Repository navigation
Add a feature for a fallback behavior if a project leaf dir is moved as part of an automated process #15762
travelingflwr
started this conversation in
Ideas
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
What happened
A thread whose workspace folder is a leaf directory (
C:\Users\xxxxx\Claude_Work\cases\xxxxxxxxx) became unopenable after that folder was legitimately deleted as part of the work the thread was doing:The thread had no other fault — full history intact, work complete. Recovery required manually recreating the deleted directory by hand; T3 does not fall back, prompt, or self-heal. The folder has to be restored exactly where the user just deliberately removed it, which also silently undoes the cleanup the thread existed to perform.
Why this isn't an edge case
The workspace folder being deleted is a normal outcome of some tasks. Here the thread's job was to archive a support case, which moves
cases/<id>/tocases/Closed/<id>/. The thread doing its job correctly is what breaks the thread.It also can't complete that job unaided: on Windows the workspace folder appears to be held open for the thread's lifetime, so
shutil.move/Remove-Itemon it fails withWinError 32 — the process cannot access the file because it is being used by another process. Deleting it from an outside shell succeeds, and then the thread can't reopen.Repro
<some>/<leaf-dir><leaf-dir>Expected
One of:
Xno longer exists; usingY")Hard-failing on a missing directory makes the transcript unreachable at exactly the moment it's most useful — right after the operation that removed the folder.
Environment: Windows 11 Enterprise 22631, T3 Code, Claude Code harness
All reactions