-
Notifications
You must be signed in to change notification settings - Fork 7
Description
Summary
EPERM file system error needs classification fix. Backend Platform 428 errors continue from previous reports, affecting multiple commands.
| Metric | Value |
|---|---|
| Time range | last 24 hours |
| Total errors | 134 |
| CLI bugs | 1 |
| Backend issues | 1 |
| User errors (working as designed) | 6 |
| Unique users affected | 62 |
| Internal user occurrences | 11 |
Errors requiring action
1. File system permission error lacks error classification — CLI bug
| Error code | None (unclassified) |
| Occurrences | 7 (0 internal) |
| Users affected | 7 |
| Command | agents pull |
| Platforms | Darwin |
| PostHog | View in error tracking |
| Existing issue | None |
| Recurring | No |
Error: (from PostHog — the actual error message users see)
Error: EPERM: operation not permitted, scandir '/Users/<redacted>/.Trash'
Stack trace: (only showing frames from packages/cli/src/)
at globby scanning files in agent directory
Root cause: The agents pull command uses globby to scan the agents directory, but when globby encounters permission issues accessing system directories like .Trash, it throws an unhandled EPERM error:
// packages/cli/src/core/resources/agent/config.ts:46
const files = await globby(`*.${CONFIG_FILE_EXTENSION_GLOB}`, {
cwd: agentsDir,
absolute: true,
});The error occurs when the agentsDir is set to a problematic path or when globby encounters permission issues. This should be wrapped with proper error classification.
Suggested fix: In packages/cli/src/core/resources/agent/config.ts:46, wrap the globby call in a try-catch block and throw a proper FileReadError or SystemError with an appropriate error code.
Backend issues (not CLI fixes)
Brief table of errors caused by the backend/server, not the CLI:
| Error | Occurrences | Users | Command | PostHog |
|---|---|---|---|---|
| Backend Platform apps required (428) | 13 | 8 | entities push, deploy, functions deploy |
View details |
These 428 errors indicate users are trying to use new CLI features with older dashboard-created apps that don't support the Backend Platform. This is a backend migration issue, not a CLI bug.
User errors (working as designed)
Brief table of expected user errors where the CLI validation is correct:
| Error | Occurrences | Users | Command | PostHog |
|---|---|---|---|---|
| No Base44 project found | 25 | 16 | link, deploy |
View details |
| Project with ID not found/ejectable | 6 | 6 | eject |
View details |
| Project already linked | 9 | 9 | link |
View details |
| App not configured | 9 | 5 | dashboard open |
View details |
| Invalid schema in config file | 9 | 6 | deploy |
View details |
| Deno dependency not found | 5 | 3 | dev |
View details |
These errors represent correct CLI validation behavior. No volume is unusually high.
Recurring errors
Based on comparison with previous daily reports:
| Error | Days recurring | Existing issue | Tracked? |
|---|---|---|---|
| Backend Platform 428 errors | 3+ days | #414, #413 | yes |
The 428 Backend Platform errors are tracked in previous reports but continue appearing daily, suggesting this is an ongoing backend migration issue affecting user workflows.
Action items
Numbered list of concrete next steps, most important first:
- [medium]
packages/cli/src/core/resources/agent/config.ts:46— Wrapglobbycall in try-catch block and throw properFileReadErrororSystemErrorto classify EPERM file system errors - [low] Monitor Backend Platform 428 error trends — while not a CLI fix, the volume suggests this affects many users trying to migrate from dashboard-created to CLI-created apps