[🐛 Bug]: WebDriverWait polling consistency between Java and Python implementations

[guillaumepichenot]

What happened?

I have noticed that on my Python-based Seleniums runs, when using a WebDriverWait to check (let's say by CSS selector) for a non existing element in the DOM, the call never takes less than 500ms, whereas I consistently get <100ms for existing elements.

The behavior does not arise in Java implementations of the same scenario.

My best guess is that the Java and Python implementation have different sleep strategy orchestration.

Current python loop :

selenium/py/selenium/webdriver/support/wait.py

Lines 94 to 105 in 1b64798

	while True:
	try:
	value = method(self._driver)
	if value:
	return value
	except self._ignored_exceptions as exc:
	screen = getattr(exc, "screen", None)
	stacktrace = getattr(exc, "stacktrace", None)
	time.sleep(self._poll)
	if time.monotonic() > end_time:
	break
	raise TimeoutException(message, screen, stacktrace)

Current java loop :

selenium/java/src/org/openqa/selenium/support/ui/FluentWait.java

Lines 201 to 237 in 1b64798

	while (true) {
	try {
	V value = isTrue.apply(input);
	if (value != null && (Boolean.class != value.getClass() || Boolean.TRUE.equals(value))) {
	return value;
	}
	// Clear the last exception; if another retry or timeout exception would
	// be caused by a false or null value, the last exception is not the
	// cause of the timeout.
	lastException = null;
	} catch (Throwable e) {
	lastException = propagateIfNotIgnored(e);
	}
	// Check the timeout after evaluating the function to ensure conditions
	// with a zero timeout can succeed.
	if (end.isBefore(clock.instant())) {
	String message = messageSupplier != null ? messageSupplier.get() : null;
	String timeoutMessage =
	String.format(
	"Expected condition failed: %s (tried for %d second(s) with %d milliseconds"
	+ " interval)",
	message == null ? "waiting for " + isTrue : message,
	timeout.getSeconds(),
	interval.toMillis());
	throw timeoutException(timeoutMessage, lastException);
	}
	try {
	sleeper.sleep(interval);
	} catch (InterruptedException e) {
	Thread.currentThread().interrupt();
	throw new WebDriverException(e);
	}
	}

Note that the Python orchestration is :

1. Timestamp the timeout end
2. Perform the lookup
3. sleep
4. Check now vs. timeout to exit

The Java implementation is :

1. Timestamp the timeout end
2. Perform the lookup
3. Check new vs. timeout to exit
4. Sleep

The inversion of steps 3 and 4 is manifest IMO.
The Java version looks more sensible to me, as it allows for effectively having timeout=0 checks, while python implementations in practice will never get a sub pollingInterval check (for non-existing elements).

How can we reproduce the issue?

WebDriverWait(driver, 0).until(
  visibility_of_any_elements_located((By.CSS_SELECTOR, "#anyIdThatDoesNotExist"))
);


Notice you will never get <500ms (`POLL_FREQUENCY`) results.

Relevant log output

N/A

Operating System

macOs, Linux

Selenium version

4.0.25

What are the browser(s) and version(s) where you see this issue?

Chrome 129

What are the browser driver(s) and version(s) where you see this issue?

ChromeDriver 129

Are you using Selenium Grid?

No response

[github-actions]

@guillaumepichenot, thank you for creating this issue. We will troubleshoot it as soon as we can.

---

Info for maintainers

Triage this issue by using labels.

If information is missing, add a helpful comment and then I-issue-template label.

If the issue is a question, add the I-question label.

If the issue is valid but there is no time to troubleshoot it, consider adding the help wanted label.

If the issue requires changes or fixes from an external project (e.g., ChromeDriver, GeckoDriver, MSEdgeDriver, W3C), add the applicable G-* label, and it will provide the correct link and auto-close the issue.

After troubleshooting the issue, please add the R-awaiting answer label.

Thank you!

[navin772]

Hi @guillaumepichenot, I tried reproducing this and indeed the time is never less than 500ms as the POLL_FREQUENCY is set to 0.5s if not provided as a parameter.
You can set the poll_frequency parameter to 0.1 to see sub 500ms time.
Something like this:

WebDriverWait(driver, 0, poll_frequency=0.1).until(
  visibility_of_any_elements_located((By.CSS_SELECTOR, "#anyIdThatDoesNotExist"))
);

But regardless, I think the python bindings should be like Java ones as you suggested. I will open a PR for the same!

Thanks for reporting this issue!

[guillaumepichenot]

@navin772 thanks for confirming.

We can indeed work-around the issue as you highlighted.

(By the way, the bug template looked more appropriate than feature request but anyway, I guess we all agree consistency is the key argument here).

[github-actions]

This issue has been automatically locked since there has not been any recent activity since it was closed. Please open a new issue for related bugs.