Weekly update in TiDB

Last week, we landed 27 PRs in the TiDB repositories.

Added

• Support stream aggregation on TiKV
• Support long VARCHAR
• Set Range and OutputCounts in Coprocessor Response for streaming API
• Support COMMENT = string in partition definition
• Add restrict and cascade in DropTable

Fixed

• Set have_profiling to NO

Improved

• Remove invalid intervals in building Range
• Resolve locks in a batch
• Extract the same part from DNF's leaves
• Simplify the logic of HashCode
• Improve the performance of decoding decimal
• Test coverage:

• Improve the test coverage in the executor package
• Improve the test coverage in the plan package
• Improve the test coverage in the distsql package

• Panic recover:

• Add the recover mechanism for index lookup reader workers
• Add the recover mechanism for union workers
• Add the recover panic in owner.CampaignLoop

Weekly update in TiSpark

Last week, we landed 4 PRs in the TiSpark repositories.

Added

• Add the index covering optimization
• Add the broadcast join support

Improved

• Add more tests
• Add DDL Dump for using TiDB as Spark MetaStore

Weekly update in TiKV and PD

Last week, we landed 12 PRs in the TiKV and PD repositories.

Added

• Export some struct expressions for other tools
• Log panic stacks
• Update the TiKV version
• Add a Raft message flush metrics

Fixed

• Update the Rust dockerfile
• Fix lowspace metrics
• Check Regions in the background goroutine
• Handle the problem that only one row returns in stream aggregation

Improved

• Rename a confused constant variable
• Make end_point_request_max_handle_secs configurable
• Precreate some labal metrics

New contributor (Thanks!)

• TiKV: Greg Weber