If a pull request is created, we could look at the list of impacted files and extract stacks based on root subdirectories included in the list. So for example if the list of impacted files is:
api/mix.exs
api/lib/foo/bar.ex
The stacks used to fetch experts/learners for the pull request would be elixir and graphql.
The rest of the process would stay the same — for example contributors for the whole repository would be considered, not “contributors in the api subdirectory”.
Currently, we configure the stacks used for pull requests created in a repository with a comma-separated list of stacks:
For repositories containing multiple applications however, it would be great to specify stacks for each root subdirectory:
If a pull request is created, we could look at the list of impacted files and extract stacks based on root subdirectories included in the list. So for example if the list of impacted files is:
api/mix.exs
api/lib/foo/bar.ex
The stacks used to fetch experts/learners for the pull request would be
elixir
andgraphql
.The rest of the process would stay the same — for example contributors for the whole repository would be considered, not “contributors in the
api
subdirectory”.