Fix PGCR route decode for plain JSON storage (API-1N)

[owens1127]

Summary

• Fixes API-1N (incorrect header check on /pgcr/:instanceId) by detecting gzip magic bytes before decompressing
• Services StoreRawJSON writes plain JSON to pgcr; legacy raw.pgcr rows remain gzip — matches log-raw-pgcr behavior
• Adds unit tests for both gzip and plain JSON payloads

Test plan

• bun test src/services/pgcr.test.ts
• CI green
• After deploy: spot-check /pgcr/<instanceId> for a known stored PGCR

Fixes API-1N

Made with Cursor

[barecheck]

Barecheck - Code coverage report

Total: 91.52%

Your code coverage diff: 0.01% ▴

✅ All code changes are covered

[sentry]

Bug: The code expects a Buffer from the BYTEA database column but receives a string. Passing this string to TextDecoder.decode() will cause a TypeError at runtime.
_{Severity: CRITICAL}

Suggested Fix

Configure a type parser for the BYTEA type (OID 17) in src/integrations/postgres/parsers.ts. This will ensure that BYTEA columns are correctly parsed into Buffer objects, matching the expected data: Buffer type annotation and preventing the runtime error.

Prompt for AI Agent

Review the code at the location below. A potential bug has been identified by an AI
agent. Verify if this is a real issue. If it is, propose a fix; if not, explain why it's
not valid.

Location: src/services/pgcr.ts#L12-L16

Potential issue: The `getRawCompressedPGCR` function expects the `BYTEA` data column
from the database to be a `Buffer`. However, the `node-postgres` driver is not
configured with a type parser for `BYTEA` columns (type 17), so it returns a
backslash-escaped string by default. This string is passed to `decodePgcrPayload`, which
then calls `TextDecoder.decode()`. Since `TextDecoder.decode()` does not accept a
string, it will throw a `TypeError`, causing a runtime crash whenever the application
attempts to retrieve a plain JSON PGCR.

Also affects:

• src/integrations/postgres/parsers.ts
• src/services/pgcr.ts:20-23

_{Did we get this right? 👍 / 👎 to inform future reviews.}