fix flaky LittDB test#3531
Conversation
PR SummaryLow Risk Overview For truncated key/value file scenarios, truncation only picks a random non-zero length when the on-disk file has more than one byte ( For metadata preserved on restart and orphaned metadata tests, configured TTL is now at least one hour (plus a small random jitter) instead of sub-second values, so background GC is unlikely to run during the brief close/reopen cycle and disturb metadata expectations. Reviewed by Cursor Bugbot for commit 38d53c1. Bugbot is set up for automated code reviews on this repo. Configure here. |
|
The latest Buf updates on your PR. Results from workflow Buf / buf (pull_request).
|
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #3531 +/- ##
==========================================
- Coverage 59.12% 58.28% -0.85%
==========================================
Files 2213 2140 -73
Lines 182814 174431 -8383
==========================================
- Hits 108096 101660 -6436
+ Misses 64993 63725 -1268
+ Partials 9725 9046 -679
Flags with carried forward coverage won't be shown. Click here to find out more. 🚀 New features to boost your workflow:
|
wen-coding
left a comment
There was a problem hiding this comment.
Looks good, but can you update title to say which flaky test we are fixing?
Describe your changes and provide context
Fix flaky unit test