Testing API

The vscode_deno extension implements a client for the vscode Testing API and when using a version of Deno that supports the testing API, tests in your project will be displayed within your IDE for Deno enabled projects.

Test display

When both the editor and the version of Deno support the testing API, the Test Explorer view will activate represented by a beaker icon, which will provide you with a side panel of tests that have been discovered in your project.

Also, next to tests identified in the code, there will be decorations which allow you to run and see the status of each test, as well as there will be entries in the command pallette for tests.

Discovering tests

Currently, Deno will only discover tests that are part of the "known" modules inside a workspace. A module becomes "known" when it is opened in the editor, or another module which imports that module is "known" inside the editor.

In the future, tests will be discovered in a similar fashion to the way the deno test subcommand discovers tests as part of the root of the workspace.

Running tests

You can run tests from the Test Explorer view, from the decorations next to the tests when viewing the test code, or via the command pallette. You can also use the filter function in the Text Explorer view to exclude certain tests from a test run.

Currently, Deno only supports the "run" test capability. We will be adding a debug run mode as well as a coverage run mode in the future. We will also be integrating the benchmarking tests as a tag, so they can be run (or excluded) from your test runs.

The Deno language server does not spin up a new CLI subprocess. It instead spawns a new thread and JavaScript runtime per test module to execute the tests.

Test output

Any console.log() that occurs in your tests will be sent to the test output window within vscode.

When a test fails, the failure message, including the stack trace, will be available when inspecting the test results in vscode.

How tests are structured

Test will be displayed in the Test Explorer at the top level with the module that contains the test. Inside the module will be all the tests that have been discovered, and if you are using test steps, they will be included under the test.

In most cases, the Deno language server will be able to statically identify tests, but if you are generating tests dynamically, Deno may not be aware of them until runtime. In these cases it may not be possible to filter these tests out of a run, but they will be added to the explorer view as they are encountered.

Configuration

By default, tests are executed in a similar fashion to if you were to use deno test --allow-all on the command line. These default arguments can be changed by setting the Deno > Testing: Args option in your user or workspace settings (or deno.testing.args if you are configuring manually). Add individual arguments here which you would have used with the deno test subcommand.

Based on other settings that you have, those options will be automatically merged into the "command line" used when running tests unless explicitly provided in the Deno > Testing: Args setting. For example if you have a Deno: Import Map (deno.importMap) set, the value of that will be used unless you have provided an explicit --import-map value in the testing args setting.

Known limitations and caveats

Because of the way the Deno test runner runs, it is not possible to exclude (or explicitly include) a test step. While the vscode UI will allow you to do this, by for example, choosing to run a specific test step, all test steps in that test will be run (but vscode will not update the results for them). So if there are other side effects in the test case, they may occur.