-
Type:
Task
-
Resolution: Fixed
-
Priority:
Major - P3
-
Affects Version/s: None
-
Component/s: None
-
None
-
Fully Compatible
-
Execution Team 2024-04-01, Execution Team 2024-04-15
-
None
-
None
-
None
-
None
-
None
-
None
-
None
Putting my personal running list of issues I have with time-series code into a ticket.
- Conflicting symbols forĀ commitTimeseriesBucketsAtomically across these namespaces, which is confusing and makes code hard to trace.
- write_ops_exec
- timeseries
- MakeNewBucketFromWriteBatchWithMeta no longer makes from WriteBatch, update name of test.
- Pull out some of the structs in bucket_catalog.h (like Stripe, StripeNumber, etc) into a separate header to break dependencies.
- Should Stripe have an internal id? I can think of this being useful primarily for logging, debugging.
- performOrderedTimeseriesWritesAtomically needs documentation.
- Comment for claimWriteBatch refers to old function name getResult instead of getWriteBatchResult which confused me for a while.
- write_ops_exec::performDelete should describe the reason for and what is its return type of 'long long' in the header.
- Document getFleCrudProcessed in header.
- Explain preference for v2 over v3 buckets in WriteBatch header.
- Fix docs bug in bucket format in time-series arch guide.
- Explain bit-packing of Bucket and WriteBatch.