[Dev] Rename default branch to 'main'

[asfimport]

Parent issue, see subtasks

Reporter: Neal Richardson / @nealrichardson

Subtasks:

• [Dev] Update GitHub Actions workflows that hardcode master as default #31143
• [Dev] Update archery to work with either master or main as default branch #31144
• [Dev] Update release scripts to use default branch #31145
• [Dev] Update crossbow templates to use master or main #31146
• [Dev] Update apache/arrow-site GitHub Actions deploy.yml website deployment workflow to support being triggered when pushing to main #31147
• [Dev] Ask INFRA to switch default branch to main #20117
• [Dev] Update URLs in the repo to point to main #31148
• [Website] Remove support for triggering GitHub Actions apache/arrow-site website deployment workflow when pushing to master #31412
• [Website] Remove mentions of master branch from Apache Arrow website content #20161
• [Dev] Remove "master" from default_branch property of Target class in core.py after migration to "main" as the default Git branch #32804
• [Dev] Update the pull request merge script to work with master or main #33005
• [Dev] Replace hard-coded string "master" with "main" in dev/archery/archery/crossbow/core.py after default branch migration  #33215
• [Release] Update release scripts with main as new default branch #34256
• [Docs][Dev] Update git links/branches from master to main for external projects #34257

_{Note: This issue was originally created as ARROW-15689. Please see the migration documentation for further details.}

[jorisvandenbossche]

I opened https://issues.apache.org/jira/browse/INFRA-24214 to ask for the rename, and #34218 to rename all occurences in our own repo (but short term those urls also get redirected by github)

[assignUser]

I have added a few follow ups. Tagging @kevingurney and @lafiona so they are aware where this is tracked and how it is moving forward!

[lafiona]

Thank you @jorisvandenbossche for requesting INFRA to switch the default branch name! And thank you @assignUser for tagging me. I've requested ownership of one of the follow up tasks to move things forward.

[kevingurney]

I've taken ownership of #31412 !

[jorisvandenbossche]

That reminds me that we also at some point need to ask INFRA for switching the branch name on the arrow-site repo as well. Are we ready to do that? Or are there still tasks to be done in preparation?

[assignUser]

It looks like the site workflows were already prepared to work with main instead of master. I don't see anything else depending on the default branch there but I am not super familiar with that repo so not 100% certain.

[kevingurney]

@jorisvandenbossche and @assignUser - I'll verify that there aren't any remaining dependencies in the arrow-site workflows as part of my investigation of #31412.

[kevingurney]

Following up on this discussion:

1. I verified that the apache/arrow-site deployment workflows in .github/workflows/deploy.yml are already prepared to work with "main" or "master". See line 23 and line 51. Note that the Jira issue ARROW-15988, which is referred to in the comments on these lines, is now being tracked as [Website] Remove support for triggering GitHub Actions apache/arrow-site website deployment workflow when pushing to master #31412 in GitHub Issues. However, this issue can't be addressed until the default branch has been renamed to "main". Therefore, I'll follow up with a PR to address this as soon as Apache INFRA has renamed the default branch of the apache/arrow-site repo to "main".

2. I also did an investigation of [Website] Remove mentions of master branch from Apache Arrow website content #20161, which I captured in this comment. From a CI / deployment workflow stand point, none of the uses of "master" listed in the comment linked to above should be blocking renaming the default branch of the apache/arrow-site repo to "main". However, since some of the hard-coded uses of "master" in the website content can't be removed until the default branch has been renamed, there will be a short period of time where some of the website content will be referring to "master" even when the default branch name has been changed to "main". This would be a temporary state, and I can create a PR to address [Website] Remove mentions of master branch from Apache Arrow website content #20161 as soon as the default branch has been renamed if the community is OK with this approach.

3. I also did a high level search for references to arrow-site in the apache/arrow code base and none of the references appear to be dependent on "master".

---

Summary:

Given the analysis above, I believe we are ready to request INFRA to rename the default branch for the apache/arrow-site repo to "main". Neither #20161 nor #31412 should be considered blocking from a website CI / deployment workflow stand point.

Once the rename is complete, I can follow up with PRs to remove remaining uses of "master" from the website content and deployment workflows, addressing #20161 and #31412.

[assignUser]

Nice!

there will be a short period of time where some of the website content will be referring to "master" even when the default branch name has been changed to "main".

The links should continue to work as github forwards them afaik. So I agree that we are good to go, I have opened an infra ticket.

[jorisvandenbossche]

https://github.com/apache/arrow-site is now also using main

[kevingurney]

With the rename complete, I'll proceed with addressing #20161 and #31412.

[kevingurney]

I've opened PR apache/arrow-site#323 to address #31412.