Configuration

To do anything useful, once installed, you should configure Behat Code Coverage according to your project needs.

Each of the settings described below has a direct equivalent in PHPUnit, if your codebase uses both testing frameworks you may wish to ensure the settings are aligned between the two tools.

Settings are configured in your project’s behat.yml or behat.yml.dist.

Driver

In order to generate code coverage data, you must have a code coverage driver installed - using either Xdebug or the PCOV extensions is recommended, although you can also make use of PHP’s built in PHPDBG. Drivers are detected at runtime, they do not need to be configured.

Filter

A filter is the mechanism to include (or exclude) directories or files from the generated coverage reports. Technically this is an optional configuration setting, but practically speaking most projects will want to configure this to report only on their own source files rather than e.g. test files or vendor files.

For directories, list them by key. Optionally, you can specify a prefix or suffix:

default:
    extensions:
        DVDoug\Behat\CodeCoverage\Extension:
            filter:
                include:
                    directories:
                        'src':
                            suffix: 'Controller.php'
                        'src/Repository':
                            prefix: 'API'
                exclude:
                    directories:
                        'tests/data' ~

For files, list them by name:

default:
    extensions:
        DVDoug\Behat\CodeCoverage\Extension:
            filter:
                include:
                    files:
                        - 'bootstrap.php'
                        - 'index.php'
                exclude:
                    files:
                        - 'src/ExcludeMe.php'
                        - 'src/ExcludeMeToo.php'

Filters have two additional settings, to control how files that have not been covered should be handled in the report. By default uncovered files are included but not processed.

default:
    extensions:
        DVDoug\Behat\CodeCoverage\Extension:
            filter:
                includeUncoveredFiles: true # defaults to true
                processUncoveredFiles: false # defaults to false
  • Included means that the file is incorporated into the reporting, showing up as 0% covered. This allows you to see files that you haven’t written any tests for yet. If you choose not to include uncovered files, then only files that have been loaded by your test suite (whether directly or indirectly) will be shown on the report.

  • Uncovered files, by definition, will not been loaded into the PHP runtime environment during the execution of the test suite. This means that they have not been analysed by the coverage driver for detection of things like executable vs non-executable lines, dead code detection or calculation of branches and paths. By opting uncovered files into processing, the files will be loaded via an include() call and passed through the driver in exactly the same way as covered files are. However, not all files can be safely loaded in this way, for instance include()ing a script may have runtime side-effects. Because Behat Code Coverage cannot know the structure of your codebase and which files are safe to include() and which are not, by default uncovered file processing is disabled for safety reasons.

    Note

    The processUncoveredFiles setting has been removed in php-code-coverage v10, configuring it here is deprecated

Reports

Behat Code Coverage allows you to generate reports in any/all of the report formats provided by php-code-coverage. Reports are configured under the reports key.

Text

The text report is the easiest way to get started, it simply outputs the report results to the screen after each test run. It is configured by setting the text key. The default values are outlined below:

default:
    extensions:
        DVDoug\Behat\CodeCoverage\Extension:
            reports:
                text:
                    showColors: false
                    showOnlySummary: false
                    showUncoveredFiles: false
                    lowUpperBound: 50
                    highLowerBound: 90
  • If showColors is true, the results will be output in a colour-coded format, red for low coverage, amber/yellow for a medium amount of coverage and green for high coverage. If desired, the thresholds for each colour can be configured via lowUpperBound and highLowerBound.

  • If you have a large codebase, outputting the coverage data for each and every individual file to the CLI may be too be noisy to be helpful. If so, you can set showOnlySummary to true which will output only a project-level overview.

  • By default, when showing data for individual files the text report does not show data for uncovered files (even if data collection enabled under filter). This can be changed if desired by setting showUncoveredFiles to true. Note that this is a report-specific display option only, choosing not to show uncovered files in the text report has no impact on whether they are included in other types of report.

HTML

The HTML report is the most common format of report. As well as a summary report for the project providing high-level data, it also includes a detailed overview of each file showing the coverage on a function by function, line by line basis. It is configured by setting the html key. The default values are outlined below:

default:
    extensions:
        DVDoug\Behat\CodeCoverage\Extension:
            reports:
                html:
                    target: <directory>  # no default value, you must specify
                    lowUpperBound: 50
                    highLowerBound: 90
  • The mandatory target key specifies the target directory to place the report files.

  • The HTML report is heavily colour-coded format using red for low coverage, amber/yellow for a medium amount of coverage and green for high coverage. If desired, the thresholds for each colour can be configured via lowUpperBound and highLowerBound.

Clover

Originating from the Java world, Clover-format reports are a standard way of exchanging coverage data between tools. It is configured by setting the clover key. The default values are outlined below:

default:
    extensions:
        DVDoug\Behat\CodeCoverage\Extension:
            reports:
                clover:
                    target: <file>  # no default value, you must specify
                    name: ''
  • The mandatory target key specifies the destination filename to use for the report. The reports are in XML format, an .xml file extension is suggested.

  • Optionally, you can configure the name of your project via the name key.

Cobertura

Also originating from the Java world, Cobertura-format reports are becoming a standard way of exchanging coverage data between tools. It is configured by setting the cobertura key. The default values are outlined below:

default:
    extensions:
        DVDoug\Behat\CodeCoverage\Extension:
            reports:
                cobertura:
                    target: <file>  # no default value, you must specify
                    name: ''
  • The mandatory target key specifies the destination filename to use for the report. The reports are in XML format, an .xml file extension is suggested.

  • Optionally, you can configure the name of your project via the name key.

Crap4j

An older, discontinued tool from the Java world. You can generate Crap4j-compatible reports by setting the crap4j key. The default values are outlined below:

default:
    extensions:
        DVDoug\Behat\CodeCoverage\Extension:
            reports:
                crap4j:
                    target: <file>  # no default value, you must specify
                    name: ''
  • The mandatory target key specifies the destination filename to use for the report. The reports are in XML format, an .xml file extension is suggested.

  • Optionally, you can configure the name of your project via the name key.

PHP “.cov”

A PHP or “.cov” report is a raw serialisation of internal php-code-coverage state, allowing for full fidelity of data to be preserved. They can be manipulated by the phpcov tool, for instance to combine reports from multiple testing tools. You can generate PHP “.cov” reports by setting the php key.

default:
    extensions:
        DVDoug\Behat\CodeCoverage\Extension:
            reports:
                php:
                    target: <file>  # no default value, you must specify
  • The mandatory target key specifies the destination filename to use for the report. The reports are actually PHP, but a .cov file extension is customary.

PHPUnit XML

You can generate PHPUnit XML reports by setting the xml key.

default:
    extensions:
        DVDoug\Behat\CodeCoverage\Extension:
            reports:
                xml:
                    target: <directory>  # no default value, you must specify
  • The mandatory target key specifies the target directory to use for the report.

Branch and path coverage

When using Xdebug as a coverage driver, it has the ability to generate branch and path coverage data as well as the traditional line-based data. More information on this topic is available at https://doug.codes/php-code-coverage.

default:
    extensions:
        DVDoug\Behat\CodeCoverage\Extension:
            branchAndPathCoverage: true

By default branchAndPathCoverage is true when running under Xdebug, false otherwise.

Caching

Since analysing source code files to generate coverage reports is computationally expensive, Behat Code Coverage makes use of a cache to ameliorate this.

default:
    extensions:
        DVDoug\Behat\CodeCoverage\Extension:
            cache: <directory>

The default cache directory is sys_get_temp_dir() . '/behat-code-coverage-cache'. You may wish to relocate this this to be inside your project workspace.