One of the problems of a team lead is that people on the team (sometimes even including myself) often create JUnit tests without any testing functionality.
It's easily done since the developers use their JUnit test as a harness to launch the part of the application they are coding, and then either deliberately or forgetfully just check it in without any assert tests or mock verifies.
Then later it gets forgotten that the tests are incomplete, yet they pass and produce great code coverage. Running up the application and feeding data through it will create high code coverage stats from Cobertura or Jacoco and yet nothing is tested except its ability to run without blowing up - and I've even seen that worked-around with big try-catch blocks in the test.
Is there a reporting tool out there which will test the tests, so that I don't need to review the test code so often?
I was temporarily excited to find Jester which tests the tests by changing the code under test (e.g. an if clause) and re-running it to see if it breaks the test.
However this isn't something you could set up to run on a CI server - it requires set-up on the command line, can't run without showing its GUI, only prints results onto the GUI and also takes ages to run.
PIT is the standard Java mutation tester. From their site:
Mutation testing is conceptually quite simple.
Faults (or mutations) are automatically seeded into your code, then your tests are run. If your tests fail then the mutation is killed, if your tests pass then the mutation lived.
...
Traditional test coverage (i.e line, statement, branch etc) measures only which code is executed by your tests. It does not check that your tests are actually able to detect faults in the executed code. It is therefore only able to identify code the is definitely not tested.
The most extreme example of the problem are tests with no assertions. Fortunately these are uncommon in most code bases. Much more common is code that is only partially tested by its suite. A suite that only partially tests code can still execute all its branches (examples).
As it is actually able to detect whether each statement is meaningfully tested, mutation testing is the gold standard against which all other types of coverage are measured.
The quality of your tests can be gauged from the percentage of mutations killed.
It has a corresponding Maven plugin to make it simple to integrate as part of a CI build. I believe the next version will also include proper integration with Maven site reports too.
Additionally, the creator/maintainer is pretty active here on StackOverflow, and is good about responding to tagged questions.
As far as possible, write each test before implementing the feature or fixing the bug the test is supposed to deal with. The sequence for a feature or bug fix becomes:
Write a test.
Run it. At this point it will fail if it is a good test. If it does
not fail, change, replace, or add to it.
When you have a failing test, implement the feature it is supposed
to test. Now it should pass.
You have various options:
You probably could use some code analysis tool like checkstyle to verify that each test has an assertion. Or alternatively use a JUnit Rule to verify this, but both is easily tricked and works only on a superficial level.
Mutation testing as Jester does is again a technical solution which would work, and it seems #Tom_G has a tool that might work. But these tools are (in my experience) extremely slow, because the work by changing the code, running tests, analyzing result over and over again. So even tiny code bases take lots of time and I wouldn't even think about using it in a real project.
Code Reviews: such bad tests are easily caught by code reviews, and they should be part of every development process anyway.
All this still only scratches on the surface. The big question you should ponder is: why do developers feel tempted to create code just to start a certain part of the application? Why don't they write tests for what they want to implement, so there is almost no need for starting parts of the application. Get some training for automated unit testing and especially TDD/BDD, i.e. a process where you write the tests first.
In my experience it is very likely that you will hear things like: We can't test this because .... You need to find the real reason why the developers, can't or don't want to write these tests, which might or might not be the reasons they state. Then fix those reasons and those abominations of tests will go away all on their own.
What you are looking for is indeed mutation testing.
Regarding tool support, you might also want to look at the Major
mutation framework (mutation-testing.org), which is quite efficient and configurable. Major
uses a compiler-integrated mutator and gives you great control over
what should be mutated and tested. As far as I know Major does not yet
produce graphical reports but rather data (csv) files that you can
process or visualize in any way you want.
Sounds like you need to consider a coverage tool like Jacoco, the gradle plugin provides report on coverage. I also use the EclEmma Eclipse plugin to obtain the same results, but with a fairly nice integration in the IDE.
In my experience, Jacoco has provided acceptable numbers even when there are no-op unit test. As it seems able to accurately determine the tested code paths. No-op test get low or 0% coverage scores and the score increase as the test become more complete.
Update
To address the down-voter. Perhaps a more appropriate tool to address this is PMD. Can be used in an IDE or build system. With proper configuration and rule development it could be used to find these incomplete unit tests. I have used it in the past to find methods missing certain security related annotation in the past.
Related
Maybe my question is a newbie one, but I can not really understand the circumstances under which I would use junit?
Whether I write simple applications or larger ones I test them with the System.out statements and it seams quite easy to me.
Why create test-classes with JUnit, unnecessary folders in the project if we still have to call the same methods, check what they return and we then have an overhead of annotating everything?
Why not write a class and test it at once with System.out but not create Test-classes?
PS. I have never worked on large projects I am just learning.
So what is the purpose?
That's not testing, that's "looking manually at output" (known in the biz as LMAO). More formally it's known as "looking manually for abnormal output" (LMFAO). (See note below)
Any time you change code, you must run the app and LMFAO for all code affected by those changes. Even in small projects, this is problematic and error-prone.
Now scale up to 50k, 250k, 1m LOC or more, and LMFAO any time you make a code change. Not only is it unpleasant, it's impossible: you've scaled up the combinations of inputs, outputs, flags, conditions, and it's difficult to exercise all possible branches.
Worse, LMFAO might mean visiting pages upon pages of web app, running reports, poring over millions of log lines across dozens of files and machines, reading generated and delivered emails, checking text messages, checking the path of a robot, filling a bottle of soda, aggregating data from a hundred web services, checking the audit trail of a financial transaction... you get the idea. "Output" doesn't mean a few lines of text, "output" means aggregate system behavior.
Lastly, unit and behavior tests define system behavior. Tests can be run by a continuous integration server and checked for correctness. Sure, so can System.outs, but the CI server isn't going to know if one of them is wrong–and if it does, they're unit tests, and you might as well use a framework.
No matter how good we think we are, humans aren't good unit test frameworks or CI servers.
Note: LMAO is testing, but in a very limited sense. It isn't repeatable in any meaningful way across an entire project or as part of a process. It's akin to developing incrementally in a REPL, but never formalizing those incremental tests.
We write tests to verify the correctness of a program's behaviour.
Verifying the correctness of a program's behaviour by inspecting the content of output statements using your eyes is a manual, or more specifically, a visual process.
You could argue that
visual inspection works, I check that the code does what it's meant to
do, for these scenarios and once I can see it's correct we're good to
go.
Now first up, it's great to that you are interested in whether or not the code works correctly. That's a good thing. You're ahead of the curve! Sadly, there are problems with this as an approach.
The first problem with visual inspection is that you're a bad welding accident away from never being able to check your code's correctness again.
The second problem is that the pair of eyes used is tightly coupled with the brain of the owner of the eyes. If the author of the code also owns the eyes used in the visual inspection process, the process of verifying correctness has a dependency on the knowledge about the program internalised in the visual inspector's brain.
It is difficult for a new pair of eyes to come in and verify the correctness of the code simply because they are not partnered up with brain of the original coder. The owner of the second pair of eyes will have to converse with original author of the code in order to fully understand the code in question. Conversation as a means of sharing knowledge is notoriously unreliable. A point which is moot if the Original Coder is unavailable to the new pair eyes. In that instance the new pair of eyes has to read the original code.
Reading other people's code that is not covered by unit tests is more difficult than reading code that has associated unit tests. At best reading other peoples code is tricky work, at its worst this is the most turgid task in software engineering. There's a reason that employers, when advertising job vacancies, stress that a project is a greenfield (or brand new) one. Writing code from scratch is easier than modifying existing code and thereby makes the advertised job appear more attractive to potential employees.
With unit testing we divide code up into its component parts. For each component we then set out our stall stating how the program should behave. Each unit test tells a story of how that part of the program should act in a specific scenario. Each unit test is like a clause in a contract that describes what should happen from the client code's point of view.
This then means that a new pair of eyes has two strands of live and accurate documentation on the code in question.
First they have the code itself, the implementation, how the code was done; second they have all of the knowledge that the original coder described in a set of formal statements that tell the story of how this code is supposed to behave.
Unit tests capture and formally describe the knowledge that the original author possessed when they implemented the class. They provide a description of how that class behaves when used by a client.
You are correct to question the usefulness of doing this because it is possible to write unit tests that are useless, do not cover all of the code in question, become stale or out of date and so on. How do we ensure that unit tests not only mimics but improves upon the process of a knowledgeable, conscientious author visually inspecting their code's output statements at runtime? Write the unit test first then write the code to make that test pass. When you are finished, let the computers run the tests, they're fast they are great at doing repetitive tasks they are ideally suited to the job.
Ensure test quality by reviewing them each time you touch off the code they test and run the tests for each build. If a test fails, fix it immediately.
We automate the process of running tests so that they are run each time we do a build of the project. We also automate the generation of code coverage reports that details what percentage of code that is covered and exercised by tests. We strive for high percentages. Some companies will prevent code changes from being checked in to source code control if they do not have sufficient unit tests written to describe any changes in behaviour to the code. Typically a second pair of eyes will review code changes in conjunction with the author of the changes. The reviewer will go through the changes ensure that the changes understandable and sufficiently covered by tests. So the review process is manual, but when the tests (unit and integration tests and possibly user acceptance tests) pass this manual review process the become part of the automatic build process. These are run each time a change is checked in. A continuous-integration server carries out this task as part of the build process.
Tests that are automatically run, maintain the integrity of the code's behaviour and help to prevent future changes to the code base from breaking the code.
Finally, providing tests allows you to aggressively re-factor code because you can make big code improvements safe in the knowledge that your changes do not break existing tests.
There is a caveat to Test Driven Development and that is that you have to write code with an eye to making it testable. This involves coding to interfaces and using techniques such as Dependency Injection to instantiate collaborating objects. Check out the work of Kent Beck who describes TDD very well. Look up coding to interfaces and study design-patterns
When you test using something like System.out, you're only testing a small subset of possible use-cases. This is not very thorough when you're dealing with systems that could accept a near infinite amount of different inputs.
Unit tests are designed to allow you to quickly run tests on your application using a very large and diverse set of different data inputs. Additionally, the best unit tests also account for boundary cases, such as the data inputs that lie right on the edge of what is considered valid.
For a human being to test all of these different inputs could take weeks whereas it could take minutes for a machine.
Think of it like this: You're also not "testing" something that will be static. Your application is most likely going through constant changes. Therefore, these unit tests are designed to run at different points in the compile or deployment cycle. Perhaps the biggest advantage is this:
If you break something in your code, you'll know about it right now, not after you deployed, not when a QA tester catches a bug, not when your clients have cancelled. You'll also have a better chance of fixing the glitch immediately, since it's clear that the thing that broke the part of the code in question most likely happened since your last compile. Thus, the amount of investigative work required to fix the problem is greatly reduced.
I added some other System.out can NOT do:
Make each test cases independent (It's important)
JUnit can do it: each time new test case instance will be created and #Before is called.
Separate testing code from source
JUnit can do it.
Integration with CI
JUnit can do it with Ant and Maven.
Arrange and combine test cases easily
JUnit can do #Ignore and test suite.
Easy to check result
JUnit offers many Assert methods (assertEquals, assertSame...)
Mock and stub make you focus on the test module.
JUnit can do: Using mock and stub make you setup correct fixture, and focus on the test module logic.
Unit tests ensure that code works as intended. They are also very helpful to ensure that the code still works as intended in case you have to change it later to build new functionalities to fix a bug. Having a high test coverage of your code allows you to continue developing features without having to perform lots of manual tests.
Your manual approach by System.out is good but not the best one.This is one time testing that you perform. In real world, requirements keep on changing and most of the time you make a lot of modificaiotns to existing functions and classes. So… not every time you test the already written piece of code.
there are also some more advanced features are in JUnit like like
Assert statements
JUnit provides methods to test for certain conditions, these methods typically start with asserts and allow you to specify the error message, the expected and the actual result
Some of these methods are
fail([message]) - Lets the test fail. Might be used to check that a certain part of the code is not reached. Or to have failing test before the test code is implemented.
assertTrue(true) / assertTrue(false) - Will always be true / false. Can be used to predefine a test result, if the test is not yet implemented.
assertTrue([message,] condition) - Checks that the boolean condition is true.
assertEquals([message,] expected, actual) - Tests whether two values are equal (according to the equals method if implemented, otherwise using == reference comparison). Note: For arrays, it is the reference that is checked, and not the contents, use assertArrayEquals([message,] expected, actual) for that.
assertEquals([message,] expected, actual, delta) - Tests whether two float or double values are in a certain distance from each other, controlled by the delta value.
assertNull([message,] object) - Checks that the object is null
and so on. See the full Javadoc for all examples here.
Suites
With Test suites, you can in a sense combine multiple test classes into a single unit so you can execute them all at once. A simple example, combining the test classes MyClassTest and MySecondClassTest into one Suite called AllTests:
import org.junit.runner.RunWith;
import org.junit.runners.Suite;
import org.junit.runners.Suite.SuiteClasses;
#RunWith(Suite.class)
#SuiteClasses({ MyClassTest.class, MySecondClassTest.class })
public class AllTests { }
The main advantage of JUnit is that it is automated rather than you manually having to check with your print outs. Each test you write stays with your system. This means that if you make a change that has an unexpected side effect your test will catch it and fail rather than you having to remember to manually test everything after each change.
JUnit is a unit testing framework for the Java Programming Language. It is important in the test driven development, and is one of a family of unit testing frameworks collectively known as xUnit.
JUnit promotes the idea of "first testing then coding", which emphasis on setting up the test data for a piece of code which can be tested first and then can be implemented . This approach is like "test a little, code a little, test a little, code a little..." which increases programmer productivity and stability of program code that reduces programmer stress and the time spent on debugging.
Features
JUnit is an open source framework which is used for writing & running tests.
Provides Annotation to identify the test methods.
Provides Assertions for testing expected results.
Provides Test runners for running tests.
JUnit tests allow you to write code faster which increasing quality
JUnit is elegantly simple. It is less complex & takes less time.
JUnit tests can be run automatically and they check their own results and provide immediate feedback. There's no need to manually comb through a report of test results.
JUnit tests can be organized into test suites containing test cases and even other test suites.
Junit shows test progress in a bar that is green if test is going fine and it turns red when a test fails.
I have slightly different perspective of why JUnit is needed.
You can actually write all test cases yourself but it's cumbersome. Here are the problems:
Instead of System.out we can add if(value1.equals(value2)) and return 0 or -1 or error message. In this case, we need a "main" test class which runs all these methods and checks results and maintains which test cases failed and which are passed.
If you want to add some more tests you need to add them to this "main" test class as well. Changes to existing code. If you want to auto detect test cases from test classes, then you need to use reflection.
All your tests and your main class to run tests are not detected by eclipse and you need to write custom debug/run configurations to run these tests. You still don't see those pretty green/red colored outputs though.
Here is what JUnit is doing:
It has assertXXX() methods which are useful for printing helpful error messages from the conditions and communicating results to "main" class.
"main" class is called runner which is provided by JUnit, so we don't have to write any. And it detects the test methods automatically by reflection. If you add new tests with #Test annotation then they are automatically detected.
JUnit has eclipse integration and maven/gradle integration as well, so it is easy to run tests and you will not have to write custom run configurations.
I'm not an expert in JUnit, so that's what I understood as of now, will add more in future.
You cannot write any test case without using testing framework or else you will have to write your testing framewok to give justice to your test cases.
Here are some info about JUnit Framework apart from that you can use TestNG framework .
What is Junit?
Junit is widely used testing framework along with Java Programming Language. You can use this automation framework for both unit testing and UI testing.It helps us define the flow of execution of our code with different Annotations. Junit is built on idea of "first testing and then coding" which helps us to increase productivity of test cases and stability of the code.
Important Features of Junit Testing -
It is open source testing framework allowing users to write and run test cases effectively.
Provides various types of annotations to identify test methods.
Provides different Types of Assertions to verify the results of test case execution.
It also gives test runners for running tests effectively.
It is very simple and hence saves time.
It provides ways to organize your test cases in form of test suits.
It gives test case results in simple and elegant way.
You can integrate jUnit with Eclipse, Android Studio, Maven & Ant, Gradle and Jenkins
JUNIT is the method that is usually accepted by java developer.
Where they can provide similar expected input to the function and decide accordingly that written code is perfectly written or if test case fails then different approach may also need to implement.
JUNIT will make development fast and will ensure the 0 defects in the function.
JUNIT : OBSERVE AND ADJUST
Here is my perspective of JUNIT.
JUNIT can be used to,
1)Observe a system behaviour when a new unit is added in that system.
2)Make adjustment in the system to welcome the "new" unit in the system.
What? Exactly.
Real life eg.
When your relative visits your college hostel room,
1) You will pretend to be more responsible.
2) You will keep all things where they should be, like shoes in shoe rack not on chair, clothes in cupboard not on chair.
3) You will get rid of all the contraband.
4) you will start cleanUp in every device you posses.
In programming terms
System: Your code
UNIT: new functionality.
As JUNIT framework is used for JAVA language so JUNIT = JAVA UNIT (May be).
Suppose a you already have a well bulletproof code, but a new requirement came and you have to add the new requirement in your code. This new requirement may break your code for some input(testcase).
Easy way to adapt this change is using unit testing (JUNIT).
For that you should write multiple testcases for your code when you are building your codebase. And whenever a new requirement comes you just run all the test cases to see if any test case fails.
If No then you are a BadA** artist and you are ready to deploy the new code.
If any of the testcases fail then you change your code and again run testcases until you get the green status.
We have a massive project with almost no unit tests at all. I would like to ensure from now on that the developers commit new features (or bugs!) without minimal coverage for corresponding unit tests.
What are some ways to enforce this?
We use many tools, so perhaps I can use a plugin (jira, greenhopper, fisheye, sonar, hudson). I was also thinking perhaps a Subversion pre-commit hook, the Commit Acceptance Plugin for jira, or something equivalent.
Thoughts?
Sonar (wonderful tool by the way) with Build breaker plugin can break your Hudson build when some metrics don't meet specified rules. You can setup such a rule in Sonar that would trigger an alert (eventually causing the build to fail) when the coverage is below given point. The only drawback is that you probably want the coverage to grow, so you must remember to increase the alert level every day to the current value.
What you want to do is determine what is new code, and verify that the new code is covered by some test.
Determining code coverage in general can be accomplished with any of a variety of test coverage tools. Many test coverage tools can simply reinstrument your entire application and then you can run tests to determine coverage.
Our (Semantic Designs') line of Test Coverage tools can determine, from a changed-file list, just the individual files that need to re-instrumented, and with careful test organization, just the tests that need to be reexecuted. This will minimize the cost of re-running your tests, and you'll still end
with the same overall coverage data. (Actually, these tools detect what tests need to be made based on changes at the method level).
Once you have test coverage data, what you want to know is the the specifically new code is covered by some tests. You can do this sloppily with just test coverage data if you know which files changed, by insisting the changed files have 100% coverage. That probably doesn't work in practice.
You could instead take advantage of SD's Smart Differencer tools to give a more precise answer. These tools compare two language files, and indicate where the changes are using the language syntax (e.g., expression, statement, declaration, method body, not just changed source lines) and conceptual editing operations (move, copy, delete, insert, rename-identifier-within-block). SmartDifferencer deltas tend to be both smaller and finer than what you would get from a plain diff tool.
It is easy to extract from the SmartDifferencer's output a list of lines changed. One could compute the intersection of that, per file, with the lines covered by the test coverage data. If the changed-lines are not all entirely within the set of covered lines, then "new" code hasn't been tested and you can raise a flag, stop a checkin, or whatever to signal that your checking policy has been violated.
The TestCoverage and SmartDifferencer tools don't come out-of-the-box with this computation done for you, but it should be a pretty easy script to implement.
if you use maven - cobertura plugin can be a good choice ( and not so annoying for developers as svn hook )
http://mojo.codehaus.org/cobertura-maven-plugin/usage.html
I am at the stage now where I have a fairly good understanding of programming/development, using Java.
Could anyone tell me the best way for me to start using testing packages? I have looked at Hibernate but not surer where to go with it...
I use Eclipse 3.5 on Mac OS X. Is it a case of writing scripts to test methods? What is unit-testing? etc.
Where do I begin?
Many thanks. Alex
What is Unit Testing
Unit testing is writing code (i.e. test code) that passes known inputs into code under test and then validating the code under test returns expected outputs. It's the most granular testing you can perform on an application. To make it easier, usually a unit testing framework is used. For Java, JUnit is the most popular, but TestNG is also notable.
Getting Started
Unit testing frameworks provide tools for test execution, validation and results reporting. For your setup, Eclipse has built in support for JUnit. Eclipse is able to automatically detect tests, compile tests and code under test, execute tests, and report results within the IDE. Furthermore, failures are reported as clickable stack trace information that loads the corresponding file at the given line number.
Mock Objects
That you're also working with Hibernate, suggests you also investigate a mock object framework as well - such as jMock. Mock objects are usually substituted as part of a code under tests's composition and serve two purposes: (1) returning known outputs and (2) recording they've been called and how so that unit tests can introspect that information as part of validation.
The ability to use Mock objects to make testing easier is predicated on dependency injection. That is other entities that compose the object under test. The idea is decoupling dependencies (e.g. Hibernate) to focus on testing algorithms that manipulate that data you're working with.
Database
However, if you've got code that is not easily refactored, or perhaps you want to validate database code, you can also test Hibernate interaction as well. In that case you want a database in a known state. Three approaches come to mind:
Restoring a database backup at the beginning of each test execution.
Use dbunit, which provides its own mechanisms for maintain state.
Transactional locking with rollback. Wrap the entire case is wrapped with a try{} finally{}, where the latter always rolls back the transaction.
James Shore ("a thought leader in the Agile software development community") has a series of screen casts of him demonstrating Test Driven Development, using Eclipse.
http://jamesshore.com/Blog/Lets-Play/
While there are many ways to start testing, there is no "best" way so there's no point in looking for that as a starting point.
Search the web for a good tutorial on junit and do it. That will be the absolute best way to get started IMO. Don't get sidetracked with code coverage or integrating with Hudson or any of the other tasks that are on the periphery to testing. Focus on writing a handful (or 10) if tests first.
Once you understand the basics you can start looking at other tools to see if they meet your needs any better or worse than junit.
First up: Hibernate is not a testing package.
Now that's out of the way, I'd suggest you take a look at JUnit. Read up on unit testing first so you know what it is (the Wikipedia entry is a good place to start), then try the JUnit cookbook. Write some unit tests for a small piece of your code to see how it works, then move on to bigger chunks.
While you are at it, take a look at other development tools like Cobertura (for finding out how good your test coverage is) and static analysis tools like Findbugs and Checkstyle. These all integrate nicely with Ant and probably Eclipse, too.
If you are interested in improving your coding standards and build systems then I highly recommend using Ant, JUnit, Cobertura, Checkstyle and Findbugs together with a continuous integration server (e.g. Hudson or CruiseControl) and a version control system (e.g. git). With a toolkit like that you can't go wrong.
There are other frameworks out there (TestNG, Mockito etc) so take a look at them, too, and decide which you prefer (EDIT: And which work nicely together. Mockito + JUnit is a good combination.)
I have written JUnit tests for my class, and would like it to tell me if there is any part of my code that is not unit tested. Is there a way to do this?
Yes, coverage tools like cobertura or emma.
They create reports that show every line in the source code and whether it was executed or not (and aggregated statistics as well).
Of course, they can only show you if the code was run. There is no way to tell if the unit test contained assertions to confirm that the result was correct.
You need some code coverage tools. See here (http://java-source.net/open-source/code-coverage) for some
If you look at the first one I think it does what you need
Cobertura is a free Java tool that calculates the percentage of code accessed by tests. It can be used to identify which parts of your Java program are lacking test coverage. It is based on jcoverage. Features of Cobertura:
Can be executed from ant or from the
command line.
If you use Eclipse, you can also try EclEmma, which shows you which lines of source were covered by your test. This is sometimes more useful than running a coverage tool like Cobertura because you can run a single test from inside Eclipse and then get immediate feedback on what was covered.
Your headline and your actual question differ. The tools mentioned in the other answers can tell you, which part of the code were not tested (=not executed at all). Making "make sure that all parts of code is unit tested" is a different thing. The coverage tools can tell you whether all lines/instructions have been executed, but they don't guarantee that everything is tested functionally (all constellations of data, all execution paths, etc.). This requires some brain power.
In my opinion, test coverage often gives a wrong feeling of safety. E.g. testing trivial getters increases coverage a lot but is rather useless.
If you are using IntelliJ then there is a button titled
"Run With Coverage"
I'm learning JUnit. Since my app includes graphical output, I want the ability to eyeball the output and manually pass or fail the test based on what I see. It should wait for me for a while, then fail if it times out.
Is there a way to do this within JUnit (or its extensions), or should I just throw up a dialog box and assertTrue on the output? It seems like it might be a common problem with an existing solution.
Edit: If I shouldn't be using JUnit for this, what should I be using? I want to manually verify the build every so often, and unit test automatically, and it'd be great if the two testing frameworks got along.
Manually accepting/rejecting a test defeats the purpose of using an automated test framework. JUnit is not made for this kind of stuff. Unless you find a way to create and inject a mockup of the object representing your output device, you should consider alternatives (don't really know any there sorry).
I once wrote automated tests for a video decoding component. I dumped the decoded data to a file using some other decoder as a reference, and then compared the output of my decoder to that using the PSNR of each pair of images. This is not 100% self contained (needs external files as resources), but was automated at least, and worked fine for me.
Although you could probably code that, that is not what JUnit is about. It is about automated tests, not guided manual tests. Generally that "does it look right" test is regarded as an integration test, as it is something that is very hard to automate correctly in a way that doesn't break for trivial changes all the time.
Take a look at Abbot to give you a more robust way to test your GUI.
Unit tests shouldn't require human intervention. If you need a user to take an action then I think you're doing it wrong.
If you need a human to verify things, then don't do this as part of your unit tests. Just make it a required step for your test department to carry out when QA'ing builds. (this still works of your QA department is just you.)
I recommend using your unit tests for the Models if using MVC, or any utility method (i.e. with Swing it's common to have color mapping methods). If you have a good set of unit tests on things like model behavior, if you have a UI bug it'll help narrow your search.
Visual based unit tests are very difficult, at a company I worked at they had tried these visual tests but slight differences in video cards could produce failed tests. In the end this is where a good Q/A team is required.
Take a look at FEST-Swing. It provides an easy way to automatically test your GUIs.
The other thing you'll want to do is separate your the code which does the bulk of the work from your gui code as much as possible. You can then write unit tests on this work code without having to deal with the user interface. You'll also find that you'll run these tests much more frequently, as they can run quickly.