Tag: testing

Sonar Gerrit Plugin Release

Initial release

I am happy to announce a first release of my Sonar Gerrit plugin. This plugin reports Sonar violations on your patchsets to your Gerrit server. Sonar analyses full project, but only files included in patchset are commented on Gerrit. Please forward to project page for installation instructions.

This plugin is intended to use with Gerrit Trigger plugin for Jenkins CI server. Together they provide a great tool for automatic static code analysis.

How does it work?

At the moment you push a patchset to Gerrit, Jenkins is notified with a ssh event. It fetches a code with a patchset and it builds your changes. It quits when build or tests fail.

But if it succeeds, Sonar analase your project in a post-build action. This is a place where my Sonar Gerrit plugin shines. It asks Gerrit for changed files before analysis and after Sonar analysis is finished, plugin reports comments on these files as a Gerrit reviewer. Currently plugin always reports +1 for Code Review, as it's still in development. However, you should always treat these comments as hints to improve, not as direct errors.

Extras

I've released also a second plugin: Sonar File Alerts plugin. This plugin raises alerts on file level in Sonar. It extends default behaviour, which raises alerts only at root project level. It is useful when you create alert rules in Sonar like "Code Coverage < 60". Each file is checked against this rule!

If you use Sonar File Alerts plugin and an alert will be generated on some file, then a comment will be published on this file on Gerrit.

Feedback

Please provide a feedback on these plugins. Feel free to submit issues on github or comment. It's still an early stage so your input is very welcome!

Sonar Gerrit Plugin Release

Initial release

I am happy to announce a first release of my Sonar Gerrit plugin. This plugin reports Sonar violations on your patchsets to your Gerrit server. Sonar analyses full project, but only files included in patchset are commented on Gerrit. Please forward to project page for installation instructions.

This plugin is intended to use with Gerrit Trigger plugin for Jenkins CI server. Together they provide a great tool for automatic static code analysis.

How does it work?

At the moment you push a patchset to Gerrit, Jenkins is notified with a ssh event. It fetches a code with a patchset and it builds your changes. It quits when build or tests fail.

But if it succeeds, Sonar analase your project in a post-build action. This is a place where my Sonar Gerrit plugin shines. It asks Gerrit for changed files before analysis and after Sonar analysis is finished, plugin reports comments on these files as a Gerrit reviewer. Currently plugin always reports +1 for Code Review, as it's still in development. However, you should always treat these comments as hints to improve, not as direct errors.

Extras

I've released also a second plugin: Sonar File Alerts plugin. This plugin raises alerts on file level in Sonar. It extends default behaviour, which raises alerts only at root project level. It is useful when you create alert rules in Sonar like "Code Coverage < 60". Each file is checked against this rule!

If you use Sonar File Alerts plugin and an alert will be generated on some file, then a comment will be published on this file on Gerrit.

Feedback

Please provide a feedback on these plugins. Feel free to submit issues on github or comment. It's still an early stage so your input is very welcome!

Error generating web.xml file with IntelliJ IDEA

If you use IntelliJ IDEA for your Grails development you might encounter this error running integration tests:

Error Error generating web.xml file (Use --stacktrace to see the full trace)

The reason for this is that IDEA adds classpath by default on creating integration test run configuration. Unfortunately, sometimes it causes strange errors like this one. Follow these steps to resolve:

  1. Open Run → Edit Configurations... (or press Alt-Shift-F10)
  2. Select your configuration that fails
  3. Uncheck Add --classpath checkbox
  4. You are done! Run.

How to use mocks in controller tests

Even since I started to write tests for my Grails application I couldn't find many articles on using mocks. Everyone is talking about tests and TDD but if you search for it there isn't many articles.

Today I want to share with you a test with mocks for a simple and complete scenario. I have a simple application that can fetch Twitter tweets and present it to user. I use REST service and I use GET to fetch tweets by id like this: http://api.twitter.com/1/statuses/show/236024636775735296.json. You can copy and paste it into your browser to see a result.

My application uses Grails 2.1 with spock-0.6 for tests. I have TwitterReaderService that fetches tweets by id, then I parse a response into my Tweet class.

TwitterController plays main part here. Users call show action along with id of a tweet. This action is my subject under test. I've implemented some basic functionality. It's easier to focus on it while writing tests.

Let's start writing a test from scratch. Most important thing here is that I use mock for my TwitterReaderService. I do not construct new TwitterReaderService(), because in this test I test only TwitterController. I am not interested in injected service. I know how this service is supposed to work and I am not interested in internals. So before every test I inject a twitterReaderServiceMock into controller:

Now it's time to think what scenarios I need to test. This line from TwitterReaderService is the most important:

You must think of this method like a black box right now. You know nothing of internals from controller's point of view. You're only interested what can be returned for you:

  • a TwitterError can be thrown
  • null can be returned
  • Tweet instance can be returned

This list is your test blueprint. Now answer a simple question for each element: "What do I want my controller to do in this situation?" and you have plan test:

  • show action should redirect to index if TwitterError is thrown and inform about error
  • show action should redirect to index and inform if tweet is not found
  • show action should show found tweet

That was easy and straightforward! And now is the best part: we use twitterReaderServiceMock to mock each of these three scenarios!

In Spock there is a good documentation about interaction with mocks. You declare what methods are called, how many times, what parameters are given and what should be returned. Remember a black box? Mock is your black box with detailed instruction, e.g.: I expect you that if receive exactly one call to readTweet with parameter '1' then you should throw me a TwitterError. Rephrase this sentence out loud and look at this:

This is a valid interaction definition on mock! It's that easy! Here is a complete test that fails for now:

You may notice 0 * _._ notation. It says: I don't want any other mocks or any other methods called. Fail this test if something is called! It's a good practice to ensure that there are no more interactions than you want.

Ok, now I need to implement controller logic to handle TwitterError.

My tests passes! We have two scenarios left. Rule stays the same: TwitterReaderService returns something and we test against it. So this line is the heart of each test, change only returned values after >>:

Here is a complete test for three scenarios and controller that passes it.

The most important thing here is that we've tested controller-service interaction without logic implementation in service! That's why mock technique is so useful. It decouples your dependencies and let you focus on exactly one subject under test. Happy testing!

Hibernate Envers with Grails 2.1.0

Our client requires that every entity in his Grails application has to be audited with detailed information: who and when did exactly what? It is a perfect opportunity to use Hibernate Envers! It turned out not as straightforward as I thought, so I want to share my experience with you.

Introduction

I use latest release of Grails with version 2.1.0. I've created a sample application on github for you as an example - SpOOnman/enversdemo. All files and techniques mentioned here can be found inside it. If you have other experiences with Grails and Envers working together, please share in comments.

Grails Envers plugin

First I've searched for a Grails plugin. I've found one outdated on google code and a (more modern) on github. This plugin is created by Lucas Ward and it has great introductionary blog post by author. This is a great place to start.

Unfortunately Lucas Ward's plugin supports Grails up 1.3.7 and it comes with outdated Envers version. I've searched through forks and I've found most up-to-date fork by Matija Folnovic: https://github.com/mfolnovic/grails-envers-plugin. It supports Grails 2.1 out of the box. Edit: Matija pointed out that credit should go to Jay Hogan and Damir Murat for their work on upgrading Envers plugin to Grails 2.

Matija's plugin is not packaged nor published so you have to package it by yourself:

Copy grails-envers-0.3-SNAPSHOT.zip (and rename to envers-0.3-SNAPSHOT.zip) to your application's lib diretory or publish to your company maven repository if you have one. Last step is to include a plugin in your BuildConfig.groovy:

Audit with Envers!

It's really easy to audit all entities with Envers. All you have to do is to use @Audit annotation on entities. SpOOnman/enversdemo audits Hotels and Bookings:

Envers plugin comes with some handy methods. For example Hotel.findAllRevisions() returns list of previous revisions. Once again I recommend a great blog post by Lucas Ward with many examples.

Grails and transactions

Grails has some glitches working with Envers. Envers is closely tied to database transactions. This means that it runs its listeners at the transaction end, when session is flushed. This is not always the case in Grails and it's the cause of some problems. You need to make some small modifications to your application to be sure that Envers works fine.

  • controllers - you need to annotate your controllers (or separate controller methods) with @Transactional and for every save you need to flush a session manually, e.g. save(flush: true). Otherwise Envers may not be notified.
  • services in Grails are transactional by default. However you mey explicitily set them with static transactional = true to be sure that this is not altered. I recommend to use session flushing on every save as well.
  • BootStrap.groovy script is not transactional. If you want Envers to audit your changes in this script you need to wrap your inserts with transactions. You can achieve it by using withTransaction closure. You need session flushing too. Take a look at SpOOnman/enversdemo BootStrap.groovy for an example.
  • integration tests wrap every test in a separate transaction that is rolled back at the end of a test, hence Envers won't work here. You need to disable wrapped transactions with static transactional = false. Remember that your changes are commited to database. To fulfill a contract of an empty database for every test you need to take care of cleaning up a database manually. Here is my sample integration test.

Add User to revision entities

By default Envers comes with a DefaultRevisionEntity class and creates its instance for every change on audited instance. It keeps information about entity revision, date and modification type. It lacks information about a user that made a change. I want to extend it.

There is already a code inside a plugin (here and here) but it's not packaged with a plugin itself. I guess it's because it adds dependency to Spring Security. This not a problem for me since I already use Spring Security plugin already in my application. I decided to add missing classes inside my application.

First I need a domain class - UserRevisionEntity. New instance of this class is created for every revision saved by Envers. It is a simple domain class with some extra annotations:

Notice that I've made currentUser field nullable. There may be some actions that doesn't require user to be logged in. Also, there is no user while executing BootStrap.groovy script.

Annotations @RevisionNumber and @RevisionTimestamp are required by Envers as described in documentation. @RevisionEntity annotation configures Envers to use non-default listener - SpringSecurityRevisionListener. It looks simple:

SpringSecurityRevisionListener's newRevision method fills each UserRevisionEntity instances with currently logged User.

Configuration

Envers offers some configuration options listed on documentation page. Options can be set in persistence.xml. However in Grails there is no persistence.xml. You can add it (and Grails supports it), but I've found other way to configure properties. I use System.setProperties to configure Envers and I place it in grails-app/conf/spring/resources.groovy. This is an example to change Envers tables' prefix and suffix:

Testing

Testing Envers makes sense only with Grails integration tests. As I've mentioned earlier you must disable transactions that wrap your tests in order for Envers listeners to work. Also make sure you use withTransaction, session flushing and you've marked your controllers as @Transactional. There rules all stand in integration tests too. Example integration test can look like this:

Summary

Hibernate Envers does a great job auditing my application. It is easy to use, despite all the small glitches I've mentioned here. Special thanks to plugin authors that package Envers library in a plugin that can be used out of the box. I really recommend it! It would be great to hear all your thoughts on Envers and Grails working together so I can improve this post for others too.

Update 12.04.2013

Lucas Ward has posted updaate entry on Grails Hibernate Envers plugin here: http://www.lucasward.net/2013/04/grails-envers-plugin-update.html. Thanks!

A Mockito catch

Suppose we have such classes and interfaces

public class AddOrganizationAction implements Action {}
public class AddPersonToOrganizationAction implements Action {}


public interface DispatchAsync {
     void execute( Action action, AsyncCallback callback );
}

We're using the best mocking framework ;) Suppose we want to verify that code under test will call execute() with proper Action - AddOrganizationAction.
I found that many developers (including me!) check such condition with

verify(async).execute(any(AddOrganizationAction.class), any(AsyncCallback.class));

In such case AsyncCallback is not important for us. We just want to ensure that AddOrganizationAction will be passed. We run test and it's green. But suddenly if we put the code below into test it will be green too!

verify(async).execute(any(AddPersonToOrganizationAction .class), any(AsyncCallback.class));

Why? Because any() matcher doesn't check the instance of passed object to be equal to declared class (AddOrganizationAction in this case). Any() checks if passed object conforms to method signature. In this case any Action's child will do. And we have an erroneous test!
The proper matcher we'd like to use is isA() matcher that checks if passed object is instance of declared class (which means instance of class or it's children).

So the proper test should contain

verify(async).execute(isA(AddOrganizationAction.class), any(AsyncCallback.class));

Go now and search for any() usages and think about changing it to isA(). In 1 of 10 cases whenever I change all tests in a testcase to isA() usage, I find an error in implementation of logic under test. Luckily I know the catch and now you do :)

Ok. It's not really a catch but ignorance of all us developers that we don't read entire documentantation :)

Testing ServiceMix 4.3 services with PaxExam

In the project that I'm currently working on we are developing a set of extensions for ServiceMix - more or less some monitoring stuff. It's pretty simple using SMX ExchangeListeners - you implement the interface, register it via Spring DM and voila - you have everything you need. But how are you going to test it? Of course, we write unit tests, and functional tests, but all of them are leaving NMR/OSGi stuff behind - you just don't include osgi-beans.xml's. This leaves us with untested code - my baaaaad ;) Fortunately, here comes a rescue - it's PaxExam. It's a framework for testing OSGi services in various containers - most notably felix & equinox. But making it play nicely with latest & greates ServiceMix 4.3 was not so easy - so I what to share it. I made a sample project, it's on my github It consists of 2 modules: first contains service & listener under test, and the second one contains The Test. I won't go into details about the service bundle - it's just cxf-bc endpoint passing invocations to camel router, which sets output message. The listener is also very simple. The point is that we have 2 of most widely used smx components, with OSGi packaging and OSGi-config configuration. This leaves us with itests module. It's core is BaseIntegrationTest, which contains Pax Exam configuration and some utility methods. Pax Exam configuration was the most difficult part for me. The reason is that servicemix consists of many, many dependant projects, each of them having own release cycle, and Pax Exam has it's own. Anyway, here is what works for me (note - it'll work with sun jdk 1.6 only I think). JUnit test (which is of course no unit test btw) using Pax Exam has to have static configuration method, which tells testing framework how to prepare container - which one, which parameters, which bundles to install. It looks like this: So, what should go into configuration method? First, we have to set up system packages. We have to exclude StAX API and some SOAP APIs from system packages: The system property "org.ops4j.pax.runner.platform.ee" tells PaxRunner where to look for system packages. I copied them from SMX distro - except that I had to add two Sun packages containing Schema and JAXP implementations - otherwise XPath & schema failed to work. Still, to make SchemaFactory work I had to manually add system property specifying implementation class. Next thing is rather standard PaxExam configuration: OSGi container, logging profile, spring and spring DM setup (note - you have to use 1.2 version, not the latest one): Two notes here. A profile is a set of bundles predefined to be used in PaxExam. Also note that we don't use System.setProperty, but PaxExam DSL expression. The reason is that the OSGi container is run separately, so we have to distinguish system properties used to start it, and properties that will be accessible to bundles inside it. Then comes felix install configuration. I wanted to mimic SMX configuration conventions (files in etc dir, ended with .cfg), so I copied felixInstall configuration and add appropriate system properties. Note, that we set start level to 2. In Karaf deafatul start level is 60, in PaxRunner - it's 5. But I decided to stay with Pax way - don't have that many bundles to need so fine grained configuration. Now, karaf configuration. It's also pretty standard, just remember that we need Aries bluepring in version 0.2-incubation, and not 0.1 (as PaxExam has it) or latest 0.3 Now, time for Servicemix components configuration. We'll use nice PaxExam feature - feature scan. This means that we can specify Karaf features package, and PaxExam will install appropriate bundles. So, we choose servicemix features, specify servicemix-camel, camel-nmr and servicemix-cxf-bc, and ...? And it turns out that servicemix-cxf-bc has unresolved dependencies :/ The reason is that it was (at least it seems so ;)) tested with full distro, which contains also activemq-blueprint feature. Since we don't want activemq for our test setup, we'll install required bundles manually: Now, everything is ready for the final step - installation of services bundle: mavenBundle("pl.touk.smx4", "paxExamSample-service") To make testing easier I also created few util methods. Two interesting ones are: getBean and sendMessage. getBean uses ServiceTracker to get Spring/Blueprint bean registered as a OSGi service. Note, that it is done asynchronously - that's why we need Thread.sleep for ServiceTracker. sendMessage wraps interaction with NMR - it creates inOut exchange, populates it and send synchronously via channel. Well... this setup is not particularly easy, but once we get through it, the tests look quite nice: This sends test message down the NMR, check that output is ok, and that it was caught by Listener. What's left to do? You can see some nasty explicit dependencies in configuration - should be taken from maven. Next goal would also be to run SOAPUI tests for testing end to end interaction. But let's leave it for next entry. To sum up: I think PaxExam is a great tool, although the setup is not so easy. Tests setup is quite long - actually it's faster to build the service bundle and run update on Servicemix to see changes. That's why I would rather recommend using PaxExam for regression tests, run on Hudson (whooops, Jenkins ;))