[ https://issues.apache.org/jira/browse/PARQUET-1687?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Gabor Szadovszky updated PARQUET-1687: -------------------------------------- Fix Version/s: 1.11.0 format-2.8.0 > Update release process > ---------------------- > > Key: PARQUET-1687 > URL: https://issues.apache.org/jira/browse/PARQUET-1687 > Project: Parquet > Issue Type: Task > Components: parquet-format, parquet-mr, parquet-site > Reporter: Gabor Szadovszky > Assignee: Gabor Szadovszky > Priority: Major > Labels: pull-request-available > Fix For: 1.11.0, format-2.8.0 > > Time Spent: 20m > Remaining Estimate: 0h > > Our current tagging policy in the release process requires to use the same > tag for all the release candidates which means at RC2 we remove the tag from > RC1 head and adds again to the RC2 head and so on. It is not a good practice. > Hard to track RCs and rewriting git history is usually not a good idea. > Update the related scripts and docs so we will create RC tags (e.g. > {{apache-parquet-1.11.0-rc6}}) first and add the final release tag (e.g. > {{apache-parquet-1.11.0}}) after the vote passes. > It is also very confusing that we always use SNAPSHOT version of the next > release even if the previous one is not released yet. For example we have not > released {{1.11.0}} yet (we are after the failed vote of 1.11.0-rc6) still, > we use the version number {{1.12.0-SNAPSHOT}}. This is very confusing. > Update the release scripts and docs to use the same {{SNAPSHOT}} version at > RC phase and only step to the next version number after the vote passed and > the final release tag is added. -- This message was sent by Atlassian Jira (v8.3.4#803005)