bug#58782: 29.0.50; tramp-kubernetes - connection cache not cleared

[Michael Albinus]

Michael Albinus <michael.albinus@gmx.de> writes:

Hi Filipp,

>> Kubernetes from Docker Desktop
>> (https://www.docker.com/products/docker-desktop/), one pod
>> "myapp-67cd597654-wxdjn" in namespace "main".
>>
>> emacs -Q
>>
>> 1) shell command: kubectl config set-context --current --namespace=main
>> 2) C-x C-f /kubernetes:myapp-67cd597654-wxdjn:/
>> 3) M-x tramp-cleanup-this-connection
>> 4) shell command: kubectl config set-context --current --namespace=
>> 5) Again C-x C-f /kubernetes:myapp-67cd597654-wxdjn:/
>>
>> I get an error, and completion cache (C-x C-f /kubernetes: TAB) is
>> cleared properly.
>
> I'm not sure what you mean with "I get an error".
>
>> But if at the step 5 I instead go to existing dired buffer opened at
>> step 2, and type, say, "i" on a subdirectory, I get the same error,
>> but the cache is NOT cleared - this is the bug.
>
> Again: which error?
>
>> The workaround is to set tramp-completion-use-cache to nil.  But
>> perhaps there exists a better fix.
>
> This is only for hot name completion. I don't see how it fits to your
> use case.
>
> (As usual, please show all messages you get, and pls show the Tramp
> debug buffer with sufficient tramp-verbose)

Hmm, no response. Do you still feel we shall follow this report? If yes,
pls provide more information.

>> Thanks.

Best regards, Michael.