Chromedriver on windows (uc_driver.exe) directory not customizable

[cmp-nct]

I've been digging a bit, trying to solve the problem and there appears to be no way to do this.
I do not want to permanently re-download the driver but I have a testing environment with multiple chrome versions and I want to run them simultaneously.
The driver SB uses is '\Lib\site-packages\seleniumbase\drivers\uc_driver.exe' or DRIVER_DIR = os.path.dirname(os.path.realpath(drivers.__file__))
Or it uses chromedriver,msedgedriver etc, but always the same vanilla type name in the python package directory.

So the driver file is not version tagged and the location appears to be hardcoded to the package directory of the drivers python file.

Is there a specific reason why this was done so restrictive and in a folder that is normally only managed by python package installation ?
Any intention to customize this or version tag it ?

[mdmintz]

Duplicate of #3087 (comment).

There are 2 places SeleniumBase checks for drivers:

1. The seleniumbase/drivers folder.
2. Each folder of os.environ["PATH"], in order.

If you don't want to use the seleniumbase/drivers folder for driver storage, then you can download chromedriver manually to a folder on your System PATH. The version of chromedriver must match your version of Chrome, or else SeleniumBase will automatically download a matching driver to the seleniumbase/drivers folder, which can't be changed. You can, however, change os.environ["PATH"] at runtime so that you can add a folder to the front that isn't already there.

There's generally only one main/default version of Chrome on a machine, so you would only need a single driver to pair with that version of Chrome.

If you have multiple versions of Chrome that need to be used at the same time, you can either use a different Python virtual environment for each one (as each virtual environment will have it's own drivers/ folder), OR, you can set driver_version="keep" (after getting the latest driver), and then you can use that driver with not only the latest Chrome, but also earlier versions of Chrome. See #2849 (comment) for details.

[cmp-nct]

Thanks for that hint on driver_version="keep"

I'll try the PATH trick (that was kind of my last resort to try today).
There is also a potential race condition when two drivers are started simultaneously that needs locking.

In addition there is a bug currently, when you have a "lingering" driver process running on windows and you start any new driver, the extraction of the uc_driver.exe is going to fail because the file shows as 'in use'. In my case it just stalled there.

Imho the best solution forward is to either tag the driver executable with a version, so they are unique per version, or to expose the path by API so the location can be customized.

[mdmintz]

Although you can have multiple UC browsers running at the same time, only one UC browser can be launched at one time. You'll need to activate thread-locking to achieve that.
(Hint: Include sys.argv.append("-n") so that the current thread-locking system activates. The built-in thread-locking system is what prevents multiple threads from updating the same resource at the same time, such as multiple threads trying to patch the same driver at the same time.)

There's an example of using ThreadPoolExecutor in the UC Mode docs:

import sys
from concurrent.futures import ThreadPoolExecutor
from seleniumbase import Driver
sys.argv.append("-n")  # Tell SeleniumBase to do thread-locking as needed

def launch_driver(url):
    driver = Driver(uc=True)
    try:
        driver.get(url=url)
        driver.sleep(2)
    finally:
        driver.quit()

urls = ['https://seleniumbase.io/demo_page' for i in range(3)]
with ThreadPoolExecutor(max_workers=len(urls)) as executor:
    for url in urls:
        executor.submit(launch_driver, url)

Thread-locking is automatic from the pytest formats.
(See SeleniumBase/help_docs/syntax_formats.md for all formats.)