Failed initial process produce invalid state

[sebsgr]

If the initial process fails (e.g., due to an invalid repository URL or token), the status command does not list the code machine. I believe this is the correct behavior, but when I try to rerun the init command, Codchi encounters an error (Code machine '...' already exists).

Environment:

codchi v0.2.1
d397ca96519512b1d117ca0c373b380006b3fc4b

Steps to reproduce

$ codchi.exe init java-ide https://oauth2:XXXXXXX@XXXXXXXX
Code forge: Gitlab
Error: Couldn't access repository. If the repository is private you need to provide the correct credentials. If it's a local repository, make sure it is checked out locally in the machine at the provided path. See <https://codchi.dev/docs/start/usage.html#creating-a-machine> on how to access private repositories.

$ codchi.exe status
┌─────────┬────────────┬──────────┐
│ Machine ┆ Status     ┆ Running? │
╞═════════╪════════════╪══════════╡
│ demo    ┆ Up to date ┆ ❌       │
├╌╌╌╌╌╌╌╌╌┼╌╌╌╌╌╌╌╌╌╌╌╌┼╌╌╌╌╌╌╌╌╌╌┤
│ test    ┆ Up to date ┆ ❌       │
└─────────┴────────────┴──────────┘

java-ide is not listed

$ codchi.exe init java-ide https://oauth2:XXXXXXX@XXXXXXXX
Error: Code machine 'java-ide' already exists.

[htngr]

$ codchi.exe init java-ide https://oauth2:XXXXXXX@XXXXXXXX

Btw: Please use --token oauth2:xxx for private repos

[sebsgr]

Thanks, but the main issue is still unsolved.

If I use invalid URLs or tokens, the process stalls while retrieving available modules. After stopping the process with CTRL+C, the same problem still occurs.

[htngr]

Should be fixed in d8a0acf (v0.2.3). Until then you can manually codchi delete and then init again. If that fails wsl --unregister, delete %APPDATA%\codchi\machine\java-ide and then init again.