Error message "the system cannot find the path specified" upon trying to start the kernel #22551
Comments
|
Because the Spyder bug report tool does not allow me to attach files, I continue the description of the issue, including attached files, here. This is the Conda environment for which the problem occurs. This is the Conda environment for which the problem does not occur. In the hope that it will be helpful, I have already used debug mode to generate the relevant log files. They are here. Potentially relevant evidence and suspected causeWhile generating the environment listing files above, something which I suspect to be relevant happened. Upon activating the problematic environment using Mamba, a message was displayed which contained the same error text which had appeared in Spyder, plus what appears to be a fragment of a file path. Perhaps relevantly, this file path is cut off at the space character in my user folder name. This distinctly reminds me of certain previous experiences (#21937). Could that space character be causing problems again? Because the message upon activating the environment in the command line appears to have no direct connection to Spyder, I tried to determine whether the environment itself would work without Spyder. It does appear to work: it is possible, in the affected environment, to start Python, import OpenCV, and correctly run programs previously known to work in that environment on other computers. |
|
Hey @meltingSnowdrift, thanks for reporting and all the supporting info you collected, that's really important. I find especially significant this part:
So, it seems a library in your env (OpenCL) is generating an error message on env activation. The problem is that that message is caught by Spyder (because we activate envs before starting a kernel for them), which shows it as a kernel start error. Just to be sure, could you post a screenshot of what you're seeing in Spyder when starting a console for that env? Thanks! |
|
I am quite aware that, if the cause of the problem is a library, and something resembling the problem happens outside Spyder, it is arguable that the problem is not a problem with Spyder. The only reason I reported this as a Spyder issue at all, given the evidence so far, is because it is possible to correctly run Python programs in the affected environment outside Spyder, while it is not possible to do so within Spyder. To the extent that this is a Spyder problem at all, I suspect that it may prove to be about how Spyder handles non-fatal errors from libraries in the environment. Here is the screenshot you requested. |
Yeah, that's reasonable. Good thing is that we can deal with it on our side.
Thanks! That's useful too. |
ccordoba12
added
type:Bug
component:IPython Console
and removed
status:Awaiting Followup
labels
Description
What steps will reproduce the problem?
This problem occurs whenever Spyder tries to start a kernel using a certain Conda environment which I have just created. It does not occur when starting the kernel in another Conda environment: in that other environment, Spyder works normally.
Edit: See the comment below for the rest of the description of this issue.
Versions
Dependencies
Environment
Environment
The text was updated successfully, but these errors were encountered: