Crates.io | weval |
lib.rs | weval |
version | 0.3.3 |
source | src |
created_at | 2022-12-10 02:18:58.36201 |
updated_at | 2024-11-20 23:08:06.220199 |
description | The WebAssembly partial evaluator |
homepage | |
repository | https://github.com/bytecodealliance/weval |
max_upload_size | |
id | 733670 |
size | 405,128 |
weval
partially evaluates WebAssembly snapshots to turn interpreters into
compilers (see Futamura
projection
for more).
weval
binaries are available via releases on this repo or via an npm
package.
Usage of weval is like:
$ weval weval -w -i program.wasm -o wevaled.wasm
which runs Wizer on program.wasm
to obtain a snapshot, then processes any
weval requests (function specialization requests) in the resulting heap image,
appending the specialized functions and filling in function pointers in
wevaled.wasm
.
See the API in include/weval.h
for more.
Bump the version in Cargo.toml
and cargo check
to ensure Cargo.lock
is
updated as well.
Bump the tag version (TAG
constant) in npm/weval/index.js
.
Bump the npm package version in npm/weval/package.json
.
Run npm i
in npm/weval/
to ensure the package-lock.json
file is
updated.
Commit all of this as a "version bump" PR.
Push it to main
and ensure CI completes successfully.
Tag as v0.x.y
and push that tag.
cargo publish
from the root.
npm publish
from npm/weval/
.
The theory behind weval is described in the author's blog post here, covering partial evaluation and Futumura projections as well as how weval's main transform works.
weval is in use to provide ahead-of-time compilation of JavaScript by wevaling a build of the SpiderMonkey interpreter, providing 3-5x speedups over the generic interpreter. Please let us know if you use it elsewhere!