Release 0.24: Native Bun support

Today we're announcing the availability of the Primate 0.24 preview release. This release introduces native Bun support, enabling you to serve your applications significantly faster than before.

Bun is a new runtime based on JavaScriptCore and written in Zig. It has recently had its first stable version, and has seen an uptick in interest ever since.

Native Bun support

When Bun 1.0 came out, we were quick to test running Primate using Bun's NPM compatibility. As it turned out, running bun x primate works as expected and runs a Primate app without issues.

However, the real promise of Bun lies in its fast HTTP library implementation, Bun.serve. Bun's NPM compatibility is great, but to get real speed gains over Node, we have to use its native standard library.

As it turns out, Primate already uses a standard library of its own, rcompat. As Primate is not the only project using it, we decided to implement native Bun support into it and thus have Primate and other projects benefit from it. This release uses the latest release of rcompat, which uses Bun natively wherever possible and falls back to Node/NPM otherwise.

In addition to the core package, we have also patched @primate/store to use Bun's native bun:sqlite module.

Benchmarks

To illustrate the difference in serving Primate with native Bun, we created a threeway benchmark: running Primate with NPM (npx primate), running it with Bun's NPM compatibility (bun x primate), and running it natively with Bun (bun --bun x primate).

Using the ab testing tool, we tested two scenarios, serving plain-text content and an SSR-rendered Svelte page.

Our command for benchmarking was ab -c 350 -n 100000 [url]. That is, send 100000 requests with a concurrency of 350 requests at a time.

The app we used for testing was the Primate template app. To replicate the benchmarks, clone that repository and run the app.

In the following benchmarks, the quantities are:

• TT Time taken for tests (seconds)
• RPS Requests per second
• TPR Time per request (milliseconds) (mean, across all concurrent requests)
• TR Transfer rate, in Kbytes/second

Plain-text

Command used ab -c 350 -n 100000 http://localhost:6161/benchmark

Svelte

Command used ab -c 350 -n 100000 http://localhost:6161/benchmark/svelte

Caveats

In upcoming releases, we'll continue to improve Primate's support for Bun, using native modules and functions wherever possible. Currently, if you rely on @primate/ws for WebSockets in your application, you would have to stick to NPM. Bun's native WebSocket implementation is radically different from the ws package which @primate/ws uses, and unifying them was deemed out of scope for this release. Please reach out if you specifically need WebSockets in your application so we can better prioritize our work.

Other changes

Consult the full changelog for a list of all relevant changes.

Next on the road

Some of the things we plan to tackle in the upcoming weeks are,

• Add projections and relations to stores
• Multidriver transactions
• Introduce IDE TypeScript support
• Add a command hook that would allow modules to register command line namespaces, to be able to run npx primate [namespace] [command] [flags]
• Use this new hook to create database migrations for SQL-flavored databases
• Extend create-primate with the ability to add new routes
• Add hydration and liveview support for @primate/vue
• Support the multipart/form-data content type
• Flesh out stores with default values, additional predicates and relations between tables/collections
• Add more type variants

This list isn't exhaustive or binding. None, some or all of these features may be included in 0.25, and other features may be prioritized according to feedback.

Fin

If you like Primate, consider joining our channel #primate on irc.libera.chat.

Otherwise, have a blast with the new version!