TDD on a client-server application - unit-testing

I’m having problem to grasp how to do TDD when building a client-server system.
The simple katas (Fizzbuzz etc) are easy to understand, but when my client needs to send the server a file using TCP sockets and get a response back from the server I’m getting confused regarding how to test that.
I had a project building a file-sync system. The client will monitor a folder and every time a change happens (new file, file deletion etc.) the server should update automatically.
The client can have many devices, for example I can have a copy of the folder in two different computers and they all should sync perfectly.
I started the project with tests, but once I reached the part of speaking with the server I got stuck and didn’t understand how to implement tests.
Most of the things I find regarding TDD are the simple stuff. I would love your advice on this slightly more complex application.

I’m having problem to grasp how to do TDD when building a client-server system.
The reference you want is Growing Object Oriented Software, Guided by Tests
I started the project with tests, but once I reached the part of speaking with the server I got stuck and didn’t understand how to implement tests.
Basic idea: you are trying to work towards a design where you can separate the complicated code from the code that is hard/expensive to test.
This often means three "modules"
A really simple module that knows how to talk to the network
A complicated module that knows how to prepare messages for the network, and how to interpret the responses (and timeouts)
A module that can coordinate the interaction of the two modules above.
The first module, you "test" using code review, acceptance testing, and taking advantage of the fact that it doesn't change very often (because it is so simple).
The second module, you use lots of programmer tests to make sure that the logic correctly handles all of the different messages that can pass through it
The third module, you concentrate on testing the protocol. Here, we'll often use a substitute implementation (aka a mock or some other flavor of test double) for one or both of the first two modules.
In a language with types like Java or C#, the need for substitutes will often mean that the first two modules will need to implement some interface, and the third module will depend on those interfaces rather than having direct dependencies on the implementations.
You'll likely also need some code in your composition root that wires together the actual implementations.
For a good take on separating the networking client from the networking logic/protocol, see Cory Benfield 2016.
It might also be useful to review:
Boundaries, by Gary Bernhardt
At the Boundaries, Applications Aren't Object Oriented by Mark Seemann

Put the client code that is working with the socket into a separate class that can be injected in the "business code". For your tests inject a mock instead, verifying that the API of the "client socket adapter" is called in the apropriate way. Mocking libraries make this easy.
Put the server code that is working with the socket into a separate class and design an internal API for the "business code" that the "server socket adapter" is calling. Ignore the adapter in your tests and call the API of the business code directly.
You might want to read about the Ports & Adapter architecture (sometimes also called the "Hexagonal Model").

Related

Distributing unit testing across virtual machines

I've spent the last few days looking around for an existing solution to a functional testing problem, but I am out of ideas and would appreciate some SO help!
I've got a preexisting suite of functional networking tests currently written in C++ using Boost.Test and Google Test, but might become rewritten into Rust soon. These generally take the following form:
unit test fixture {
1. Start a thread representing "the server" which goes and listens on some localhost port for incoming network connections.
2. Do client stuff representing "the client" to that localhost port.
3. Join the server thread, fetching an errors or problems.
4. Exit with success or failure.
}
This is great, and it works well. However it only tests loopback and in the real world the server component is probably in its own process running behind a NAT routed network, so it's not particularly realistic and therefore not really testing the code. What I think I'm looking for is some method of splitting the server thread part off into its own process, and then some method of getting the server test process and the client test process to work together to run the functional tests. Ideally speaking the server and client processes would run in separate "machines", this is something I can automate using OpenVZ scripting fairly easily so consider that problem out of scope, though it makes forking the process non-ideal.
I had been thinking that in this age of Web 2.0 et al surely this is a very common functional test scenario, and therefore that established patterns and test frameworks would abound. I mean, I as an old timer thinks "DCOM" as my first thought on how to solve this, though that's a 1990s Microsoft only solution. And maybe there is some modern and portable equivalent, and I am not searching for the right terms here, so here is my first question:
Is there any standard functional testing library or framework which extends Google Test or Boost.Test etc which lets you easily choose at runtime whether the server and client parts of each functional test should run as threads or as processes or best of all, as processes inside their own virtual machine with its own network stack?
This test scenario is surely common as muck. But let's assume it isn't, and no such tool exists. If it doesn't, then we need to extend Boost.Test or Google Test with some extra support. Firstly, we need to associate with each test fixture a supporting "server" part test fixture, and for the threaded test scenario we need to always run the server and client test fixtures concurrently. So, my second question:
Is there any way of strongly associating two test fixtures in any of the popular C++ or Rust unit testing frameworks where the two fixtures are seen as two halves of the same test, and always executed concurrently?
This leaves the second part: how to get a unit test framework to execute only the client parts in one process and only the server parts in the other process, and to do both always concurrently and in sync with one another, and moreover to merge the junit XML output from both parts into a single test result. So:
Is there any alternative functional testing approach, methodology, or open source solution which is better suited for distributed network functional testing than unit test frameworks such as Google Test or Boost.Test? Preferably something libvirt aware so it can orchestrate virtual machines as part of the testing setup and teardown? For example, is there some Jenkins plugin or something which could use Jenkins slaves in each OpenVZ container to orchestrate the concurrent execution of the multiple parts of each of the functional tests? Or is old fashioned CORBA still the least worst solution here? Is there maybe some way of automatically wrapping up test fixtures into a REST HTTP API?
I did do a quick review of the major integration testing frameworks, so Citrus, STAF and Twister. I'll be honest in saying they all seem way overkill for what I want which is a quick and easy way of making the existing functional test suite use a more realistic network routing than loopback. That's all I really want essentially, and I don't care how it's done so long as the check and requires still appear in Jenkins. Over to you Stackoverflow!
My thanks in advance for any help.
I have had similar requirements but I am come from the Java side of the world. What you can easily do is having a distributed management of nodes / machines using jGroups.
Once you understand how it works you can build a distributed system of nodes by just using 100 lines of code. With this system you can span and control child processes on each of those systems and check output and everything yourself. Should only cost you a day to take a jGroup example and get this running.
Once you have the infrastructure copy code and execute it as independent process your control is easy. Then use some of those nodes and get Selenium up and use a number of browser windows and execute scripts (or use Sikuli) and do your testing. Since the Selenium process is again Java you can generate all kind of reports you print to console or send directly to the cluster since those processes can join the cluster using jGroups too.
Such a system can be implemented in a week and it is really under your control. Very simple to do and very extendable.
Also you can provide plugins for Jenkins, Jira or Quality Center to interact with it and trigger test execution and management.

What part of project to test first?

I want to write first unit test in my life.
At present, I am developing new ASP.NET MVC 5 project. This is simple workflow system. My project contains 4 layers:
Presentaton layer (an MVC-project)
Infrastructure layer (which contains Repositories and ORM)
Domain (which contains POCO classes and interfaces of busines logic)
Service layer (which is implemented domain interfaces)
I believe, that I need test Service layer firstly. Is that right? Which layer should I test first?
There's no single correct approach, but the most common techniques are
Top-Down, also called Outside-In. Here, you start at the outside layer and work your way in.
Bottom-Up. Here, you start with the constituent building blocks and assembly them to a working system.
As Code Complete describes, using dual approaches interchangeably can actually be beneficial, because the stuff you learn from doing one thing, helps you better understand what you need to do in the other end, and vice versa. I often do a bit of Outside-In, then some Bottom-Up, then some more Outside-In, etc.
As per Mark Seemann's answer, you can test from UI layer first and finish with the data layer, or in reverse.
Who is responsible for the project? Which part of the project is business critical? Rather than test across each layer "horizontally", test through all of the layers for a particular piece of functionality "vertically".
This gives you the benefit of coverage based on business priorities and you can apply any testability changes you need to make or techniques across all of the layers as you start to test each piece of functionality.
Since you have written your code already, be prepared to refactor some code to make it more testable (for example setting up Dependency Injection to isolate code for unit testing) and make note of these changes to help design for testability in future.

Java EE -- using the same stateful object for several users

Even though I've been in Java SE for quite some time now, I started EE & web w/ Java only about a month ago, so pardon if the question seems a bit noobish...
So here's the situation: I'm trying to write a JS based multi-player game with real-time interaction (let's say chess in this example, though it really doesn't matter what particular game it is, could be tennis or w/ever). The clients would interact with the server through JS calls, sending the move etc. Now, while I could just receive the move from one client & pass it straight on to the other player, not maintaining the game state on the server would mean putting a huge sign out saying "user JS scripts welcome" (and that's out of experience -- "hacked" a crapload of that kind myself). This brings me to my problem -- how do I share a stateful object between several sessions? One idea that came to mind was a singleton storing a Hashmap of stateful beans & then each session could retrieve the bean by it's hash, but I've no idea how right that is (and it seems rather complex for a fairly common thing like that). Tieing it to application scope seems overkill as well...
P.S. I do understand that the object would need concurrency managing etc, I just can't seem to put my finger on how to get it shared...
EDIT: I'm sorry I didn't mention it before -- using Glassfish, EE6.
You have a business process scenario which is defined according to Seam framework documentation as follows
The business process spans multiple interactions with multiple users, so this state is shared between multiple users, but in a well-defined manner. The current task determines the current business process instance, and the lifecycle of the business process is defined externally using a process definition language, so there are no special annotations for business process demarcation.
Here you can see a Seam business process management Tutorial
Notice Seam uses JBoss BPM behind the scenes to handle its business process context. If you just want to use plain JBoss BPM capabilities, you can see here how to integrate with JBoss
See also JBoss BPM User guide
Solved. Shared it via ServletContext, which I initially thought wouldn't work 'cause FacesServlet is a separate one, thought it has smthn like a different container either.

Unit/integration testing Asterisk configuration

Unit and integration testing is usually performed as part of a development process, of course. I'm looking for ways to use this methodology in configuration of an existing system, in this case the Asterisk soft PBX.
In the case of Asterisk, the configuration file is as much a programming language as anything else, complete with loops, jumps, conditionals, etc., and can get quite complex. Changes to the configuration often suffers from the same problems as changes to a complex software product - it can be hard to foresee all the effects without tests in place. It's made worse by the fact that the nature of the system is to communicate with external entities, i.e. make phone calls.
I have a few ideas about testing the system using call files (to create specific calls between extensions) while watching the manager interface for generated events. A test could then watch for an expected result, i.e. dialling *99# should result in the Voicemail application getting called.
The flaws are obvious - it doesn't test the actual result, only what the system thinks is the result, and it probably requires some modification of the system under test. It's also really hard to write these tests robustly enough to only trigger on the expected output, especially if the system is in use (i.e. there are other calls in progress).
Is what I want, a testing system for Asterisk, impossible? If not, do you have any ideas about ways to go about this in a reasonable manner? I'm willing to put a fair amount of development time into this and release the result under a friendly license, but I'm unsure about the best way to approach it.
This is obviously an old question, so there's a good chance that when the original answers were posted here that Asterisk did not support unit / integration testing to the extent that it does today (although the Unit Test Framework API went in on 12/22/09, so that, at least, did exist).
The unit testing framework (David's e-mail from the dev list here) lets you execute unit tests directly within Asterisk. Tests are registered with the framework and can be executed / viewed through the CLI. Since this is all part of Asterisk, the tests are compiled into the executable. You do have to configure Asterisk with the --enable-dev-mode option, and mark the tests for compilation using the menuselect tool (some applications, like app_voicemail, automatically register tests - but they're the minority).
Writing unit tests is fairly straight-forward - and while it (obviously) isn't as fully featured as a commercial unit test framework, it gets the job done and can be enhanced as needed.
That most likely isn't what the majority of Asterisk users are going to want to use - although Asterisk developers are highly encouraged to check it out. Both users and developers are probably interested in integration tests, which the Asterisk Test Suite provides. At its core, the Test Suite is a python script that executes other scripts - be they lua, python, etc. The Test Suite comes with a set of python and lua libraries that help to orchestrate and execute multiple Asterisk instances. Test writers can use third party applications such as SIPp or Asterisk interfaces (AMI, AGI) or a combination thereof to test the hosted Asterisk instance(s).
There are close to 200 tests now in the Test Suite, with more being added on a fairly regular basis. You could obviously write your own tests that exercise your Asterisk configuration and have them managed by the Test Suite - if they're generic enough, you could submit them for inclusion in the Test Suite as well.
Note that the Test Suite can be a bit tricky to set up - Leif wrote a good blog post on setting up the Test Suite here.
Well, it depends on what you are testing. There are a lot of ways to handle this sort of thing. My preference is to use Asterisk Call Files bundled with dialplan code. EG: Create a callfile to dial some public number, once it is answered, hop back to the specified dialplan context and perform all of my testing logic (play soundfiles, listen for keypresses, etc.)
I wrote an Asterisk call file library which makes this sort of testing EXTREMELY easy. It has a lot of documentation / examples too, check it out here: http://pycall.org/. That may help you.
Good luck!
You could create a set of specific scenarios and use Asterisk's MixMonitor command to record these calls. This would enable you to establish a set of sound recordings that were normative for your system for these tests, and use an automated sound file comparison tool (Perhaps something from comparing-sound-files-if-not-completely-identical?) to examine the results. Just an idea.
Unit testing as opposed to integration testing means your code is supposed to be architectured so the logic itself is insulated from external dependencies. You said "the configuration file is as much a programming language as anything else" but that's the thing --- real languages has not just control flow but abstraction capabilities, which allow you to write the logic in a way that can be unit tested. That's why I keep logic outside of asterisk as much as possible.
For integration testing, script linphonec to drive your application, and grep the asterisk console to see what it's doing.
You can use docker, and fire up temporary asterisk instances for each test.

architecture/design advise for a test program

I am trying to build a test program in c++ to automate testing for a specific application. The testing will involve sending requests which have a field 'CommandType' and some other fields to a server
The commandType can be 'NEW', 'CHANGE' or 'DELETE'
The tests can be
Send a bunch of random requests with no pattern
Send 100 'NEW' requests, then a huge amount of 'CHANGE' requests followed by 200 'DELETE' requests
Send 'DELETE' requests followed by 'CHANGE' requests
... and so on
How can I design my software (what kind of modules or layers) so that adding any new type of test case is easy and modular?
EDIT: To be more specific, this test will be to only test one specific application that gets requests of the type described above and handles them. This will be a client application that will send the requests to the server.
I would not create your own framework. There are many already written that follow a common pattern and can likely accomodate your needs elegantly.
The xUnit framework in all incarnations I have seen allows you to add new test cases without having to edit the code that runs the tests. For example, CppUnit provides a macro that when added to a test case will auto-register the test case with a global registry (through static initialization I assume). This allows you to add new test cases without cracking open and editing the thing that runs them.
And don't let the "unit" in xUnit and CppUnit make you think it is inappropriate. I've used the xUnit framework for all different kinds of testing.
I would separate out each individual test into it's own procedure or, if it requires code beyond a function or two, it's own source file. Then in my main routine I'd do something like:
void main()
{
run_test_1();
run_test_2();
//...
run_test_N();
}
Alternatively, I'd recommend leveraging the Boost Test Library and following their conventions.
I'm assuming you're not talking about creating unit tests.
IMHO, Your question is too vague to provide useful answers. Is this to test a specific application or are you trying to make something generic enough to test as many different applications as is possible? Where do these applications live? Are they client server apps, web apps, etc.?
If it's more than one application that you want your tool to test, you'll need an architecture that creates a protocol in between the testing tool and the applications such that you can convert the instructions your tool and consumers of your tool can understand, into instructions that the application being tested can understand. I've done similar things in the past but I've only ever had to worry about maybe 5 different "applications" so it was a pretty simple matter of summing up all the unique functionality of the apps and then creating an interfact that supports them all.
I wouldn't presume that NEW, CHANGE, and DELETE would be your only command types either. A lot of testing involves data cleanup, test reporting, etc. And applications all handle this their own special ways.
use C++ unit testing framework , Read this for Detail and examples