mgoerens
commented
5 months ago @komish I removed the DRAFT status as I believe this is now mature enough to be reviewed again and eventually merged with the current set of capabilities. I intend to continue iterate over this work, but this PR is already long enough, and I'd rather create smaller PR in the future.
This PR has seen significant changes since you reviewed it first:
- Most impacting change is that I account for another type of "Submission", when a user submits only the OWNERS file.
- Consequently, I added the
is_valid_certification_submissionandis_valid_owners_submissionmethods. - Reworked tests
This is the first step towards solving #296. The code added by this commit is currently not put in use in the certification pipeline and is a partial duplicate of the "checkprcontent" package.
The Submission data structure can collect and run early validation checks on all information concerning a user's submission (i.e. a GitHub Pull Request). In its current state, given an api_url (link to a given PR), it can collect the list of modified files, and extract the category, organization, chart's name and version, as well as running basic checks, such as SemVer compatibility of the provided version.
Eventually, this data structure is meant to encapsulate more information concerning the user's submission, and be dumped to a json file (using JSONEncode / JSONDecoder, more on that in a later PR) to be passed around the certification pipeline jobs as an artifact. This early (and incomplete) version is intended to trigger discussion around the design choices and general direction taken to solve #296.