Implement GZIP decompression with deflate algorithm

Question

I need to decompress Base64-encoded gzip or deflate data from JavaScript without adding much to the bundle size (this is why I can't just use pako ). I found a nice, small library for decompressing deflate data called UZIP .

It appears that gzip is just deflate with a 10-byte header and an 8-byte footer. The data is guaranteed not to not have any corruption and not have any extra headers beyond those ten bytes. With all this in mind, I assumed that I could just Base64 decode the data, convert it to a Uint8Array , and slice off the header and footer.

const data = // some Base64-encoded, gzipped string
const dataBytes = new Uint8Array(atob(data).split('').map(char => char.charCodeAt(0)));
const decompressed = UZIP.inflate(dataBytes.slice(10, -8));

This code hangs indefinitely (presumably because UZIP has bad error handling). I tried using pako and zlib on both Node.js and the browser, but both complained about an invalid header. For reference, using zlib.gunzip after decoding from Base64, as well as running pako.inflate (which automatically detects the compression type), both properly decompress the gzipped data.

Is it possible to extract the raw deflate data from a gzip string? If not, are there any lightweight libraries that can decompress both gzip and deflate ?

Answer 1

I found that the deflate spec does not include headers, but zlib does, and many libraries call zlib decompression inflate , while calling true inflation inflateRaw . I changed my method calls from inflate to inflateRaw and it immediately worked.