The fault-injection test for a restore that runs out of disk space, TestRestoreReportsDiskFull in internal/vaultik/fault_injection_test.go, is still skipped with "blocked on #163". That issue is closed: since commit d9f0220, restore removes a file whose write or close fails. So the test either passes now or shows that a full disk still leaves a truncated file at the target. Either way, the disk-full case of the 1.0 fault-injection tests (#72, scenario 5) is not being checked.
Definition of done
The t.Skip in TestRestoreReportsDiskFull is gone, and its doc comment says what the test checks, with no "skipped until" wording.
The test runs and passes: restore fails with an error saying the disk is full (out of space), and no partial file is left at the target path.
If the test fails as written, restore is fixed, not the test: no assertion is weakened or dropped. If the test itself is wrong (for example the simulated full disk does not fail the way a real one does), the PR says why in one line and corrects the test so it still checks both points in item 2.
TODO.md is not touched (a parallel unit edits it).
make check passes.
Model: opus-5-5
The fault-injection test for a restore that runs out of disk space, `TestRestoreReportsDiskFull` in `internal/vaultik/fault_injection_test.go`, is still skipped with "blocked on https://git.eeqj.de/sneak/vaultik/issues/163". That issue is closed: since commit `d9f0220`, restore removes a file whose write or close fails. So the test either passes now or shows that a full disk still leaves a truncated file at the target. Either way, the disk-full case of the 1.0 fault-injection tests (https://git.eeqj.de/sneak/vaultik/issues/72, scenario 5) is not being checked.
## Definition of done
1. The `t.Skip` in `TestRestoreReportsDiskFull` is gone, and its doc comment says what the test checks, with no "skipped until" wording.
2. The test runs and passes: restore fails with an error saying the disk is full (out of space), and no partial file is left at the target path.
3. If the test fails as written, restore is fixed, not the test: no assertion is weakened or dropped. If the test itself is wrong (for example the simulated full disk does not fail the way a real one does), the PR says why in one line and corrects the test so it still checks both points in item 2.
4. `TODO.md` is not touched (a parallel unit edits it).
5. `make check` passes.
Model: opus-5-5
clawbot
added this to the 1.0.0 milestone 2026-10-01 19:54:22 +02:00
clawbot
self-assigned this 2026-10-01 19:54:22 +02:00
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
The fault-injection test for a restore that runs out of disk space,
TestRestoreReportsDiskFullininternal/vaultik/fault_injection_test.go, is still skipped with "blocked on #163". That issue is closed: since commitd9f0220, restore removes a file whose write or close fails. So the test either passes now or shows that a full disk still leaves a truncated file at the target. Either way, the disk-full case of the 1.0 fault-injection tests (#72, scenario 5) is not being checked.Definition of done
t.SkipinTestRestoreReportsDiskFullis gone, and its doc comment says what the test checks, with no "skipped until" wording.TODO.mdis not touched (a parallel unit edits it).make checkpasses.Model: opus-5-5
clawbot referenced this issue2026-10-01 19:58:58 +02:00