Ru Strings Bundle Laravel Package

Technical Evaluation

Architecture fit. The package is a Symfony2 bundle, not a Laravel package, so the "fit" question is really about which layer a TPM can salvage. It ships three artifacts — a kernel-registered Bundle class, a Symfony DI service definition, and a Twig extension exposing declension/pluralization filters — and Laravel can consume none of them natively. Laravel's analogues are a ServiceProvider, Illuminate\Container, and Blade directives/helpers. The one durable asset is the framework-agnostic string logic: Russian grammatical case declension and numeric concord (1 день / 2 дня / 5 дней). For planning purposes, treat the deliverable as reference source code under MIT, not a dependency.

Integration feasibility. Low as an install, moderate as an extraction. The viable path is bounded and well understood: isolate the declension/concord classes, strip Symfony and Twig references, re-home them behind a Laravel service provider, and expose a Blade directive plus a helper. A mid-level Laravel developer can do this in one to two days. What makes it a fork-and-own decision rather than a procurement decision is that no upstream will ever merge our patches.

Technical risk. High, and concentrated in maintenance and linguistic correctness rather than security. The repo has 0 stars, 0 dependents, a Russian-only description, and no visible community — no issue triage, no CVE feed, no external validation of the grammar tables. Symfony2 went EOL in 2015, so the code is likely PHP 5.3–5.6 era and may use constructs removed in PHP 7/8 (each(), create_function(), PSR-0 autoloading). MIT keeps the legal friction near zero, which is exactly why extraction is the right framing. The unresolved risk is whether the rules are correct: Russian has well-known traps at 11–14, on zero and fractions, and on gendered and animate nouns, and there is no test suite to tell us.

Key questions.

• What PHP constraint is in composer.json, and does the code run under PHP 8.2+?
• Is there any test suite at all, and does it cover 11–14, zero, and fractional counts?
• Does the package separate case declension (именительный → родительный → дательный …) from numeric concord, or conflate them? Concord is the commodity half; declension is the expensive half.
• Is declension algorithmic (suffix rules) or a lookup dictionary? A dictionary silently caps our coverage.
• Who on the team can give native-speaker linguistic sign-off, and what is that reviewer's cost per release?

Integration Approach

Stack fit. Poor on every axis except the string logic: Blade vs. Twig, Illuminate\Translation vs. Symfony's TranslatorInterface, and a container that will not read Symfony bundle DI config. Installing this verbatim would drag Twig and Symfony DI into a Laravel app — an unacceptable footprint for a filter that produces one line of Russian text.

Migration path. Two routes, neither of which ends in composer require it-blaster/ru-strings-bundle:

1. Extract — copy the declension/concord classes into app/Support/RussianStrings/ or a private Composer package, strip framework references, write tests, register a RussianStringsServiceProvider, and expose @ru_decline and ru_plural($count, $forms).
2. Replace — use trans_choice with Laravel's range selectors ({0} … |{1} … |[2,4] … |[5,*] …) for numeric concord, which eliminates the dependency for the majority of use cases. Only genuine case declension of proper nouns still needs external logic.

A TPM should sequence these as a decision gate, not a parallel bet: prove the replace path is insufficient before funding the extract path.

Compatibility. Expect friction and budget for it: PHP-version deprecations, PSR-0-era autoloading that may not resolve under Composer 2, and no service provider to register. Nothing here is drop-in, and the integration estimate should carry an explicit "unknown unknowns in the source audit" line item.

Sequencing.

1. Source audit — read the classes, confirm algorithm vs. dictionary, read composer.json constraints. (Half a day.)
2. Decision gate — if only numeric concord is needed, stop and use trans_choice. Close the ticket.
3. If declension is required, extract the minimal classes into a first-party, MIT-attributed package.
4. Build a regression corpus: nouns across all three genders and animacy, plus 0 / 1 / 2–4 / 5–20 / 11–14 / 21 / 1.5.
5. Wrap in a service provider and Blade directive; document usage in English.
6. Native-speaker sign-off on the corpus before any user-facing copy ships.

Licensing and attribution. MIT permits the fork. TPM action: retain the original copyright notice in the extracted files and record the provenance in the repo README so a future engineer knows why this code exists and where it came from.

Operational Impact

Maintenance. Ownership transfers to us in full at the moment of fork. There is no upstream to rebase against, so every PHP minor upgrade and Laravel major bump becomes an internal ticket. Size it as a small internal utility with a named owner, not as a dependency with a vendor — and put it on the same review cadence as other first-party support code.

Support. Effectively zero. With 0 stars and 0 dependents there is no community to escalate to, no security feed, and no guarantee anyone is watching the repository. Every production defect in Russian copy becomes an internal bug with an internal SLA. A TPM should state this plainly in the adoption record so it is not rediscovered during an incident.

Scaling. A non-issue technically. This is in-memory string manipulation with negligible CPU and memory; it will not be a throughput bottleneck at any realistic request volume, and per-row use in list rendering does not require caching. The scaling risk is correctness at scale: a wrong form ships silently to thousands of rendered pages before anyone notices, and the blast radius grows with traffic.

Failure modes.

• Wrong form at the 11–14 and teen boundaries ("11 яблок" vs. "11 яблока") — visibly broken Russian on high-traffic surfaces.
• Zero and fractional counts falling through to a default form.
• Case declension failing on proper nouns, indeclinable surnames, hyphenated or non-Cyrillic input — with no error raised, only bad output.
• PHP-version breakage on upgrade, surfacing as fatals if code is adopted verbatim.
• Blade directive name collision or escaping issues if wired carelessly.

Ramp-up. Low for engineering, high for the organization. A mid-level Laravel developer can extract and wire the logic in a day or two; the underestimated cost is validation, which the package does nothing to reduce. The gate is a native-Russian reviewer signing off on a regression corpus, and that reviewer's availability — not engineering hours — is the real schedule constraint. TPM recommendation: do not adopt the bundle. Satisfy numeric concord with Laravel's native trans_choice, and only fork the declension logic if personalized Russian copy is a funded roadmap item — in which case fund the fork and the linguistic review together, as one workstream.