InvisibleProgrammer commented on PR #4252: URL: https://github.com/apache/hive/pull/4252#issuecomment-1522384762
I started to play with the porting, let me share some extra information and details: There are two modules that we 'copy' into Hive: `mr` and `hive-metastore`. When we do a Hive-Iceberg upgrade, it is worth checking if is there anything that should be ported from there as well - or otherwise, we can get unexpected behavior. I was able to narrow down the promising commits to 7: To gather them, I got the git commits between the 1.1.0 and 1.2.1 tags but only for the mr and hive-metastore folders with those commands: ``` zsoltmiskolczi@zsmiskolczi-MBP16 iceberg % git log apache-iceberg-1.1.0...apache-iceberg-1.2.1 --pretty=tformat:"%H %s %cs" --reverse -- hive-metastore 6b8f7e0e31a81029b478e7757aba749f5ed27f42 Hive: Set the database owner on namespace creation (#6045) 2022-11-28 6725dc4d5168f41180d4e4acb73fde5b78abf062 Hive: Merge identical catch branch (#6477) 2022-12-22 fede493d59f17ff2bfc0744b296d90bd36130386 Hive: Lock hardening (#6451) 2023-01-11 df87b2e214c7576184e85e1be145069fc0c4d871 Hive: Make UGI current user the owner of new Hive objects (#6324) 2023-01-18 881be5e5d3746da0d8f1a837d2d1bb0f83776b81 Build: Fix minor error-prone warnings (#6629) 2023-01-23 81bf8d30766b1b129b87abde15239645cb127046 Core: Avoid creating new metadata file on registerTable (#6591) 2023-01-25 333227fbd13821365cec1bdbfcb9314a239bea0f Hive: Refactor commit lock mechanism from HiveTableOperations (#6648) 2023-02-03 ba4818b871d36eb34fcdd20dc7f824447f5daa4a Use UGI shortUserName as the default owner of Hive objects (#6955) 2023-03-07 zsoltmiskolczi@zsmiskolczi-MBP16 iceberg % git log apache-iceberg-1.1.0...apache-iceberg-1.2.1 --pretty=tformat:"%H %s %cs" --reverse -- mr b5102a6f6603b4ec776c849b2c0e5f212366f11d Build: Bump jackson-annotations from 2.14.0 to 2.14.1 (#6280) 2022-11-28 fede493d59f17ff2bfc0744b296d90bd36130386 Hive: Lock hardening (#6451) 2023-01-11 ab6ba6801cddd70d261357b514103fd8e95775f9 Build: Bump jackson-annotations from 2.14.1 to 2.14.2 (#6687) 2023-02-16 ``` As you can see, there are 8 commits but one of them modified both folders. I started to cherry-pick them but I haven't finished with it yet. I'll share the result when I'm finished. According to my current knowledge, there are two commits that are hard to handle: `fede493d59f17ff2bfc0744b296d90bd36130386` and `333227fbd13821365cec1bdbfcb9314a239bea0f`. Both of them modify `HiveTableOperations.java` and this file contains a lot of differences between Hive-Iceberg and Iceberg. We have to make sure that we don't introduce new issues during the porting: there can be extra development in the Hive repository that is completely different from the Iceberg part. And also, based on my experience, there can be old changes that haven't been ported at all. @zhangbutao , what do you thing about that? -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
