@basenative/wrangler-preset

Pin once, every project moves together.

A pinned Wrangler version plus a small, dependency-free fragment generator for wrangler.toml. Designed to keep every DuganLabs repo (basenative, duganlabs, greenput, pendingbusiness, ralph-station, t4bs, warren-sys, warrendugan) on the same Wrangler at the same time.

Why this exists

Across the org, projects had Wrangler versions ranging from ^3.114 to ^4.85. Drift caused:

• Inconsistent compatibility_date semantics.
• Different defaults for nodejs_compat and friends.
• Each repo re-deciding how to wire D1 / KV / R2.
• Wrangler upgrade pain spread out over weeks.

This package pins one Wrangler version, exposes one set of recommended flags, and ships a tiny, typed builder so configs look the same in every repo.

Install

pnpm add @basenative/wrangler-preset
# wrangler is pulled in transitively at the version this preset pins.

You no longer need a direct wrangler dependency in your repo. Use bn-wrangler (or pnpm exec bn-wrangler) and you'll always get the pinned version.

CLI

# Pinned wrangler — pass through any wrangler args:
bn-wrangler --version
bn-wrangler pages dev dist
bn-wrangler deploy
bn-wrangler d1 migrations apply DB --remote

# Preset diagnostics:
bn-wrangler --bn-version
bn-wrangler --bn-help

Programmatic API

defaults

Recommended baseline. Spread it into your config:

import { defaults, mergeWrangler, toToml } from '@basenative/wrangler-preset';

const config = mergeWrangler(
  { name: 'my-app', main: 'src/index.js' },
  defaults,
);

console.log(toToml(config));

bindings.d1(name, dbName, dbId) / kv(name, id) / r2(name, bucket) / do(name, className, scriptName?)

Each returns a small fragment object you can mergeWrangler(...) into a config:

import { defaults, bindings, mergeWrangler, toToml } from '@basenative/wrangler-preset';

const cfg = mergeWrangler(
  { name: 'my-app', main: 'src/index.js' },
  defaults,
  bindings.d1('DB', 'my-app-db', '<uuid>'),
  bindings.kv('CACHE', '<kv-id>'),
  bindings.r2('ASSETS', 'my-app-assets'),
  bindings.do('ROOM', 'Room'),
);

await fs.writeFile('wrangler.toml', toToml(cfg));

mergeWrangler(base, ...frags)

Deep-merges scalars (later wins) and concatenates array-of-tables (d1_databases, kv_namespaces, r2_buckets, durable_objects.bindings).

toToml(config)

Zero-dependency TOML serializer covering the subset Wrangler uses: scalars, arrays of scalars, tables, and arrays of tables.

Generate from the template

cp node_modules/@basenative/wrangler-preset/templates/wrangler.toml.tmpl wrangler.toml
# then edit __PROJECT_NAME__ etc.

Adoption checklist (per repo)

1. Remove any direct wrangler dep from the repo's package.json.
2. pnpm add @basenative/wrangler-preset.
3. Replace direct wrangler ... calls in scripts with bn-wrangler ....
4. Set compatibility_date from defaults.compatibility_date (or copy the template).
5. Bump everyone together by upgrading this package: one PR per quarter, not eight.

Links

• BaseNative monorepo — https://github.com/DuganLabs/basenative
• Issues — https://github.com/DuganLabs/basenative/issues
• Docs — https://github.com/DuganLabs/basenative#readme

License

Apache-2.0