Nonblocking-Logger - A High-Performance Rust Logging Library

A high-performance logging library for Rust with format string support and WebAssembly compatibility. Nonblocking-Logger provides fast, efficient logging with minimal overhead.

🚀 Key Advantages Over Other Logging Crates

High Performance

• Fast I/O: Direct writing to targets with minimal overhead
• Low latency: Logs are written immediately
• Simple and reliable: Clean, straightforward implementation

Format String Support Like println!

• Familiar API: Use info!("User {} logged in", user_id) just like println!
• Full format specifiers: Support for {:.2}, {:?}, {:#x}, etc.
• Lazy evaluation: Expensive computations only run when needed with debug_lazy!()

WebAssembly Ready

• Browser console logging: Works seamlessly in WebAssembly targets
• Same API everywhere: Write once, run on native and web platforms

Performance Optimized

• Minimal dependencies: Only simple-datetime-rs for time formatting
• Efficient I/O: Fast writing with minimal overhead
• Memory efficient: String formatting with minimal allocations

Developer Experience

• Multiple output targets: stdout, stderr, files, or any Write implementer
• Custom time formats: Use standard datetime format strings
• Level-specific formatting: Different formats for different log levels
• Comprehensive examples: 10+ examples covering all use cases

Features

• Macro-based logging with formatting - info!(), debug!(), error!() etc. with format string support
• Date and time formatting - use standard datetime format strings for timestamps
• Multiple output targets - write to stdout, stderr, files, or any Write implementer
• Custom Write targets - add any Write + Send + Sync implementer as a logging target
• Custom log string formats for different levels
• Environment variable configuration - configure log levels via environment variables
• Lazy evaluation - avoid expensive computations when log level is insufficient
• WebAssembly (WASM) support - works in browsers with console logging
• Lightweight - minimal dependencies (only simple-datetime-rs)
• Simple API - easy to use and configure

Quick Start

The library provides macros with format string support just like println!:

use nonblocking_logger::{info, error, debug, warning, trace};

fn main() {
    let user_id = 42;
    let status = "active";
    
    // Simple messages
    info!("Application started");
    error!("Something went wrong!");
    
    // With formatting (like println!)
    info!("User {} has status: {}", user_id, status);
    warning!("CPU usage is {:.1}%", 85.5);
    debug!("Processing data: {:?}", vec![1, 2, 3]);
}

Macro-based Logging

The library provides macros that work like println! but for logging, with full format string support.

Available Macros

use nonblocking_logger::{log, info, debug, warning, error, trace};

// Basic logging macros with format string support
info!("Application started");
error!("Something went wrong!");
warning!("This is a warning");
debug!("Debug information");
trace!("Very detailed trace");

// With formatting (like println!)
let user_id = 42;
let status = "active";
info!("User {} has status: {}", user_id, status);
warning!("CPU usage is {:.1}%", 85.5);
debug!("Processing data: {:?}", vec![1, 2, 3]);

// Generic logging macro (always outputs, no level filtering)
use nonblocking_logger::log;
log!("Custom message with value: {}", 42);

Logger-Specific Macros

When you have a specific logger instance, use the *_with macros:

use nonblocking_logger::{log_with, info_with, error_with, debug_with, Logger};

fn main() {
    let logger = Logger::new().stdout();
    
    // Use specific logger with macros
    log_with!(logger, "Processing {} items", 5);
    info_with!(logger, "User {} connected", 42);
    error_with!(logger, "Failed to process: {}", "timeout");
    debug_with!(logger, "Debug info: {:?}", vec![1, 2, 3]);
}

Lazy Evaluation Macros

For expensive computations, use the lazy evaluation macros:

use nonblocking_logger::{debug_lazy, info_lazy, warning_lazy, error_lazy, trace_lazy};

fn expensive_computation() -> String {
    // Some expensive work
    "result".to_string()
}

// Lazy evaluation - closure only runs if log level allows
debug_lazy!(|| "Expensive computation result: {}", expensive_computation());
info_lazy!(|| "User {}: {}", 42, expensive_computation());
warning_lazy!(|| "CPU usage: {:.1}%", 85.5);
error_lazy!(|| "Error code: {}", 500);
trace_lazy!(|| "Step {} completed", "validation");

Logger-Specific Lazy Macros

For lazy evaluation with specific logger instances:

use nonblocking_logger::{log_lazy_with, info_lazy_with, debug_lazy_with, Logger};

fn main() {
    let logger = Logger::new().stdout();
    
    // Lazy evaluation with specific logger
    log_lazy_with!(logger, || "Expensive: {}", expensive_computation());
    info_lazy_with!(logger, || "User {}: {}", 42, expensive_computation());
    debug_lazy_with!(logger, || "Debug: {:?}", expensive_computation());
}

Format String Support

All macros support the same format string syntax as println!:

let number = 42;
let float = 3.14159;
let text = "hello";

// Basic formatting
info!("Integer: {}, Float: {:.2}, Text: {}", number, float, text);

// Different format specifiers
info!("Hex: 0x{:x}, Binary: 0b{:b}, Octal: 0o{:o}", number, number, number);
info!("Scientific: {:.2e}, Percentage: {:.1}%", float * 100.0, float * 100.0);

// Debug formatting for complex data
let data = vec![1, 2, 3, 4, 5];
let metadata = std::collections::HashMap::from([
    ("version", "1.0.0"),
    ("environment", "development"),
]);
debug!("Data: {:?}, Metadata: {:?}", data, metadata);

Try the macro example:

# Run the included example
cargo run --example macro_logging

When to Use Each Method

Use Macros when:

• You want format string support like println!
• You're building any application
• You want the most convenient API

Use Logger-Specific Macros (*_with) when:

• You have a specific logger instance with custom configuration
• You want format string support with a particular logger
• You need different loggers for different parts of your application
• You want to combine macro convenience with logger-specific settings

Use Function calls when:

• You prefer function calls over macros

Use Logger API when:

• You need custom time formats or level-specific formatting
• You want to write to multiple targets (stdout + file, etc.)
• You need different loggers for different parts of your application
• You want full control over configuration

Always-Output vs Level-Filtered Logging

The library provides two types of logging methods:

Always-Output Methods (no level filtering):

• log!(format, ...) - Always outputs regardless of logger level
• log_lazy!(|| format, ...) - Always outputs with lazy evaluation
• logger.log(message) - Always outputs regardless of logger level
• logger.log_lazy(message_fn) - Always outputs with lazy evaluation

Level-Filtered Methods (respect logger level):

• info!(format, ...), error!(format, ...), etc. - Filtered by level
• info_lazy!(|| format, ...), error_lazy!(|| format, ...), etc. - Filtered by level
• logger.info(message), logger.error(message), etc. - Filtered by level
• logger.info_lazy(message_fn), logger.error_lazy(message_fn), etc. - Filtered by level

When to use each:

• Use always-output methods when you want to ensure a message is always logged (e.g., critical system events, audit logs)
• Use level-filtered methods for normal application logging where you want to control verbosity via log levels

Example:

use nonblocking_logger::{log, info, log_lazy, info_lazy, warning, error, log_with, info_with, Logger, LogLevel};

fn main() {
    // Global macros (use default logger)
    info!("This info message will NOT appear");
    log!("This message will ALWAYS appear");
    
    // Logger-specific macros (use custom logger)
    let custom_logger = Logger::with_level(LogLevel::Debug).stdout();
    info_with!(custom_logger, "This will appear with debug level");
    log_with!(custom_logger, "This will ALWAYS appear with custom logger");
}

Time Format Strings

The time formatting uses standard datetime format strings for maximum flexibility:

• %Y-%m-%d %H:%M:%S - Date and time (default): "2025-09-14 19:50:08"
• %H:%M:%S - Time only: "19:50:08"
• %Y-%m-%d - Date only: "2025-09-14"
• %Y-%m-%d %H:%M:%S%.3f - With milliseconds: "2025-09-14 19:50:08.123"
• %Y-%m-%d %H:%M:%S.%f - With fractional seconds: "2025-09-14 19:50:08.123456"
• %Y-%m-%d %H:%M:%S %z - With timezone: "2025-09-14 19:50:08 Z"

Examples:

• "%Y-%m-%d %H:%M:%S" → "2025-09-14 19:50:08"
• "%H:%M:%S" → "19:50:08"
• "%Y-%m-%d" → "2025-09-14"
• "%Y-%m-%d %H:%M:%S%.3f" → "2025-09-14 19:50:08.123"
• "%Y-%m-%d %H:%M:%S.%f" → "2025-09-14 19:50:08.123456"

Environment Variable Configuration

The library supports configuring log levels through environment variables, following Rust conventions:

Basic Configuration

Set the global log level using either RUST_LOG or LOG_LEVEL:

# Using RUST_LOG (Rust convention)
export RUST_LOG=debug

# Using LOG_LEVEL (fallback)
export LOG_LEVEL=info

Supported Log Levels

• error or err - Error messages only
• warning or warn - Warning and error messages
• info - Informational, warning, and error messages (default)
• debug - Debug, info, warning, and error messages
• trace - All messages including trace

Case Insensitive

Log levels are case-insensitive:

export RUST_LOG=DEBUG    # Same as debug
export RUST_LOG=WARN     # Same as warn
export RUST_LOG=Error    # Same as error

API Reference

Global Convenience Functions

Simple string logging without formatting (also available as function calls):

Basic Logging Functions

• error(message: &str) -> io::Result<()> - Log error message (filtered by level)
• warning(message: &str) -> io::Result<()> - Log warning message (filtered by level)
• info(message: &str) -> io::Result<()> - Log info message (filtered by level)
• debug(message: &str) -> io::Result<()> - Log debug message (filtered by level)
• trace(message: &str) -> io::Result<()> - Log trace message (filtered by level)
• log(message: &str) -> io::Result<()> - Log message (always outputs, no level filtering)

Lazy Evaluation Functions

• error_lazy<F>(message_fn: F) -> io::Result<()> - Log error with lazy evaluation (filtered by level)
• warning_lazy<F>(message_fn: F) -> io::Result<()> - Log warning with lazy evaluation (filtered by level)
• info_lazy<F>(message_fn: F) -> io::Result<()> - Log info with lazy evaluation (filtered by level)
• debug_lazy<F>(message_fn: F) -> io::Result<()> - Log debug with lazy evaluation (filtered by level)
• trace_lazy<F>(message_fn: F) -> io::Result<()> - Log trace with lazy evaluation (filtered by level)
• log_lazy<F>(message_fn: F) -> io::Result<()> - Log with lazy evaluation (always outputs, no level filtering)

Initialization Functions

• get_global_logger() -> &'static Logger - Get the global logger instance

Logger

The logger struct with fluent API for custom formatting or multiple targets.

Methods

• new() - Create a new logger with default settings
• with_level(level) - Create a logger with specific log level
• from_env() - Create a logger with level from environment variables
• time_format(format) - Set time format using datetime format string
• no_time_prefix() - Disable time prefix
• format(format) - Set a single custom format for all levels
• format_for_level(level, format) - Set custom format for specific level
• add_target(target) - Add custom Write target (any Write + Send + Sync implementer)
• custom(target) - Set custom Write target (replaces all existing targets)
• stdout() - Add stdout as target
• stderr() - Add stderr as target
• file(path) - Add file as target
• log(message) - Log message (always outputs, no level filtering)
• log_lazy(message_fn) - Log with lazy evaluation (always outputs, no level filtering)
• error(message) - Log error message (filtered by level)
• warning(message) - Log warning message (filtered by level)
• info(message) - Log info message (filtered by level)
• debug(message) - Log debug message (filtered by level)
• trace(message) - Log trace message (filtered by level)

LogLevel

Log levels in order of priority (higher levels include lower ones):

• Error - Error messages
• Warning - Warning messages
• Info - Informational messages
• Debug - Debug messages
• Trace - Trace messages

LogMessage

A structured log message with level.

Methods

• new(level, message) - Create new log message
• error(message) - Create error message
• warning(message) - Create warning message
• info(message) - Create info message
• debug(message) - Create debug message
• trace(message) - Create trace message

Format String Placeholders

• {time} - Time prefix (formatted using the time_format setting)
• {level} - Log level (ERROR, WARN, INFO, DEBUG, TRACE)
• {message} - The actual log message

How Time Formatting Works

The time_format() method sets the datetime format string used to format the {time} placeholder. This allows you to:

1. Set a global time format that applies to all log levels
2. Use different time formats in custom level-specific formats
3. Mix time formats with custom formatting

// Global time format affects all levels
let logger = Logger::new()
    .time_format("%H:%M:%S")  // Sets time format for {time} placeholder
    .stdout();

// Custom level formats can use the {time} placeholder
let logger = Logger::new()
    .time_format("%Y-%m-%d %H:%M:%S")
    .format_for_level(LogLevel::Error, "🔥 {time} ERROR: {message}".to_string())
    .format_for_level(LogLevel::Info, "ℹ️  {time} INFO: {message}".to_string())
    .stdout();

// Or build the format string dynamically (for example, with a TLS prefix)
const LOG_PREFIX: &str = "worker-1";

let logger = Logger::new()
    .time_format("%Y-%m-%d %H:%M:%S")
    .format(format!("{{time}} [{{level}}][{}] {{message}}", LOG_PREFIX))
    .stdout();

WebAssembly (WASM) Support

This library supports WebAssembly targets (wasm32-unknown-unknown) for use in web browsers. When compiled for WASM:

• Logs are written directly to the browser's developer console
• File logging is not available (browsers don't allow direct file access)
• Same API as the native version for easy porting

WASM Usage

use nonblocking_logger::{error, warning, info, debug, trace, log, log_lazy};

fn main() {
    // Log messages (will appear in browser console)
    error!("This is an error");
    info!("This is info");
    warning!("This is a warning");
    debug!("Debug information");
    trace!("Trace information");
    
    // Always-output logging (no level filtering)
    log!("This message will always appear");
    log_lazy!(|| "This lazy message will always appear");
    
    // With formatting
    let user_id = 42;
    let status = "active";
    info!("User {} has status: {}", user_id, status);
    error!("Error code: {}", 500);
}

Using AsyncWrite with Logging

While the library doesn't have built-in AsyncWrite support, you can easily convert AsyncWrite to Write for logging. In most cases, short log message write operations are fast and don't benefit from async I/O, making synchronous logging more efficient. See the complete example:

cargo run --example async_write_logging

This example shows how to:

• Convert any AsyncWrite + Unpin to a sync Write trait
• Bridge async and sync I/O seamlessly using futures::block_on
• Create custom AsyncWrite implementations
• Use async streams with synchronous logging infrastructure

License

MIT