mirror of
https://github.com/aljazceru/turso.git
synced 2025-12-25 12:04:21 +01:00
cc643362a4
This kind of fault does not semantically represent anything real, since we already have fault injection for every concrete IO operation like reading, writing, syncing and so forth. Moreover, having this feature is the direct cause of the false positive simulator failure as reported in issue #2727. There, a "run_once fault" happened immediately after we fsynced following an INSERT, which caused the simulator to think the INSERT failed, and later a sim assertion failed because the on-disk database had 1 more row than it thought it would.