I found this: https://stackoverflow.com/questions/63944480/maxretryerror-selenium (but that shouldn't be happening when using SeleniumBase)
but here an error
idk why
Also, Support for Python 3.5 is expected to be dropped in SeleniumBase 3.5.0 - Expecting zero impact because there have been no Python 3.5 downloads in the past week.
https://github.com/seleniumbase/SeleniumBase/releases/tag/v3.5.2 is out. Reduces required Python dependencies. Leads to faster installs on CI/CD.
2) As long as we’ve used SB, we’ve experienced situations where an element is clicked but the page doesn’t register the click and thus fails the following function call. For example, clicking a next page button and then asserting the title of the next page results in a failed assertion since the element never registered the click. Since this is in the middle of a much larger test, I’d rather it not rerun the entire test, and only try to click the “next” button again (assuming that it didnt register). To be clear, I think SB is working correctly in these scenarios (since running in slow mode solves the issue). I know that you have conditions built into clicking to ensure elements have been clicked, but sometimes, I think SB finds and clicks buttons (too quickly 😅👍) before the page can even register clicks. Is there a way to have a faster “slow” mode or would you consider building in a feature to try the previous function if the current fails, rather than only the option of a full test rerun on failures?
Hi @grezald For number 1, SeleniumBase had a reduction in optional dependencies that were installed during
For 2, you can control how slow is slow_mode. Use:
pip install, so if your tests always install dependencies before running tests, you'll see a big improvement: See seleniumbase/SeleniumBase#1406 (a significant savings)For 2, you can control how slow is slow_mode. Use:
--demo_sleep=SEC (It uses the demo_mode timeout for slow_mode)
@grezald You can see that dependency drop here: https://github.com/seleniumbase/SeleniumBase/pull/1410/files#diff-4d7c51b1efe9043e44439a949dfd92e5827321b34082903477fd04876edb7552
@grezald OK, great! Perhaps the performance boost was due to one of many different optimizations that took place. Could be anything from changes in default Chromium/Firefox command-line options, or even improvements made in the latest version of selenium. But whatever it was, looks like it was a great improvement!
pytest test_regression.py --alluredir=report --window-size=1920,1080
this is not working
Using sbase 3.5.8 with behave: I'm trying to use behave's before_all() functionality (https://behave.readthedocs.io/en/stable/api.html?highlight=before_all#environment-file-functions). If I define this function within the environment.py module (to just print some text to stdout), it successfully prints the line, but then throws a "HOOK-ERROR in before_scenario: AttributeError: 'Context' object has no attribute 'sb'". Note that the error is about "before_scenario" despite the fact that I'm only defining "before_all". I am using same environment.py module as located in seleniumbase behave examples (aside from newly defined "before_all(context)"). Is it possible to use these behave functions (before_all, before_feature, before_scenario, etc.)?
@sarchule SeleniumBase already overrode the default
If overriding them again, you would need to preserve the existing SeleniumBase versions of those methods.
behave methods like before_all, before_feature, before_scenario, etc, at the end of this file: https://github.com/seleniumbase/SeleniumBase/blob/master/seleniumbase/behave/behave_sb.pyIf overriding them again, you would need to preserve the existing SeleniumBase versions of those methods.
9 replies
Michael, hello! SeleniumBase is cool! For example, I want to display not only standard data in an html report. What if I want to display the value of some variable from my code in the report. Is it even possible?
I understand that it is possible to output the values of variables in the description of the error that Pytest made.
But, how to display the value of the variable in the html-report itself, and not in the error description?
I understand that it is possible to output the values of variables in the description of the error that Pytest made.
But, how to display the value of the variable in the html-report itself, and not in the error description?
Hi @EvgeniiBiriukov The
You can also use
pytest-html report is a separate plugin with a separate maintainer. There may be a few possible workarounds, such as using print() commands to output the data as the test runs.You can also use
self.save_data_as(data, file_name, destination_folder=None) to save that data to a file.
My presentation for SeleniumConf2022 is available: https://www.youtube.com/watch?v=0P9UpQ9yWm0&list=PL9Z-JgiTsOYRfoG_mcRBlTUIFPIknhQ6S
When using SeleniumBase "sb" pytest fixture syntax, does anyone know how to set up Visual Studio Code to allow intellisense and "go to definition" features? I've tried various extensions (e.g., python, pytest-fixtures, etc.) without success. I know this isn't a specific SeleniumBase issue, but hoping someone has figured out the magic to making this happen.
Hi @mdmintz ,
We want to access button element that contains text in its child span, having issues due to there s no id that we can access to that button. Is there any way?
We can actually get the span, but can not access to the parent button and cant click on it, snippet attached
@mburaksayici Use the
https://github.com/seleniumbase/SeleniumBase/blob/master/examples/test_contains_selector.py
:contains("TEXT") selector:https://github.com/seleniumbase/SeleniumBase/blob/master/examples/test_contains_selector.py
@denizdugru Yes, it can, through https://github.com/wkeeling/selenium-wire (installed separately):
from seleniumbase import BaseCase
from seleniumwire import webdriver # the selenium-wire webdriver
class WireTestCase(BaseCase):
def get_new_driver(self, *args, **kwargs):
options = webdriver.ChromeOptions()
options.add_experimental_option("excludeSwitches", ["enable-automation"])
options.add_experimental_option("useAutomationExtension", False)
return webdriver.Chrome(options=options)
def test_simple(self):
self.open("https://seleniumbase.io/demo_page")
for request in self.driver.requests:
print(request.url)
@EvgeniiBiriukov See if https://stackoverflow.com/a/31482681/7058266 works for you.
@EvgeniiBiriukov There were a few other answers on https://stackoverflow.com/questions/28111539/can-we-zoom-the-browser-window-in-python-selenium-webdriver
@EvgeniiBiriukov See if https://stackoverflow.com/a/31482681/7058266 works for you.
Michael, thanks! It works correctly.
Hi there! I don't know if I've said enough previously but thank you so much for your help working through things. Can you help me figure out capabilities changes with selenium 4? I use Gridlastic (selenium grid cloud provider) and with selenium 4, they are looking for this to enable video recording:
{
"gridlastic:options":
{
"video": true,
}
}I tried adding a caps.json file to test setting that, and it got added as this (cut out some lines for brevity):
{
"cloud:options":
{
"gridlastic:options":
{
"video": true,
}
}
}Is there a way to currently pass in capabilities that don't get put under cloud:options? Or can you add functionality that checks the caps provided to SB, and if it is in the format of {string}:{string} pass them straight through without putting it under cloud:options?
See the note "(check with your cloud vendor for the appropriate prefix)" on this page: https://www.selenium.dev/documentation/webdriver/getting\_started/upgrade\_to\_selenium\_4/#capabilities
@taylorm:matrix.org You'd probably use something similar to a SeleniumBase cap file: https://github.com/seleniumbase/SeleniumBase/blob/master/examples/capabilities/selenoid_cap_file.py
And then pass in the file with
--cap_file=CAP_FILE.py
I tried with a similar cap file as suggested, but actually, if I understand these lines correctly, then seleniumbase will nest all capabilities provided under cloud:options except for the ones handled specifically here (selenoid:options, screenResolution, version, platform). These handled ones get put at the top level, instead of under cloud:options. I was poking at browser_launcher on how it could be handled differently and came up with this (L1, L13-14, L29-31 here):
custom_options = {}
for key in desired_caps.keys():
capabilities[key] = desired_caps[key]
if key == "selenoid:options":
selenoid = True
selenoid_options = desired_caps[key]
elif key == "screenResolution":
screen_resolution = desired_caps[key]
elif key == "version" or key == "browserVersion":
browser_version = desired_caps[key]
elif key == "platform" or key == "platformName":
platform_name = desired_caps[key]
elif re.match("[a-zA-Z0-9]*:[a-zA-Z0-9]*", key):
custom_options[key] = desired_caps[key]
if selenium4:
chrome_options.set_capability("cloud:options", capabilities)
if selenoid:
snops = selenoid_options
chrome_options.set_capability("selenoid:options", snops)
if screen_resolution:
scres = screen_resolution
chrome_options.set_capability("screenResolution", scres)
if browser_version:
br_vers = browser_version
chrome_options.set_capability("browserVersion", br_vers)
if platform_name:
plat_name = platform_name
chrome_options.set_capability("platformName", plat_name)
if custom_options:
for key in custom_options:
chrome_options.set_capability(key, custom_options[key])open to your thoughts - this will handle multiple custom capabilities (meaning a cap that is string:string) should I submit a PR?
@taylorm:matrix.org Could work. Just make sure to test it on the latest Python, AND 3.6 for pre-selenium-4 compatibility. Then submit a PR. Be sure to run
flake8. If everything's good, I'll make a release for this. I'm working on a few big things, such as seleniumbase/SeleniumBase#1440