Conventional commits usage and further definition

[philknows]

This issue follows up on #1773 and recently merged #5342 for further defining usage of conventional commit structure specifically for our Lodestar monorepo.

Reference comment: #5419 (review)

• Add structure: type(scope?): subject and brief explanations for each
• Define when to use each type
• Should we advocate for using scope in every PR title? If yes, how? The github actions checker currently does not enforce this usage.
• Modify examples of PR titles with bullet points instead of quoting
• Define when and how to indicate breaking changes on commit level and further discuss how to display those in the Changelog which should be targeted towards non-developer user/operator of Lodestar. (see comment)

[dapplion]

What's the purpose of (scope)? Having to pick a scope every time sounds annoying to me

[nflaig]

What's the purpose of (scope)?

I think scope could be an easy way to give the commit more context, especially because Lodestar monorepo has several packages and subcategories

• gives more clarity on what part of the code was modified, meaning the commit message itself does not need to do that and can focus on the change being done
• makes commit easier filterable/searchable, would only work if this patterns is consistently applied
• depending on how we generate changelogs, most tools group by scope, this makes changelog easier to read
• generally, it should be easy to define a scope for most commits based on
  • package (or subcategory)
  • hard fork
  • EIP

The scope is not required by the PR title checker as discussed here #5342 (comment), so usage would be based on personal preference. For me, it sometimes makes it easier to convey all required information in the commit message.

[philknows]

Completed by #5342