%matplotlib QT5 doesn't work

[gonghenghai]

Environment data

• VS Code version: 1.62.3
• Jupyter Extension version (available under the Extensions sidebar): after v2021.10.1001409655
• Python Extension version (available under the Extensions sidebar): v2021.11.1422169775
• OS (Windows | Mac | Linux distro) and version: windows
• Python and/or Anaconda version: 3.9.5
• Type of virtual environment used (N/A | venv | virtualenv | conda | ...): venv
• Jupyter server running: Local

Expected behaviour

Execute the cells

Actual behaviour

Just waiting

[DonJayamanne]

Thanks for filing this issue, & I"m sorry yo'ure running into this. We've been working on some perf related fixes and the like in our current development branch.

• Do you experience similar perf issues if you do not run the first cell %matplotlib
• I.e. can you confirm that when using notebooks without that cell magic then cells run fast.

Please could you

• Install VS Code insiders from here https://code.visualstudio.com/insiders/
• Install the Jupyter extension (it will automatically install the latest development version, starts with 2021.11.
• Enable verbose logging
  • Go into your settings file (use the command 'Preferences: Open Settings UI`)
  • Search for Jupyter logging
  • Change the logging level for Jupyter extension to verbose

See here

* Reload VS Code

• Replicate the above problem
• Go into your Jupyter output panel (use the command Jupyter: View Output)
  • Copy everything from there, and please paste the contents into this issue

[faultdiagnosistoolbox]

I will try to contribute with the information, I asked the original question at https://stackoverflow.com/questions/70056478/vscode-python-interactive-window-freezes-when-using-matplotlib

First I can confirm that when not running %matplotlib (or even running %matplotlib inline) all works as expected. It is only when I have the external window that I see the problem.

I've installed the code-insider edition and the python (including the jupyter) extension version Jupyter v2021.11.1001498571.

Then I

1. Activate verbose jupyter logging
2. Restart vscode insider
3. Create a jupyter interactive window
4. Run %matplotlib
5. run 1+1 (finishes in 0.2s), run 1+2 (finishes in 1m 39s)

I have attached the debug output to this message.

Thanks for investigating the issue, Erik

debug_jupyter.txt

[DonJayamanne]

@faultdiagnosistoolbox Looks like you have failed to attach the logs correctly. Please could you re-attach them FYI - I'm unable to replicate this same issue.

• Also, please do let me know what version of Matplot lib you have installed.
• Perhaps a pip list would help (so i can get a similar env setup locally), if using conda, then please export the env to a yml file)

[DonJayamanne]

It is only when I have the external window that I see the problem.

Oh, ok, I have not tried this.

[DonJayamanne]

Tried, still unable to repro. I must be missing something

This is the code I"m using

%matplotlib
import matplotlib.pyplot as plt
import numpy as np

x = np.linspace(0, 10, 100)

plt.plot(x, np.sin(x))
plt.plot(x, np.cos(x))

plt.show()

& this is the output I get in VS Code Using matplotlib backend: <object object at 0x00000129A3DDB4E0> Basically the plot doesn't open up, i get an empty Window. However I'm able to run subsequent cells. Seems to be the case in VS Code stable as well.

When running the same code in Jupyter i the the output using matplotlib backend: TkAgg.

@faultdiagnosistoolbox @gonghenghai For now here's the work around:

• %matplotlib doesn't work in VS Code, you need to use %matplotlib widget
• Please try that, (FYI you'll need to install ipympl package.

More info here https://matplotlib.org/stable/users/explain/interactive.html?highlight=vscode#jupyter-notebooks-jupyterlab & https://matplotlib.org/stable/users/explain/backends.html?highlight=ipympl

[faultdiagnosistoolbox]

OK, strange. I have tried on 4 machines. It fails identically on two Macs (iMac, Macbook Pro) and my Ubuntu 20.04 Linux laptop. On my office Linux, also Ubuntu 20.04, workstation I can't reproduce.

I thought I attached the debug log. I have copy-pasted it below, hope that is the correct way.

Erik

Python Extension Verison: 2021.11.1422169775.
Info 18:53:13: Experimentation service retrieved: [object Object]
User belongs to experiment group 'jupyterTestcf'
User belongs to experiment group 'jupyterEnhancedDataViewer'
Info 18:53:13: ZMQ install verified.
Verbose 18:53:13: Get Custom Env Variables, Class name = m (started execution), Arg 1: undefined
Verbose 18:53:13: Get Custom Env Variables, Class name = m (started execution), Arg 1: "~/tmp/vsc_test"
Verbose 18:53:13: Get Active Interpreter, Class name = P (started execution), Arg 1: "~/tmp/vsc_test"
Verbose 18:53:13: Get Custom Env Variables, Class name = m, completed in 2ms, has a truthy return value, Arg 1: undefined
Error 18:53:13: Failed to parse env file /home/erifr93/tmp/vsc_test/.env EntryNotFound: Unable to read file '/home/erifr93/tmp/vsc_test/.env' (Error: Unable to resolve nonexistent file '/home/erifr93/tmp/vsc_test/.env')
    at Ei.doReadFileStream (vscode-file://vscode-app/usr/share/code-insiders/resources/app/out/vs/workbench/workbench.desktop.main.js:576:25010)
    at async Ei.doReadFile (vscode-file://vscode-app/usr/share/code-insiders/resources/app/out/vs/workbench/workbench.desktop.main.js:576:24026)
Verbose 18:53:13: Get Custom Env Variables, Class name = m, completed in 32ms, has a truthy return value, Arg 1: "~/tmp/vsc_test"
Verbose 18:53:13: Get Active Interpreter, Class name = P (started execution), Arg 1: undefined
Verbose 18:53:13: Get Interpreters, Class name = P (started execution), 
Info 18:53:13: Experiment status for python is {"enabled":true,"optInto":[],"optOutFrom":[]}
Verbose 18:53:14: Get Active Interpreter, Class name = P, completed in 245ms, has a truthy return value, Arg 1: undefined
Verbose 18:53:14: Create activated Env, Class name = k (started execution), Arg 1: ""
Verbose 18:53:14: Getting activated env variables, Class name = D (started execution), Arg 1: undefined, Arg 2: "~/sw/pyenv/39/bin/python", Arg 3: true
Verbose 18:53:14: Getting activated env variables ourselves, Class name = D (started execution), Arg 1: undefined, Arg 2: "~/sw/pyenv/39/bin/python"
Verbose 18:53:14: Getting env activation commands, Class name = D (started execution), Arg 1: undefined, Arg 2: "~/sw/pyenv/39/bin/python"
Verbose 18:53:14: Create ProcessService, Class name = p (started execution), Arg 1: undefined
Verbose 18:53:14: Get Custom Env Variables, Class name = m (started execution), Arg 1: undefined
Verbose 18:53:14: Cached data exists getEnvironmentVariables, <No Resource>
Verbose 18:53:14: Getting activated env variables from Python, Class name = D (started execution), Arg 1: undefined, Arg 2: "~/sw/pyenv/39/bin/python"
Verbose 18:53:14: Get Active Interpreter, Class name = P, completed in 319ms, has a truthy return value, Arg 1: "~/tmp/vsc_test"
Verbose 18:53:14: Get Custom Env Variables, Class name = m, completed in 0ms, has a truthy return value, Arg 1: undefined
Verbose 18:53:14: Create ProcessService, Class name = p, completed in 1ms, has a truthy return value, Arg 1: undefined
Verbose 18:53:14: Getting env activation commands, Class name = D, completed in 1ms, has a falsy return value, Arg 1: undefined, Arg 2: "~/sw/pyenv/39/bin/python"
Verbose 18:53:14: Getting activated env variables ourselves, Class name = D, completed in 3ms, has a falsy return value, Arg 1: undefined, Arg 2: "~/sw/pyenv/39/bin/python"
Verbose 18:53:14: Got env vars ourselves faster, but empty ~/sw/pyenv/39/bin/python
Verbose 18:53:14: Got env vars ourselves ~/sw/pyenv/39/bin/python in 3ms
Verbose 18:53:14: Getting activated env variables from Python, Class name = D, completed in 113ms, has a truthy return value, Arg 1: undefined, Arg 2: "~/sw/pyenv/39/bin/python"
Verbose 18:53:14: Got env vars with python ~/sw/pyenv/39/bin/python in 114ms
Verbose 18:53:14: Getting activated env variables, Class name = D, completed in 114ms, has a truthy return value, Arg 1: undefined, Arg 2: "~/sw/pyenv/39/bin/python", Arg 3: true
Verbose 18:53:14: Create activated Env, Class name = k, completed in 115ms, has a truthy return value, Arg 1: ""
Info 18:53:14: Process Execution: > ~/sw/pyenv/39/bin/python -m pip list
> ~/sw/pyenv/39/bin/python -m pip list
Verbose 18:53:14: Get Interpreters, Class name = P, completed in 477ms, has a truthy return value, 
Info 18:53:36: Creating controller for interactive with interpreter ~/sw/pyenv/39/bin/python
Info 18:53:36: Starting interactive window with controller ID ms-toolsai.jupyter/.jvsc74a57bd0b2d30df487f2f295440af491f15be1736706da7a0a50c540ff6414a45436c63c./home/erifr93/sw/pyenv/39/python./home/erifr93/sw/pyenv/39/python.-m#ipykernel_launcher (Interactive)
Verbose 18:53:36: Get Kernelspec root path, Class name = m (started execution), Arg 1: {"_isCancelled":false,"_emitter":null}
Verbose 18:53:36: Get Jupyter Paths, Class name = m (started execution), Arg 1: {"_isCancelled":false,"_emitter":null}
Verbose 18:53:36: Get Custom Env Variables, Class name = m (started execution), 
Verbose 18:53:36: Cached data exists getEnvironmentVariables, <No Resource>
Verbose 18:53:36: Get Interpreters, Class name = P (started execution), Arg 1: undefined
Verbose 18:53:36: Get Interpreters, Class name = P (started execution), Arg 1: undefined
Verbose 18:53:36: Get Active Interpreter, Class name = P (started execution), Arg 1: "Interactive-1.interactive"
Info 18:53:36: Setting setActiveController for vscode-interactive:Interactive-1.interactive
Verbose 18:53:36: Intiailize notebook communications for editor vscode-interactive:Interactive-1.interactive
Verbose 18:53:36: Resolving notebook UI Comms (resolve) for vscode-interactive:Interactive-1.interactive
Verbose 18:53:36: initialize CommonMessageCoordinator
Verbose 18:53:37: Get Custom Env Variables, Class name = m, completed in 28ms, has a truthy return value, 
Verbose 18:53:37: Get Active Interpreter, Class name = P, completed in 25ms, has a truthy return value, Arg 1: "Interactive-1.interactive"
Verbose 18:53:37: IPyWidgetMessageDispatcher.initialize
Verbose 18:53:37: Get Jupyter Paths, Class name = m, completed in 32ms, has a truthy return value, Arg 1: {"_isCancelled":false,"_emitter":null}, Return Value: []
Verbose 18:53:37: Get Kernelspec root path, Class name = m, completed in 33ms, has a truthy return value, Arg 1: {"_isCancelled":false,"_emitter":null}, Return Value: ["/usr/share/jupyter/kernels","/usr/local/share/jupyter/kernels","/home/erifr93/.local/share/jupyter/kernels"]
Verbose 18:53:37: Created and initailized CommonMessageCoordinator
Verbose 18:53:37: Attach Coordinator for vscode-interactive:Interactive-1.interactive
Verbose 18:53:37: Get Interpreters, Class name = P, completed in 33ms, has a truthy return value, Arg 1: undefined
Verbose 18:53:37: Getting Jupyter KernelSpec Root Path, Class name = m (started execution), 
Verbose 18:53:37: Get Active Interpreter, Class name = P (started execution), Arg 1: undefined
Verbose 18:53:37: Get Kernelspec root path, Class name = m (started execution), 
Verbose 18:53:37: Get Jupyter Paths, Class name = m (started execution), Arg 1: undefined
Verbose 18:53:37: Get Kernelspec root path, Class name = m (started execution), Arg 1: {"_isCancelled":false,"_emitter":null}
Verbose 18:53:37: Get Jupyter Paths, Class name = m (started execution), Arg 1: {"_isCancelled":false,"_emitter":null}
Verbose 18:53:37: Get Interpreters, Class name = P, completed in 36ms, has a truthy return value, Arg 1: undefined
Verbose 18:53:37: Getting Jupyter KernelSpec Root Path, Class name = m (started execution), 
Verbose 18:53:37: Get Active Interpreter, Class name = P (started execution), Arg 1: undefined
Verbose 18:53:37: Get Kernelspec root path, Class name = m (started execution), 
Verbose 18:53:37: Get Jupyter Paths, Class name = m (started execution), Arg 1: undefined
Verbose 18:53:37: Getting Jupyter KernelSpec Root Path, Class name = m, completed in 4ms, has a truthy return value, , Return Value: "/home/erifr93/.local/share/jupyter/kernels"
Verbose 18:53:37: Get Active Interpreter, Class name = P, completed in 3ms, has a truthy return value, Arg 1: undefined
Verbose 18:53:37: Getting Jupyter KernelSpec Root Path, Class name = m, completed in 2ms, has a truthy return value, , Return Value: "/home/erifr93/.local/share/jupyter/kernels"
Verbose 18:53:37: Get Active Interpreter, Class name = P, completed in 2ms, has a truthy return value, Arg 1: undefined
Verbose 18:53:37: Get Jupyter Paths, Class name = m, completed in 4ms, has a truthy return value, Arg 1: undefined, Return Value: []
Verbose 18:53:37: Get Jupyter Paths, Class name = m, completed in 4ms, has a truthy return value, Arg 1: {"_isCancelled":false,"_emitter":null}, Return Value: []
Verbose 18:53:37: Get Jupyter Paths, Class name = m, completed in 2ms, has a truthy return value, Arg 1: undefined, Return Value: []
Verbose 18:53:37: Get Kernelspec root path, Class name = m, completed in 5ms, has a truthy return value, , Return Value: ["/usr/share/jupyter/kernels","/usr/local/share/jupyter/kernels","/home/erifr93/.local/share/jupyter/kernels"]
Verbose 18:53:37: Get Kernelspec root path, Class name = m, completed in 4ms, has a truthy return value, Arg 1: {"_isCancelled":false,"_emitter":null}, Return Value: ["/usr/share/jupyter/kernels","/usr/local/share/jupyter/kernels","/home/erifr93/.local/share/jupyter/kernels"]
Verbose 18:53:37: Get Kernelspec root path, Class name = m, completed in 3ms, has a truthy return value, , Return Value: ["/usr/share/jupyter/kernels","/usr/local/share/jupyter/kernels","/home/erifr93/.local/share/jupyter/kernels"]
Info 18:53:37: Starting Notebook in kernel.ts id = .jvsc74a57bd0b2d30df487f2f295440af491f15be1736706da7a0a50c540ff6414a45436c63c./home/erifr93/sw/pyenv/39/python./home/erifr93/sw/pyenv/39/python.-m#ipykernel_launcher
Info 18:53:37: Creating raw notebook for vscode-interactive:Interactive-1.interactive
Info 18:53:37: Getting preferred kernel for vscode-interactive:Interactive-1.interactive
Info 18:53:37: Computing working directory vscode-interactive:Interactive-1.interactive
Verbose 18:53:37: Connecting to raw session for vscode-interactive:Interactive-1.interactive with connection .jvsc74a57bd0b2d30df487f2f295440af491f15be1736706da7a0a50c540ff6414a45436c63c./home/erifr93/sw/pyenv/39/python./home/erifr93/sw/pyenv/39/python.-m#ipykernel_launcher
Info 18:53:37: Starting raw kernel 39 (Python 3.9.5)
Info 18:53:37: installMissingDependencies ~/sw/pyenv/39/bin/python
Info 18:53:37: IPykernel found previously in this enviornment ~/sw/pyenv/39/bin/python
Verbose 18:53:37: Install Missing Dependencies, Class name = k, completed in 2ms, has a falsy return value, Return Value: undefined
Info 18:53:37: Kernel launching with ports 9010,9011,9012,9013,9014. Start port is 9000
Verbose 18:53:37: Launching kernel in kernelProcess.ts, Class name = E (started execution), Arg 1: "~/tmp/vsc_test", Arg 2: ""
Verbose 18:53:37: Launching kernel daemon, Class name = m (started execution), Arg 1: undefined, Arg 2: "~/tmp/vsc_test", Arg 3: "python395jvsc74a57bd0b2d30df487f2f295440af491f15be1736706da7a0a50c540ff6414a45436c63c", Arg 4: "~/sw/pyenv/39/bin/python"
Verbose 18:53:37: Get kernel daemon, Class name = y (started execution), Arg 1: undefined, Arg 2: undefined, Arg 3: "~/sw/pyenv/39/bin/python"
Verbose 18:53:37: Create daemon, Class name = k (started execution), Arg 1: {"daemonModule":"vscode_datascience_helpers.kernel_launcher_daemon","interpreter":{"sysPrefix":"/home/erifr93/sw/pyenv/39","envType":"Venv","envName":"39","envPath":"/home/erifr93/sw/pyenv/39","path":"/home/erifr93/sw/pyenv/39/bin/python","architecture":3,"sysVersion":"3.9.5 (default, May 29 2021, 14:32:59) \n[GCC 9.3.0]","version":{"raw":"3.9.5","major":3,"minor":9,"patch":5,"build":[],"prerelease":["final","0"]},"displayName":"Python 3.9.5 64-bit ('39': venv)"},"dedicated":true}
Verbose 18:53:37: Create activated Env, Class name = k (started execution), Arg 1: ""
Verbose 18:53:37: Getting activated env variables, Class name = D (started execution), Arg 1: undefined, Arg 2: "~/sw/pyenv/39/bin/python", Arg 3: true
Verbose 18:53:37: Getting activated env variables ourselves, Class name = D (started execution), Arg 1: undefined, Arg 2: "~/sw/pyenv/39/bin/python"
Verbose 18:53:37: Getting activated env variables from Python, Class name = D (started execution), Arg 1: undefined, Arg 2: "~/sw/pyenv/39/bin/python"
Verbose 18:53:37: Getting activated env variables, Class name = D (started execution), Arg 1: undefined, Arg 2: "~/sw/pyenv/39/bin/python", Arg 3: true
Verbose 18:53:37: Getting activated env variables ourselves, Class name = D (started execution), Arg 1: undefined, Arg 2: "~/sw/pyenv/39/bin/python"
Verbose 18:53:37: Getting activated env variables from Python, Class name = D (started execution), Arg 1: undefined, Arg 2: "~/sw/pyenv/39/bin/python"
Info 18:53:37: Launching kernel daemon for Python 3.9.5 64-bit ('39': venv) # ~/sw/pyenv/39/bin/python
Info 18:53:37: No custom variables for Kernel as interpreter is not conda, but is Venv
Verbose 18:53:37: Getting activated env variables ourselves, Class name = D, completed in 8ms, has a falsy return value, Arg 1: undefined, Arg 2: "~/sw/pyenv/39/bin/python"
Verbose 18:53:37: Getting activated env variables ourselves, Class name = D, completed in 7ms, has a falsy return value, Arg 1: undefined, Arg 2: "~/sw/pyenv/39/bin/python"
Verbose 18:53:37: Got env vars ourselves faster, but empty ~/sw/pyenv/39/bin/python
Verbose 18:53:37: Got env vars ourselves faster, but empty ~/sw/pyenv/39/bin/python
Verbose 18:53:37: Got env vars ourselves ~/sw/pyenv/39/bin/python in 10ms
Verbose 18:53:37: Getting activated env variables from Python, Class name = D, completed in 10ms, has a truthy return value, Arg 1: undefined, Arg 2: "~/sw/pyenv/39/bin/python"
Verbose 18:53:37: Got env vars ourselves ~/sw/pyenv/39/bin/python in 9ms
Verbose 18:53:37: Getting activated env variables from Python, Class name = D, completed in 9ms, has a truthy return value, Arg 1: undefined, Arg 2: "~/sw/pyenv/39/bin/python"
Verbose 18:53:37: Got env vars with python ~/sw/pyenv/39/bin/python in 10ms
Verbose 18:53:37: Got env vars with python ~/sw/pyenv/39/bin/python in 9ms
Verbose 18:53:37: Getting activated env variables, Class name = D, completed in 10ms, has a truthy return value, Arg 1: undefined, Arg 2: "~/sw/pyenv/39/bin/python", Arg 3: true
Verbose 18:53:37: Getting activated env variables, Class name = D, completed in 11ms, has a truthy return value, Arg 1: undefined, Arg 2: "~/sw/pyenv/39/bin/python", Arg 3: true
Verbose 18:53:37: Create activated Env, Class name = k, completed in 12ms, has a truthy return value, Arg 1: ""
Info 18:53:37: Creating daemon process for ~/sw/pyenv/39/bin/python with env variables count 86
Info 18:53:37: Process Execution: > ~/sw/pyenv/39/bin/python -m vscode_datascience_helpers.daemon --daemon-module=vscode_datascience_helpers.kernel_launcher_daemon -v
> ~/sw/pyenv/39/bin/python -m vscode_datascience_helpers.daemon --daemon-module=vscode_datascience_helpers.kernel_launcher_daemon -v
Verbose 18:53:37: Loading kernelspec from ~/sw/pyenv/39/share/jupyter/kernels/python3/kernel.json for ~/sw/pyenv/39/bin/python
Verbose 18:53:37: Daemon is alive
Verbose 18:53:37: Create daemon, Class name = k, completed in 158ms, has a truthy return value, Arg 1: {"daemonModule":"vscode_datascience_helpers.kernel_launcher_daemon","interpreter":{"sysPrefix":"/home/erifr93/sw/pyenv/39","envType":"Venv","envName":"39","envPath":"/home/erifr93/sw/pyenv/39","path":"/home/erifr93/sw/pyenv/39/bin/python","architecture":3,"sysVersion":"3.9.5 (default, May 29 2021, 14:32:59) \n[GCC 9.3.0]","version":{"raw":"3.9.5","major":3,"minor":9,"patch":5,"build":[],"prerelease":["final","0"]},"displayName":"Python 3.9.5 64-bit ('39': venv)"},"dedicated":true}
Verbose 18:53:37: Get kernel daemon, Class name = y, completed in 160ms, has a truthy return value, Arg 1: undefined, Arg 2: undefined, Arg 3: "~/sw/pyenv/39/bin/python"
Info 18:53:37: Starting kernel from scratch
Verbose 18:53:37: Python Daemon (pid: 13470): Execute rpc method exec_module in DS Daemon
Verbose 18:53:37: Python Daemon (pid: 13470): Execute rpc method exec_module from /home/erifr93/sw/pyenv/39/bin/python
Verbose 18:53:37: Python Daemon (pid: 13470): Exec module in DS Kernel Launcher Daemon ipykernel_launcher with args ['--ip=127.0.0.1', '--stdin=9013', '--control=9011', '--hb=9010', '--Session.signature_scheme="hmac-sha256"', '--Session.key=b"cf49b15b-5038-4bb3-be78-e4e0cf7f6ae3"', '--shell=9012', '--transport="tcp"', '--iopub=9014', '--f=/tmp/tmp-13273IS2sAD1rc5ry.json']
Verbose 18:53:37: Python Daemon (pid: 13470): Exec in DS Kernel Launcher Daemon (observable) ipykernel_launcher with args ['--ip=127.0.0.1', '--stdin=9013', '--control=9011', '--hb=9010', '--Session.signature_scheme="hmac-sha256"', '--Session.key=b"cf49b15b-5038-4bb3-be78-e4e0cf7f6ae3"', '--shell=9012', '--transport="tcp"', '--iopub=9014', '--f=/tmp/tmp-13273IS2sAD1rc5ry.json']
Verbose 18:53:37: Python Daemon (pid: 13470): Exec in DS Kernel Launcher Daemon (observable) ['/home/erifr93/sw/pyenv/39/bin/python', '-m', 'ipykernel_launcher', '--ip=127.0.0.1', '--stdin=9013', '--control=9011', '--hb=9010', '--Session.signature_scheme="hmac-sha256"', '--Session.key=b"cf49b15b-5038-4bb3-be78-e4e0cf7f6ae3"', '--shell=9012', '--transport="tcp"', '--iopub=9014', '--f=/tmp/tmp-13273IS2sAD1rc5ry.json']
Verbose 18:53:37: Python Daemon (pid: 13470): Exec in DS Kernel Launcher Daemon (observable)
Verbose 18:53:37: Python Daemon (pid: 13470): Kernel launched, with PID 13472
Verbose 18:53:37: Python Daemon (pid: 13470): Waiting for Kernel to die 13472
Verbose 18:53:37: Launching kernel daemon, Class name = m, completed in 171ms, has a truthy return value, Arg 1: undefined, Arg 2: "~/tmp/vsc_test", Arg 3: "python395jvsc74a57bd0b2d30df487f2f295440af491f15be1736706da7a0a50c540ff6414a45436c63c", Arg 4: "~/sw/pyenv/39/bin/python", Return Value: <Return value cannot be serialized for logging>
Verbose 18:53:37: Launching kernel in kernelProcess.ts, Class name = E, completed in 173ms, has a truthy return value, Arg 1: "~/tmp/vsc_test", Arg 2: ""
Verbose 18:53:37: Hiding default kernel spec Python 3 (ipykernel), python
Verbose 18:53:37: Hiding default kernel spec Python 3 (ipykernel), python
Info 18:53:37: Registering dummy command feature
Verbose 18:53:37: Registering commtarget jupyter.widget
Verbose 18:53:37: IPyWidgetMessageDispatcher.initialize
Verbose 18:53:37: KernelProcess output: Content-Length: 361
Content-Type: application/vscode-jsonrpc; charset=utf8

{"jsonrpc": "2.0", "method": "log", "params": {"level": "INFO", "msg": "write to stderr: /home/erifr93/sw/pyenv/39/lib/python3.9/site-packages/traitlets/traitlets.py:2202: FutureWarning: Supporting extra quotes around strings is deprecated in traitlets 5.0. You can use 'hmac-sha256' instead of '\"hmac-sha256\"' if you require traitlets >=5.\n", "pid": 13470}}
Verbose 18:53:37: Python Daemon (pid: 13470): write to stderr: /home/erifr93/sw/pyenv/39/lib/python3.9/site-packages/traitlets/traitlets.py:2202: FutureWarning: Supporting extra quotes around strings is deprecated in traitlets 5.0. You can use 'hmac-sha256' instead of '"hmac-sha256"' if you require traitlets >=5.

Verbose 18:53:37: KernelProcess output: Content-Length: 336
Content-Type: application/vscode-jsonrpc; charset=utf8

{"jsonrpc": "2.0", "method": "output", "params": {"source": "stderr", "out": "/home/erifr93/sw/pyenv/39/lib/python3.9/site-packages/traitlets/traitlets.py:2202: FutureWarning: Supporting extra quotes around strings is deprecated in traitlets 5.0. You can use 'hmac-sha256' instead of '\"hmac-sha256\"' if you require traitlets >=5.\n"}}Content-Length: 115
Content-Type: application/vscode-jsonrpc; charset=utf8

{"jsonrpc": "2.0", "method": "log", "params": {"level": "INFO", "msg": "write to stderr:   warn(\n", "pid": 13470}}Content-Length: 90
Content-Type: application/vscode-jsonrpc; charset=utf8

{"jsonrpc": "2.0", "method": "output", "params": {"source": "stderr", "out": "  warn(\n"}}Content-Length: 373
Content-Type: application/vscode-jsonrpc; charset=utf8

{"jsonrpc": "2.0", "method": "log", "params": {"level": "INFO", "msg": "write to stderr: /home/erifr93/sw/pyenv/39/lib/python3.9/site-packages/traitlets/traitlets.py:2157: FutureWarning: Supporting extra quotes around Bytes is deprecated in traitlets 5.0. Use 'cf49b15b-5038-4bb3-be78-e4e0cf7f6ae3' instead of 'b\"cf49b15b-5038-4bb3-be78-e4e0cf7f6ae3\"'.\n", "pid": 13470}}Content-Length: 348
Content-Type: application/vscode-jsonrpc; charset=utf8

{"jsonrpc": "2.0", "method": "output", "params": {"source": "stderr", "out": "/home/erifr93/sw/pyenv/39/lib/python3.9/site-packages/traitlets/traitlets.py:2157: FutureWarning: Supporting extra quotes around Bytes is deprecated in traitlets 5.0. Use 'cf49b15b-5038-4bb3-be78-e4e0cf7f6ae3' instead of 'b\"cf49b15b-5038-4bb3-be78-e4e0cf7f6ae3\"'.\n"}}Content-Length: 115
Content-Type: application/vscode-jsonrpc; charset=utf8

{"jsonrpc": "2.0", "method": "log", "params": {"level": "INFO", "msg": "write to stderr:   warn(\n", "pid": 13470}}Content-Length: 90
Content-Type: application/vscode-jsonrpc; charset=utf8

{"jsonrpc": "2.0", "method": "output", "params": {"source": "stderr", "out": "  warn(\n"}}
Warn 18:53:37: StdErr from Kernel Process /home/erifr93/sw/pyenv/39/lib/python3.9/site-packages/traitlets/traitlets.py:2202: FutureWarning: Supporting extra quotes around strings is deprecated in traitlets 5.0. You can use 'hmac-sha256' instead of '"hmac-sha256"' if you require traitlets >=5.

Verbose 18:53:37: Python Daemon (pid: 13470): write to stderr:   warn(

Warn 18:53:37: StdErr from Kernel Process   warn(

Verbose 18:53:37: Python Daemon (pid: 13470): write to stderr: /home/erifr93/sw/pyenv/39/lib/python3.9/site-packages/traitlets/traitlets.py:2157: FutureWarning: Supporting extra quotes around Bytes is deprecated in traitlets 5.0. Use 'cf49b15b-5038-4bb3-be78-e4e0cf7f6ae3' instead of 'b"cf49b15b-5038-4bb3-be78-e4e0cf7f6ae3"'.

Warn 18:53:38: StdErr from Kernel Process /home/erifr93/sw/pyenv/39/lib/python3.9/site-packages/traitlets/traitlets.py:2157: FutureWarning: Supporting extra quotes around Bytes is deprecated in traitlets 5.0. Use 'cf49b15b-5038-4bb3-be78-e4e0cf7f6ae3' instead of 'b"cf49b15b-5038-4bb3-be78-e4e0cf7f6ae3"'.

Verbose 18:53:38: Python Daemon (pid: 13470): write to stderr:   warn(

Warn 18:53:38: StdErr from Kernel Process   warn(

Info 18:53:38: Raw session started and connected
Started kernel 39 (Python 3.9.5)
Info 18:53:38: Finished connecting 58bd132f-89c9-4144-b179-138a76754e20
Info 18:53:38: UpdateWorkingDirectoryAndPath in Kernel
Info 18:53:38: Executing (status idle) silently Code = %config Completer.use_jedi = False
Info 18:53:38: Executing silently Code (completed) = %config Completer.use_jedi = False
Info 18:53:38: Executing (status idle) silently Code = %config InlineBackend.figure_formats = ['png']
Info 18:53:38: Executing silently Code (completed) = %config InlineBackend.figure_formats = ['png']
Info 18:53:38: Executing (status idle) silently Code = import sys\nprint(sys.executable)
Info 18:53:38: Executing silently Code (completed) = import sys\nprint(sys.executable)
Info 18:53:38: Waiting for idle on (kernel): 73d9a265-6627-4fbf-9df6-f3d2c45070be -> idle
Info 18:53:38: Finished waiting for idle on (kernel): 73d9a265-6627-4fbf-9df6-f3d2c45070be -> idle
Verbose 18:53:38: IPyWidgetMessageDispatcher.initialize
Info 18:53:38: IPyWidgetScriptSource.initialize
Verbose 18:53:38: IPyWidgetMessageDispatcher.initialize
Verbose 18:53:38: IPyWidgetMessageDispatcher.initialize
Info 18:53:41: Jupyter completions for % at pos 0:1 with trigger: %
   %%!,%%HTML,%%SVG,%alias,%alias_magic,%autoawait,%autocall,%automagic,%autosave,%%bash,%bookmark,%%capture,%cat,%cd,%clear,%colors,%conda,%config,%connect_info,%cp,%debug,%%debug,%dhist,%dirs,%doctest_mode,%ed,%edit,%env,%%file,%gui,%hist,%history,%%html,%%javascript,%%js,%killbgscripts,%%latex,%ldir,%less,%lf,%lk,%ll,%load,%load_ext,%loadpy,%logoff,%logon,%logstart,%logstate,%logstop,%ls,%lsmagic,%lx,%macro,%magic,%man,%%markdown,%matplotlib,%mkdir,%more,%mv,%notebook,%page,%pastebin,%pdb,%pdef,%pdoc,%%perl,%pfile,%pinfo,%pinfo2,%pip,%popd,%pprint,%precision,%prun,%%prun,%psearch,%psource,%pushd,%pwd,%pycat,%pylab,%%pypy,%%python,%%python2,%%python3,%qtconsole,%quickref,%recall,%rehashx,%reload_ext,%rep,%rerun,%reset,%reset_selective,%rm,%rmdir,%%ruby,%run,%save,%sc,%%script,%set_env,%%sh,%store,%%svg,%sx,%%sx,%system,%%system,%tb,%time,%%time,%timeit,%%timeit,%unalias,%unload_ext,%who,%who_ls,%whos,%%writefile,%xdel,%xmode
Info 18:53:41: Jupyter results for % at pos 0:1 with trigger: %
   %%!,%%HTML,%%SVG,%alias,%alias_magic,%autoawait,%autocall,%automagic,%autosave,%%bash,%bookmark,%%capture,%cat,%cd,%clear,%colors,%conda,%config,%connect_info,%cp,%debug,%%debug,%dhist,%dirs,%doctest_mode,%ed,%edit,%env,%%file,%gui,%hist,%history,%%html,%%javascript,%%js,%killbgscripts,%%latex,%ldir,%less,%lf,%lk,%ll,%load,%load_ext,%loadpy,%logoff,%logon,%logstart,%logstate,%logstop,%ls,%lsmagic,%lx,%macro,%magic,%man,%%markdown,%matplotlib,%mkdir,%more,%mv,%notebook,%page,%pastebin,%pdb,%pdef,%pdoc,%%perl,%pfile,%pinfo,%pinfo2,%pip,%popd,%pprint,%precision,%prun,%%prun,%psearch,%psource,%pushd,%pwd,%pycat,%pylab,%%pypy,%%python,%%python2,%%python3,%qtconsole,%quickref,%recall,%rehashx,%reload_ext,%rep,%rerun,%reset,%reset_selective,%rm,%rmdir,%%ruby,%run,%save,%sc,%%script,%set_env,%%sh,%store,%%svg,%sx,%%sx,%system,%%system,%tb,%time,%%time,%timeit,%%timeit,%unalias,%unload_ext,%who,%who_ls,%whos,%%writefile,%xdel,%xmode
Info 18:53:43: Execute Cells request 1
Info 18:53:43: Execute Cell 1 vscode-interactive:Interactive-1.interactive
Info 18:53:43: Cell 1 executed with state Success
Info 18:53:45: Execute Cells request 2
Info 18:53:45: Execute Cell 2 vscode-interactive:Interactive-1.interactive
Info 18:53:45: Cell 2 executed with state Success
Info 18:53:48: Execute Cells request 3
Info 18:53:48: Execute Cell 3 vscode-interactive:Interactive-1.interactive
Info 18:55:27: Loading webview. View is notset
Info 18:55:27: Loading web view...
Info 18:55:27: Webview panel created.
Info 18:55:27: Web view react rendered
Info 18:55:27: Executing (status idle) silently Code = # Query Jupyter server for the info about a dataframe\nimport json as _VSCODE_json\nimport builtins as
Info 18:55:27: Executing (status idle) silently Code = # Query Jupyter server for the info about a dataframe\nimport json as _VSCODE_json\nimport builtins as
Info 18:55:27: Cell 3 executed with state Success
Info 18:55:27: Executing silently Code (completed) = # Query Jupyter server for the info about a dataframe\nimport json as _VSCODE_json\nimport builtins as
Info 18:55:27: Executing (status idle) silently Code = _rwho_ls = %who_ls\nprint(_VSCODE_getVariableTypes(_rwho_ls))
Info 18:55:27: Executing silently Code (completed) = # Query Jupyter server for the info about a dataframe\nimport json as _VSCODE_json\nimport builtins as
Info 18:55:27: Executing (status idle) silently Code = _rwho_ls = %who_ls\nprint(_VSCODE_getVariableTypes(_rwho_ls))
Info 18:55:27: Executing (status busy) silently Code = _rwho_ls = %who_ls\nprint(_VSCODE_getVariableTypes(_rwho_ls))
Info 18:55:27: Executing silently Code (completed) = _rwho_ls = %who_ls\nprint(_VSCODE_getVariableTypes(_rwho_ls))
Info 18:55:27: Executing silently Code (completed) = _rwho_ls = %who_ls\nprint(_VSCODE_getVariableTypes(_rwho_ls))
Info 18:55:27: Executing silently Code (completed) = _rwho_ls = %who_ls\nprint(_VSCODE_getVariableTypes(_rwho_ls))
Info 18:55:27: Searching for token colors ...
Info 18:55:27: Attempting search for colors ...
Info 18:55:27: Loading colors from /usr/share/code-insiders/resources/app/extensions/theme-defaults/themes/dark_plus.json ...
Info 18:55:27: Attempting search for colors ...
Info 18:55:27: Loading base colors from /usr/share/code-insiders/resources/app/extensions/theme-defaults/themes/dark_plus.json ...
Info 18:55:27: Using colors to generate CSS ...
Info 18:56:29: Searching for token colors ...
Info 18:56:29: Attempting search for colors ...
Info 18:56:29: Loading colors from /usr/share/code-insiders/resources/app/extensions/theme-defaults/themes/dark_plus.json ...
Info 18:56:29: Executing (status idle) silently Code = _rwho_ls = %who_ls\nprint(_VSCODE_getVariableTypes(_rwho_ls))
Info 18:56:29: Executing (status idle) silently Code = _rwho_ls = %who_ls\nprint(_VSCODE_getVariableTypes(_rwho_ls))
Info 18:56:29: Executing (status idle) silently Code = _rwho_ls = %who_ls\nprint(_VSCODE_getVariableTypes(_rwho_ls))
Info 18:56:29: Executing (status idle) silently Code = _rwho_ls = %who_ls\nprint(_VSCODE_getVariableTypes(_rwho_ls))
Info 18:56:29: Executing silently Code (completed) = _rwho_ls = %who_ls\nprint(_VSCODE_getVariableTypes(_rwho_ls))
Info 18:56:29: Attempting search for colors ...
Info 18:56:29: Loading base colors from /usr/share/code-insiders/resources/app/extensions/theme-defaults/themes/dark_plus.json ...
Info 18:56:29: Executing silently Code (completed) = _rwho_ls = %who_ls\nprint(_VSCODE_getVariableTypes(_rwho_ls))
Info 18:56:29: Executing silently Code (completed) = _rwho_ls = %who_ls\nprint(_VSCODE_getVariableTypes(_rwho_ls))
Info 18:56:29: Using colors to generate CSS ...
Info 18:56:29: Executing silently Code (completed) = _rwho_ls = %who_ls\nprint(_VSCODE_getVariableTypes(_rwho_ls))```

[DonJayamanne]

@faultdiagnosistoolbox Yes, I don't expect that to work, Please try %matplotlib widget

[faultdiagnosistoolbox]

OK, that is unfortunate but it has worked before (and is also working now when I downgrade the Jupyter extension version)?

In any case, thanks for investigating the issue. Unfortunately this (regrettably) makes me have to chose another tool than VSCode since the widget backend is not what I'm looking for (it is working though).

[DonJayamanne]

OK, that is unfortunate but it has worked before (and is also working now when I downgrade the Jupyter extension version)?

What version have you tried, I tried this at my end and the plot didn't work (see here https://github.com/microsoft/vscode-jupyter/issues/8322#issuecomment-978500903). I.e. plot window (external window app) is blank & I can run cells.

[faultdiagnosistoolbox]

I'm using VSCode version 1.62.3 and Jupyter extension v2021.9.1101343141. Everything else is up-to-date as far as I can tell. Then I can run %matplotlib and interactively run commands, produce plots, etc. Currently I'm sitting on a Mac, but I have the exact same experience on a Linux (Ubuntu 20.04) laptop. I've not had the chance to try on a Windows machine.

Edit: I saw now that you asked for version of matplotlib earlier in the thread. I'm currently on 3.5.0 (running python 3.9). Edit 2: I don't know if it matters, but I am running the Qt5Agg matplotlib backend (%matplotlib qt5, requires pyqt5 package)

[rchiodo]

We might want to consider reverting this change and try it out: https://github.com/microsoft/vscode-jupyter/pull/7959/commits/545b139ff1f4c6935fc5f63173374fac2e331bcd

[rchiodo]

Going to try this today to see if I can get qt5 to work.

[rchiodo]

This works fine for me. See GIF:

I got this working by

• Creating brand new conda environment with conda create -n condaEnvTest python==3.9
• Installing the appropriate stuff into that conda environment from conda-forge. Here's the yml for my environment: condaEnvTest.yml.txt
• Running the code in the GIF

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt

%matplotlib qt5

x = np.linspace(0, 20, 200)
plt.plot(x, np.sin(x))
plt.show()

[rchiodo]

So if it's not working, it may be a qt5/matlplotlib issue.

[rchiodo]

Verification steps:

• Create new environment
• Install matplotlib into it
• Run the code above
• QT5 window should popup

[faultdiagnosistoolbox]

I have attached a gif that illustrates the problem for me. I am running this on an Ubuntu machine but it works exactly the same on my Mac. If I downgrade the jupyter extension to v2021.9.1101343141 it all works.

The only difference to your set of instructions is that I do a 1+1 command after (I think any instructions will do to illustrate).

[rchiodo]

@faultdiagnosistoolbox it looks like you're not using conda. The set of packages installed likely has a big impact on this. Running other code after the plot works fine for me using conda, but not with a venv. I recommend using the yml file I included above to create an environment where this works (or maybe use it to compare against what's in your venv).

[faultdiagnosistoolbox]

It is true, I don't normally use conda. Do I understand correctly that you get the same behavior as me in venv but it works in a conda environment?

In my example above I created a minimal venv with only the required packages (all latest versions) to excite the problem. I will try to create a virtual environment based on the yml-file you created to see if there are any changes in behavior.

[rchiodo]

Do I understand correctly that you get the same behavior as me in venv but it works in a conda environment?

For me if I create a virtual env from scratch, the creation of the plot hangs the kernel forever (so not exact same results).

[faultdiagnosistoolbox]

I have no conda installation to experiment with (I could install if that is helpful).

I do think think that a minimal package installation is useful for deugging, but I installed all package according to the yml-file provided (except packages that were not pypi available on linux). No difference, same exact problem.

It do seems like venv/cpython is key to excite the problem. I don't know enough about the differences between conda/cpython+venv here to help I'm sorry to say.

[rchiodo]

Unfortunately, I think the end result is that QT5 or other backends are just not supported/use at your own risk. We don't really have the time to try and test every combination of python environments where it might work.

[rchiodo]

For my testing, the ipympl library was key to get it to work.

[faultdiagnosistoolbox]

I fully understand that the combinations of packages, backends, python distributions etc makes this complicated. What did you do with ipyml to get it to work?

I would be sad to have to leave VSCode, I like it in many ways, but this is a show-stopper for me unfortunately. I just wonder if my REPL use-case is that fringe? It is the python distribution from python.org after all with the standard (I think) interactive backend for matplotlib. If I use the tkinter backend, I experience the exact same problem.

In any case, I sincerely thank you for your efforts and energy trying to help me with this.

[rchiodo]

What did you do with ipyml to get it to work?

When I was building/testing with the conda environment, I first used %matplotlib widget to verify that the environment was correct. %matplotlib widget requires the python package ipympl. When I tried to repro a second time, the conda environment didn't work either until I also added ipympl. I haven't the foggiest why. I haven't debugged the kernel to see why it gets stuck in the venv or when I don't have ipympl.

What is about the QT5 backend that you need? Perhaps we can solve your problem a different way.

[faultdiagnosistoolbox]

It's not really the QT5 backend that is important, for example the tk works just as fine. What I do need is to have all my plots in external windows where I can zoom, resize etc. In my workflow I often have a number of plot windows open (with ID:s so that when I rerun code the plots get updated and stays in the same position on my screen). All works fine from a pure ipython prompt.

[rchiodo]

Is the key part having it in another window? As in if it was in another VS code tab that's too much in the way?

We discussed implementing a plotting window using the %matplotlib widget that you could do that with, but it would have to be in another tab in VS code.

[faultdiagnosistoolbox]

(I hope I don't come across as whiny and overly picky; not my intention at all)

For me and my workflow, external windows is key since I work with multiple monitors and typically populate one screen with plots and another screen with code. It is also essential that I can see all plots at once since I compare end evaluate them against each other. I guess some of my preferences comes from my old Matlab days, but this is how I'm efficient.

[rchiodo]

No no worries. This kind of feedback is golden. It's much appreciated.

VS code won't likely support multiple monitors for some time. But this is a good way to help push that.

And now I understand why QT5 is so crucial for your workflow.

[faultdiagnosistoolbox]

Also on a laptop, with only one screen, external windows is equally useful for me. So it is not entirely a multiple-monitor issue.

[pgA3D]

I concur with @faultdiagnosistoolbox, my workflow is similar and this bug is a show stopper. Those are the 3 essentials for my workflow that the Qt5 backend (and probably others) provides :

• Having the plots open in a separate windows such that we can compare them,
• being able to adjust the figures axes (max and min value)
• be able to zoom into the figure interactively (using the mouse).

[rchiodo]

@pgA3D all 3 of those would be possible inside of VS code. The 'separate' window would just be limited to another tab.

We kinda support this, but it's rather limited:

If we expanded our current plotviewer to be more like the QT control (but still be inside VS code), would that be usable?

@faultdiagnosistoolbox really needed the windows outside of VS code. That's a lot harder to accomplish. But improving our plotviewer should be pretty easy.

[faultdiagnosistoolbox]

First I want to say I haven't the insight to realize all problems but what I'm looking for is what were available before Jupyter extension v2021.9.1101343141. Then, it worked as in current versions of, e.g., PyCharm where it works perfectly. What I'm looking for is similarly to what I would get if I pasted to an ipython prompt, directly utilizing the matplotlib backends and no tight integration of the plot windows into the IDE. This I get if I do %matplotlib tk or %matplotlib qt in PyCharm (on with non-recent Jupyter extensions).

Integration of plots in the IDE in tabs is not a solution for me unfortunately.

[rchiodo]

I did some more digging. It works for me ONLY if I do this first:

1. %matplotlib widget
2. do something to get a plot
3. restart the kernel
4. %matplotlib qt5

The plot we render in step 2 establishes a comm channel to the kernel. This channel goes away in step 3, but the javascript listening to events is still there in step 4. Somehow this makes it work.

I'm reopening because this might be fixable.

Can somebody else try those steps?

[rchiodo]

The reason this might have been working with older versions is because our webview might have always loaded the extra javascript code. The new VS code based UI will only load it when a widget gets rendered.

[rchiodo]

Oh and @DonJayamanne in case you're curious, I did revert that perf change for widget support. It had no effect. I have to load the widget for a plot to get it to work. It might also be something that widget leaves behind that makes it work.

[faultdiagnosistoolbox]

I tried the steps suggested by @rchiodo and unfortunately I don't get it to work. I tried with tk and qt backend.

What happens is that when I try to open the plot the second time, after %matplotlib qt5, it freezes (or more acurately it produces a plot after about 50 seconds).

Edit: after the successful attempt by pgAD3 I gave it another try. I can swear that I did the same procedure now but now I got it to kind of work. I got a new plot window the second time; but trying to do a simple 1+1 in the console made it freeze again.

[pgA3D]

I tried as well the steps from @rchiodo , and it works on my side.

I do %matplotlib widget import matplotlib.pyplot as plt plt.figure()

then restart the kernel

%matplotlib qt5 import matplotlib.pyplot as plt plt.figure()

And I get a figure instantaneously, no hanging afterward.

[Groberts93]

I can reproduce the steps too. I note that also typing any character in the interactive command line has the same "unblocking" effect on previous cells. This is on native Ubuntu 20.04.

@rchiodo with regard to your question in the other thread about my use case -- my work is dependent on a library (MNE-Python) which requires the Qt backend for interactive plotting of large electrophys datasets. This setup allows me to plot my own matplotlib figures in the same environment with minimal friction, use the full screen space for large composite plots etc.

My hard requirement is that I can use the MNE data viewer, which is unfortunately tied to Qt for now. My feature preference is plotting in separate windows.

At present I'm using an old version of the Jupyter extension (v2021.10.1001393861) which works fine with qt magics. I can continue to work with this setup for the time being.

[rchiodo]

More investigation notes:

Running this code will also make it work (with or without restarting the kernel)

import ipywidgets as widgets
widgets.IntSlider()

So it has something to do with us activating our widget manager.

[faultdiagnosistoolbox]

I can confirm that activating the widget manager made it work better also for me. First I thought it worked as desired, but the problems re-occurred after a few plotting commands. I did

1. activate widget manager
2. %matplotlib qt5
3. necessary imports and then plot a sinus-curve

Then iterating 3 (or other sinple 1+1 commands) makes it freeze for me, usually the second time but it varies.

[bermeom]

I did anything, and it doesn't work, so I have no choice but to install an old version (September version - 1.61.2). Thanks @faultdiagnosistoolbox and @gonghenghai, for bringing up the issue and I hope the vscode-jupyter team will consider fixing this issue soon as this feature is very important to everyone who works analyzing data (data scientist, data analyst, data engineer) .

[rchiodo]

I've submitted what I think is a fix. Hopefully it will come out in our 'prerelease/insiders' build tonight (in about 7 hours from when I write this comment)

If anybody can try it, that'd be great.

[pgA3D]

thanks @rchiodo I got the pre-release v2022.1.1001666841 from the VScode extension tab (is that the correct one?), and it seems to work perfectly fine.

[rchiodo]

Yep that's the one. Glad to hear it's working.

[rchiodo]

Closing as fixed.

Verification steps:

1. Open python file
2. Run a cell with: %matplotlib qt5 (might require installing Pyqt5)
3. Do normal stuff after that
4. Run a matplotlib plot. It should open a separate QT window like so:

5. Run a bunch of code after that and make sure you can still talk to the kernel
6. Try opening more plots.

[faultdiagnosistoolbox]

I can confirm that it also works for me. Thank you!

[bermeom]

Yes, it is working !! Thank you.

[Groberts93]

As others have said, this is working now. Thank you very much for fixing this.

[karahanyilmazer]

It also works for me. Thank you very much, this is of huge help!

[DonJayamanne]

@faultdiagnosistoolbox @bermeom @Groberts93 @karahanyilmazer @pgA3D We're making a few changes to this part of the code and would like to verify that it does not break your usage of qt5. I have tested this at my end and it works as expected.

• Please could you update to the latest version of VS Code (1.83)
• Download this file and extract it
• Go into VS Code and use the command Extensions: Install from VSIX
• Select the extracted VSIX
• Reload VS Code
• Go into extensions panel and serach Jupyter and verify the version installed is v2023.10.100
• Test the code again and verify you can use qt5 and the kernel does not hang when you attempt to make additional requests

Thank you, and really appreciate your help in this space. ms-toolsai-jupyter-insiders.vsix.zip