How to diagnose problems when a linter fails.

Sometimes you need to get under-the-hood to diagnose why a linter is failing on a particular file or to more easily reproduce an issue you are seeing with a tool.

Analyzing Linter Failures

When a linter fails, Trunk will generate a linter failure report that contains all the information needed to understand what went wrong and why. prettier, for example, will fail on improperly formatted HTML; let's see what that looks like:

 0:0  failure  prettier error (details: .trunk/out/lYa9w.yaml)  prettier

This tells us that prettier failed when Trunk ran it on trunk/hello_world.html and that we can find a full report of the invocation in .trunk/out/lYa9w.yaml. This report will include:

  1. the command that was run,

  2. the environment (e.g. cwd, environment variables),

  3. the invocation's stdout, stderr, and exit code, and

  4. repro instructions.

All of this combined is generally sufficient to fix either the file or the tool to get useful results out of the linter.

title: "prettier error"
  - command: |
      /home/horton/.cache/trunk/linters/prettier/2.3.2-368f87d0e434ae207c0a3622371f91cc/node_modules/.bin/prettier --stdin-filepath trunk/hello_world.html
      HOME: /home/horton
      PATH: /home/horton/.cache/trunk/linters/node/16.14.2/bin:/home/horton/.cache/trunk/linters/prettier/2.3.2-368f87d0e434ae207c0a3622371f91cc/node_modules/.bin
    run_from: /tmp/trunk-Bi6hWP
      env -i -C /tmp/trunk-Bi6hWP HOME=/home/horton
      --stdin-filepath trunk/hello_world.html
    exit_code: 2
    stdout: (none)
    stderr: |
      [error] trunk/hello_world.html: SyntaxError: Unexpected closing tag "head". It may happen when the tag has already been closed by another tag. For more info see (3:2)
      [error]   1 | <html>
      [error]   2 |  head>
      [error] > 3 |  </head>
      [error]     |  ^^^^^^^
      [error]   4 |  <body>
      [error]   5 |    <h1>Hello World<h1>
      [error]   6 |  </body>

Debugging Linter Failures in CI

If linter failures occur while Trunk Check is running in CI, you'll see something like this:

To debug these, you'll need get the linter failure report(s) from the Github Actions output. Click on "Details" to see a screen like this:

Then click on "View more details on trunk-io"

You'll then see your check run summary. Click "<your repo name>/pull_request" to see the CI logs, and scroll all the way to the bottom.

You'll then see a list of the failures that occurred, as well as all of the linter failure reports in collapsibles. You can then follow the instructions in Analyzing Linter Failures to resolve the problem.


When trunk check is run, a dynamic set of check actions are generated and executed in parallel. An engineer developing custom integrations may find it helpful to examine in detail exactly what trunk check does and sees when it runs a given tool. The output of the check run will include an ACTIONS section with a execution report for each action that was run as well as cache hit information.

For example, to run a report on all actions taken on

trunk check --verbose

Last updated