Breakpoints not triggering when launch configuration uses promptString input

[jasonpolites]

Does this issue occur when all extensions are disabled?: Yes

• VS Code Version:

Version: 1.100.0 (user setup)
Commit: 19e0f9e
Date: 2025-05-07T12:48:53.763Z
Electron: 34.5.1
ElectronBuildId: 11369351
Chromium: 132.0.6834.210
Node.js: 20.19.0
V8: 13.2.152.41-electron.0
OS: Windows_NT x64 10.0.19045

• OS Version:

Host OS:
Microsoft Windows [Version 10.0.19045.5737]

Guest OS:
WSL version: 2.3.26.0
Kernel version: 5.15.167.4-1
Ubuntu 20.04.6 LTS

Steps to Reproduce:

I am using Mocha to run javascript tests

In order to run a single test in a suite, mocha supports a "grep" syntax:

mocha --grep "Test Name"

My goal is to create a launch configuration in launch.json which allows me to debug a specific test. To achieve this, I am using the ${input:...} syntax in my launch.json:

"configurations": [
    {
      "name": "Mocha Tests (Specific Test)",
      "type": "node",
      "request": "launch",
      "program": "${workspaceFolder}/client/node_modules/mocha/bin/_mocha",
      "args": [
        "${workspaceFolder}/client/test",
        "--no-timeouts",
        "--colors",
        "--grep",
        "\"${input:pickTest}\""
      ],
      "internalConsoleOptions": "openOnSessionStart",
      "skipFiles": [
        "<node_internals>/**"
      ]
    }
],
"inputs": [
  {
    "id": "pickTest",
    "description": "Choose a test name",
    "type": "promptString"
  }
]

This does indeed execute the specific test based on the user input, however breakpoints do not trigger.

If I alter the args to hard-code a specific test, the breakpoints do trigger. For example:

This works

"args": [
  "${workspaceFolder}/client/test",
  "--no-timeouts",
  "--colors",
  "--grep",
  "\"Test Name\""
],

This does not work

"args": [
  "${workspaceFolder}/client/test",
  "--no-timeouts",
  "--colors",
  "--grep",
  "\"${input:pickTest}\""
],

Is debugging supported when using promptString? If so, is there a secret trick I'm missing?

[connor4312]

/jsDebugLogs

[vs-code-engineering]

Please collect trace logs using the following instructions:

If you're able to, add "trace": true to your launch.json and reproduce the issue. The location of the log file on your disk will be written to the Debug Console. Share that with us.

⚠️ This log file will not contain source code, but will contain file paths. You can drop it into https://microsoft.github.io/vscode-pwa-analyzer/index.html to see what it contains. If you'd rather not share the log publicly, you can email it to connor@xbox.com

[jasonpolites]

The trace output does have some data in it that I'd rather not put on the public internet, so I will email the file to connor@xbox.com (baller email address btw)

I had a quick look and didn't spot anything obvious, but I don't really know what I'm looking for.

There is one bit of information that may be helpful. Upon seeing your comment, I went to repro the issue and wasn't able to.. suddenly it worked! Then I discovered the repro case.

Mocha has the familiar "describe=>it" syntax:

describe ('foo', () => {
  it('bar', () => {
    // Test code here
  });
});

If the breakpoint is within the describe callback, it will trigger. If it's within the it callback, it will not:

describe ('foo', () => {

  // X<== Break point here works

  it('bar', () => {
    // X<== Break point here does NOT work
  });
});

Potentially unrelated, but I did see a bunch of these line:

{"seq":4,"type":"event","event":"loadedSource","body":{"reason":"new","source":{"name":"<node_internals>/internal/per_context/primordials"...

Which was unexpected given:

"skipFiles": [
  "<node_internals>/**"
]

Maybe this is normal?

Edit

In case it wasn't clear, the inner breakpoint will trigger if I simply run the whole test suite. It is only with the ${input:...} that it does not trigger

[connor4312]

Looks like the test CLI is executed correctly. Not sure why it would be different with pickString, the args the debugger gets is

[
  "/home/jasonpolites/projects/home-budget/client/test",
  "--no-timeouts",
  "--colors",
  "--grep",
  "\"call onProgress\""
]

You probably don't need to be wrapping your test name with quotes, that could be breaking it although you said it works by manually specifying "\"Test Name\"".

[jasonpolites]

You found it.

The grep failed due to the extra quotes. My (incorrect) assumption was that because the command line execution requires quotes, the launch.json arg should have quotes. It's still counter-intuitive to me, the command line that the debug console says it's executing looks like this:

... --no-timeouts --colors --grep Test Name

I would expect that should be:

... --no-timeouts --colors --grep "Test Name"

but.. whatever. Removing these "extra" quotes fixes the issue. Apologies for the wild goose chase. Closing with this comment.

Thanks!

[connor4312]

Node.js programs are spawned via child_process.spawn and avoid the shell (unless you were to use the node-terminal launch type which does in fact run in the shell)