Q on WebAssembly (wasm-exec, WAGI)

Q — the toy language whose interpreter RunR runs as a container-based runtime — on WebAssembly: the same interpreter (examples/runtimes/q/interpreter, Go, standard library only) compiled with TinyGo to a WASI 0.2 command component and run on Spin's WAGI executor in CGI mode (q -cgi /main.q: the request in the environment and on stdin, the response as CGI on stdout). The script is the same five lines; it answers the three triggers.

In RunR: a runtime from an uploaded interpreter#

1. The region needs WebAssembly support (Settings → Regions → (region) → Isolation → WebAssembly).

2. Compile the interpreter and upload it into the region (Settings → Images, or the CLI):

   cd examples/runtimes/q/interpreter
   tinygo build -target=wasip2 -no-debug -o q.wasm .     # TinyGo 0.42 with wasm-tools
   runr images push q.wasm --region eu-1 --name q-interpreter:1.0

   Do this as a platform administrator (outside an organization): platform runtimes use platform uploads. Upload it into every region the runtime is offered in.

3. Register the runtime — Settings → Runtimes → Add runtime with runtime.yaml, runr runtimes register -f examples/wasm/q-exec/runtime.yaml from the repository root, or the API from examples/wasm/q-exec:

   jq -n --rawfile manifest runtime.yaml '{manifest: $manifest}' > body.json
   curl -sS -X POST "$API/runtimes" -H "Authorization: Bearer $TOKEN" \
     -H 'Content-Type: application/json' -H 'Accept: application/json' -d @body.json

   Registration checks that q-interpreter:1.0 is uploaded in each of the runtime's regions.

4. Create a project on Q (Wasm), keep main.q, deploy, call it:

   curl "https://<project>.<region>.<domain>/?name=RunR"
   # Hello, RunR! (from Q)

Every build packages the uploaded interpreter with the project's main.q (mapped read-only into the component) into a Spin app — no compiler runs per function.

In RunR: a wasm-exec project#

Without registering a runtime, a FaaS project on WASI (Wasm exec) can carry the interpreter itself:

• with the uploaded module: wasm.upload (one line, q-interpreter:1.0), main.q, and a spin.toml whose executor runs the interpreter — executor = { type = "wagi", argv = "q -cgi /main.q" }, source = "main.wasm", files = ["main.q"];
• built from source: the interpreter's .go files and go.mod next to main.q (the build runs TinyGo), plus the same spin.toml.

Locally#

spin build && spin up      # Spin 4.1 or later, TinyGo 0.42 with wasm-tools
curl "http://127.0.0.1:3000/?name=RunR"
# Hello, RunR! (from Q)