Merges PR #35

src/blockchain_converter/CMakeLists.txt

@@ -35,6 +35,32 @@ set(blockchain_converter_private_headers)
 bitmonero_private_headers(blockchain_converter
 	  ${blockchain_converter_private_headers})
 
+set(blockchain_import_sources
+  blockchain_import.cpp
+  )
+
+set(blockchain_import_private_headers
+  import.h
+  fake_core.h
+  )
+
+bitmonero_private_headers(blockchain_import
+	  ${blockchain_import_private_headers})
+
+set(blockchain_export_sources
+  blockchain_export.cpp
+  )
+
+set(blockchain_export_private_headers
+  import.h
+  blockchain_export.h
+  )
+
+bitmonero_private_headers(blockchain_export
+	  ${blockchain_export_private_headers})
+
+
+
 if (BLOCKCHAIN_DB STREQUAL DB_LMDB)
 bitmonero_add_executable(blockchain_converter
   ${blockchain_converter_sources}
@@ -52,3 +78,33 @@ set_property(TARGET blockchain_converter
 	PROPERTY
 	OUTPUT_NAME "blockchain_converter")
 endif ()
+
+bitmonero_add_executable(blockchain_import
+  ${blockchain_import_sources}
+  ${blockchain_import_private_headers})
+
+target_link_libraries(blockchain_import
+  LINK_PRIVATE
+    cryptonote_core
+    ${CMAKE_THREAD_LIBS_INIT})
+
+add_dependencies(blockchain_import
+	version)
+set_property(TARGET blockchain_import
+	PROPERTY
+	OUTPUT_NAME "blockchain_import")
+
+bitmonero_add_executable(blockchain_export
+  ${blockchain_export_sources}
+  ${blockchain_export_private_headers})
+
+target_link_libraries(blockchain_export
+  LINK_PRIVATE
+    cryptonote_core
+    ${CMAKE_THREAD_LIBS_INIT})
+
+add_dependencies(blockchain_export
+	version)
+set_property(TARGET blockchain_export
+	PROPERTY
+	OUTPUT_NAME "blockchain_export")

src/blockchain_converter/blockchain_converter.cpp

@@ -38,52 +38,179 @@
 #include "cryptonote_core/blockchain.h"
 #include "blockchain_db/lmdb/db_lmdb.h"
 #include "cryptonote_core/tx_pool.h"
+#include "common/command_line.h"
+#include "serialization/json_utils.h"
+#include "include_base_utils.h"
+#include "version.h"
 #include <iostream>
 
+// CONFIG
+static bool opt_batch   = true;
+static bool opt_testnet = false;
+
+// number of blocks per batch transaction
+// adjustable through command-line argument according to available RAM
+static uint64_t db_batch_size = 20000;
+
+
+namespace po = boost::program_options;
+
 using namespace cryptonote;
+using namespace epee;
 
 struct fake_core
 {
-  tx_memory_pool m_pool;
   Blockchain dummy;
+  tx_memory_pool m_pool;
 
   blockchain_storage m_storage;
 
-
+#if !defined(BLOCKCHAIN_DB)
-  fake_core(const boost::filesystem::path &path) : m_pool(dummy), dummy(m_pool), m_storage(&m_pool)
+  // for multi_db_runtime:
+  fake_core(const boost::filesystem::path &path, const bool use_testnet) : dummy(m_pool), m_pool(&dummy), m_storage(m_pool)
+#else
+  // for multi_db_compile:
+  fake_core(const boost::filesystem::path &path, const bool use_testnet) : dummy(m_pool), m_pool(dummy), m_storage(&m_pool)
+#endif
   {
     m_pool.init(path.string());
-    m_storage.init(path.string(), false);
+    m_storage.init(path.string(), use_testnet);
   }
 };
 
 int main(int argc, char* argv[])
 {
-  std::string dir = tools::get_default_data_dir();
+  uint64_t height = 0;
-  boost::filesystem::path default_data_path {dir};
+  uint64_t start_block = 0;
-  if (argc >= 2 && !strcmp(argv[1], "--testnet")) {
+  uint64_t end_block = 0;
-    default_data_path /= "testnet";
+  uint64_t num_blocks = 0;
+
+  boost::filesystem::path default_data_path {tools::get_default_data_dir()};
+  boost::filesystem::path default_testnet_data_path {default_data_path / "testnet"};
+
+
+  po::options_description desc_cmd_only("Command line options");
+  po::options_description desc_cmd_sett("Command line options and settings options");
+  const command_line::arg_descriptor<uint32_t> arg_log_level   =  {"log-level", "", LOG_LEVEL_0};
+  const command_line::arg_descriptor<uint64_t> arg_batch_size  =  {"batch-size", "", db_batch_size};
+  const command_line::arg_descriptor<bool>     arg_testnet_on  = {
+    "testnet"
+      , "Run on testnet."
+      , opt_testnet
+  };
+  const command_line::arg_descriptor<uint64_t> arg_block_number =
+  {"block-number", "Number of blocks (default: use entire source blockchain)",
+    0};
+
+  command_line::add_arg(desc_cmd_sett, command_line::arg_data_dir, default_data_path.string());
+  command_line::add_arg(desc_cmd_sett, command_line::arg_testnet_data_dir, default_testnet_data_path.string());
+  command_line::add_arg(desc_cmd_sett, arg_log_level);
+  command_line::add_arg(desc_cmd_sett, arg_batch_size);
+  command_line::add_arg(desc_cmd_sett, arg_testnet_on);
+  command_line::add_arg(desc_cmd_sett, arg_block_number);
+
+  command_line::add_arg(desc_cmd_only, command_line::arg_help);
+
+  const command_line::arg_descriptor<bool> arg_batch  =  {"batch",
+    "Batch transactions for faster import", true};
+
+  // call add_options() directly for these arguments since command_line helpers
+  // support only boolean switch, not boolean argument
+  desc_cmd_sett.add_options()
+    (arg_batch.name, make_semantic(arg_batch), arg_batch.description)
+    ;
+
+  po::options_description desc_options("Allowed options");
+  desc_options.add(desc_cmd_only).add(desc_cmd_sett);
+
+
+  po::variables_map vm;
+  bool r = command_line::handle_error_helper(desc_options, [&]()
+  {
+    po::store(po::parse_command_line(argc, argv, desc_options), vm);
+    po::notify(vm);
+
+    return true;
+  });
+  if (!r)
+    return 1;
+
+  int log_level = command_line::get_arg(vm, arg_log_level);
+  opt_batch     = command_line::get_arg(vm, arg_batch);
+  db_batch_size = command_line::get_arg(vm, arg_batch_size);
+
+  if (command_line::get_arg(vm, command_line::arg_help))
+  {
+    std::cout << CRYPTONOTE_NAME << " v" << MONERO_VERSION_FULL << ENDL << ENDL;
+    std::cout << desc_options << std::endl;
+    return 1;
   }
 
-  fake_core c(default_data_path);
+  if (! opt_batch && ! vm["batch-size"].defaulted())
+  {
+    std::cerr << "Error: batch-size set, but batch option not enabled" << ENDL;
+    return 1;
+  }
+  if (! db_batch_size)
+  {
+    std::cerr << "Error: batch-size must be > 0" << ENDL;
+    return 1;
+  }
+
+  log_space::get_set_log_detalisation_level(true, log_level);
+  log_space::log_singletone::add_logger(LOGGER_CONSOLE, NULL, NULL);
+  LOG_PRINT_L0("Starting...");
+
+  std::string src_folder;
+  opt_testnet = command_line::get_arg(vm, arg_testnet_on);
+  auto data_dir_arg = opt_testnet ? command_line::arg_testnet_data_dir : command_line::arg_data_dir;
+  src_folder = command_line::get_arg(vm, data_dir_arg);
+  boost::filesystem::path dest_folder(src_folder);
+
+  num_blocks = command_line::get_arg(vm, arg_block_number);
+
+  if (opt_batch)
+  {
+    LOG_PRINT_L0("batch:   " << std::boolalpha << opt_batch << std::noboolalpha
+        << "  batch size: " << db_batch_size);
+  }
+  else
+  {
+    LOG_PRINT_L0("batch:   " << std::boolalpha << opt_batch << std::noboolalpha);
+  }
+  LOG_PRINT_L0("testnet: " << std::boolalpha << opt_testnet << std::noboolalpha);
+
+  fake_core c(src_folder, opt_testnet);
+
+  height = c.m_storage.get_current_blockchain_height();
+  if (! num_blocks || num_blocks > height)
+    end_block = height - 1;
+  else
+    end_block = start_block + num_blocks - 1;
 
   BlockchainDB *blockchain;
+  blockchain = new BlockchainLMDB(opt_batch);
+  dest_folder /= blockchain->get_db_name();
+  LOG_PRINT_L0("Source blockchain: " << src_folder);
+  LOG_PRINT_L0("Dest blockchain:   " << dest_folder.string());
+  LOG_PRINT_L0("Opening LMDB: " << dest_folder.string());
+  blockchain->open(dest_folder.string());
 
-  blockchain = new BlockchainLMDB();
+  if (opt_batch)
-
+    blockchain->batch_start();
-  boost::filesystem::path db_path(default_data_path);
+  uint64_t i = 0;
-
+  for (i = start_block; i < end_block + 1; ++i)
-  db_path /= blockchain->get_db_name();
-
-  blockchain->open(db_path.string());
-
-  for (uint64_t height, i = 0; i < (height = c.m_storage.get_current_blockchain_height()); ++i)
   {
-    if (i % 10 == 0)
+    // block: i  height: i+1  end height: end_block + 1
+    if ((i+1) % 10 == 0)
     {
-      std::cout << "\r                   \r" << "block " << i << "/" << height
+      std::cout << "\r                   \r" << "height " << i+1 << "/" <<
-         << " (" << (i+1)*100/height<< "%)" << std::flush;
+        end_block+1 << " (" << (i+1)*100/(end_block+1)<< "%)" << std::flush;
     }
+    // for debugging:
+    // std::cout << "height " << i+1 << "/" << end_block+1
+    //   << " ((" << i+1 << ")*100/(end_block+1))" << "%)" << ENDL;
+
     block b = c.m_storage.get_block(i);
     size_t bsize = c.m_storage.get_block_size(i);
     difficulty_type bdiff = c.m_storage.get_block_cumulative_difficulty(i);
@@ -94,8 +221,8 @@ int main(int argc, char* argv[])
     c.m_storage.get_transactions(b.tx_hashes, txs, missed);
     if (missed.size())
     {
-      std::cout << std::endl;
+      std::cout << ENDL;
-      std::cerr << "Missed transaction(s) for block at height " << i << ", exiting" << std::endl;
+      std::cerr << "Missed transaction(s) for block at height " << i + 1 << ", exiting" << ENDL;
       delete blockchain;
       return 1;
     }
@@ -103,17 +230,40 @@ int main(int argc, char* argv[])
     try
     {
       blockchain->add_block(b, bsize, bdiff, bcoins, txs);
+
+      if (opt_batch)
+      {
+        if ((i < end_block) && ((i + 1) % db_batch_size == 0))
+        {
+          std::cout << "\r                   \r";
+          std::cout << "[- batch commit at height " << i + 1 << " -]" << ENDL;
+          blockchain->batch_stop();
+          blockchain->batch_start();
+          std::cout << ENDL;
+          blockchain->show_stats();
+        }
+      }
     }
     catch (const std::exception& e)
     {
-      std::cout << std::endl;
+      std::cout << ENDL;
-      std::cerr << "Error adding block to new blockchain: " << e.what() << std::endl;
+      std::cerr << "Error adding block to new blockchain: " << e.what() << ENDL;
       delete blockchain;
       return 2;
     }
   }
+  if (opt_batch)
+  {
+    std::cout << "\r                   \r" << "height " << i << "/" <<
+      end_block+1 << " (" << (i)*100/(end_block+1)<< "%)" << std::flush;
+    std::cout << ENDL;
+    std::cout << "[- batch commit at height " << i << " -]" << ENDL;
+    blockchain->batch_stop();
+  }
+  std::cout << ENDL;
+  blockchain->show_stats();
+  std::cout << "Finished at height: " << i << "  block: " << i-1 << ENDL;
 
-  std::cout << std::endl;
   delete blockchain;
   return 0;
 }

src/blockchain_converter/blockchain_export.cpp

@@ -0,0 +1,386 @@
+// Copyright (c) 2014-2015, The Monero Project
+//
+// All rights reserved.
+//
+// Redistribution and use in source and binary forms, with or without modification, are
+// permitted provided that the following conditions are met:
+//
+// 1. Redistributions of source code must retain the above copyright notice, this list of
+//    conditions and the following disclaimer.
+//
+// 2. Redistributions in binary form must reproduce the above copyright notice, this list
+//    of conditions and the following disclaimer in the documentation and/or other
+//    materials provided with the distribution.
+//
+// 3. Neither the name of the copyright holder nor the names of its contributors may be
+//    used to endorse or promote products derived from this software without specific
+//    prior written permission.
+//
+// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY
+// EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
+// MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL
+// THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+// SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
+// PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+// INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+// STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF
+// THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+#include <algorithm>
+#include <cstdio>
+#include <fstream>
+#include <boost/iostreams/copy.hpp>
+#include <atomic>
+
+#include <boost/archive/binary_oarchive.hpp>
+#include <boost/archive/binary_iarchive.hpp>
+#include <boost/iostreams/stream_buffer.hpp>
+#include <boost/iostreams/stream.hpp>
+#include <boost/iostreams/device/back_inserter.hpp>
+#include <boost/iostreams/filtering_streambuf.hpp>
+#include <boost/iostreams/filter/bzip2.hpp>
+#include "common/command_line.h"
+#include "version.h"
+#include "blockchain_export.h"
+#include "cryptonote_core/cryptonote_boost_serialization.h"
+
+#include "import.h"
+
+static int max_chunk = 0;
+static size_t height;
+
+namespace po = boost::program_options;
+
+using namespace cryptonote;
+using namespace epee;
+
+bool BlockchainExport::open(const boost::filesystem::path& dir_path)
+{
+  if (boost::filesystem::exists(dir_path))
+  {
+    if (!boost::filesystem::is_directory(dir_path))
+    {
+      LOG_PRINT_RED_L0("export directory path is a file: " << dir_path);
+      return false;
+    }
+  }
+  else
+  {
+    if (!boost::filesystem::create_directory(dir_path))
+    {
+      LOG_PRINT_RED_L0("Failed to create directory " << dir_path);
+      return false;
+    }
+  }
+
+  std::string file_path = (dir_path / BLOCKCHAIN_RAW).string();
+  m_raw_data_file = new std::ofstream();
+  m_raw_data_file->open(file_path , std::ios_base::binary | std::ios_base::out| std::ios::trunc);
+  if (m_raw_data_file->fail())
+    return false;
+
+  m_output_stream = new boost::iostreams::stream<boost::iostreams::back_insert_device<buffer_type>>(m_buffer);
+  m_raw_archive = new boost::archive::binary_oarchive(*m_output_stream);
+  if (m_raw_archive == NULL)
+    return false;
+
+  return true;
+}
+
+void BlockchainExport::flush_chunk()
+{
+  m_output_stream->flush();
+  char buffer[STR_LENGTH_OF_INT + 1];
+  int chunk_size = (int) m_buffer.size();
+  if (chunk_size > BUFFER_SIZE)
+  {
+    LOG_PRINT_L0("WARNING: chunk_size " << chunk_size << " > BUFFER_SIZE " << BUFFER_SIZE);
+  }
+  sprintf(buffer, STR_FORMAT_OF_INT, chunk_size);
+  m_raw_data_file->write(buffer, STR_LENGTH_OF_INT);
+  if (max_chunk < chunk_size)
+  {
+    max_chunk = chunk_size;
+  }
+  long pos_before = m_raw_data_file->tellp();
+  std::copy(m_buffer.begin(), m_buffer.end(), std::ostreambuf_iterator<char>(*m_raw_data_file));
+  m_raw_data_file->flush();
+  long pos_after = m_raw_data_file->tellp();
+  long num_chars_written = pos_after - pos_before;
+  if ((int) num_chars_written != chunk_size)
+  {
+    LOG_PRINT_RED_L0("INTERNAL ERROR: num chars wrote NEQ buffer size. height = " << height);
+  }
+
+  m_buffer.clear();
+  delete m_raw_archive;
+  delete m_output_stream;
+  m_output_stream = new boost::iostreams::stream<boost::iostreams::back_insert_device<buffer_type>>(m_buffer);
+  m_raw_archive = new boost::archive::binary_oarchive(*m_output_stream);
+}
+
+void BlockchainExport::serialize_block_to_text_buffer(const block& block)
+{
+  *m_raw_archive << block;
+}
+
+void BlockchainExport::buffer_serialize_tx(const transaction& tx)
+{
+  *m_raw_archive << tx;
+}
+
+void BlockchainExport::buffer_write_num_txs(const std::list<transaction> txs)
+{
+  int n = txs.size();
+  *m_raw_archive << n;
+}
+
+void BlockchainExport::write_block(block& block)
+{
+  serialize_block_to_text_buffer(block);
+  std::list<transaction> txs;
+
+  uint64_t block_height = boost::get<txin_gen>(block.miner_tx.vin.front()).height;
+
+  // put coinbase transaction first
+  transaction coinbase_tx = block.miner_tx;
+  crypto::hash coinbase_tx_hash = get_transaction_hash(coinbase_tx);
+#if SOURCE_DB == DB_MEMORY
+  const transaction* cb_tx_full = m_blockchain_storage->get_tx(coinbase_tx_hash);
+#else
+  transaction cb_tx_full = m_blockchain_storage->get_db()->get_tx(coinbase_tx_hash);
+#endif
+
+#if SOURCE_DB == DB_MEMORY
+  if (cb_tx_full != NULL)
+  {
+    txs.push_back(*cb_tx_full);
+  }
+#else
+  // TODO: should check and abort if cb_tx_full equals null_hash?
+  txs.push_back(cb_tx_full);
+#endif
+
+  // now add all regular transactions
+  BOOST_FOREACH(const auto& tx_id, block.tx_hashes)
+  {
+#if SOURCE_DB == DB_MEMORY
+    const transaction* tx = m_blockchain_storage->get_tx(tx_id);
+#else
+    transaction tx = m_blockchain_storage->get_db()->get_tx(tx_id);
+#endif
+
+#if SOURCE_DB == DB_MEMORY
+    if(tx == NULL)
+    {
+      if (! m_tx_pool)
+        throw std::runtime_error("Aborting: tx == NULL, so memory pool required to get tx, but memory pool isn't enabled");
+      else
+      {
+        transaction tx;
+        if(m_tx_pool->get_transaction(tx_id, tx))
+          txs.push_back(tx);
+        else
+          throw std::runtime_error("Aborting: tx not found in pool");
+      }
+    }
+    else
+      txs.push_back(*tx);
+#else
+    txs.push_back(tx);
+#endif
+  }
+
+  // serialize all txs to the persistant storage
+  buffer_write_num_txs(txs);
+  BOOST_FOREACH(const auto& tx, txs)
+  {
+    buffer_serialize_tx(tx);
+  }
+
+  // These three attributes are currently necessary for a fast import that adds blocks without verification.
+  bool include_extra_block_data = true;
+  if (include_extra_block_data)
+  {
+#if SOURCE_DB == DB_MEMORY
+    size_t block_size = m_blockchain_storage->get_block_size(block_height);
+#else
+    size_t block_size = m_blockchain_storage->get_db()->get_block_size(block_height);
+#endif
+#if SOURCE_DB == DB_MEMORY
+    difficulty_type cumulative_difficulty = m_blockchain_storage->get_block_cumulative_difficulty(block_height);
+#else
+    difficulty_type cumulative_difficulty = m_blockchain_storage->get_db()->get_block_cumulative_difficulty(block_height);
+#endif
+#if SOURCE_DB == DB_MEMORY
+    uint64_t coins_generated = m_blockchain_storage->get_block_coins_generated(block_height);
+#else
+    // TODO TEST to verify that this is the equivalent. make sure no off-by-one error with block height vs block number
+    uint64_t coins_generated = m_blockchain_storage->get_db()->get_block_already_generated_coins(block_height);
+#endif
+
+    *m_raw_archive << block_size;
+    *m_raw_archive << cumulative_difficulty;
+    *m_raw_archive << coins_generated;
+  }
+}
+
+bool BlockchainExport::BlockchainExport::close()
+{
+  if (m_raw_data_file->fail())
+    return false;
+
+  m_raw_data_file->flush();
+  delete m_raw_archive;
+  delete m_output_stream;
+  delete m_raw_data_file;
+  return true;
+}
+
+
+#if SOURCE_DB == DB_MEMORY
+bool BlockchainExport::store_blockchain_raw(blockchain_storage* _blockchain_storage, tx_memory_pool* _tx_pool, boost::filesystem::path& output_dir, uint64_t use_block_height)
+#else
+bool BlockchainExport::store_blockchain_raw(Blockchain* _blockchain_storage, tx_memory_pool* _tx_pool, boost::filesystem::path& output_dir, uint64_t use_block_height)
+#endif
+{
+  uint64_t use_block_height2 = 0;
+  m_blockchain_storage = _blockchain_storage;
+  m_tx_pool = _tx_pool;
+  LOG_PRINT_L0("Storing blocks raw data...");
+  if (!BlockchainExport::open(output_dir))
+  {
+    LOG_PRINT_RED_L0("failed to open raw file for write");
+    return false;
+  }
+  block b;
+  LOG_PRINT_L0("source blockchain height: " <<  m_blockchain_storage->get_current_blockchain_height());
+  LOG_PRINT_L0("requested block height: " << use_block_height);
+  if ((use_block_height > 0) && (use_block_height < m_blockchain_storage->get_current_blockchain_height()))
+    use_block_height2 = use_block_height;
+  else
+  {
+    use_block_height2 = m_blockchain_storage->get_current_blockchain_height();
+    LOG_PRINT_L0("using block height: " << use_block_height2);
+  }
+  for (height=0; height < use_block_height2; ++height)
+  {
+    crypto::hash hash = m_blockchain_storage->get_block_id_by_height(height);
+    m_blockchain_storage->get_block_by_hash(hash, b);
+    write_block(b);
+    if (height % NUM_BLOCKS_PER_CHUNK == 0) {
+      flush_chunk();
+    }
+  }
+  if (height % NUM_BLOCKS_PER_CHUNK != 0)
+  {
+    flush_chunk();
+  }
+
+  LOG_PRINT_L0("longest chunk was " << max_chunk << " bytes");
+  return BlockchainExport::close();
+}
+
+
+int main(int argc, char* argv[])
+{
+  uint32_t log_level = 0;
+  uint64_t block_height = 0;
+  std::string import_filename = BLOCKCHAIN_RAW;
+
+  boost::filesystem::path default_data_path {tools::get_default_data_dir()};
+  boost::filesystem::path default_testnet_data_path {default_data_path / "testnet"};
+
+  po::options_description desc_cmd_only("Command line options");
+  po::options_description desc_cmd_sett("Command line options and settings options");
+  const command_line::arg_descriptor<uint32_t> arg_log_level   =  {"log-level",  "", log_level};
+  const command_line::arg_descriptor<uint64_t> arg_block_height =  {"block-number", "", block_height};
+  const command_line::arg_descriptor<bool>     arg_testnet_on  = {
+    "testnet"
+      , "Run on testnet."
+      , false
+  };
+
+
+  command_line::add_arg(desc_cmd_sett, command_line::arg_data_dir, default_data_path.string());
+  command_line::add_arg(desc_cmd_sett, command_line::arg_testnet_data_dir, default_testnet_data_path.string());
+  command_line::add_arg(desc_cmd_sett, arg_log_level);
+  command_line::add_arg(desc_cmd_sett, arg_block_height);
+  command_line::add_arg(desc_cmd_sett, arg_testnet_on);
+
+  command_line::add_arg(desc_cmd_only, command_line::arg_help);
+
+  po::options_description desc_options("Allowed options");
+  desc_options.add(desc_cmd_only).add(desc_cmd_sett);
+
+  po::variables_map vm;
+  bool r = command_line::handle_error_helper(desc_options, [&]()
+  {
+    po::store(po::parse_command_line(argc, argv, desc_options), vm);
+    po::notify(vm);
+    return true;
+  });
+  if (! r)
+    return 1;
+
+  if (command_line::get_arg(vm, command_line::arg_help))
+  {
+    std::cout << CRYPTONOTE_NAME << " v" << MONERO_VERSION_FULL << ENDL << ENDL;
+    std::cout << desc_options << std::endl;
+    return 1;
+  }
+
+  log_level    = command_line::get_arg(vm, arg_log_level);
+  block_height = command_line::get_arg(vm, arg_block_height);
+
+  log_space::get_set_log_detalisation_level(true, log_level);
+  log_space::log_singletone::add_logger(LOGGER_CONSOLE, NULL, NULL);
+  LOG_PRINT_L0("Starting...");
+  LOG_PRINT_L0("Setting log level = " << log_level);
+
+  bool opt_testnet = command_line::get_arg(vm, arg_testnet_on);
+
+  std::string m_config_folder;
+
+  auto data_dir_arg = opt_testnet ? command_line::arg_testnet_data_dir : command_line::arg_data_dir;
+  m_config_folder = command_line::get_arg(vm, data_dir_arg);
+  boost::filesystem::path output_dir {m_config_folder};
+  output_dir /= "export";
+  LOG_PRINT_L0("Export directory: " << output_dir.string());
+
+  // If we wanted to use the memory pool, we would set up a fake_core.
+
+#if SOURCE_DB == DB_MEMORY
+  // blockchain_storage* core_storage = NULL;
+  // tx_memory_pool m_mempool(*core_storage); // is this fake anyway? just passing in NULL! so m_mempool can't be used anyway, right?
+  // core_storage = new blockchain_storage(&m_mempool);
+
+  blockchain_storage* core_storage = new blockchain_storage(NULL);
+  LOG_PRINT_L0("Initializing source blockchain (in-memory database)");
+  r = core_storage->init(m_config_folder, opt_testnet);
+#else
+  // Use Blockchain instead of lower-level BlockchainDB for two reasons:
+  // 1. Blockchain has the init() method for easy setup
+  // 2. exporter needs to use get_current_blockchain_height(), get_block_id_by_height(), get_block_by_hash()
+  //
+  // cannot match blockchain_storage setup above with just one line,
+  // e.g.
+  //   Blockchain* core_storage = new Blockchain(NULL);
+  // because unlike blockchain_storage constructor, which takes a pointer to
+  // tx_memory_pool, Blockchain's constructor takes tx_memory_pool object.
+  LOG_PRINT_L0("Initializing source blockchain (BlockchainDB)");
+  Blockchain* core_storage = NULL;
+  tx_memory_pool m_mempool(*core_storage);
+  core_storage = new Blockchain(m_mempool);
+  r = core_storage->init(m_config_folder, opt_testnet);
+#endif
+
+  CHECK_AND_ASSERT_MES(r, false, "Failed to initialize source blockchain storage");
+  LOG_PRINT_L0("Source blockchain storage initialized OK");
+  LOG_PRINT_L0("Exporting blockchain raw data...");
+
+  BlockchainExport be;
+  r = be.store_blockchain_raw(core_storage, NULL, output_dir, block_height);
+  CHECK_AND_ASSERT_MES(r, false, "Failed to export blockchain raw data");
+  LOG_PRINT_L0("Blockchain raw data exported OK");
+}

src/blockchain_converter/blockchain_export.h

@@ -0,0 +1,87 @@
+// Copyright (c) 2014-2015, The Monero Project
+//
+// All rights reserved.
+//
+// Redistribution and use in source and binary forms, with or without modification, are
+// permitted provided that the following conditions are met:
+//
+// 1. Redistributions of source code must retain the above copyright notice, this list of
+//    conditions and the following disclaimer.
+//
+// 2. Redistributions in binary form must reproduce the above copyright notice, this list
+//    of conditions and the following disclaimer in the documentation and/or other
+//    materials provided with the distribution.
+//
+// 3. Neither the name of the copyright holder nor the names of its contributors may be
+//    used to endorse or promote products derived from this software without specific
+//    prior written permission.
+//
+// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY
+// EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
+// MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL
+// THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+// SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
+// PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+// INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+// STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF
+// THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+#pragma once
+
+#include <boost/archive/binary_oarchive.hpp>
+#include <boost/iostreams/stream_buffer.hpp>
+#include <boost/iostreams/stream.hpp>
+#include <boost/iostreams/device/back_inserter.hpp>
+#include "cryptonote_core/cryptonote_basic.h"
+#include "cryptonote_core/blockchain_storage.h"
+#include "cryptonote_core/blockchain.h"
+#include "blockchain_db/blockchain_db.h"
+#include "blockchain_db/lmdb/db_lmdb.h"
+
+// CONFIG: choose one of the three #define's
+//
+// DB_MEMORY is a sensible default for users migrating to LMDB, as it allows
+// the exporter to use the in-memory blockchain while the other binaries
+// work with LMDB, without recompiling anything.
+//
+#define SOURCE_DB DB_MEMORY
+// #define SOURCE_DB DB_LMDB
+// to use global compile-time setting (DB_MEMORY or DB_LMDB):
+// #define SOURCE_DB BLOCKCHAIN_DB
+
+using namespace cryptonote;
+
+class BlockchainExport
+{
+public:
+#if SOURCE_DB == DB_MEMORY
+  bool store_blockchain_raw(cryptonote::blockchain_storage* cs, cryptonote::tx_memory_pool* txp,
+      boost::filesystem::path& output_dir, uint64_t use_block_height=0);
+#else
+  bool store_blockchain_raw(cryptonote::Blockchain* cs, cryptonote::tx_memory_pool* txp,
+      boost::filesystem::path& output_dir, uint64_t use_block_height=0);
+#endif
+
+protected:
+#if SOURCE_DB == DB_MEMORY
+  blockchain_storage* m_blockchain_storage;
+#else
+  Blockchain* m_blockchain_storage;
+#endif
+
+  tx_memory_pool* m_tx_pool;
+  typedef std::vector<char> buffer_type;
+  std::ofstream * m_raw_data_file;
+  boost::archive::binary_oarchive * m_raw_archive;
+  buffer_type m_buffer;
+  boost::iostreams::stream<boost::iostreams::back_insert_device<buffer_type>>* m_output_stream;
+
+  // open export file for write
+  bool open(const boost::filesystem::path& dir_path);
+  bool close();
+  void write_block(block& block);
+  void serialize_block_to_text_buffer(const block& block);
+  void buffer_serialize_tx(const transaction& tx);
+  void buffer_write_num_txs(const std::list<transaction> txs);
+  void flush_chunk();
+};

src/blockchain_converter/blockchain_import.cpp

@@ -0,0 +1,751 @@
+// Copyright (c) 2014-2015, The Monero Project
+//
+// All rights reserved.
+//
+// Redistribution and use in source and binary forms, with or without modification, are
+// permitted provided that the following conditions are met:
+//
+// 1. Redistributions of source code must retain the above copyright notice, this list of
+//    conditions and the following disclaimer.
+//
+// 2. Redistributions in binary form must reproduce the above copyright notice, this list
+//    of conditions and the following disclaimer in the documentation and/or other
+//    materials provided with the distribution.
+//
+// 3. Neither the name of the copyright holder nor the names of its contributors may be
+//    used to endorse or promote products derived from this software without specific
+//    prior written permission.
+//
+// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY
+// EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
+// MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL
+// THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+// SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
+// PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+// INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+// STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF
+// THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+#include <atomic>
+#include <cstdio>
+#include <algorithm>
+#include <fstream>
+
+#include <boost/filesystem.hpp>
+#include <boost/iostreams/stream.hpp>
+#include <boost/archive/binary_iarchive.hpp>
+#include "cryptonote_core/cryptonote_basic.h"
+#include "cryptonote_core/cryptonote_format_utils.h"
+#include "cryptonote_core/cryptonote_boost_serialization.h"
+#include "serialization/json_utils.h" // dump_json()
+#include "include_base_utils.h"
+#include "common/command_line.h"
+#include "version.h"
+
+#include <lmdb.h> // for db flag arguments
+
+#include "import.h"
+#include "fake_core.h"
+
+// CONFIG
+static bool opt_batch   = true;
+static bool opt_verify  = true; // use add_new_block, which does verification before calling add_block
+static bool opt_resume  = true;
+static bool opt_testnet = true;
+
+// number of blocks per batch transaction
+// adjustable through command-line argument according to available RAM
+static uint64_t db_batch_size = 20000;
+
+
+namespace po = boost::program_options;
+
+using namespace cryptonote;
+using namespace epee;
+
+
+int parse_db_arguments(const std::string& db_arg_str, std::string& db_engine, int& mdb_flags)
+{
+  std::vector<std::string> db_args;
+  boost::split(db_args, db_arg_str, boost::is_any_of("#"));
+  db_engine = db_args.front();
+  boost::algorithm::trim(db_engine);
+
+  if (db_args.size() == 1)
+  {
+    return 0;
+  }
+  else if (db_args.size() > 2)
+  {
+    std::cerr << "unrecognized database argument format: " << db_arg_str << ENDL;
+    return 1;
+  }
+
+  std::string db_arg_str2 = db_args[1];
+  boost::split(db_args, db_arg_str2, boost::is_any_of(","));
+  for (auto& it : db_args)
+  {
+    boost::algorithm::trim(it);
+    if (it.empty())
+      continue;
+    LOG_PRINT_L1("LMDB flag: " << it);
+    if (it == "nosync")
+    {
+      mdb_flags |= MDB_NOSYNC;
+    }
+    else if (it == "nometasync")
+    {
+      mdb_flags |= MDB_NOMETASYNC;
+    }
+    else if (it == "writemap")
+    {
+      mdb_flags |= MDB_WRITEMAP;
+    }
+    else if (it == "mapasync")
+    {
+      mdb_flags |= MDB_MAPASYNC;
+    }
+    else
+    {
+      std::cerr << "unrecognized database flag: " << it << ENDL;
+      return 1;
+    }
+  }
+  return 0;
+}
+
+
+int count_blocks(std::string& import_file_path)
+{
+  boost::filesystem::path raw_file_path(import_file_path);
+  boost::system::error_code ec;
+  if (!boost::filesystem::exists(raw_file_path, ec))
+  {
+    LOG_PRINT_L0("import file not found: " << raw_file_path);
+    throw std::runtime_error("Aborting");
+  }
+  std::ifstream import_file;
+  import_file.open(import_file_path, std::ios_base::binary | std::ifstream::in);
+
+  uint64_t h = 0;
+  if (import_file.fail())
+  {
+    LOG_PRINT_L0("import_file.open() fail");
+    throw std::runtime_error("Aborting");
+  }
+  LOG_PRINT_L0("Scanning blockchain from import file...");
+  char buffer1[STR_LENGTH_OF_INT + 1];
+  block b;
+  transaction tx;
+  bool quit = false;
+  uint64_t bytes_read = 0;
+  int progress_interval = 10;
+
+  while (! quit)
+  {
+    int chunk_size;
+    import_file.read(buffer1, STR_LENGTH_OF_INT);
+    if (!import_file) {
+      std::cout << "\r                   \r";
+      LOG_PRINT_L1("End of import file reached");
+      quit = true;
+      break;
+    }
+    h += NUM_BLOCKS_PER_CHUNK;
+    if (h % progress_interval == 0)
+    {
+      std::cout << "\r                   \r" << "block height: " << h <<
+        std::flush;
+    }
+    bytes_read += STR_LENGTH_OF_INT;
+    buffer1[STR_LENGTH_OF_INT] = '\0';
+    chunk_size = atoi(buffer1);
+    if (chunk_size > BUFFER_SIZE)
+    {
+      std::cout << "\r                   \r";
+      LOG_PRINT_L0("WARNING: chunk_size " << chunk_size << " > BUFFER_SIZE " << BUFFER_SIZE
+          << "  height: " << h);
+      throw std::runtime_error("Aborting: chunk size exceeds buffer size");
+    }
+    if (chunk_size > 100000)
+    {
+      std::cout << "\r                   \r";
+      LOG_PRINT_L0("WARNING: chunk_size " << chunk_size << " > 100000" << "  height: "
+          << h);
+    }
+    else if (chunk_size <= 0) {
+      std::cout << "\r                   \r";
+      LOG_PRINT_L0("ERROR: chunk_size " << chunk_size << " <= 0" << "  height: " << h);
+      throw std::runtime_error("Aborting");
+    }
+    // skip to next expected block size value
+    import_file.seekg(chunk_size, std::ios_base::cur);
+    if (! import_file) {
+      std::cout << "\r                   \r";
+      LOG_PRINT_L0("ERROR: unexpected end of import file: bytes read before error: "
+          << import_file.gcount() << " of chunk_size " << chunk_size);
+      throw std::runtime_error("Aborting");
+    }
+    bytes_read += chunk_size;
+    std::cout << "\r                   \r";
+    LOG_PRINT_L3("Total bytes scanned: " << bytes_read);
+  }
+
+  import_file.close();
+
+  std::cout << ENDL;
+  std::cout << "Done scanning import file" << ENDL;
+  std::cout << "Total bytes scanned: " << bytes_read << ENDL;
+  std::cout << "Height: " << h << ENDL;
+
+  return h;
+}
+
+template <typename FakeCore>
+int import_from_file(FakeCore& simple_core, std::string& import_file_path)
+{
+#if !defined(BLOCKCHAIN_DB)
+  static_assert(std::is_same<fake_core_memory, FakeCore>::value || std::is_same<fake_core_lmdb, FakeCore>::value,
+      "FakeCore constraint error");
+#endif
+#if !defined(BLOCKCHAIN_DB) || (BLOCKCHAIN_DB == DB_LMDB)
+  if (std::is_same<fake_core_lmdb, FakeCore>::value)
+  {
+    // Reset stats, in case we're using newly created db, accumulating stats
+    // from addition of genesis block.
+    // This aligns internal db counts with importer counts.
+    simple_core.m_storage.get_db()->reset_stats();
+  }
+#endif
+  boost::filesystem::path raw_file_path(import_file_path);
+  boost::system::error_code ec;
+  if (!boost::filesystem::exists(raw_file_path, ec))
+  {
+    LOG_PRINT_L0("import file not found: " << raw_file_path);
+    return false;
+  }
+
+  uint64_t source_height = count_blocks(import_file_path);
+  LOG_PRINT_L0("import file blockchain height: " << source_height);
+
+  std::ifstream import_file;
+  import_file.open(import_file_path, std::ios_base::binary | std::ifstream::in);
+
+  uint64_t h = 0;
+  if (import_file.fail())
+  {
+    LOG_PRINT_L0("import_file.open() fail");
+    return false;
+  }
+  char buffer1[STR_LENGTH_OF_INT + 1];
+  char buffer_block[BUFFER_SIZE];
+  block b;
+  transaction tx;
+  int quit = 0;
+  uint64_t bytes_read = 0;
+
+  uint64_t start_height = 1;
+  if (opt_resume)
+    start_height = simple_core.m_storage.get_current_blockchain_height();
+
+  // Note that a new blockchain will start with a height of 1 (block number 0)
+  // due to genesis block being added at initialization.
+
+  // CONFIG
+  // TODO: can expand on this, e.g. with --block-number option
+  uint64_t stop_height = source_height;
+
+  // These are what we'll try to use, and they don't have to be a determination
+  // from source and destination blockchains, but those are the current
+  // defaults.
+  LOG_PRINT_L0("start height: " << start_height << "  stop height: " <<
+      stop_height);
+
+  bool use_batch = false;
+  if (opt_batch)
+  {
+    if (simple_core.support_batch)
+      use_batch = true;
+    else
+      LOG_PRINT_L0("WARNING: batch transactions enabled but unsupported or unnecessary for this database engine - ignoring");
+  }
+
+  if (use_batch)
+    simple_core.batch_start();
+
+  LOG_PRINT_L0("Reading blockchain from import file...");
+  std::cout << ENDL;
+
+  // Within the loop, we skip to start_height before we start adding.
+  // TODO: Not a bottleneck, but we can use what's done in count_blocks() and
+  // only do the chunk size reads, skipping the chunk content reads until we're
+  // at start_height.
+  while (! quit)
+  {
+    int chunk_size;
+    import_file.read(buffer1, STR_LENGTH_OF_INT);
+    if (! import_file) {
+      std::cout << "\r                   \r";
+      LOG_PRINT_L0("End of import file reached");
+      quit = 1;
+      break;
+    }
+    bytes_read += STR_LENGTH_OF_INT;
+    buffer1[STR_LENGTH_OF_INT] = '\0';
+    chunk_size = atoi(buffer1);
+    if (chunk_size > BUFFER_SIZE)
+    {
+      LOG_PRINT_L0("WARNING: chunk_size " << chunk_size << " > BUFFER_SIZE " << BUFFER_SIZE);
+      throw std::runtime_error("Aborting: chunk size exceeds buffer size");
+    }
+    if (chunk_size > 100000)
+    {
+      LOG_PRINT_L0("WARNING: chunk_size " << chunk_size << " > 100000");
+    }
+    else if (chunk_size < 0) {
+      LOG_PRINT_L0("ERROR: chunk_size " << chunk_size << " < 0");
+      return 2;
+    }
+    import_file.read(buffer_block, chunk_size);
+    if (! import_file) {
+      LOG_PRINT_L0("ERROR: unexpected end of import file: bytes read before error: "
+          << import_file.gcount() << " of chunk_size " << chunk_size);
+      return 2;
+    }
+    bytes_read += chunk_size;
+    LOG_PRINT_L3("Total bytes read: " << bytes_read);
+
+    if (h + NUM_BLOCKS_PER_CHUNK < start_height + 1)
+    {
+      h += NUM_BLOCKS_PER_CHUNK;
+      continue;
+    }
+    if (h > stop_height)
+    {
+      LOG_PRINT_L0("Specified height reached - stopping.  height: " << h << "  block: " << h-1);
+      quit = 1;
+      break;
+    }
+
+    try
+    {
+      boost::iostreams::basic_array_source<char> device(buffer_block, chunk_size);
+      boost::iostreams::stream<boost::iostreams::basic_array_source<char>> s(device);
+      boost::archive::binary_iarchive a(s);
+
+      int display_interval = 1000;
+      int progress_interval = 10;
+      for (int chunk_ind = 0; chunk_ind < NUM_BLOCKS_PER_CHUNK; chunk_ind++)
+      {
+        h++;
+        if (h % display_interval == 0)
+        {
+          std::cout << "\r                   \r";
+          LOG_PRINT_L0("loading block height " << h);
+        }
+        else
+        {
+          LOG_PRINT_L3("loading block height " << h);
+        }
+        try {
+          a >> b;
+        }
+        catch (const std::exception& e)
+        {
+          std::cout << "\r                   \r";
+          LOG_PRINT_RED_L0("exception while de-archiving block, height=" << h);
+          quit = 1;
+          break;
+        }
+        LOG_PRINT_L2("block prev_id: " << b.prev_id << ENDL);
+
+        if (h % progress_interval == 0)
+        {
+          std::cout << "\r                   \r" << "block " << h-1
+            << std::flush;
+        }
+
+        std::vector<transaction> txs;
+
+        int num_txs;
+        try
+        {
+          a >> num_txs;
+        }
+        catch (const std::exception& e)
+        {
+          std::cout << "\r                   \r";
+          LOG_PRINT_RED_L0("exception while de-archiving tx-num, height=" << h);
+          quit = 1;
+          break;
+        }
+        for(int tx_num = 1; tx_num <= num_txs; tx_num++)
+        {
+          try {
+                a >> tx;
+          }
+          catch (const std::exception& e)
+          {
+            LOG_PRINT_RED_L0("exception while de-archiving tx, height=" << h <<", tx_num=" << tx_num);
+            quit = 1;
+            break;
+          }
+          // if (tx_num == 1) {
+          //   std::cout << "coinbase transaction" << ENDL;
+          // }
+          // crypto::hash hsh = null_hash;
+          // size_t blob_size = 0;
+          // NOTE: all tx hashes except for coinbase tx are available in the block data
+          // get_transaction_hash(tx, hsh, blob_size);
+          // LOG_PRINT_L0("tx " << tx_num << "  " << hsh << " : " << ENDL);
+          // LOG_PRINT_L0(obj_to_json_str(tx) << ENDL);
+
+          // add blocks with verification.
+          // for Blockchain and blockchain_storage add_new_block().
+          if (opt_verify)
+          {
+            if (tx_num == 1) {
+              continue; // coinbase transaction. no need to insert to tx_pool.
+            }
+            // crypto::hash hsh = null_hash;
+            // size_t blob_size = 0;
+            // get_transaction_hash(tx, hsh, blob_size);
+            tx_verification_context tvc = AUTO_VAL_INIT(tvc);
+            bool r = true;
+            r = simple_core.m_pool.add_tx(tx, tvc, true);
+            if (!r)
+            {
+              LOG_PRINT_RED_L0("failed to add transaction to transaction pool, height=" << h <<", tx_num=" << tx_num);
+              quit = 1;
+              break;
+            }
+          }
+          else
+          {
+            // for add_block() method, without (much) processing.
+            // don't add coinbase transaction to txs.
+            //
+            // because add_block() calls
+            // add_transaction(blk_hash, blk.miner_tx) first, and
+            // then a for loop for the transactions in txs.
+            if (tx_num > 1)
+            {
+              txs.push_back(tx);
+            }
+          }
+        }
+
+        if (opt_verify)
+        {
+          block_verification_context bvc = boost::value_initialized<block_verification_context>();
+          simple_core.m_storage.add_new_block(b, bvc);
+
+          if (bvc.m_verifivation_failed)
+          {
+            LOG_PRINT_L0("Failed to add block to blockchain, verification failed, height = " << h);
+            LOG_PRINT_L0("skipping rest of import file");
+            // ok to commit previously batched data because it failed only in
+            // verification of potential new block with nothing added to batch
+            // yet
+            quit = 1;
+            break;
+          }
+          if (! bvc.m_added_to_main_chain)
+          {
+            LOG_PRINT_L0("Failed to add block to blockchain, height = " << h);
+            LOG_PRINT_L0("skipping rest of import file");
+            // make sure we don't commit partial block data
+            quit = 2;
+            break;
+          }
+        }
+        else
+        {
+          size_t block_size;
+          difficulty_type cumulative_difficulty;
+          uint64_t coins_generated;
+
+          a >> block_size;
+          a >> cumulative_difficulty;
+          a >> coins_generated;
+
+          std::cout << "\r                   \r";
+          LOG_PRINT_L2("block_size: " << block_size);
+          LOG_PRINT_L2("cumulative_difficulty: " << cumulative_difficulty);
+          LOG_PRINT_L2("coins_generated: " << coins_generated);
+
+          try
+          {
+            simple_core.add_block(b, block_size, cumulative_difficulty, coins_generated, txs);
+          }
+          catch (const std::exception& e)
+          {
+            std::cout << "\r                   \r";
+            LOG_PRINT_RED_L0("Error adding block to blockchain: " << e.what());
+            quit = 2; // make sure we don't commit partial block data
+            break;
+          }
+        }
+
+        if (use_batch)
+        {
+          if (h % db_batch_size == 0)
+          {
+            std::cout << "\r                   \r";
+            std::cout << ENDL << "[- batch commit at height " << h << " -]" << ENDL;
+            simple_core.batch_stop();
+            simple_core.batch_start();
+            std::cout << ENDL;
+#if !defined(BLOCKCHAIN_DB) || (BLOCKCHAIN_DB == DB_LMDB)
+            simple_core.m_storage.get_db()->show_stats();
+#endif
+          }
+        }
+      }
+    }
+    catch (const std::exception& e)
+    {
+      std::cout << "\r                   \r";
+      LOG_PRINT_RED_L0("exception while reading from import file, height=" << h);
+      return 2;
+    }
+  } // while
+
+  import_file.close();
+
+  if (use_batch)
+  {
+    if (quit > 1)
+    {
+      // There was an error, so don't commit pending data.
+      // Destructor will abort write txn.
+    }
+    else
+    {
+      simple_core.batch_stop();
+    }
+#if !defined(BLOCKCHAIN_DB) || (BLOCKCHAIN_DB == DB_LMDB)
+    simple_core.m_storage.get_db()->show_stats();
+#endif
+    if (h > 0)
+      LOG_PRINT_L0("Finished at height: " << h << "  block: " << h-1);
+  }
+  std::cout << ENDL;
+  return 0;
+}
+
+int main(int argc, char* argv[])
+{
+  std::string import_filename = BLOCKCHAIN_RAW;
+#if defined(BLOCKCHAIN_DB) && (BLOCKCHAIN_DB == DB_MEMORY)
+  std::string default_db_engine = "memory";
+#else
+  std::string default_db_engine = "lmdb";
+#endif
+
+  uint32_t log_level = LOG_LEVEL_0;
+  std::string dirname;
+  std::string db_arg_str;
+
+  boost::filesystem::path default_data_path {tools::get_default_data_dir()};
+  boost::filesystem::path default_testnet_data_path {default_data_path / "testnet"};
+
+  po::options_description desc_cmd_only("Command line options");
+  po::options_description desc_cmd_sett("Command line options and settings options");
+  const command_line::arg_descriptor<uint32_t> arg_log_level   =  {"log-level",  "", log_level};
+  const command_line::arg_descriptor<uint64_t> arg_batch_size  =  {"batch-size", "", db_batch_size};
+  const command_line::arg_descriptor<bool>     arg_testnet_on  = {
+    "testnet"
+      , "Run on testnet."
+      , false
+  };
+  const command_line::arg_descriptor<bool>     arg_count_blocks = {
+    "count-blocks"
+      , "Count blocks in import file and exit"
+      , false
+  };
+  const command_line::arg_descriptor<std::string> arg_database = {
+    "database", "available: memory, lmdb"
+      , default_db_engine
+  };
+  const command_line::arg_descriptor<bool> arg_verify =  {"verify",
+    "Verify blocks and transactions during import", true};
+  const command_line::arg_descriptor<bool> arg_batch  =  {"batch",
+    "Batch transactions for faster import", true};
+  const command_line::arg_descriptor<bool> arg_resume =  {"resume",
+    "Resume from current height if output database already exists", true};
+
+  command_line::add_arg(desc_cmd_sett, command_line::arg_data_dir, default_data_path.string());
+  command_line::add_arg(desc_cmd_sett, command_line::arg_testnet_data_dir, default_testnet_data_path.string());
+  command_line::add_arg(desc_cmd_sett, arg_log_level);
+  command_line::add_arg(desc_cmd_sett, arg_batch_size);
+  command_line::add_arg(desc_cmd_sett, arg_testnet_on);
+  command_line::add_arg(desc_cmd_sett, arg_database);
+
+  command_line::add_arg(desc_cmd_only, arg_count_blocks);
+  command_line::add_arg(desc_cmd_only, command_line::arg_help);
+
+  // call add_options() directly for these arguments since
+  // command_line helpers support only boolean switch, not boolean argument
+  desc_cmd_sett.add_options()
+    (arg_verify.name, make_semantic(arg_verify), arg_verify.description)
+    (arg_batch.name,  make_semantic(arg_batch),  arg_batch.description)
+    (arg_resume.name, make_semantic(arg_resume), arg_resume.description)
+    ;
+
+  po::options_description desc_options("Allowed options");
+  desc_options.add(desc_cmd_only).add(desc_cmd_sett);
+
+  po::variables_map vm;
+  bool r = command_line::handle_error_helper(desc_options, [&]()
+  {
+    po::store(po::parse_command_line(argc, argv, desc_options), vm);
+    po::notify(vm);
+    return true;
+  });
+  if (! r)
+    return 1;
+
+  log_level     = command_line::get_arg(vm, arg_log_level);
+  opt_verify    = command_line::get_arg(vm, arg_verify);
+  opt_batch     = command_line::get_arg(vm, arg_batch);
+  opt_resume    = command_line::get_arg(vm, arg_resume);
+  db_batch_size = command_line::get_arg(vm, arg_batch_size);
+
+  if (command_line::get_arg(vm, command_line::arg_help))
+  {
+    std::cout << CRYPTONOTE_NAME << " v" << MONERO_VERSION_FULL << ENDL << ENDL;
+    std::cout << desc_options << std::endl;
+    return 1;
+  }
+
+  if (! opt_batch && ! vm["batch-size"].defaulted())
+  {
+    std::cerr << "Error: batch-size set, but batch option not enabled" << ENDL;
+    exit(1);
+  }
+  if (! db_batch_size)
+  {
+    std::cerr << "Error: batch-size must be > 0" << ENDL;
+    exit(1);
+  }
+
+  std::vector<std::string> db_engines {"memory", "lmdb"};
+
+  opt_testnet = command_line::get_arg(vm, arg_testnet_on);
+  auto data_dir_arg = opt_testnet ? command_line::arg_testnet_data_dir : command_line::arg_data_dir;
+  dirname = command_line::get_arg(vm, data_dir_arg);
+  db_arg_str = command_line::get_arg(vm, arg_database);
+
+  log_space::get_set_log_detalisation_level(true, log_level);
+  log_space::log_singletone::add_logger(LOGGER_CONSOLE, NULL, NULL);
+  LOG_PRINT_L0("Starting...");
+  LOG_PRINT_L0("Setting log level = " << log_level);
+
+  boost::filesystem::path file_path {dirname};
+  std::string import_file_path;
+
+  import_file_path = (file_path / "export" / import_filename).string();
+
+  if (command_line::has_arg(vm, arg_count_blocks))
+  {
+    count_blocks(import_file_path);
+    exit(0);
+  }
+
+
+  std::string db_engine;
+  int mdb_flags = 0;
+  int res = 0;
+  res = parse_db_arguments(db_arg_str, db_engine, mdb_flags);
+  if (res)
+  {
+    std::cerr << "Error parsing database argument(s)" << ENDL;
+    exit(1);
+  }
+
+  if (std::find(db_engines.begin(), db_engines.end(), db_engine) == db_engines.end())
+  {
+    std::cerr << "Invalid database engine: " << db_engine << std::endl;
+    exit(1);
+  }
+
+  LOG_PRINT_L0("database: " << db_engine);
+  LOG_PRINT_L0("verify:  " << std::boolalpha << opt_verify << std::noboolalpha);
+  if (opt_batch)
+  {
+    LOG_PRINT_L0("batch:   " << std::boolalpha << opt_batch << std::noboolalpha
+        << "  batch size: " << db_batch_size);
+  }
+  else
+  {
+    LOG_PRINT_L0("batch:   " << std::boolalpha << opt_batch << std::noboolalpha);
+  }
+  LOG_PRINT_L0("resume:  " << std::boolalpha << opt_resume  << std::noboolalpha);
+  LOG_PRINT_L0("testnet: " << std::boolalpha << opt_testnet << std::noboolalpha);
+
+  std::cout << "import file path: " << import_file_path << ENDL;
+  std::cout << "database path:    " << file_path.string() << ENDL;
+
+  try
+  {
+
+  // fake_core needed for verification to work when enabled.
+  //
+  // NOTE: don't need fake_core method of doing things when we're going to call
+  // BlockchainDB add_block() directly and have available the 3 block
+  // properties to do so. Both ways work, but fake core isn't necessary in that
+  // circumstance.
+
+  // for multi_db_runtime:
+#if !defined(BLOCKCHAIN_DB)
+  if (db_engine == "lmdb")
+  {
+    fake_core_lmdb simple_core(dirname, opt_testnet, opt_batch, mdb_flags);
+    import_from_file(simple_core, import_file_path);
+  }
+  else if (db_engine == "memory")
+  {
+    fake_core_memory simple_core(dirname, opt_testnet);
+    import_from_file(simple_core, import_file_path);
+  }
+  else
+  {
+    std::cerr << "database engine unrecognized" << ENDL;
+    exit(1);
+  }
+
+  // for multi_db_compile:
+#else
+  if (db_engine != default_db_engine)
+  {
+    std::cerr << "Invalid database engine for compiled version: " << db_engine << std::endl;
+    exit(1);
+  }
+#if BLOCKCHAIN_DB == DB_LMDB
+  fake_core_lmdb simple_core(dirname, opt_testnet, opt_batch, mdb_flags);
+#else
+  fake_core_memory simple_core(dirname, opt_testnet);
+#endif
+
+  import_from_file(simple_core, import_file_path);
+#endif
+
+  }
+  catch (const DB_ERROR& e)
+  {
+    std::cout << std::string("Error loading blockchain db: ") + e.what() + " -- shutting down now" << ENDL;
+    exit(1);
+  }
+
+  // destructors called at exit:
+  //
+  // ensure db closed
+  //   - transactions properly checked and handled
+  //   - disk sync if needed
+  //
+  // fake_core object's destructor is called when it goes out of scope. For an
+  // LMDB fake_core, it calls Blockchain::deinit() on its object, which in turn
+  // calls delete on its BlockchainDB derived class' object, which closes its
+  // files.
+}

src/blockchain_converter/fake_core.h

@@ -0,0 +1,143 @@
+// Copyright (c) 2014-2015, The Monero Project
+//
+// All rights reserved.
+//
+// Redistribution and use in source and binary forms, with or without modification, are
+// permitted provided that the following conditions are met:
+//
+// 1. Redistributions of source code must retain the above copyright notice, this list of
+//    conditions and the following disclaimer.
+//
+// 2. Redistributions in binary form must reproduce the above copyright notice, this list
+//    of conditions and the following disclaimer in the documentation and/or other
+//    materials provided with the distribution.
+//
+// 3. Neither the name of the copyright holder nor the names of its contributors may be
+//    used to endorse or promote products derived from this software without specific
+//    prior written permission.
+//
+// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY
+// EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
+// MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL
+// THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+// SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
+// PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+// INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+// STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF
+// THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+#pragma once
+
+#include <boost/filesystem.hpp>
+#include "cryptonote_core/blockchain.h" // BlockchainDB and LMDB
+#include "cryptonote_core/blockchain_storage.h" // in-memory DB
+
+using namespace cryptonote;
+
+
+#if !defined(BLOCKCHAIN_DB) || BLOCKCHAIN_DB == DB_LMDB
+
+struct fake_core_lmdb
+{
+  Blockchain m_storage;
+  tx_memory_pool m_pool;
+  bool support_batch;
+  bool support_add_block;
+
+  // for multi_db_runtime:
+#if !defined(BLOCKCHAIN_DB)
+  fake_core_lmdb(const boost::filesystem::path &path, const bool use_testnet=false, const bool do_batch=true, const int mdb_flags=0) : m_pool(&m_storage), m_storage(m_pool)
+  // for multi_db_compile:
+#else
+  fake_core_lmdb(const boost::filesystem::path &path, const bool use_testnet=false, const bool do_batch=true, const int mdb_flags=0) : m_pool(m_storage), m_storage(m_pool)
+#endif
+  {
+    m_pool.init(path.string());
+    m_storage.init(path.string(), use_testnet, mdb_flags);
+    if (do_batch)
+      m_storage.get_db()->set_batch_transactions(do_batch);
+    support_batch = true;
+    support_add_block = true;
+  }
+  ~fake_core_lmdb()
+  {
+    m_storage.deinit();
+  }
+
+  uint64_t add_block(const block& blk
+                            , const size_t& block_size
+                            , const difficulty_type& cumulative_difficulty
+                            , const uint64_t& coins_generated
+                            , const std::vector<transaction>& txs
+                            )
+  {
+    return m_storage.get_db()->add_block(blk, block_size, cumulative_difficulty, coins_generated, txs);
+  }
+
+  void batch_start()
+  {
+    m_storage.get_db()->batch_start();
+  }
+
+  void batch_stop()
+  {
+    m_storage.get_db()->batch_stop();
+  }
+
+};
+#endif
+
+#if !defined(BLOCKCHAIN_DB) || BLOCKCHAIN_DB == DB_MEMORY
+
+struct fake_core_memory
+{
+  blockchain_storage m_storage;
+  tx_memory_pool m_pool;
+  bool support_batch;
+  bool support_add_block;
+
+  // for multi_db_runtime:
+#if !defined(BLOCKCHAIN_DB)
+  fake_core_memory(const boost::filesystem::path &path, const bool use_testnet=false) : m_pool(&m_storage), m_storage(m_pool)
+#else
+  // for multi_db_compile:
+  fake_core_memory(const boost::filesystem::path &path, const bool use_testnet=false) : m_pool(m_storage), m_storage(&m_pool)
+#endif
+  {
+    m_pool.init(path.string());
+    m_storage.init(path.string(), use_testnet);
+    support_batch = false;
+    support_add_block = false;
+  }
+  ~fake_core_memory()
+  {
+    LOG_PRINT_L3("fake_core_memory() destructor called - want to see it ripple down");
+    m_storage.deinit();
+  }
+
+  uint64_t add_block(const block& blk
+                            , const size_t& block_size
+                            , const difficulty_type& cumulative_difficulty
+                            , const uint64_t& coins_generated
+                            , const std::vector<transaction>& txs
+                            )
+  {
+    // TODO:
+    // would need to refactor handle_block_to_main_chain() to have a direct add_block() method like Blockchain class
+    throw std::runtime_error("direct add_block() method not implemented for in-memory db");
+    return 2;
+  }
+
+  void batch_start()
+  {
+    LOG_PRINT_L0("WARNING: [batch_start] opt_batch set, but this database doesn't support/need transactions - ignoring");
+  }
+
+  void batch_stop()
+  {
+    LOG_PRINT_L0("WARNING: [batch_stop] opt_batch set, but this database doesn't support/need transactions - ignoring");
+  }
+
+};
+
+#endif

src/blockchain_converter/import.h

@@ -0,0 +1,37 @@
+// Copyright (c) 2014-2015, The Monero Project
+//
+// All rights reserved.
+//
+// Redistribution and use in source and binary forms, with or without modification, are
+// permitted provided that the following conditions are met:
+//
+// 1. Redistributions of source code must retain the above copyright notice, this list of
+//    conditions and the following disclaimer.
+//
+// 2. Redistributions in binary form must reproduce the above copyright notice, this list
+//    of conditions and the following disclaimer in the documentation and/or other
+//    materials provided with the distribution.
+//
+// 3. Neither the name of the copyright holder nor the names of its contributors may be
+//    used to endorse or promote products derived from this software without specific
+//    prior written permission.
+//
+// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY
+// EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
+// MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL
+// THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+// SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
+// PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+// INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+// STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF
+// THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+#pragma once
+
+// TODO: bounds checking is done before writing to buffer, but buffer size
+// should be a sensible maximum
+#define BUFFER_SIZE 1000000
+#define NUM_BLOCKS_PER_CHUNK 1
+#define STR_LENGTH_OF_INT 9
+#define STR_FORMAT_OF_INT "%09d"
+#define BLOCKCHAIN_RAW "blockchain.raw"

src/blockchain_db/blockchain_db.h

@@ -326,7 +326,7 @@ public:
   void show_stats();
 
   // open the db at location <filename>, or create it if there isn't one.
-  virtual void open(const std::string& filename) = 0;
+  virtual void open(const std::string& filename, const int db_flags = 0) = 0;
 
   // make sure implementation has a create function as well
   virtual void create(const std::string& filename) = 0;

src/blockchain_db/lmdb/db_lmdb.cpp

@@ -149,6 +149,20 @@ inline void lmdb_db_open(MDB_txn* txn, const char* name, int flags, MDB_dbi& dbi
 namespace cryptonote
 {
 
+// If m_batch_active is set, a batch transaction exists beyond this class, such
+// as a batch import with verification enabled, or possibly (later) a batch
+// network sync.
+//
+// For some of the lookup methods, such as get_block_timestamp(), tx_exists(),
+// and get_tx(), when m_batch_active is set, the lookup uses the batch
+// transaction. This isn't only because the transaction is available, but it's
+// necessary so that lookups include the database updates only present in the
+// current batch write.
+//
+// A regular network sync without batch writes is expected to open a new read
+// transaction, as those lookups are part of the validation done prior to the
+// write for block and tx data, so no write transaction is open at the time.
+
 void BlockchainLMDB::add_block( const block& blk
               , const size_t& block_size
               , const difficulty_type& cumulative_difficulty
@@ -621,7 +635,7 @@ BlockchainLMDB::BlockchainLMDB(bool batch_transactions)
   m_height = 0;
 }
 
-void BlockchainLMDB::open(const std::string& filename)
+void BlockchainLMDB::open(const std::string& filename, const int mdb_flags)
 {
   LOG_PRINT_L3("BlockchainLMDB::" << __func__);
 
@@ -661,7 +675,7 @@ void BlockchainLMDB::open(const std::string& filename)
   size_t mapsize = 1LL << 34;
   if (auto result = mdb_env_set_mapsize(m_env, mapsize))
     throw0(DB_ERROR(std::string("Failed to set max memory map size: ").append(mdb_strerror(result)).c_str()));
-  if (auto result = mdb_env_open(m_env, filename.c_str(), 0, 0644))
+  if (auto result = mdb_env_open(m_env, filename.c_str(), mdb_flags, 0644))
     throw0(DB_ERROR(std::string("Failed to open lmdb environment: ").append(mdb_strerror(result)).c_str()));
 
   // get a read/write MDB_txn
@@ -1260,14 +1274,6 @@ uint64_t BlockchainLMDB::get_tx_block_height(const crypto::hash& h) const
   LOG_PRINT_L3("BlockchainLMDB::" << __func__);
   check_open();
 
-  // If m_batch_active is set, a batch transaction exists beyond this class,
-  // such as a batch import with verification enabled, or possibly (later) a
-  // batch network sync.
-  //
-  // A regular network sync without batching would be expected to open a new
-  // read transaction here, as validation is done prior to the write for block
-  // and tx data.
-
   txn_safe txn;
   txn_safe* txn_ptr = &txn;
   if (m_batch_active)

src/blockchain_db/lmdb/db_lmdb.h

@@ -114,7 +114,7 @@ public:
   BlockchainLMDB(bool batch_transactions=false);
   ~BlockchainLMDB();
 
-  virtual void open(const std::string& filename);
+  virtual void open(const std::string& filename, const int mdb_flags=0);
 
   virtual void create(const std::string& filename);
 

src/cryptonote_core/blockchain.cpp

@@ -226,7 +226,7 @@ uint64_t Blockchain::get_current_blockchain_height() const
 //------------------------------------------------------------------
 //FIXME: possibly move this into the constructor, to avoid accidentally
 //       dereferencing a null BlockchainDB pointer
-bool Blockchain::init(const std::string& config_folder, bool testnet)
+bool Blockchain::init(const std::string& config_folder, const bool testnet, const int db_flags)
 {
   LOG_PRINT_L3("Blockchain::" << __func__);
   CRITICAL_REGION_LOCAL(m_blockchain_lock);
@@ -246,7 +246,7 @@ bool Blockchain::init(const std::string& config_folder, bool testnet)
   const std::string filename = folder.string();
   try
   {
-    m_db->open(filename);
+    m_db->open(filename, db_flags);
   }
   catch (const DB_OPEN_FAILURE& e)
   {

src/cryptonote_core/blockchain.h

@@ -81,7 +81,7 @@ namespace cryptonote
 
     Blockchain(tx_memory_pool& tx_pool);
 
-    bool init(const std::string& config_folder, bool testnet = false);
+    bool init(const std::string& config_folder, const bool testnet = false, const int db_flags = 0);
     bool deinit();
 
     void set_checkpoints(checkpoints&& chk_pts) { m_checkpoints = chk_pts; }
@@ -145,6 +145,11 @@ namespace cryptonote
     void set_enforce_dns_checkpoints(bool enforce);
     bool update_checkpoints(const std::string& file_path, bool check_dns);
 
+    BlockchainDB* get_db()
+    {
+      return m_db;
+    }
+
   private:
     typedef std::unordered_map<crypto::hash, size_t> blocks_by_id_index;
     typedef std::unordered_map<crypto::hash, transaction_chain_entry> transactions_container;