Laravel Firebird Laravel Package

Product Decisions This Supports

• Build vs. buy for Firebird access: a correct, maintained Firebird driver for Laravel is weeks of grammar/compiler work plus perpetual catch-up with Laravel's annual majors. This package makes "integrate" the obvious call and keeps the team focused on product surface instead of SQL dialects.
• Modernize-without-migrate strategy: Firebird underpins a lot of vertical/on-prem software (POS, ERP, accounting, municipal, healthcare, Delphi/VB estates). This driver is what lets Laravel serve as the modernization layer — APIs, self-service portals, admin tooling, integrations — on top of a system of record we are not going to re-platform this year.
• Unblocks table-stakes features: insertGetId() via RETURNING means Eloquent model creation and auto-increment keys behave normally, which unlocks standard CRUD, factories/seeders, and anything assuming generated IDs. The new date/time where clauses unblock date-range filtering and dashboards; lateral joins and improved union limit/offset unblock reporting and analytics screens; multi-column distinct and corrected standard offset … rows / fetch first … rows only pagination fix known correctness bugs.
• Forces a roadmap conversation: v4 is a compatibility reset bound to PHP ^8.2, Laravel 12/13, and Firebird 4/5. Adopting it is a platform-calendar commitment, not a Composer bump — the upgrade plan and the feature roadmap need to be sequenced together.
• Risk posture: v4.0.0-beta.0 is explicitly pre-final. Treat it as an evaluation/planning input; do not put it behind a customer-facing commitment until 4.0.0 ships.

When to Consider This Package

Consider it when:

• You have a Firebird system of record and want Laravel for new surface area without a data migration.
• Your feature backlog depends on pagination, date-range filtering, aggregations, or reporting queries against Firebird — v4 closes most of the historical gaps here.
• You are already on PHP 8.2+, Laravel 12/13, and Firebird 4 or 5, or can finance that triple upgrade within a reasonable window.
• You need Eloquent-style model persistence with generated keys rather than hand-rolled sequences and triggers.

Look elsewhere when:

• You are pinned to Firebird 2.5/3.x, PHP <8.2, or Laravel <12. This is the single biggest gating factor in v4 — stay on the 3.x line and budget the upgrade deliberately.
• You need a GA dependency today. Wait for 4.0.0 final or pin 3.x; a beta driver is a poor foundation for a launch date.
• The project is greenfield with no Firebird estate — use MySQL or Postgres, which Laravel supports natively with no third-party maintenance risk.
• You require a vendor SLA or a named support contract; this is a community package, so assess maintainer bus factor and your internal ability to patch the grammar.
• Your logic lives in Firebird stored procedures and triggers that Eloquent can't express — the driver is not the constraint either way.

How to Pitch It (Stakeholders)

Executives: This is the connector that lets our Laravel work talk to the Firebird database we already run, so we can build new apps on existing data instead of funding a migration. The new 4.0 line brings the driver onto current Laravel and PHP and closes the gaps that made ordinary features — creating records, pagination, filtering, reporting — awkward or unreliable. It is in beta, so this is a planning input, not a commitment: we should schedule the PHP/Laravel/Firebird upgrades now so we are ready the day 4.0 goes final.

Engineering: v4.0.0-beta.0 is a compatibility reset, not just a feature drop. Baseline is PHP ^8.2, Laravel 12.x/13.x, Firebird 4 and 5; support for PHP <8.2, Laravel <12, and older Firebird versions is dropped, so plan a coordinated upgrade or stay on 3.x. On the query side we gain insertGetId() using RETURNING (proper model persistence), lateral joins, improved union compilation including union limit/offset, date- and time-based where clauses, additional comparison operators for newer Firebird versions, multi-column distinct, and a switch to standard limit/offset syntax (offset … rows, fetch first … rows only). It also fixes exists alias quoting that was causing case-conversion problems and rewrites drop if exists so it no longer depends on legacy table-exists compilation. The surface area is broad — validate against a Firebird 4/5 replica, with specific attention to pagination and model creation, before promoting anything off beta.