Backends

The following code owner backends are supported:

  • find-owners: Code owner backend that supports the syntax of the find-owners plugin.
  • proto: Code owner backend that supports a proto-based syntax. The proto syntax is not final yet and backwards incompatible changes are likely to happen. This is why this backend is experimental for now and should not be used in production.

Which backend is used can be configured globally, per repository or per branch.

Code owner config files

Code owner config files are stored in the source tree of the repository and define the code owners for a path.

The code owners that are defined in a code owner config file apply to the directory that contains the code owner config file, and all its subdirectories (except if a subdirectory contains a code owner config file that disables the inheritance of code owners from the parent directories).

In which files code owners are defined depends on the configured code owner backend:

BackendPrimary code owner config files
find-ownersOWNERS
protoOWNERS_METADATA

In addition, there can be secondary code owner config files, which may be imported into other code owner config files. These files have no effect on their own, but only when they are directly or indirectly imported into a primary code owner config file.

BackendSecondary code owner config files
find-owners<prefix>_OWNERS, OWNERS_<extension>
proto<prefix>_OWNERS_METADATA, OWNERS_METADATA_<extension>

Primary and secondary code owner config files are validated by the @PLUGIN@ plugin when they are changed to ensure that they are always parsable and valid.

By configuring a file extension for code owner config files it is possible to use a different set of code owner config files:

BackendPrimary code owner config filesSecondary code owner config files
find-ownersOWNERS.<configured-file-extension><prefix>_OWNERS.<configured-file-extension>, OWNERS_<extension>.<configured-file-extension>
protoOWNERS_METADATA.<configured-file-extension><prefix>_OWNERS_METADATA.<configured-file-extension>, OWNERS_METADATA_<extension>.<configured-file-extension>

In this case only the primary and secondary code owner config files with the configured file extension are considered as code owner config files. This means code owner config files without file extension or with other file extensions are not interpreted and also not validated.

Configuring a file extension for code owner config files is useful if a repository is forked and the fork should use a different set of code owner config files than the upstream repository. For this use case it is important that the code owner config files from the upstream repository are not validated, as they may use a different syntax or reference non-resolvable accounts, and hence would always be detected as invalid.

As some projects want to allow arbitrary file extensions for code owner config files, it is possible to enable arbitrary file extensions for code owner config files by configuration. If arbitrary file extensions are enabled, the following files are consideres as secondary code owner config files in addition to the once described above:

BackendAdditional secondary code owner config files
find-ownersOWNERS.<arbitrary-file-extension>
protoOWNERS_METADATA.<arbitrary-file-extension>

With this setup code owner config files with any file extension are validated.

NOTE: Enabling arbitrary file extensions for code owner config files conflicts with configuring a certain file extension for code owner config files in order to ignore upstream code owner config files, as in this case the upstream code owner config files should not be validated. This is why arbitrary file extensions should not be enabled if a certain file extension for code owner config files was already configured.

NOTE: Code owner config files can only import other code owner config files, but not arbitrary files. This ensures that code owner config files only import files that have gone through validation and hence are known to be valid. If arbitrary files could be imported, the imported files may not be parsable since they were not validated. As a result of this, looking up code owners could break, which would block submission of all changes for which the invalid file is relevant. This would likely be considered as a serious outage, hence importing arbitrary files is disallow so that this cannot happen.


Back to @PLUGIN@ documentation index

Part of Gerrit Code Review