view8-rs

A fast V8 bytecode decompiler written in Rust. Parses the text output of --print-bytecode and reconstructs a pseudo-JavaScript representation of each function.

Why Rust?

The original View8 Python tool works well but struggles with large V8 dump files. For a file with ~308K bytecode lines:

Tool	Peak RAM
Python view8	~900 GB (GC + dicts + string copies)
view8-rs	~30 MB (flat structs, no GC)

Install

cargo install --path view8-rs

Or build manually:

git clone https://github.com/awasthi21/view8-rs
cd view8-rs/view8-rs
cargo build --release
# binary at: target/release/view8_rs

Usage

First, generate a V8 bytecode dump:

node --print-bytecode your_script.js > bytecode.txt 2>&1

Then decompile it:

view8_rs bytecode.txt output.js

Example output

function func_start_0x7f1a2b3c(a0, a1) {
    ACCU = a0
    r0 = ACCU
    ACCU = r0["length"]
    if (!ACCU) goto +12
    ACCU = r0.call(r1, ...)
    return ACCU
}

How it works

1. Parse — reads the entire dump into memory line-by-line, resolving FixedArray and SharedFunctionInfo blocks
2. Translate — maps each V8 opcode to a readable pseudo-JS expression (accumulator model)
3. Resolve — replaces ConstPool[N] references with their actual string/value from the constant pool
4. Export — writes one function block per SharedFunctionInfo to the output file

Dependencies

• regex — bytecode line parsing
• lazy_static — compiled regex pool
• memmap2 — fast file I/O

License

MIT