It looks like the module actually loads faster with the embedded wasm file, despite it being larger and requiring base64-decoding.
Before 0697afd:
〉deno task bench
Task bench deno bench -A
cpu: AMD Ryzen 7 3700X 8-Core Processor
runtime: deno 1.35.3 (x86_64-pc-windows-msvc)
file:///C:/Users/Cody/code/dax/benchmarks/mod.bench.ts
benchmark time (avg) (min … max) p75 p99 p995
------------------------------------------------------ -----------------------------
loadWasmModule 83.62 ms/iter (81.72 ms … 86.68 ms) 83.93 ms 86.68 ms 86.68 ms
After 0697afd:
benchmark time (avg) (min … max) p75 p99 p995
------------------------------------------------------ -----------------------------
loadWasmModule 80.41 ms/iter (79.28 ms … 83.79 ms) 80.47 ms 83.79 ms 83.79 ms
This also removes the need to --allow-write or --allow-net:
#!/usr/bin/env -S deno run --allow-env --allow-read
// ⬆️ New 🎉
// old:
// #!/usr/bin/env -S deno run --allow-env --allow-read --allow-write --allow-net
import $ from "https://raw.githubusercontent.com/NfNitLoop/dax/0697afdbee91a866288214327fad04cc63a57b67/mod.ts"
await $`pwd`.printCommand()
This relates to conversation in #31
It looks like the module actually loads faster with the embedded wasm file, despite it being larger and requiring base64-decoding.
Before 0697afd:
After 0697afd:
This also removes the need to
--allow-write
or--allow-net
: