Click here to Skip to main content
15,883,606 members
Articles / Programming Languages / Visual Basic

Are you Covered? Measuring the Quality of .NET Code with NCover

17 Jan 2008CPOL5 min read 31.6K   13   1
This article describes using NCover to measure the effectiveness of your automated tests, shows how to track coverage over time, and presents strategies for increasing the effectiveness of your tests.

This article is in the Product Showcase section for our sponsors at CodeProject. These articles are intended to provide you with information on products and services that we consider useful and of value to developers.

Image 1

This article will tell you more about NCover, the premier code coverage tool for the .NET framework. Since its creation 4 years ago, NCover has been used worldwide by over 100,000 developers.

Code quality is one of those timeless topics that developers and software architects have struggled with for decades. The accepted wisdom is that the earlier a software defect is found, the cheaper it is to fix.

Most software development methodologies focus on finding defects as early as possible in the development cycle. Agile development is a great approach for increasing software quality. One of the core tenets of agile development is building a library of automated tests that the code must pass every time it changes. This approach has been used by development teams for years. It has proven to be effective at improving quality while working for both small and large teams.

Let’s assume you’re using agile methodologies and that you’ve written an automated set of tests. Once you start to analyze the results, an obvious set of questions comes to mind:

  • What portions of my code have not been tested?
  • How can I decrease the amount of untested code?
  • Is my code becoming more or less tested over time?

For a software development team, answering these questions can mean the difference between a high-quality project and a disaster. NCover answers these questions.

Code Coverage Analysis, applied while running unit tests over your code, shows the developer which portions of his/her code were exercised during the test and gives a concrete measurement of unit test penetration. By tracking these statistics over time, you gain a concrete measurement of code quality during the development cycle.


Figure 1. Test coverage should increase over time as you build more and better tests into your code.

What is Code Coverage?

Code coverage analysis is a measurement of how much of a running program’s code was executed. Specifically, code coverage usually shows how many times each line of code was executed. Code coverage analysis is often described as telling you how many times each line of code was executed, but there are actually many forms of coverage metrics, such as function coverage, statement coverage, branch coverage, exit point coverage, and conditional coverage. NCover supports sequence point coverage (which is a special form of statement coverage) and branch coverage.

NCover’s definition of coverage is as follows:

  • Sequence Point Coverage – Sequence point coverage looks at the number of points within the program where the debugger can pause execution. Usually this corresponds to a line of code, however some compilers allow you to pause at several places within an expression, so the sequence points could be slightly higher than the number of lines or statements.
  • Branch Coverage – Branch coverage looks at the execution path of a method and determines the number of places where the code can make a decision to "branch" off. This corresponds to statements like if, while, for, foreach, switch, etc... The branch points are used to segment a method into contiguous blocks of code. Those code blocks are then used to calculate coverage.

So, why would you use the two forms of coverage?

Branch coverage is a more accurate measure of coverage because it compensates for method complexity. Since branch coverage is generated from the underlying opcodes, it often does not map cleanly to source code. That means it’s difficult to take branch coverage reports and determine how to write tests that will improve coverage. On the other hand, since sequence points directly reference debug symbols, it’s easy to relate sequence points to source code. The bottom line is:

  • Use branch coverage as your measure of quality, but
  • Use sequence point coverage to determine which code statements need more testing.

When Should you Use Code Coverage?

Early and often is the short answer. The heart of an iterative development environment is a continuous integration server. Each time a developer checks code into the revision control system (or on a regularly scheduled cycle), the build server should check out the latest code, do a full build of the product, run the test suite and report test success/failure and code coverage metrics. If you aren’t already using a continuous build server, CruiseControl.NET by ThoughtWorks is a terrific (and free) option. A continuous build process facilitates the early detection of code errors and immediate reporting of code coverage also makes it clear to everyone on the team if the quality of the tests are increasing or declining.

How do I Implement Code Coverage?

NCover performs analysis on code by instrumenting it behind the scenes while the tests are running. If using NUnit as the unit testing tool, simply prefix the NUnit command line with the NCover command.

NUnit command line:

nunit-console MyTests.dll

NUnit with NCover:

NCover.Console nunit-console MyTests.dll

It's easy enough to create code coverage data, but analysis is where you actually begin to improve your tests. We include NCoverExplorer with NCover for this purpose. NCoverExplorer highlights those lines of your code that are uncovered, and does statistical analysis of your coverage data as well. It can even run a program and collect coverage data for you, so many developers never touch the command line version of NCover. NCover also includes an HTML reporting feature for teams that want to generate analysis reports including highlighted code directly from their continuous integration server.


Figure 2. HTML Coverage Report


Figure 3. HTML Report showing highlighted source code.


Figure 4. NCoverExplorer puts coverage analysis on your desktop.


This article, along with any associated source code and files, is licensed under The Code Project Open License (CPOL)

Written By
This member has not yet provided a Biography. Assume it's interesting and varied, and probably something to do with programming.

Comments and Discussions

QuestionIntegrating NCoverExplorer in CruiseControl.NET? Pin
ppayal5-Apr-08 21:30
ppayal5-Apr-08 21:30 

I am using for continuous integration for my project. I want to integrate NCoverExplorer also in it.

In my project I have created separate project for all unit tests and and separate project for the code to be unittested.

How should I change my file to integrate NCoverexplorer in it?

Any help?


General General    News News    Suggestion Suggestion    Question Question    Bug Bug    Answer Answer    Joke Joke    Praise Praise    Rant Rant    Admin Admin   

Use Ctrl+Left/Right to switch messages, Ctrl+Up/Down to switch threads, Ctrl+Shift+Left/Right to switch pages.