When writing plugin tests, a common task is testing various kinds of highlighting (inspections, annotators, parser error highlighting, etc.). The IntelliJ Platform provides a dedicated utility and markup format for this task.
To test the highlighting for the file currently loaded into the in-memory editor, invoke
CodeInsightTestFixture.checkHighlighting(). The parameters to the method specify which severities should be taken into account when comparing the results with the expected results: errors are always taken into account, whereas warnings, weak warnings, and infos are optional. To ignore verifying additional highlighting, set parameter
Alternatively, you can use
CodeInsightTestFixture.testHighlighting(), which loads a testdata file into the in-memory editor and highlights it as a single operation.
If you need to test inspections, they must be enabled explicitly. This is done by calling
CodeInsightTestFixture.enableInspections() in the setup method of your test or directly in a test method, before the call to
This method takes the tested file and the answer file describing expected highlighting information for each token. The answer file format is as follows:
It starts with token value, which is the actual token value from the tested file (space characters in token value are replaced with
␣). It is followed by indented list of text attribute keys, starting with the key defined for the given token, followed by fallback keys separated by
Creating an answer file from scratch would be cumbersome process, so it is recommended to provide only the tested file and executing the test. Test will fail and the expected answer file will be generated. Review the generated file carefully, and provide fixes to the syntax highlighter and the answer file if needed.
Expected Highlighting Results
The expected results of the highlighting are specified directly in the source file. The platform supports an extensive XML-like markup language for this. In its simplest form, the markup looks like this:
A more realistic example, embedded in Java test data to be highlighted:
The tag name specifies the severity of the expected highlighting. The following severities are supported:
<inject>for an injected fragment
<symbolName>for a marker that highlights an identifier according to its type
any custom severity can be referenced by its name
The tag can also have the following optional attributes.
descrexpected (hardcoded) message associated with the highlighter (if not specified, any text will match). If the message contains a quotation mark, it can be escaped by putting two backslash characters before it.
bundleMsgexpected message from a message bundle in format
[bundleName#] bundleKey [|argument]...
tooltipexpected tooltip message
TextAttributesKeyreferenced by its
effectColorexpected colors for the highlighting
effectTypeexpected effect type for the highlighting (see
fontTypeexpected font style for the highlighting (
3- bold italic)
Nested tags are supported:
Overlapping tags (annotations) are currently not supported in the test framework (but displayed correctly in the editor, albeit this is not an officially supported scenario):