I've been following https://github.com/kripken/emscripten/wiki/Using-Emscripten-on-Windows along up to now. When I run the test suite in tests/runner.py, some tests fail with a python error.
subprocess.Popen is throwing error 2 on files that clearly either exist or should be getting created mid-script (the latter being, ie, files in the C:/tmp directory, which I have created and also made certain is not read-only).
I have no idea what's going wrong or how I can fix this.
I set an environment variable EM_SAVE_DIR=1 as recommended on the github page.... Though it was recommended for debugging purposes, the fact of the matter is that it caused the files to remain both in memory and in the correct location, such that the error now only gets thrown on executables that I mistakenly didn't have anywhere on my PATH (cmake, sh, and make).
Though at some point later today, if I get all these tests worked out for runner.py, I may need to ask about making it work on visual studio, which throws error 6 on not finding a correct file in the tmp directory, and seems to be ignoring EM_SAVE_DIR=1....
Related
Background
Some time ago I seriously crashed my Windows computer while using PyCharm - I remember some errors about memory and then a hard crash with no blue screen - just black with some thin vertical lines and reboot to Windows installation / fixing screen. Since then, I had this problem, with no way I found online to fix this.
Edit : Apparently, this has nothing to do with the problem.
The problem
Whenever I open a project, or create a new one, an error appears with the Invalid Python SDK error message.
**Invalid Python SDK**
Cannot set up a Python SDKat Python 3.9 (%projectName%) (%projectPath%).The SDK seems to be invalid.
Also, this is what the work environment looks like the moment I close this message. In the Project window, the venv directory (and every directory under it) is marked as an Exclusion, and in the code, the print(f'Hi, {name}') function is marked as an unresolved reference error shown below. The program, however, executes flawlessly.
What's more, when I go to Python Interpreter settings at File -> Settings -> Project -> Python Interpreter there's a yellow bar on the bottom which says:
Non-zero exit code (4).
which after some time says:
Python packaging tools not found.
Upon installing, nothing changes, and I can't add packages from this screen (the '+' button is greyed out):
When I try to check Python interpreter paths, there are no paths shown, and I don't know what that means:
In short, all of the default Python functions like print are marked as errors, even though they work when executed. This makes coding extremely confusing, as I can't quickly distinguish between real errors and 'errors'.
The search for solution
Normally this would be a problem with interpreter set-up or path, but I've tried most of the methods proposed in other answers to similar questions. To name a few :
PyCharm shows unresolved references error for valid code
'Cannot setup a Python SDK' in PyCharm project using virtualenv after OS reinstallation
Why do I get an 'SDK seems invalid' error when setting up my Project Interpreter in PyCharm?
Invalid Python SDK Error while using python 3.4 on PyCharm
Invalid Python SDK when setting a venv
There were supposed to be links, but I don't have enough reputation on Stack Overflow to post them with the questions. These, however, can be easily looked up in Google, all of them are posted to Stack Overflow.
What I tried
I should mention that the first things I tried were removing and installing PyCharm, all user configurations and Python itself as well. I installed Python from the official site, and from the PyCharm application, both methods ended with the same result.
File -> Invalidate Caches... -> Invalidate and restart. Didn't work.
Checking file interpreter in Edit Configurations. Don't know what to make of it. The result:
Refreshing the interpreter paths. Even now, the paths yield no results.
Removing the interpreter and adding it again. No result.
Deleting the .idea folder. No result.
Deleting PyCharm user preferences under %homepath%/.PyCharm50. I don't have that folder though.
Switching interpreter back and forth. No result.
Creating a new interpreter in a different location. No result.
Marking project directory as root ProjectName -> Mark Directory as -> Sources Root and unmarking other directories as Excluded. No result.
Using no interpreter. Yeah, it doesn't mark non-errors as errors anymore. But the code doesn't work. That's not a solution for me.
Checking if venv/pyvenv.cfg has paths set correctly. These look fine to me.
Checking Windows environment variables - Path variable. It was in the user section, but wasn't in the system section. I added it, restarted but still no result.
Changing account name in Windows. My account name was 'username' and that's how my User folder is called `C:\Users\username', but I later connected it to Microsoft account and my user name is now User Name with a space and I can't really change it. My folder stayed the same. Not sure if I can fix it that way.
To the two last things I tried I should also add that I changed my Windows username from 'username' to 'user name' with a space, but that wasn't until recently.
I'm attaching the idea.log file for you to check. I replaced my real username with 'User Name' to highlight the existence of a space.
OK, that was a lucky one! I'm thus posting my comment as an answer:
The problem is caused by the non-ASCII characters in the path, and the solution is to remove them. As indicated by #TheLazyScripter this is a known issue.
I have encountered an issue where I have a file in Onedrive that I am not able to move. I have determined that, despite it saying that the action cant be completed because the folder or file is open in another program, it is not being used by another program.
I have come to this conclusion by using resource monitor to check which apps are using the folder. There are none besides file explorer.
Context as it's probably important:
The reason why this has happened is that i deleted a python venv in the file directory, resulting in my VS Code interpreter breaking. I have restored this venv file but the issue still has not been fixed. According to this post, it is the result of workspace links, however, I have never once used workspaces. The general idea that a file WITHIN that unmovable directory could be the reason though, and as a result, is possibly why this issue may be occuring. What does not make sense is why it is still happening even without VS Code running or any service connections being made as evident through resource monitor
Another thing to note is that the folder has not completely synced, however, based on past experience, this should not be an issue.
I successfully moved the file via terminal, however the issue still remains about why i could not simply move the file with file explorer.
I have a short script written in python that saves a simple .txt to the same directory the .py file is in. Problem is: It only does so on one of my two computers.
My code doesn't include a hard-coded path to write to. On my laptop, I can put the makemeanote.py in any folder and it will create the note right there. On my desktop pc, all the notes end up in System32. PATH is set exactly the same way on both machines, and both use Windows\py.exe as the executable.
Somewhat interestingly, I only get an admin-screen on the desktop pc, asking if I want to allow changes to my system by "Built: Release_master_v3.8.2_"etc., whereas on my laptop it simply runs and does its job.
No amount of un- and reinstalling has changed anything, even when I thought I had eradicated any trace of python on my hard drive. Both PCs use current Win10/64 installations.
What is happening there?
A clarification: It's not about fixing the bug, it's about understanding the inconsistent behaviour! I know I could just hard-code any directory, but that takes away the beauty, don't you think?
I finally found the cause of the weird behavior:
For some reason, the py.exe in my WINDOWS folder was set to always be run as admin. As soon as I unchecked that option, the User Account Control (UAC) check disappeared and my file finally behaved as I had expected.
During a presentation yesterday I had a colleague run one of my scripts on a fresh installation of Python 3.8.1. It was able to create and write to a csv file in his folder (proof that the csv library was working correctly), but everything else failed due to not being able to find the needed files. To try and isolate the problem and figure out why, we tried the below simple script, which also failed.
He had this test.py script in "D:/TEST", which also contained some folders and image files. Running this script printed nothing to the console. No empty list, no error message, no newline. Maybe the print() function was also not working, but I didn't get around to testing that.
import os
print(os.listdir())
This script works fine on my computer and my other colleagues computers (all Windows 10, similar hardware). I didn't have time to look into the issue more thoroughly and don't have access to his computer anymore. What could be the problem? What other things could I have him look into in order to fix this? In case this problem appears again during a future presentation, what steps could I take to figure out the cause of it?
My colleague uninstalled Python and reinstalled it. After doing this apparently the "python" command will no longer run his scripts, but using "py" instead will. Now that he is using "py" to run his scripts, it is working as expected.
I'm having the same problem on Windows and Linux. I launch any of various python 2.6 shells and run nose.py() to run my test suite. It works fine. However, the second time I run it, and every time thereafter I get exactly the same output, no matter how I change code or test files. My guess is that it's holding onto file references somehow, but even deleting the *.pyc files, I can never get the output of nose.run() to change until I restart the shell, or open another one, whereupon the problem starts again on the second run. I've tried both del nose and reload(nose) to no avail.
Solved* it with some outside help. I wouldn't consider this the proper solution, but by searching through sys.modules for all of my test_modules (which point to *.pyc files) and deling them, nose finally recognizes changes again. I'll have to delete them before each nose.run() call. These must be in-memory versions of the pyc files, as simply deleting them out in the shell wasn't doing it. Good enough for now.
Edit:
*Apparently I didn't entirely solve it. It does seem to work for a bit, and then all of a sudden it won't anymore, and I have to restart my shell. Now I'm even more confused.