Stream ends prematurely

[GerwinBosch]

The parser stops without an error when it only parsed a chunk. Example file

Example code

const bz2 = require("unbzip2-stream");
import * as fs from "fs";

async function readBzip(inFile, outFile)
  const streamIn = fs.createReadStream(inFile);
  streamIn
    .on("error", (e: any) => console.error("input Error", e))
    // .on("data", (data: any) => console.log(data))
    .on("close", () => console.log("input CLOSE"))
    .on("end", () => console.log("input end"))
    .on((readable: any) => console.log(" input readable"));
  const streamOut = fs.createWriteStream(outFile);
  streamIn
    .pipe(bz2())
    .on("error", (e: any) => console.error("bzip Error", e))
    // .on("data", (data: any) => console.log(data))
    .on("close", () => console.log("bzip close"))
    .on("end", () => console.log("bzip end"))
    .on("readable", (readable: any) => console.log(" bzip readable",readable))
    .pipe(streamOut);
  return new Promise((resolve, reject) => strmOut.on("end", () => resolve()).on("error", () => reject()));
};

Notice that the end event of bzip is thrown almost immediately

[regular]

I tried your example file with the code from README, and indeed, it stops after outputting roughly 1 MB (1048576 bytes to be precise).

[regular]

Turns out your example file is the concatenation of multiple, independently compressed files. That's legal and the original bunzip2 (the cli binary) has no trouble decoding such files. It's simply a feature that is not implemented here. The good news: it is currently being implemented.

[regular]

Fixed in unbzip2-stream@1.3.0