[jira] [Updated] (OAK-7425) Add discovery mechanism for tooling implementations

["Davide Giannella (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/OAK-7425?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Davide Giannella updated OAK-7425:
----------------------------------
    Fix Version/s: 1.14.0

> Add discovery mechanism for tooling implementations
> ---------------------------------------------------
>
>                 Key: OAK-7425
>                 URL: https://issues.apache.org/jira/browse/OAK-7425
>             Project: Jackrabbit Oak
>          Issue Type: Improvement
>          Components: segment-tar
>            Reporter: Francesco Mari
>            Assignee: Francesco Mari
>            Priority: Major
>              Labels: technical_debt
>             Fix For: 1.12.0, 1.14.0
>
>         Attachments: 001.patch
>
>
> This issue proposes an idea for discovering implementations of tooling for the Segment Store. Developing a tool for the Segment Store should include the following step.
> * The tool compiles against the {{NodeStore}} API and the API exposed through the oak-segment-tar-tool-api. In particular, the tool uses the {{ToolingSupportFactory}} and related interfaces to instantiate a NodeStore and, optionally, a {{NodeState}} for the proc tree.
> * The tool runs with an implementation-dependent uber-jar in the classpath. The uber-jar includes the {{ToolingSupportFactory}} API, its implementation, and every other class required for the implementation to work. No other JARs is required to use the {{ToolingSupportFactory}} API. The tool uses the Java's {{ServiceLoader}} to instantiate an implementation of {{ToolingSupportFactory}}. The uber-jar is the {{oak-segment-tar-tool}} module.
> The patch falls short of fully implementing the use case because {{oak-segment-tar-tool-api}} is not versioned independently from Oak. This can't happen at the moment because {{oak-store-spi}} and its dependencies are not independently versioned either. The workflow described above could still work, but only because the {{NodeStore}} and {{NodeState}} API are quite stable. A cleaner solution to dependency management is required in the long run.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)