[Bug] Incomplete jasmine stack-trace in case of utam method caused a test failure

[tanujvishnoi]

Actual behavior

Hey everyone, we're working with utam-webdriver-jasmine flavors along with TS and during spec run we have noticed the stack trace get truncated if a test failed because of web-element issue [utam waitfor- method or element is not in scope],
in-short we got a complete stack trace in case of assertion failure like which spec line and file, but other failures wont populate complete stack-traces & the error source,

• Unable to compile a Page Object

• Invalid generated code

• Runtime error

What is your expected behavior?

during any failure, the complete stack-trace should be visible indicating the spec line which actually caused this issue

Steps to reproduce

1: Create a TS-Jasmine-utam-wdio based project
2: Create few utam files with wait for methods that will fail or some non-existing locators
3: Draft spec file and use utam file and called specific methods or access locators stated in step2
4: Run the spec and observe the stack traces

Environment

• Node.js: v16.18.1
• OS: Windows
• Browser: Chrome
• Browser Version: Latest
• Type script
• • typescript: "^4.8.3",
• Jasmine
• • @wdio/jasmine-framework: "^7.24.1",

Additional context

Refer to attached truncated stack-traces
1:

2:

[olivier-martin-sf]

Hi @tanujvishno, thank you for taking the time to open this issue.

Hey everyone, we're working with utam-webdriver-jasmine flavors along with TS

May you clarify what's utam-webdriver-jasmine? We don't have such package published on npm. Is it a package that you developed internally or is it terminology meaning that you are using WebdriverIO with Jasmine?

in-short we got a complete stack trace in case of assertion failure like which spec line and file, but other failures won't populate complete stack traces & the error source

Would you be able to set up a minimal repository where the unwanted behavior can be reproduced? That would be helpful to see a test that exhibits this behavior (it can be against any Web page and doesn't have to be against a Salesforce test environment).

[tanujvishnoi]

sorry for the confusion , Yes, by utam-webdriver-jasmine I meant webdriverIO with jasmine , I'll share a repo url soon reproducing this issue

[tanujvishnoi]

@olivier-martin-sf https://github.com/tanujvishnoi/utam_demo , here is the demo code to produce truncated stack trace

[olivier-martin-sf]

Thank you so much for providing the minimal repository, we are going to look into this as soon as we have the capacity to do so (probably next week)

[olivier-martin-sf]

Hi @tanujvishnoi, I look into this issue and I would like to clarify the problem statement and expectations.

Digging deeper, if I understood the problem statement correctly, it seems that the issue appears in the context of a waitFor predicate.

• May you confirm that the first use case (Getting non-existing element) works correctly?
• May you confirm that the second use case (Getting non-existing element inside a predicate) reproduces the original issue?
• May you define what would be a correct scenario (providing an example of an expected stack trace)?

Getting non-existing element

From my observations, if I am trying to get a non-existing element, the stack trace provides the Jasmine context (with the position of the line that throws the error in the spec file). To reproduce try to define a public element with a CSS selector that doesn't match any element and try to invoke the function that gets the element from the test. For example:

{
  // other non-relevant properties
  "elements": [
    {
      "name": "userNameInput",
      "type": ["editable", "clickable"],
      "public": true,
      "selector": {
        "css": ".dsadsads"
      }
    }
  ],
  // other non-relevant properties
}

And try to invoke this from the test:

await demoPage.getUserNameInput();

This should throw an error with the following stack trace:

[0-0] Error: Can't find elements with locator 'dsadsads' inside its scope element.
[0-0]     at ElementWdioAdapter.findElements (/Users/JaneDoe/utam_demo/node_modules/wdio-utam-service/build/adapters/element-adapter.js:95:19)
[0-0]     at processTicksAndRejections (node:internal/process/task_queues:96:5)
[0-0]     at async ElementWdioAdapter.findElement (/Users/JaneDoe/utam_demo/node_modules/wdio-utam-service/build/adapters/element-adapter.js:88:17)
[0-0]     at async Demo2.getUserNameInput (/Users/JaneDoe/utam_demo/pageObjects/demo2.js:44:23)
[0-0]     at async UserContext.f1 (/Users/JaneDoe/utam_demo/tests/spec/test2.spec.ts:9:7)

As you can see the last line of the stack trace is the Jasmine UserContext.

Note: f1 is the name of the function that wraps the test, I avoid arrow functions in tests to have more explicit context in the stack trace.

Getting non-existing element inside a predicate

However if I am trying to wait for an element that doesn't exist in a waitFor predicate such as:

// other non-relevant properties
"methods:"  [
   {
      "name": "waitForNonExistentElement",
      "compose": [
        {
          "apply": "waitFor",
          "args": [
            {
              "type": "function",
              "predicate": [
                {
                  "element": "root",
                  "apply": "containsElement",
                  "args": [
                    {
                      "type": "locator",
                      "value": {
                        "css": ".fdskjfdkslj"
                      }
                    }
                  ]
                }
              ]
            }
          ]
        }
      ]
    }
]
// other non-relevant properties

And I am trying to invoke this method from my test:

await demoPage.waitForNonExistentElement();

This is the error that I am having:

[0-0] Error: Timeout while waiting for condition after 4639ms.
[0-0]     at handlePollTimeout (/Users/JaneDoe/utam_demo/node_modules/@utam/core/build/helper-methods.js:53:28)
[0-0]     at /Users/JaneDoe/utam_demo/node_modules/@utam/core/build/helper-methods.js:62:25
[0-0]     at processTicksAndRejections (node:internal/process/task_queues:96:5)

There we don't have the information about the test context (we don't know which line in our spec file generated the error).

[tanujvishnoi]

Hi @olivier-martin-sf

Thanks for sharing the examples . Yes the issue eventually reproducible for predicate . If a condition get failed inside a predicate , we did not get the full error stack
Apart from the above 2 conditions, i added one more where we expect a hidden element to get visible .
https://github.com/tanujvishnoi/utam_demo

`1) reproduce truncate error stack Case 1: Getting no existing element directly using getter
Error: Can't find elements with locator 'dsadsads' inside its scope element.
Error: Can't find elements with locator 'dsadsads' inside its scope element.
at ElementWdioAdapter.findElements (\utam_demo\node_modules\wdio-utam-service\build\adapters\element-adapter.js:95:19)
at processTicksAndRejections (node:internal/process/task_queues:96:5)
at async ElementWdioAdapter.findElement (\utam_demo\node_modules\wdio-utam-service\build\adapters\element-adapter.js:88:17)
at async Demo.getUserNameInput (\utam_demo\pageObjects\demo.js:50:23)
at async UserContext. (\utam_demo\tests\spec\test.spec.ts:12:21)

2. reproduce truncate error stack Case 2: Getting no existing element inside a method
   Error: Can't find elements with locator 'dsadsads' inside its scope element.
   Error: Can't find elements with locator 'dsadsads' inside its scope element.
   at ElementWdioAdapter.findElements (\utam_demo\node_modules\wdio-utam-service\build\adapters\element-adapter.js:95:19)
   at processTicksAndRejections (node:internal/process/task_queues:96:5)
   at async ElementWdioAdapter.findElement (\utam_demo\node_modules\wdio-utam-service\build\adapters\element-adapter.js:88:17)
   at async Demo.getUserNameInput (\utam_demo\pageObjects\demo.js:50:23)
   at async Demo.clickUserName (\utam_demo\pageObjects\demo.js:85:29)
   at async UserContext. (\utam_demo\tests\spec\test.spec.ts:23:9)

3. reproduce truncate error stack Case 3: Getting no existing inside a predicate
   Error: Timeout while waiting for condition after 10484ms.
   Error: Timeout while waiting for condition after 10484ms.
   at handlePollTimeout (\utam_demo\node_modules@utam\core\build\helper-methods.js:53:28)
   at \utam_demo\node_modules@utam\core\build\helper-methods.js:62:25
   at processTicksAndRejections (node:internal/process/task_queues:96:5)

4. reproduce truncate error stack Case 4: Wait for visibility of hidden search input
   Error: Timeout while waiting for condition after 10265ms.
   Error: Timeout while waiting for condition after 10265ms.
   at handlePollTimeout (\utam_demo\node_modules@utam\core\build\helper-methods.js:53:28)
   at \utam_demo\node_modules@utam\core\build\helper-methods.js:62:25
   at runMicrotasks ()
   at processTicksAndRejections (node:internal/process/task_queues:96:5)

5. reproduce truncate error stack Case 5: Get hidden element ,then wait for visibility
   Error: Timeout while waiting for condition after 10212ms.
   Error: Timeout while waiting for condition after 10212ms.
   at handlePollTimeout (\utam_demo\node_modules@utam\core\build\helper-methods.js:53:28)
   at \utam_demo\node_modules@utam\core\build\helper-methods.js:62:25`

[olivier-martin-sf]

Thanks, I will pull your changes.

Judging by your stack traces example, 1 & 2 provide the full stack trace but 3, 4 & 5 don't and all of them are related to timeout errors so can we narrow the problem statement to: "we don't have the full stack trace when there's a wait condition that times out?"

Do you agree with that? If yes, I will start investigating why we shallow the context in waitFor

[tanujvishnoi]

Yes absolutely agree.

[olivier-martin-sf]

After some investigation, I can confirm that I can reproduce the issue. It happens when the waitFor function is being invoked and the condition doesn't return a truthy value before the timeout.

The problem is due to the execution of an async function that is triggered by a timer as described here: https://github.com/sindresorhus/got/blob/main/documentation/async-stack-traces.md. I created an item in our backlog so that we can fix the issue in a future release.

Thank you for reporting this

[tanujvishnoi]

Hi @olivier-martin-sf just taking a follow up, please let us know once we have a fix for this issue

[irinash13]

Hello @tanujvishnoi , what company do you represent?