Nurul Huda (Apon) 25c3630547
refactor(migration): Zig 0.16.0 (#116)
Few regressions bugs are there but merging to main to continue development on main and will make release only after all known bugs related to Zig 0.16.0 has been fixed.

- [x] Upstream http.zig - Core web server and routing - [PR](https://github.com/karlseguin/http.zig/pull/191)
- [x] Upstream zli - CLI Parser - remove dependancies and use [proposed std.cli](4b4d625ca9/lib/std/cli.zig)
- [x] Upstream zig-tree-sitter - [PR](https://github.com/tree-sitter/zig-tree-sitter/pull/40)
- [x] Core Ziex Language
- [x] Core Ziex app
- [x] Core Ziex CLI
- [x] Ziex Dev Server (zx dev)
- [x] LSP
- [x] Playground
    - [x] Upgrade to Zig 0.16.0
    - [x] Fix WASM Regression
- [ ] Fix regressions
    - [ ] Runtime: File Form
    - [x] CLI: Export
    - [x] CLI: Bundle
    - [ ] Runtime: Websocket
    - [ ] DevServer: Error Overlay
    - [x] WASI Build
2026-05-29 09:22:04 +06:00
..
2026-03-29 20:01:24 +06:00
2026-02-13 01:41:13 +06:00
2026-03-05 16:20:25 +06:00
2026-03-05 14:32:07 +06:00
2026-02-13 05:05:01 +06:00
2026-03-05 16:20:25 +06:00

Benchmark

Compares SSR performance of Ziex, Leptos, SolidStart, and Next.js in Docker containers limited to 2 CPUs and 2 GB RAM.

Prerequisites

  • Docker

Usage

Clone the repository and run the benchmark script:

cd bench

# all frameworks
bash run.sh

# specific framework (e.g., Ziex)
bash run.sh ziex

Results are written to result.csv and ../site/pages/bench.zon (used to generate the benchmark chart on ziex.dev).

Measures:

  • Requests per second (req/s)
  • p50 latency
  • p99 latency
  • Idle memory (MB)
  • Peak memory (MB)

Frameworks

Framework Port Stack
Ziex 3003 Zig, native binary
Leptos 3002 Rust, Actix-web
SolidStart 3001 Bun, Vinxi
Next.js 3000 Bun, React 19