Hi,
I disagree that the traceability is a problem. There is a 1:1 mapping between a Jira and a PR and the mapping is kept in the commit log history. SVN based project have been working like this for a while. That said, I don't really have a strong opinion about this policy and I'm happy to follow what the community thinks is better, and follow the example of other Apache projects. Gyula, would you like to make the necessary changes to the contributor guidelines in the wiki? I guess the main requirements would be linear history, tagging the commits with the Jira ID, and closing off the GitHub PR automatically. gdfm On Friday 26 Jun 2015 at 13:54, Gyula Fóra <[email protected]>, wrote: Hey, I would like to revive an older discussion about you policy for merging PR-s. I think the current way of squashing all commits in the PR and removing history and authorship of the commits, is very bad and will cause serious problems for the whole project. There are multiple problems here: 1. There is absolutely no reason to have this policy. Large projects large contributor base still keep the commits of individual contributors, and they only expect the contributors to rebase on the master to keep the history linear. This way the commits in one PR are still in one place and the changes become more tracable for the smaller commits. 2. This policy will drive off many would-be contributors as no-one want to put a lot of work into a new feature when he will not be even noted as a contributor on git. I think this is actually a major concern, I didnt conduct any surveys but I think at least half of the people will not contribute this way :P 3. Since the contributors don't show on git, there is no sign of community activity which will again drive off many new contributors. (2+3) will drive off most people I believe. I think in order to get a larger community, this needs to change for future contributions. I am not sure what to do about past contributors. Best regards, Gyula
