From 5833d66f6540e7b34e10ddef37c2b67bd501994b Mon Sep 17 00:00:00 2001 From: moneromooo-monero Date: Sun, 1 Jan 2017 16:34:23 +0000 Subject: Change logging to easylogging++ This replaces the epee and data_loggers logging systems with a single one, and also adds filename:line and explicit severity levels. Categories may be defined, and logging severity set by category (or set of categories). epee style 0-4 log level maps to a sensible severity configuration. Log files now also rotate when reaching 100 MB. To select which logs to output, use the MONERO_LOGS environment variable, with a comma separated list of categories (globs are supported), with their requested severity level after a colon. If a log matches more than one such setting, the last one in the configuration string applies. A few examples: This one is (mostly) silent, only outputting fatal errors: MONERO_LOGS=*:FATAL This one is very verbose: MONERO_LOGS=*:TRACE This one is totally silent (logwise): MONERO_LOGS="" This one outputs all errors and warnings, except for the "verify" category, which prints just fatal errors (the verify category is used for logs about incoming transactions and blocks, and it is expected that some/many will fail to verify, hence we don't want the spam): MONERO_LOGS=*:WARNING,verify:FATAL Log levels are, in decreasing order of priority: FATAL, ERROR, WARNING, INFO, DEBUG, TRACE Subcategories may be added using prefixes and globs. This example will output net.p2p logs at the TRACE level, but all other net* logs only at INFO: MONERO_LOGS=*:ERROR,net*:INFO,net.p2p:TRACE Logs which are intended for the user (which Monero was using a lot through epee, but really isn't a nice way to go things) should use the "global" category. There are a few helper macros for using this category, eg: MGINFO("this shows up by default") or MGINFO_RED("this is red"), to try to keep a similar look and feel for now. Existing epee log macros still exist, and map to the new log levels, but since they're used as a "user facing" UI element as much as a logging system, they often don't map well to log severities (ie, a log level 0 log may be an error, or may be something we want the user to see, such as an important info). In those cases, I tried to use the new macros. In other cases, I left the existing macros in. When modifying logs, it is probably best to switch to the new macros with explicit levels. The --log-level options and set_log commands now also accept category settings, in addition to the epee style log levels. --- src/blockchain_utilities/bootstrap_file.cpp | 68 +++++++++++++++-------------- 1 file changed, 35 insertions(+), 33 deletions(-) (limited to 'src/blockchain_utilities/bootstrap_file.cpp') diff --git a/src/blockchain_utilities/bootstrap_file.cpp b/src/blockchain_utilities/bootstrap_file.cpp index 61bd35a6f..add2c65e6 100644 --- a/src/blockchain_utilities/bootstrap_file.cpp +++ b/src/blockchain_utilities/bootstrap_file.cpp @@ -32,6 +32,8 @@ #include "bootstrap_file.h" +#undef MONERO_DEFAULT_LOG_CATEGORY +#define MONERO_DEFAULT_LOG_CATEGORY "bcutil" namespace po = boost::program_options; @@ -59,7 +61,7 @@ bool BootstrapFile::open_writer(const boost::filesystem::path& file_path) { if (!boost::filesystem::is_directory(dir_path)) { - LOG_PRINT_RED_L0("export directory path is a file: " << dir_path); + MFATAL("export directory path is a file: " << dir_path); return false; } } @@ -67,7 +69,7 @@ bool BootstrapFile::open_writer(const boost::filesystem::path& file_path) { if (!boost::filesystem::create_directory(dir_path)) { - LOG_PRINT_RED_L0("Failed to create directory " << dir_path); + MFATAL("Failed to create directory " << dir_path); return false; } } @@ -80,14 +82,14 @@ bool BootstrapFile::open_writer(const boost::filesystem::path& file_path) if (! boost::filesystem::exists(file_path)) { - LOG_PRINT_L0("creating file"); + MDEBUG("creating file"); do_initialize_file = true; num_blocks = 0; } else { num_blocks = count_blocks(file_path.string()); - LOG_PRINT_L0("appending to existing file with height: " << num_blocks-1 << " total blocks: " << num_blocks); + MDEBUG("appending to existing file with height: " << num_blocks-1 << " total blocks: " << num_blocks); } m_height = num_blocks; @@ -138,7 +140,7 @@ bool BootstrapFile::initialize_file() uint32_t bd_size = 0; blobdata bd = t_serializable_object_to_blob(bfi); - LOG_PRINT_L1("bootstrap::file_info size: " << bd.size()); + MDEBUG("bootstrap::file_info size: " << bd.size()); bd_size = bd.size(); if (! ::serialization::dump_binary(bd_size, blob)) @@ -149,7 +151,7 @@ bool BootstrapFile::initialize_file() *output_stream_header << bd; bd = t_serializable_object_to_blob(bbi); - LOG_PRINT_L1("bootstrap::blocks_info size: " << bd.size()); + MDEBUG("bootstrap::blocks_info size: " << bd.size()); bd_size = bd.size(); if (! ::serialization::dump_binary(bd_size, blob)) @@ -172,10 +174,10 @@ void BootstrapFile::flush_chunk() m_output_stream->flush(); uint32_t chunk_size = m_buffer.size(); - // LOG_PRINT_L0("chunk_size " << chunk_size); + // MTRACE("chunk_size " << chunk_size); if (chunk_size > BUFFER_SIZE) { - LOG_PRINT_L0("WARNING: chunk_size " << chunk_size << " > BUFFER_SIZE " << BUFFER_SIZE); + MWARNING("WARNING: chunk_size " << chunk_size << " > BUFFER_SIZE " << BUFFER_SIZE); } std::string blob; @@ -196,14 +198,14 @@ void BootstrapFile::flush_chunk() long num_chars_written = pos_after - pos_before; if (static_cast(num_chars_written) != chunk_size) { - LOG_PRINT_RED_L0("Error writing chunk: height: " << m_cur_height << " chunk_size: " << chunk_size << " num chars written: " << num_chars_written); + MFATAL("Error writing chunk: height: " << m_cur_height << " chunk_size: " << chunk_size << " num chars written: " << num_chars_written); throw std::runtime_error("Error writing chunk"); } m_buffer.clear(); delete m_output_stream; m_output_stream = new boost::iostreams::stream>(m_buffer); - LOG_PRINT_L1("flushed chunk: chunk_size: " << chunk_size); + MDEBUG("flushed chunk: chunk_size: " << chunk_size); } void BootstrapFile::write_block(block& block) @@ -267,10 +269,10 @@ bool BootstrapFile::store_blockchain_raw(Blockchain* _blockchain_storage, tx_mem m_blockchain_storage = _blockchain_storage; m_tx_pool = _tx_pool; uint64_t progress_interval = 100; - LOG_PRINT_L0("Storing blocks raw data..."); + MINFO("Storing blocks raw data..."); if (!BootstrapFile::open_writer(output_file)) { - LOG_PRINT_RED_L0("failed to open raw file for write"); + MFATAL("failed to open raw file for write"); return false; } block b; @@ -280,16 +282,16 @@ bool BootstrapFile::store_blockchain_raw(Blockchain* _blockchain_storage, tx_mem // height. uint64_t block_start = m_height; uint64_t block_stop = 0; - LOG_PRINT_L0("source blockchain height: " << m_blockchain_storage->get_current_blockchain_height()-1); + MINFO("source blockchain height: " << m_blockchain_storage->get_current_blockchain_height()-1); if ((requested_block_stop > 0) && (requested_block_stop < m_blockchain_storage->get_current_blockchain_height())) { - LOG_PRINT_L0("Using requested block height: " << requested_block_stop); + MINFO("Using requested block height: " << requested_block_stop); block_stop = requested_block_stop; } else { block_stop = m_blockchain_storage->get_current_blockchain_height() - 1; - LOG_PRINT_L0("Using block height of source blockchain: " << block_stop); + MINFO("Using block height of source blockchain: " << block_stop); } for (m_cur_height = block_start; m_cur_height <= block_stop; ++m_cur_height) { @@ -315,9 +317,9 @@ bool BootstrapFile::store_blockchain_raw(Blockchain* _blockchain_storage, tx_mem std::cout << refresh_string; std::cout << "block " << m_cur_height-1 << "/" << block_stop << ENDL; - LOG_PRINT_L0("Number of blocks exported: " << num_blocks_written); + MINFO("Number of blocks exported: " << num_blocks_written); if (num_blocks_written > 0) - LOG_PRINT_L0("Largest chunk: " << m_max_chunk << " bytes"); + MINFO("Largest chunk: " << m_max_chunk << " bytes"); return BootstrapFile::close(); } @@ -338,11 +340,11 @@ uint64_t BootstrapFile::seek_to_first_chunk(std::ifstream& import_file) if (file_magic != blockchain_raw_magic) { - LOG_PRINT_RED_L0("bootstrap file not recognized"); + MFATAL("bootstrap file not recognized"); throw std::runtime_error("Aborting"); } else - LOG_PRINT_L0("bootstrap file recognized"); + MINFO("bootstrap file recognized"); uint32_t buflen_file_info; @@ -352,7 +354,7 @@ uint64_t BootstrapFile::seek_to_first_chunk(std::ifstream& import_file) throw std::runtime_error("Error reading expected number of bytes"); if (! ::serialization::parse_binary(str1, buflen_file_info)) throw std::runtime_error("Error in deserialization of buflen_file_info"); - LOG_PRINT_L1("bootstrap::file_info size: " << buflen_file_info); + MINFO("bootstrap::file_info size: " << buflen_file_info); if (buflen_file_info > sizeof(buf1)) throw std::runtime_error("Error: bootstrap::file_info size exceeds buffer size"); @@ -363,9 +365,9 @@ uint64_t BootstrapFile::seek_to_first_chunk(std::ifstream& import_file) bootstrap::file_info bfi; if (! ::serialization::parse_binary(str1, bfi)) throw std::runtime_error("Error in deserialization of bootstrap::file_info"); - LOG_PRINT_L0("bootstrap file v" << unsigned(bfi.major_version) << "." << unsigned(bfi.minor_version)); - LOG_PRINT_L0("bootstrap magic size: " << sizeof(file_magic)); - LOG_PRINT_L0("bootstrap header size: " << bfi.header_size); + MINFO("bootstrap file v" << unsigned(bfi.major_version) << "." << unsigned(bfi.minor_version)); + MINFO("bootstrap magic size: " << sizeof(file_magic)); + MINFO("bootstrap header size: " << bfi.header_size); uint64_t full_header_size = sizeof(file_magic) + bfi.header_size; import_file.seekg(full_header_size); @@ -379,7 +381,7 @@ uint64_t BootstrapFile::count_blocks(const std::string& import_file_path) boost::system::error_code ec; if (!boost::filesystem::exists(raw_file_path, ec)) { - LOG_PRINT_L0("bootstrap file not found: " << raw_file_path); + MFATAL("bootstrap file not found: " << raw_file_path); throw std::runtime_error("Aborting"); } std::ifstream import_file; @@ -388,14 +390,14 @@ uint64_t BootstrapFile::count_blocks(const std::string& import_file_path) uint64_t h = 0; if (import_file.fail()) { - LOG_PRINT_L0("import_file.open() fail"); + MFATAL("import_file.open() fail"); throw std::runtime_error("Aborting"); } uint64_t full_header_size; // 4 byte magic + length of header structures full_header_size = seek_to_first_chunk(import_file); - LOG_PRINT_L0("Scanning blockchain from bootstrap file..."); + MINFO("Scanning blockchain from bootstrap file..."); block b; bool quit = false; uint64_t bytes_read = 0; @@ -409,7 +411,7 @@ uint64_t BootstrapFile::count_blocks(const std::string& import_file_path) import_file.read(buf1, sizeof(chunk_size)); if (!import_file) { std::cout << refresh_string; - LOG_PRINT_L1("End of file reached"); + MDEBUG("End of file reached"); quit = true; break; } @@ -425,38 +427,38 @@ uint64_t BootstrapFile::count_blocks(const std::string& import_file_path) str1.assign(buf1, sizeof(chunk_size)); if (! ::serialization::parse_binary(str1, chunk_size)) throw std::runtime_error("Error in deserialization of chunk_size"); - LOG_PRINT_L3("chunk_size: " << chunk_size); + MDEBUG("chunk_size: " << chunk_size); if (chunk_size > BUFFER_SIZE) { std::cout << refresh_string; - LOG_PRINT_L0("WARNING: chunk_size " << chunk_size << " > BUFFER_SIZE " << BUFFER_SIZE + MWARNING("WARNING: chunk_size " << chunk_size << " > BUFFER_SIZE " << BUFFER_SIZE << " height: " << h-1); throw std::runtime_error("Aborting: chunk size exceeds buffer size"); } if (chunk_size > 100000) { std::cout << refresh_string; - LOG_PRINT_L0("NOTE: chunk_size " << chunk_size << " > 100000" << " height: " + MDEBUG("NOTE: chunk_size " << chunk_size << " > 100000" << " height: " << h-1); } else if (chunk_size <= 0) { std::cout << refresh_string; - LOG_PRINT_L0("ERROR: chunk_size " << chunk_size << " <= 0" << " height: " << h-1); + MDEBUG("ERROR: chunk_size " << chunk_size << " <= 0" << " height: " << h-1); 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 << refresh_string; - LOG_PRINT_L0("ERROR: unexpected end of file: bytes read before error: " + MFATAL("ERROR: unexpected end of file: bytes read before error: " << import_file.gcount() << " of chunk_size " << chunk_size); throw std::runtime_error("Aborting"); } bytes_read += chunk_size; // std::cout << refresh_string; - LOG_PRINT_L3("Number bytes scanned: " << bytes_read); + MINFO("Number bytes scanned: " << bytes_read); } import_file.close(); -- cgit v1.2.3