Skip to content

TraceLog is a highly configurable, flexible, portable, and simple to use debug logging system for Swift and Objective-C applications running on Linux, macOS, iOS, watchOS, and tvOS.

License

Notifications You must be signed in to change notification settings

tonystone/tracelog

Repository files navigation

TraceLog license: Apache 2.0

Platforms: iOS | macOS | watchOS | tvOS | Linux Swift 3.0 Version Build Status   Codecov Downloads

Introduction

TraceLog is a configurable debug logging system. It is unique in that it's configured after compilation in the runtime environment. It reads environment variables from the process context to set log levels. This allows each developer to configure log output per session based on the debugging needs of that session.

Usage

Using TraceLog is incredibly simple out of the box. Although TraceLog is highly configurable, to get started all you have to do is add the pod to your project, import TraceLog to the files that require logging and start adding log statements where you need them. TraceLog initializes itself and does everything else for you.

Swift

For Swift TraceLog comes with the following basic Logging functions (Note: hidden parameters and defaults were omitted for simplicity).

    logError  (tag: String?, message: @escaping () -> String)
    logWarning(tag: String?, message: @escaping () -> String)
    logInfo   (tag: String?, message: @escaping () -> String)
    logTrace  (tag: String?, level: UInt, message: @escaping () -> String)
    logTrace  (level: UInt, @escaping message: () -> String)

Using it is as simple as calling one of the methods depending on the current type of message you want to log, for instance, to log a simple informational message.

    logInfo { "A simple informational message" }

Since the message parameter is a closure that evaluates to a String, any expression that results in a string message can be used.

   logInfo {
        "A simple informational message: " +
        " Another expression or constant that evaluates to a string"
   }

We used closures for several reasons; one is that the closure will not be evaluated (and you won't incur the overhead) if logging is disabled or if the log level for this call is higher than the current log level set. And two, more complex expressions can be put into the closure to make decisions on the message to be printed based on the current context of the call. Again, these complex closures will not get executed in the cases mentioned above. For instance:

    logInfo {
         if let unwrappedOptionalString = optionalString {
            return "Executing with \(unwrappedOptionalString)..."
         } else {
            return "Executing..."
         }
    }

Log methods take an optional tag that you can use to group related messages and also be used to determine whether this statement gets logged based on the current environment configuration. It no tag is set, the file name of the calling method is used as the tag.

   logInfo("MyCustomTag") {
        "A simple informational message"
   }

There are several trace levels for logTrace that can be used. If you don't pass a level, you get level 3, otherwise, specify a level when making the logTrace call. For example, here is a trace level 1 and a level 3 call.

   logTrace {
        "A simple trace level 1 message"
   }

   logTrace(3) {
        "A simple trace level 3 message"
   }

You can of course also pass a tag like the rest of the log calls.

    logTrace("MyCustomTag", level: 3) {
         "A simple trace level message"
    }

That is all there is to adding logging to your Swift application!

Objective-C

As with Swift using TraceLog with Objective-C is extremely simple out of the box. The Objective-C implementation consists of the following primary logging methods. Each has a format specifier (like NSLog) and an optional variable number of arguments that represent your placeholder replacement values.

    LogError  (format,...)
    LogWarning(format,...)
    LogInfo   (format,...)
    LogTrace  (level,format,...)

Using it is as simple as calling one of the methods depending on the current type of message you want to log, for instance, to log a simple informational message.

    LogInfo(@"A simple informational message");

You can also call it as you would NSlog by using the format specifier and parameters for placeholder replacement.

    LogInfo(@"A simple informational message: %@", @"Another NSString or expression that evaluates to an NSString");

More complex expressions can be put into the placeholder values by using Objective-C blocks that return a printable NSObject. These can be used to make decisions on the message to be printed based on the current context of the call. These complex blocks will not get executed (and you won't incur the overhead) if logging is disabled or if the log level for this call is higher than the current log level set. For instance:

    LogInfo(@"Executing%@...",
        ^() {
            if (optionalString != nil) {
                return [NSString stringWithFormat: @" with %@", optionalString];
            } else {
                return @"";
            }
        }()
    );

There is a special version of Log methods take an optional tag that you can use to group related messages and also be used to determine whether this statement gets logged based on the current environment configuration. These methods begin with C (e.g. CLogInfo).

    CLogInfo(@"MyCustomTag", @"A simple informational message");

There are several trace levels for LogTrace that can be used. For example, here is a trace level 3 call.

    LogTrace(3, @"A simple trace level message");

You can of course also pass a tag by using the CLog version of the call.

    CLogTrace(3, @"MyCustomTag", @"A simple trace level message");

Configuration

TraceLog is configured via variables that are either set in the runtime environment or passed on startup of the application.

These variables consist of the following:

    LOG_ALL=<LEVEL>
    LOG_TAG_<TAGNAME>=<LEVEL>
    LOG_PREFIX_<TAGPREFIX>=<LEVEL>

Log output can be configured globally using the LOG_ALL variable, by TAG name using the LOG_TAG_<TAGNAME> variable pattern, and/or by a TAG prefix by using the LOG_PREFIX_<TAGPREFIX> variable pattern.

Each environment variable set is set with a level as the value. The following levels are available in order of display priority. Each level encompasses the level below it with TRACE4 including the output from every level. The lowest level setting is OFF which turns logging off for the level set.

Levels:

    TRACE4
    TRACE3
    TRACE2
    TRACE1
    INFO
    WARNING
    ERROR
    OFF

Note: Log level OFF does not disable TraceLog completely, it only suppresses log messages for the time it is set in the environment during your debug session. To completely disable TraceLog please see Runtime Overhead for more information.

Multiple Environment variables can be set at one time to get the desired level of visibility into the workings of the app. Here are some examples.

Suppose you wanted the first level of TRACE logging for your Networking module which has a class prefix of NT and you wanted to see only errors and warnings for the rest of the application. You would set the following:

    LOG_ALL=WARNING
    LOG_PREFIX_NT=TRACE1

More specific settings override less specific so in the above example, the less specific setting is LOG_ALL which is set to WARNING. The tag prefix is specifying a particular collection of tags that start with the string CS, so this is more specific and overrides the LOG_ALL. If you choose to name a particular tag, that will override the prefix settings.

For instance, in the example above, if we decided for one tag in the Networking module, we needed more output, we could set the following:

    LOG_ALL=WARNING
    LOG_PREFIX_NT=TRACE1
    LOG_TAG_NTSerializer=TRACE4

This outputs the same as the previous example except for the NTSerializer tag which is set to TRACE4 instead of using the less specific TRACE1 setting in LOG_PREFIX.

Configuration (Environment)

TraceLog can be configured via the environment either manually using export or via Xcode. For TraceLog to read the environment on startup, you must call its configure method at the beginning of your application.

   TraceLog.configure()

To set up the environment using Xcode, select "Edit Scheme" from the "Set the active scheme" menu at the top left. That brings up the menu below.

Configuration (In code)

TraceLog.configure() accepts an optional parameter called environment which you can pass the environment. This allows you to set the configuration options at startup time (note: this ignores any variables passed in the environment).

Here is an example of setting the configuration via TraceLog.configure().

TraceLog.configure(environment: ["LOG_ALL": "TRACE4",
                                 "LOG_PREFIX_NS" : "ERROR",
                                 "LOG_TAG_TraceLog" : "TRACE4"])

Note: Although the environment is typically set once at the beginning of the application, TraceLog.configure can be called anywhere in your code as often as required to reconfigured the logging levels.

Configuring Log Writers

TraceLog can be configured with multiple custom log writers who do the job of outputting the log statements to the desired location. By default, it configures itself with a ConsoleWriter which outputs to stdout. You can change the writers at any time and chain multiple of them to output to different locations such as to a remote logging servers, syslog, etc.

Writers must implement the Writer protocol. To install them, simply call configure with an array of Writers.

let networkWriter = NetworkWriter(url: URL("http:https://mydomain.com/log"))

TraceLog.configure(writers: [ConsoleWriter(), networkWriter])

Since writers are instantiated before the call, you are free to initialize them with whatever makes sense for the writer type be created. For instance in the case of the network writer the URL of the logging endpoint.

Runtime Overhead

The Swift implementation was designed to take advantage of swift compiler optimizations and will incur no overhead when compiled with optimization on (-O) and TRACELOG_DISABLED is defined.

The Objective-C implementation was designed to take advantage of the preprocessor and when compiled with TRACELOG_DISABLED defined, will incur no overhead in the application.

For XCode TRACELOG_DISABLED can be set in the project target. For Swift Package Manager pass a swiftc directive to swift build as in the following example.

swift build -Xswiftc -DTRACELOG_DISABLED

Minimum Requirements

Build Environment

Platform Swift Swift Build Xcode
Linux 3.0
OSX 3.0 Xcode 8.0

Minimum Runtime Version

iOS OS X tvOS watchOS Linux
8.0 10.10 9.0 2.0 Ubuntu 14.04, 16.04, 16.10

Note:

To build and run on Linux we have a a preconfigure Vagrant file located at https://github.com/tonystone/vagrant-swift

See the README for instructions.

Installation (Swift Package Manager)

TraceLog now supports dependency management via Swift Package Manager on All Apple OS variants as well as Linux.

Please see Swift Package Manager for further information.

Installation (CocoaPods)

TraceLog is available through CocoaPods. Currently Swift is the default so to install it, simply add the following line to your Podfile:

Swift

pod "TraceLog"

Please note that this is a change from the 1.x versions of TraceLog in where Objective-C was the default. Now to use Objective-C, you must specify the subspec as follows:

Objective-C

pod "TraceLog/ObjC"

Mixed Environments

TraceLog was designed to work in mixed environments so you can have Swift pod/modules using TraceLog as well as Objective-C pods/libraries in the same application. The configuration settings you set will set the values for both. If you have an application that contains mixed Swift and Objective-C code you can include both into your application.

For example:

pod "TraceLog"
pod "TraceLog/ObjC"

See the "Using CocoaPods" guide for more information.

Author

Tony Stone (https://github.com/tonystone)

License

TraceLog is released under the Apache License, Version 2.0