wizpanda / user-agent-utils

Java utility library for processing user-agent strings. Can be used to handle http requests in real-time or to analyse log files.

User Agent Utils

A Java utility library for processing user-agent strings. Can be used to handle HTTP requests in real-time or to analyze log files.

Difference from the original repo

Maintained by Wiz Panda, India.

Since the original repo is not being maintained, we took time to clean this up and we are now maintaining it with the following changes:

1. All codes are formatted to use Spaces instead of Tabs
2. Using Gradle to build, test & release instead of MVN
3. groupId has been changed from eu.bitwalker to com.wizpanda
4. artifactId has been changed from UserAgentUtils to user-agent-utils

Advice

A lot of user-agents lie about their identity. That makes UA sniffing a guessing game. On websites it is a better approach to detect features using a javascript library like Modernizr (http://modernizr.com/). Instead of using browser switches, build sites and services that are cross-platform/cross-device compatible.

Javadoc

http://bitwalker.eu/user-agent-utils-snapshot/javadoc/index.html

Installation

This library has been deployed at https://bintray.com/wizpanda/maven/user-agent-utils

Gradle

compile "com.wizpanda:user-agent-utils:0.0.3"

Maven

<dependency>
   <groupId>com.wizpanda</groupId>
   <artifactId>user-agent-utils</artifactId>
   <version>0.0.3</version>
</dependency>

FAQ

1. Can I add new browsers or operating systems?

Unfortunately not directly. This library uses enums to represent browsers and operating systems to make using it easy and readable. Unfortunately the use of enums also has some disadvantages. You can make a pull request on trunk for a new browser or operating system if all tests are still passing. The purpose of this library is not to detect every possible version, so you’ll have to provide the business value for a change.

2. How can I improve the performance of the user-agent parsing?

When handling requests on a web-server, consider if you need both the browser and operating system or if one of the two gives you enough information. As the parsing uses an ordered tree structure, performance depends a lot on the type of clients making the requests. The most common browsers and operating systems will be parsed faster than exotic ones. Consider caching the results temporarily.

3. Are there other libraries like this?

When I started this project in 2010 there were no other libraries offering the features I was looking for. Currently there are both open source as well as commercial solutions.

Change Log

See CHANGELOG.md. For old change logs, see this commit.