Erroneous Behavior

after getting dropped to shell from git repos pull step failing and deleting the failing repo topgrade has cached the repos from the glob in config

Expected Behavior

topgrade should invalidate the github repo directories cache if dropping to shell or retrying on repo pull step

Steps to reproduce

1. have glob repos in your config example ~/Projects/*
2. have an invalid (e.g deleted) remote in a git repo in side of said folder
3. run topgrade
4. drop to shell
5. delete repo with invalid remote
6. exit shell -> topgrade fails again but this time with OS error directory not found

Possible Cause (Optional)

need to invalidate repo cache in if dropped to shell or retrying (directories might have change since last run)

Problem persists without calling from topgrade

not applicable

Did you run topgrade through Remote Execution

not applicable

Configuration file (Optional)

[git]
max_concurrency = 10
repos = ["~/Projects/*/"]

Additional Details

• Operation System/Version
  NixOS 24.05pre590113.1536926ef562 (Uakari)

• Installation
  NixPkgs

• Topgrade version (topgrade -V)
  Topgrade 14.0.1

Verbose Output (topgrade -v)