google-chrome version is UNKNOWN with webdriver_manager 3.5.3 (Win) #307
Comments
|
@np-8 thank you for your report! It's very interesting because I've tried to find all the ways to determine google chrome on windows and I have made a lot of tests. Please add some additional info which will help to find a solution for your issue. Please give answers to these questions to let us see the full picture.
|
|
I did some debugging on this. So basically, the Here are the outputs on my machine: (1)(2)(3)(4)(5)My Chrome versionMy
I have update logo at the top right corner of the Chrome window:
Before clicking "Update", the version was What might be the problemThe first problem is that the path returned with That does not entirely explain why the returned Chrome version is |
|
In addition to above, @aleksandr-kotlyar:
|
|
Summarizing previous, the The File "C:\somevenvpath\lib\site-packages\selenium\webdriver\remote\errorhandler.py", line 247, in check_response
raise exception_class(message, screen, stacktrace)
selenium.common.exceptions.SessionNotCreatedException: Message: session not created: This version of ChromeDriver only supports Chrome version 98
Current browser version is 97.0.4692.99 with binary path C:\Program Files (x86)\Google\Chrome\Application\chrome.exeNote that |
|
Now debugging inside the def read_version_from_cmd(cmd, pattern):
with subprocess.Popen(
cmd,
stdout=subprocess.PIPE,
stderr=subprocess.DEVNULL,
stdin=subprocess.DEVNULL,
shell=True,
) as stream:
stdout = stream.communicate()[0].decode()
version = re.search(pattern, stdout)
breakpoint()
return versionThe So the problem seems to be that nothing is returned to |
|
I'm trying to run just the powershell script part. I copy-pasted the script contents, which are: into a new script file, and ran it, and got: Edit: Fixed for a copy-paste error in the powershell script (it had an extra newline that caused the error) |
|
While running this in powershell works: ..the same command does not produce desired output when using the In [44]: cmd = r""" powershell "$ErrorActionPreference='silentlycontinue' ; (Get-Item -Path "$env:PROGRAMFILES (x86)\Google\Chrome\Application\chrome.exe").VersionInfo.FileVer
...: sion ;"""
In [45]: with subprocess.Popen(
...: cmd,
...: stdout=subprocess.PIPE,
...: stderr=subprocess.DEVNULL,
...: stdin=subprocess.DEVNULL,
...: shell=True,
...: ) as stream:
...: print(stream.communicate())
(b'', None)I'm trying make a minimal example that will return the output from the powershell to python. |
|
Thank you for such detailed research! |
|
Coming back to this. Probably it is a good idea to add the following flags to the Then, I did some simple experiments using just When running in powershell, the following command works: When running this in cmd, I got this: you might already guessed from this that the problem has something to do with quoting. Reading from here I guessed that I should add the number of quotes in the subexpression. So, using four doublequotes ( |
|
Replacing single doublequotes with four doublequotes works, almost: Now the problem is that for some reason the version is read from the .. which does not work. It can be simplified even more: which just prints the command back to the console. The problem seems to be the braces ( So apparently there should be special handling of quotes when they are enclosed in braces. |
|
I have a feeling that it might be difficult trying to get command passed from cmd to Powershell in every situation as is. From this answer it looks like the |
|
Here is a powershell script I came up with that seems to do the job. First, the "all okay" case: When I run this, I will get: If there will be no match, it will print: If one of the and if everything fails, it will print: as it will find the first (and correct) version number in any of the cases. |
|
Actually, the values from the registry can also be read with powershell commands. This simplifies a possible implementation a bit. The following will print just Now it should be relatively straightforward to edit the |
|
@np-8 that's how windows_browser_apps_to_cmd works now. It stops on first result which is not $ERROR. So need just to correct list of commands, not algorithm ;) |
|
@aleksandr-kotlyar for some reason the In [1]: from webdriver_manager.utils import *
In [2]: get_browser_version_from_os("google-chrome")
[WDM] - Could not get version for google-chrome with the command: powershell "$ErrorActionPreference='silentlycontinue' ; (Get-Item -Path "$env:PROGRAMFILES\Google\Chrome\Application\chrome.exe").VersionInfo.FileVersion ; if (-not $? -or $? -match $error) { (Get-Item -Path "$env:PROGRAMFILES(x86)\Google\Chrome\AppVersionInfo.FileVersion } if (-not $? -or $? -match $error) { (Get-ItemProperty -Path Registry::HKCU\SOFTWARE\Google\Chrome\BLBeacon).version } if (-not $? -or $? -match $error) { (Get-ItemProperty -Path Registry::"HKLM\SOFTWARE\Wow6432Node\Microsoft\Windows\CurrentVersion\Uninstall\Google Chrome").version }"
[WDM] - Current google-chrome version is UNKNOWN
Out[2]: 'UNKNOWN'PS. There was actually no problem in the |
|
Same results on Windows Server 2016. As a result of this, I believe it is going and pulling the latest driver, even though it exists locally. This combined with xdist, I suspect, is making all the Jenkins runs fail. First this... Which leads to... self = <zipfile.ZipFile [closed]>
E zipfile.BadZipFile: File is not a zip file c:\python37\lib\zipfile.py:1289: BadZipFile On subsequent runs lead to this... selenium.common.exceptions.WebDriverException: Message: 'chromedriver.exe' executable may have wrong permissions. Please see https://sites.google.com/a/chromium.org/chromedriver/home subprocess.py - PermissionError: [WinError 32] The process cannot access the file because it is being used by another process Or it may be completely unrelated. I know the webdriver is capable of working, in a single instance, just not multiple. Roughly the same results with MSEdge. |
|
It does appear webdriver is looking in c:\Program Files, and not where google installs by default -- C:\Program Files (x86) Simple solution, just create a link to chrome between the two directories
Run command prompt as admin |
@geojitsu webdriver-manager is already looking in both places $env:PROGRAMFILES (x86)\Google\Chrome\Application\chrome.exe
$env:PROGRAMFILES\Google\Chrome\Application\chrome.exewhat version of webdriver-manager do you use? If it appears on the latest versions of webdriver-manager - please create new issue if it reproduces and link to this issue. Cause this issue is closed. |
Steps to reproduce:
I tested with the
webdriver_managerinstalled via pip (3.5.3) and the latest version from GitHub ( f7a0855 ).The text was updated successfully, but these errors were encountered: