SOAP::Lite
Regression reproduced; a proposed fix is being validated in draft PR #91. No upstream submission or release yet.
View progressRegression reproduced; a proposed fix is being validated in draft PR #91. No upstream submission or release yet.
View progressPublished versions, source, and maintainers checked. Next: test the released artifacts on modern Perl. No transfer of maintenance permissions.
View progressUpstream DBIXS compatibility PR #32 merged on October 15 (UTC). CPAN release inclusion has not been confirmed.
View progressThe goal is to grow capable maintainers, rather than accumulate module ownership.
MetaProvides core, Package, and representative downstream validation (#43–#48) are complete. #49 build diagnostics continue; historical build success and artifact equivalence remain unverified. Log::Any::Adapter::Screen backup, fork detachment, and transfer retry request are complete; PERLANCAR’s response is pending. DBD::ODBC GitHub Actions PR #23 is merged; upstream DBIXS compatibility PR #32 is also merged. Prefer the latest tracking comments over the original README table below.
CPAN Rescue grows maintainers by maintaining real software.
New here? Make your first CPAN Rescue contribution →
No PAUSE account or CPAN release experience required. The smallest first task takes about 30–60 minutes.
Many useful CPAN distributions need maintenance. At the same time, becoming an open-source maintainer has a surprisingly high barrier to entry.
CPAN Rescue connects these two problems.
Contributors learn maintenance by working on real distributions: investigating current state, reproducing bugs, adding regression tests, improving CI and metadata, evaluating downstream impact, preparing releases, and eventually taking stewardship when appropriate.
The project still rescues abandoned and under-maintained Perl distributions, but adoption is not the primary goal.
The goal is to create more maintainers.
CPAN Rescue uses real maintenance work as a path into stewardship.
A contributor does not need to adopt a distribution or make a long-term commitment in order to participate. Responsibility grows only when the contributor wants it and has enough context to take it on safely.
The maintainer path and responsibility policy are documented in MAINTAINER_PATH.md and tracked in #28:
This is a path, not a ranking. Stopping after any stage is a valid contribution.
Maintenance is a skill that is difficult to learn from toy projects.
A maintainer needs to learn how to:
CPAN Rescue provides real distributions where those skills can be learned with review and bounded responsibility.
The first contribution should be small enough to understand and finish without accepting ownership of a distribution.
Good first maintenance tasks may include:
The first contributor-friendly tasks are now ready. If you want to try real CPAN maintenance, start with one of these:
To take a task, comment on the issue that you would like to work on it. No PAUSE account, adoption commitment, or promise to become a maintainer is required. Questions are expected, and maintainer review/mentoring is part of the task.
The task-design work is tracked in #29, and the first end-to-end external-contributor experiment is tracked in #36.
These are real maintenance tasks, not simulated exercises. Each has a concrete outcome, a validation method, clear responsibility boundaries, and a reviewer or mentoring path.
Becoming a maintainer should not mean accepting an indefinite obligation.
CPAN Rescue treats maintenance as stewardship rather than permanent ownership. Contributors should be able to increase responsibility gradually, and maintainers should be able to hand responsibility to someone else cleanly.
The concrete mentoring, responsibility-boundary, review, and handoff policy is documented in MAINTAINER_PATH.md and tracked in #30.
For high-impact distributions, especially XS modules or distributions high in the CPAN River, mentoring and review should become more conservative as potential downstream impact increases.
CPAN Rescue began by demonstrating that conservative rescue work can be carried through to real upstream and CPAN outcomes.
Current results:
SHINGO after own-suite, metadata, signature, and downstream validationThese results now serve a second purpose: the rescue workflow is becoming a practical curriculum that can be taught and handed to other maintainers.
This table is the working task board. Start with “Next action”: it should say exactly what can be done next. “Waiting for” means there is no useful action until that event happens.
| Distribution | State | Next action | Waiting for | Tracking |
|---|---|---|---|---|
| Dist::CheckConflicts | Adoption requested; upstream PR open | When PAUSE or upstream responds, confirm permissions or address PR feedback; then establish the conservative release baseline | PAUSE adoption/co-maint response and/or upstream review of PR #13 | #1 |
| Log::Any::Adapter::Screen | Adopted | Follow repository-maintenance guidance, then decide the next maintenance change | Response/guidance on upstream issue #5 | #2 |
| Devel::CallChecker | Released 0.010 | Monitor post-release results; investigate only if a regression appears | — | #3 |
| DBD::ODBC | Adopted by WHINDS; collaboration pending | Wait for Wesley Hinds' maintenance/release direction; if welcomed, upstream the focused DBIXS compatibility fix and contribute CI only where it complements existing upstream work. Revisit sustainability/funding in #40 after learning what support the current maintainer wants | Wesley Hinds' response | #25, #40 |
| Dist::Zilla::Plugin::MetaProvides family | Candidate | Start with the core distribution; reproduce its baseline and define downstream validation before moving to Package → Class → FromFile | — | #27 |
When resuming CPAN Rescue work, pick a row with a concrete Next action and no unresolved Waiting for dependency. The tracking issue holds investigation details and evidence; this README should stay short and answer only: where are we, what do we do next, and what are we waiting on?
On 2026-09-15, CPAN Rescue reached its first two successful adoptions:
SHINGO.SHINGO after reviewing a maintenance plan that explicitly accounts for its CPAN River position.Devel::CallChecker is a low-level compatibility layer around Perl call-checker APIs used by XS code. At adoption time, PAUSE admin Neil Bowers identified 13 direct dependent distributions and 81 distributions relying on it directly or indirectly.
Before 0.010, direct dependents were exercised against the release candidate across representative Perl versions, with zero candidate regressions relative to the reviewed 0.009 baselines. Devel::CallChecker 0.010 was released on 2026-09-19 and successfully indexed by PAUSE.
This workflow — baseline reconstruction, conservative changes, artifact validation, downstream testing, release, and post-release review — is the kind of real stewardship experience CPAN Rescue aims to make teachable.
A useful distribution for maintainer incubation should have real maintenance value while allowing responsibility to be divided into understandable steps.
Good candidates typically have several of these characteristics:
ADOPTME statusNot every rescue candidate needs to become an adoption candidate, and not every contributor needs to become its maintainer.
The existing conservative maintenance workflow remains the technical foundation:
The Active rescues table remains the canonical distribution dashboard. Each tracking issue should keep a **Status:** line near the top and record ownership/adoption state, upstream links, testing evidence, downstream impact, release details, and the next concrete action.
The rescue dashboard answers what needs maintenance.
The maintainer path answers who can learn to maintain it, and what responsibility they are ready to take next.
CPAN Rescue will continue to record technical outcomes such as:
But the primary long-term question is now:
Are more people becoming capable, confident maintainers?
As the incubation model develops, the project should also track:
These should describe experience and responsibility, not create a contributor leaderboard.
Completed foundations: #28 — maintainer path, #29 — good-first-maintenance tasks, and #30 — mentoring and handoff policy.
Growing maintainers also means making maintenance sustainable after someone takes responsibility.
For infrastructure distributions with significant downstream or production use, volunteer effort alone may not always provide enough continuity, compatibility testing, review capacity, or emergency maintenance. CPAN Rescue can therefore also explore grants, recurring sponsorship, fiscal hosting, and other ways to fund shared maintenance capacity while preserving maintainer-led technical governance.
DBD::ODBC is the first concrete case being explored. Issue #40 tracks the discussion, beginning with maintainer consent, current maintenance needs, production-user evidence, CI/compatibility requirements, and whether a small funding pilot would be useful.
The aim is not to turn every rescue into a funded project. It is to learn when funding is appropriate and how it can reduce single-person dependency without creating sponsor control or adding unreasonable obligations to maintainers.
Make maintaining CPAN worth doing — and make becoming a maintainer achievable.