While talking to a web server on an embedded device, asking for gzip content encoding I get responses that are considered OK by all major browsers (Chrome, Firefox, IE) and also cURL. However, the Node zlibmodule considers the gzip stream incomplete and throws an error. The error returned by the zlib module reads: Error: unexpected end of file at Zlib._handle.onerror (zlib.js:363:17). Inspecting the gzip response it seems to be missing a footer detailing the CRC and the length of the compressed data.
A small program that reproduces this with an actual response from the device in it:
var buffer = require("buffer");
var stream = require("stream");
var fs = require("fs");
var zlib = require("zlib");
var s = new stream.Readable();
var b = new buffer.Buffer([
0x1F, 0x8B, 0x08, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x0B, 0xB2, 0xB1,
0xAF, 0xC8, 0xCD, 0x51, 0x28, 0x4B, 0x2D, 0x2A, 0xCE, 0xCC, 0xCF, 0xB3,
0x55, 0x32, 0xD4, 0x33, 0x50, 0x52, 0x48, 0xCD, 0x4B, 0xCE, 0x4F, 0xC9,
0xCC, 0x4B, 0xB7, 0x55, 0x2A, 0x2D, 0x49, 0xD3, 0xB5, 0x50, 0x52, 0xB0,
0xB7, 0xE3, 0xE5, 0xB2, 0x49, 0xCC, 0x49, 0x2C, 0xCA, 0x2D, 0x56, 0x48,
0x4E, 0x4C, 0xCE, 0x48, 0xCD, 0x4C, 0xB1, 0x55, 0xB2, 0x54, 0x52, 0x50,
0xD0, 0x07, 0x4A, 0x00, 0x00, 0x00, 0x00, 0xFF, 0xFF
]);
s.push(b);
s.push(null);
s.pipe(zlib.createGunzip()).pipe(process.stdout);Would it be possible to make the Node gzip decoder more lenient and behave more like all major browsers and cURL do?
While talking to a web server on an embedded device, asking for gzip content encoding I get responses that are considered OK by all major browsers (Chrome, Firefox, IE) and also cURL. However, the Node
zlibmodule considers the gzip stream incomplete and throws an error. The error returned by thezlibmodule reads:Error: unexpected end of file at Zlib._handle.onerror (zlib.js:363:17). Inspecting the gzip response it seems to be missing a footer detailing the CRC and the length of the compressed data.A small program that reproduces this with an actual response from the device in it:
Would it be possible to make the Node gzip decoder more lenient and behave more like all major browsers and cURL do?