T0002 lock-unwrap¶
Rule: A crate keeps its lock().unwrap() calls under a limit.
| Code | T0002 (stable, never reused) |
| Id | lock-unwrap |
| Category | state |
| Source | spirit_level |
| Fix | by hand |
Level per tier¶
| Tier | Level | Reason |
|---|---|---|
| small | off | |
| small-workspace | off | |
| medium | warn |
the count of lock().unwrap() per crate shows where shared state concentrates |
| big | warn |
the count of lock().unwrap() per crate shows where shared state concentrates |
| ai | off | |
| embedded | off | |
| ui | off | |
| egui | off | |
| units-strict | off | |
| units-paranoid | off |
Parameters¶
| Name | Type | Default | Meaning |
|---|---|---|---|
max |
integer | 0 |
calls allowed per crate before it is reported |
Why¶
Each .lock().unwrap() is a lock reached from more than one place, and a panic when another thread panicked while holding it (poison). In tpm's crates there were about 1,200 of them. A count per crate shows where shared state concentrates and whether a refactor is making it shrink.
The lint counts .lock(), .read() and .write() with no arguments followed by .unwrap() or .expect(..), outside test code, and reports a crate whose count passes max, with the first places.
How to fix¶
Give the state one owner (see static-interior-mutability). Locks that stay are taken through one small helper that decides what poison means (recover with into_inner, or fail with a message), so the count goes to zero.
Configure¶
Allow one finding in the code with // spirit-level: allow(lock-unwrap) <reason> (waivers); the code works there too, as in [lints.T0002].