fix(cli): skip the contrib_regression clone on pg-delta shadows

[avallete]

TL;DR

pg-delta shadow databases no longer run CREATE DATABASE contrib_regression TEMPLATE postgres. That clone crashes OrioleDB when the platform baseline contains an enum-indexed OrioleDB table, and the crash takes down the whole shadow. Migra still creates the database, because it uses it as the declarative diff target.

Before

flowchart LR
  shadow["pg-delta shadow"] --> clone["CREATE DATABASE TEMPLATE"]
  clone --> crash["OrioleDB crash"]

Loading

After

flowchart LR
  next["pg-delta shadow"] --> pg["postgres only"]
  migra["migra shadow"] --> clone["contrib_regression"]

Loading

Why

On an OrioleDB image, CREATE DATABASE … TEMPLATE segfaults when the template has an OrioleDB index on an enum column. A fresh Supabase database has one: auth.one_time_tokens_user_id_token_type_key. Shadow setup then fails with a connection error. pg-delta never connects to contrib_regression.

What changed

• The pg-delta declarative shadow applies the platform baseline and does not clone contrib_regression. A warm cache hit does not connect just to create it.
• The pg-delta migrations shadow applies migrations on postgres and skips the clone.
• The migra shadow still creates contrib_regression.

[github-actions]

🤖 AI Review

Both independent reviews completed. Code inspection supports the intended engine split: pg-delta skips contrib_regression, while legacy replay preserves it. Two Claude findings are confirmed as optional polish; the snapshot-ordering documentation finding is refuted. Codex reported no findings. No functional defect identified; tests were inspected but not executed.

Findings

Severity	Location	Category	Sources	Claim
⚪ NIT	apps/cli/src/command-internal/db-bootstrap/shadow-database.ts:660	api-design	claude	The new positional boolean makes the next-engine helper call less self-describing.
⚪ NIT	apps/cli/src/command-internal/db-bootstrap/shadow-database.ts:464	documentation	claude	The template helper's comment overgeneralizes that warm shadow-cache hits must still create contrib_regression. The targetUrlOverride comment also omits an explicit legacy-engine qualifier.

Findings outside the diff

• ⚪ NIT apps/cli/src/command-internal/db-bootstrap/shadow-database.ts:464 — The template helper's comment overgeneralizes that warm shadow-cache hits must still create contrib_regression. The targetUrlOverride comment also omits an explicit legacy-engine qualifier.

Refuted findings (kept for transparency, not posted as review comments)

• apps/cli/src/command-internal/db-bootstrap/shadow-database.ts:629 (documentation): The snapshotBaseline documentation is stale because it still describes running before the template database and user migrations, although pg-delta now skips the template database.
  Refuted: The shared interface documents ordering, without promising every consumer performs both later steps. Legacy replay still snapshots at line 695 before template creation at 714 and migrations at 716. Next-engine replay passes false at 760, omitting the template while preserving snapshot-before-migrations ordering. Returning after a setup-only snapshot does not contradict that invariant.

Stats

Claude findings: 3 · Codex findings: 0 · Confirmed: 2 · Refuted: 1 · Uncertain: 0

---

Models: claude-opus-5-5 + gpt-6.1-sol · Trigger: auto · Workflow run

This review runs once per PR. A maintainer can request another with a /ai-review comment.

[jgoux]

The shadow changes look correct. pg-delta shadows only use postgres, and the contrib_regression override is still limited to the legacy/migra path. One formatting issue in the test file fails the code-quality check and needs fixing before merge.