Node.js中zlib.deflate结果不符预期，网络压缩XML解压问题

Let's dig into why you're not getting the original compressed byte array when running zlib.deflate on your decompressed XML, and walk through practical steps to diagnose and fix this.

1. Compression Level & Zlib Flags Mismatch #

Your original compressed byte array starts with [120, 218] — this zlib header tells us two key details:

• 120 (0x78) confirms it uses the deflate compression method
• 218 (0xDA) signals the highest compression level (FLEVEL=3) and no preset dictionary (FDICT=0)

By default, zlib.deflate uses Z_DEFAULT_COMPRESSION (level 6). If the original data was compressed with level 9, this mismatch will produce a different byte array. Try explicitly setting the compression level to 9, and also check if the original used raw deflate (no zlib header wrapper):

const fs = require('fs');
const zlib = require('zlib');

const file = "./xmls/test.xml";
const originalSample = [120, 218, 188, 189, 121, 115, 219, 72, 150, 47, 250, 255, 251, 20, 169, 154, 59, 225, 174, 30];

// Test highest compression level (matches original header's FLEVEL=3)
const compressedLevel9 = zlib.deflateSync(fs.readFileSync(file), { level: 9 });
console.log('Level 9 first 20 bytes:', Array.from(compressedLevel9.slice(0, 20)));

// Test raw deflate (no zlib header) in case original used this format
const compressedRaw = zlib.deflateRawSync(fs.readFileSync(file), { level: 9 });
console.log('Raw deflate first 20 bytes:', Array.from(compressedRaw.slice(0, 20)));

2. Unintended Changes to the XML File #

When you wrote the decompressed buffer to disk with fs.writeFile, could you have introduced subtle changes? For example:

• Accidental encoding conversion (unlikely with a Buffer, but worth ruling out)
• Line ending normalization (Windows CRLF vs Unix LF) if the file was handled as text later
• Extra trailing bytes added by the filesystem or write operation

Verify the integrity of your decompressed XML:

1. Generate an MD5/SHA hash of the resultBuffer immediately after inflate (before writing to disk)
2. Generate a hash of the file you saved to disk
3. Compare the two — if they don't match, fix the write by using explicit binary mode:

   fs.writeFile(file, resultBuffer, { encoding: 'binary' }, () => { 
     console.log('done'); 
   });
   

3. Zlib Implementation Differences #

It’s possible the original compression was done with a different zlib version or non-JavaScript implementation (e.g., C, Python). Different zlib builds can have minor variations in compressed output (like Huffman tree construction) even with the same settings.

The critical check here: does decompressing your zlib.deflate output produce exactly the same XML as the original inflate result? If yes, the byte array difference is just an implementation detail — the compressed data is functionally correct. Only if the decompressed XML differs do you have a real issue.

4. Partial Byte Array Limitation #

Remember you’re only comparing the first 20 bytes of the compressed array. Even if these don’t match, the full compressed data might still be valid. Always test a full round-trip (compress → decompress) to confirm correctness, not just the sample bytes.

内容的提问来源于stack exchange，提问作者Bassdrop Cumberwubwubwub