fix(config): prevent base_path from flipping to ~/.forge when ~/forge still holds user data - #2925
Merged
tusharmath merged 3 commits intoApr 10, 2026
Merged
fix(config): prevent base_path from flipping to ~/.forge when ~/forge still holds user data#2925tusharmath merged 3 commits into
tusharmath merged 3 commits into
Conversation
Copy link
Copy Markdown
Collaborator
Sorry, something went wrong.
Comment thread
crates/forge_config/src/reader.rs
Comment on lines
+205
to
+212
| // Without FORGE_CONFIG set the path must be either ".forge" (new) or | ||
| // "forge" (legacy fallback when ~/forge exists on this machine). | ||
| let name = actual.file_name().unwrap(); | ||
| assert!( | ||
| name == ".forge" || name == "forge", | ||
| "Expected base_path to end with '.forge' or 'forge', got: {:?}", | ||
| name | ||
| ); |
Copy link
Copy Markdown
Contributor
There was a problem hiding this comment.
Test is now non-deterministic and depends on the host filesystem state. If ~/forge exists on the CI/CD runner or developer's machine, the test validates different behavior than if it doesn't exist. This makes test results non-reproducible and can hide regressions.
Impact: Different environments will test different code paths, reducing test reliability.
Fix: Use filesystem isolation for deterministic testing:
#[test]
fn test_base_path_prefers_legacy_when_it_exists() {
let temp = TempDir::new().unwrap();
env::set_var("HOME", temp.path());
fs::create_dir(temp.path().join("forge")).unwrap();
let actual = ConfigReader::base_path();
assert_eq!(actual.file_name().unwrap(), "forge");
}Spotted by Graphite
Is this helpful? React 👍 or 👎 to let us know.
Sorry, something went wrong.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Reviewers
1 more reviewer1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Fix ProviderNotAvailable errors caused by base_path() silently switching from ~/forge to ~/.forge when an empty ~/.forge directory was created as a side effect of the shell plugin's config-edit action.
Context
After adding the forge config migrate command (which teaches users to move from ~/forge to ~/.forge), the shell plugin's _forge_action_config_edit function was updated to call forge config path dynamically and then run mkdir -p <config_dir> to ensure the config directory exists.
The problem: forge config path returns ~/.forge/.forge.toml as the new default (since 26badbaac changed base_path() to prefer ~/.forge). So for users who still have their data in ~/forge, the shell plugin would eagerly create an empty ~/.forge directory. This caused base_path() to flip:
With base_path() now returning ~/.forge, credentials_path() pointed to ~/.forge/.credentials.json — a file that doesn't exist because all credentials are in ~/forge/.credentials.json. read_credentials() returned an empty list, every provider appeared unconfigured, and every provider lookup raised ProviderNotAvailable.
Changes
Testing
Reproduce and verify the fix: