Hi everyone,

Just adding a small follow-up to my earlier proposal email.

I'm checking that the mailing-list conversation is working correctly and
that replies are being delivered as expected.

Thanks,
Tarun


On Fri, 6 Nov 2026 at 21:59, Tarun N <[email protected]> wrote:

> Hi everyone,
>
> I'm Tarun, a CS student planning to apply for GSoC 2027 with Apache Maven.
> I wanted to share my proposal idea early and get community input before
> writing the full application.
>
> The proposal: Add a diagnostic capability to Maven that answers the
> question: "Why was this project included in or excluded from the reactor
> build?"
>
> Current situation:
>
> Users working with multi-module Maven projects often use options like -pl,
> -am, -amd, and -rf to control reactor builds. When these options are
> combined, it can be difficult to understand exactly why certain projects
> are included or excluded from the build. Currently, users often have to
> inspect the project structure and reactor behavior manually to understand
> the selection.
>
> Proposed solution:
>
> Introduce an opt-in diagnostic mode that explains reactor project
> selection decisions.
>
> For example:
>
> $ mvn install -pl module-a -am -Dreactor.selection.verbose
>
> Reactor Selection Analysis:
>
> Included projects:
>   - root-project (required as aggregator)
>   - module-a (explicitly selected)
>   - module-b (required dependency of module-a)
>
> Excluded projects:
>   - module-c (not selected)
>   - module-d (not required by selected projects)
>
> Selection reason:
>   module-a was explicitly selected using -pl.
>   module-b was included because of -am.
>
> Implementation approach:
>
> - Investigate the existing reactor selection flow, particularly
> DefaultGraphBuilder and related project graph components.
> - Trace how options such as -pl, -am, -amd, and -rf affect project
> selection.
> - Capture the reasons behind project inclusion and exclusion.
> - Present the selection decisions in a human-readable diagnostic output.
> - Ensure the feature does not change existing reactor selection or build
> behavior.
>
> I am proposing this as a diagnostic feature rather than a change to
> Maven's reactor logic. The goal is to make complex multi-module builds
> easier to understand and debug.
>
> I'd appreciate:
>
> 1. Feedback on the scope — too big, too small, or about right for a
> 12-week project?
> 2. Whether this idea overlaps with existing Maven 4 work, discussions, or
> JIRA issues.
> 3. Whether any Apache Maven committer or experienced contributor would be
> interested in mentoring me on this proposal for GSoC 2027.
> 4. Any suggestions for refining the implementation approach or identifying
> a suitable starting point in the codebase.
>
> As I'm still learning some parts of Maven's reactor internals, I would
> particularly appreciate guidance from a potential mentor who could help me
> refine the scope, understand the relevant architecture, and guide the
> implementation.
>
> Thank you for reading. Happy to discuss the idea further or share a more
> detailed write-up.
>
> Best regards,
> Tarun
> https://github.com/Tarun4201
> Sai Vidya Institute of Technology, India
>

Reply via email to