base/stringprintf.h> +#include <android-base/strings.h> + +using Range = std::pair<size_t, size_t>; + +class RangeSet { + public: + RangeSet() : blocks_(0) {} + + explicit RangeSet(std::vector<Range>&& pairs) { + CHECK_NE(pairs.size(), static_cast<size_t>(0)) << "Invalid number of tokens"; + + // Sanity check the input. + size_t result = 0; + for (const auto& range : pairs) { + CHECK_LT(range.first, range.second) + << "Empty or negative range: " << range.first << ", " << range.second; + size_t sz = range.second - range.first; + CHECK_LE(result, SIZE_MAX - sz) << "RangeSet size overflow"; + result += sz; + } + + ranges_ = pairs; + blocks_ = result; + } + + static RangeSet Parse(const std::string& range_text) { + std::vector<std::string> pieces = android::base::Split(range_text, ","); + CHECK_GE(pieces.size(), static_cast<size_t>(3)) << "Invalid range text: " << range_text; + + size_t num; + CHECK(android::base::ParseUint(pieces[0], &num, static_cast<size_t>(INT_MAX))) + << "Failed to parse the number of tokens: " << range_text; + + CHECK_NE(num, static_cast<size_t>(0)) << "Invalid number of tokens: " << range_text; + CHECK_EQ(num % 2, static_cast<size_t>(0)) << "Number of tokens must be even: " << range_text; + CHECK_EQ(num, pieces.size() - 1) << "Mismatching number of tokens: " << range_text; + + std::vector<Range> pairs; + for (size_t i = 0; i < num; i += 2) { + size_t first; + CHECK(android::base::ParseUint(pieces[i + 1], &first, static_cast<size_t>(INT_MAX))); + size_t second; + CHECK(android::base::ParseUint(pieces[i + 2], &second, static_cast<size_t>(INT_MAX))); + + pairs.emplace_back(first, second); + } + + return RangeSet(std::move(pairs)); + } + + std::string ToString() const { + if (ranges_.empty()) { + return ""; + } + std::string result = std::to_string(ranges_.size() * 2); + for (const auto& r : ranges_) { + result += android::base::StringPrintf(",%zu,%zu", r.first, r.second); + } + + return result; + } + + // Get the block number for the i-th (starting from 0) block in the RangeSet. + size_t GetBlockNumber(size_t idx) const { + CHECK_LT(idx, blocks_) << "Out of bound index " << idx << " (total blocks: " << blocks_ << ")"; + + for (const auto& range : ranges_) { + if (idx < range.second - range.first) { + return range.first + idx; + } + idx -= (range.second - range.first); + } + + CHECK(false) << "Failed to find block number for index " << idx; + return 0; // Unreachable, but to make compiler happy. + } + + // RangeSet has half-closed half-open bounds. For example, "3,5" contains blocks 3 and 4. So "3,5" + // and "5,7" are not overlapped. + bool Overlaps(const RangeSet& other) const { + for (const auto& range : ranges_) { + size_t start = range.first; + size_t end = range.second; + for (const auto& other_range : other.ranges_) { + size_t other_start = other_range.first; + size_t other_end = other_range.second; + // [start, end) vs [other_start, other_end) + if (!(other_start >= end || start >= other_end)) { + return true; + } + } + } + return false; + } + + // size() gives the number of Range's in this RangeSet. + size_t size() const { + return ranges_.size(); + } + + // blocks() gives the number of all blocks in this RangeSet. + size_t blocks() const { + return blocks_; + } + + // We provide const iterators only. + std::vector<Range>::const_iterator cbegin() const { + return ranges_.cbegin(); + } + + std::vector<Range>::const_iterator cend() const { + return ranges_.cend(); + } + + // Need to provide begin()/end() since range-based loop expects begin()/end(). + std::vector<Range>::const_iterator begin() const { + return ranges_.cbegin(); + } + + std::vector<Range>::const_iterator end() const { + return ranges_.cend(); + } + + // Reverse const iterators for MoveRange(). + std::vector<Range>::const_reverse_iterator crbegin() const { + return ranges_.crbegin(); + } + + std::vector<Range>::const_reverse_iterator crend() const { + return ranges_.crend(); + } + + const Range& operator[](size_t i) const { + return ranges_[i]; + } + + bool operator==(const RangeSet& other) const { + // The orders of Range's matter. "4,1,5,8,10" != "4,8,10,1,5". + return (ranges_ == other.ranges_); + } + + bool operator!=(const RangeSet& other) const { + return ranges_ != other.ranges_; + } + + protected: + // Actual limit for each value and the total number are both INT_MAX. + std::vector<Range> ranges_; + size_t blocks_; +}; + +static constexpr size_t kBlockSize = 4096; + +// The class is a sorted version of a RangeSet; and it's useful in imgdiff to split the input +// files when we're handling large zip files. Specifically, we can treat the input file as a +// continuous RangeSet (i.e. RangeSet("0-99") for a 100 blocks file); and break it down into +// several smaller chunks based on the zip entries. + +// For example, [source: 0-99] can be split into +// [split_src1: 10-29]; [split_src2: 40-49, 60-69]; [split_src3: 70-89] +// Here "10-29" simply means block 10th to block 29th with respect to the original input file. +// Also, note that the split sources should be mutual exclusive, but they don't need to cover +// every block in the original source. +class SortedRangeSet : public RangeSet { + public: + SortedRangeSet() {} + + // Ranges in the the set should be mutually exclusive; and they're sorted by the start block. + explicit SortedRangeSet(std::vector<Range>&& pairs) : RangeSet(std::move(pairs)) { + std::sort(ranges_.begin(), ranges_.end()); + } + + void Insert(const Range& to_insert) { + SortedRangeSet rs({ to_insert }); + Insert(rs); + } + + // Insert the input SortedRangeSet; keep the ranges sorted and merge the overlap ranges. + void Insert(const SortedRangeSet& rs) { + if (rs.size() == 0) { + return; + } + // Merge and sort the two RangeSets. + std::vector<Range> temp = std::move(ranges_); + std::copy(rs.begin(), rs.end(), std::back_inserter(temp)); + std::sort(temp.begin(), temp.end()); + + Clear(); + // Trim overlaps and insert the result back to ranges_. + Range to_insert = temp.front(); + for (auto it = temp.cbegin() + 1; it != temp.cend(); it++) { + if (it->first <= to_insert.second) { + to_insert.second = std::max(to_insert.second, it->second); + } else { + ranges_.push_back(to_insert); + blocks_ += (to_insert.second - to_insert.first); + to_insert = *it; + } + } + ranges_.push_back(to_insert); + blocks_ += (to_insert.second - to_insert.first); + } + + void Clear() { + blocks_ = 0; + ranges_.clear(); + } + + using RangeSet::Overlaps; + bool Overlaps(size_t start, size_t len) const { + RangeSet rs({ { start / kBlockSize, (start + len - 1) / kBlockSize + 1 } }); + return Overlaps(rs); + } + + // Compute the block range the file occupies, and insert that range. + void Insert(size_t start, size_t len) { + Range to_insert{ start / kBlockSize, (start + len - 1) / kBlockSize + 1 }; + Insert(to_insert); + } + + // Given an offset of the file, checks if the corresponding block (by considering the file as + // 0-based continuous block ranges) is covered by the SortedRangeSet. If so, returns the offset + // within this SortedRangeSet. + // + // For example, the 4106-th byte of a file is from block 1, assuming a block size of 4096-byte. + // The mapped offset within a SortedRangeSet("1-9 15-19") is 10. + // + // An offset of 65546 falls into the 16-th block in a file. Block 16 is contained as the 10-th + // item in SortedRangeSet("1-9 15-19"). So its data can be found at offset 40970 (i.e. 4096 * 10 + // + 10) in a range represented by this SortedRangeSet. + size_t GetOffsetInRangeSet(size_t old_offset) const { + size_t old_block_start = old_offset / kBlockSize; + size_t new_block_start = 0; + for (const auto& range : ranges_) { + // Find the index of old_block_start. + if (old_block_start >= range.second) { + new_block_start += (range.second - range.first); + } else if (old_block_start >= range.first) { + new_block_start += (old_block_start - range.first); + return (new_block_start * kBlockSize + old_offset % kBlockSize); + } else { + CHECK(false) << "block_start " << old_block_start + << " is missing between two ranges: " << this->ToString(); + return 0; + } + } + CHECK(false) << "block_start " << old_block_start + << " exceeds the limit of current RangeSet: " << this->ToString(); + return 0; + } +};
\ No newline at end of file |