使用 SharpZipLib 解压 zlib 压缩数据

Question

I am looking at the following line of c++ code, where "zlib.h" is included

int err = compress2(compressed.get(), &dest_len,
                            reinterpret_cast<const unsigned char *>(plain_text.c_str()),
                            source_len, 9);

I am trying to undo this compression using C#. I figure I should use a C# library because writing my own decompression algorithm seems wasteful and difficult. The library I have seem recommended is SharpZipLib (even by Mark Adler himself,) However. it seems like compress2 with these args is using a preset dictionary: Here is my attempt at decompressing:

Inflater inflater = new Inflater(false);
inflater.SetInput(bytes);
int ret = inflater.Inflate(outBuffer, 0, 2047);
if (ret == 0)
{
    Console.WriteLine(inflater.IsNeedingDictionary);
    Console.WriteLine(inflater.IsNeedingInput);
    Console.WriteLine(inflater.Adler);
}

This says it needs a dictionary. I know how to set the dictionary, but I am clueless about how to set it based on the zlib.h compress function. How can I create the necessary preset dictionary to decrypt?

Notes: Here is the line of code I am trying to essentially replicate in C#

int err = uncompress(uncompressed.get(), &dest_len, plain_text.get(), src_len);

Where uncompressed is empty, dest len is simply 4*src_len, and src_len is compressed length. Compressed text stored in plain_text.

Here are the detected first 2 bytes in my C# program: 78 3f (0111 1000 0011 1111)

Someone indicated that this doesn't make sense given the compress2 code given...

Sorry for any mistakes in my question, I haven't asked many questions on here.

Answer 1

No, compress2() does not use a dictionary. Your data may not be making it over intact. You should put the first several bytes of what you're trying to decompress, as seen on the C# side, in hex in your question.

Answer 2

If anyone is trying to use compress2 from zlib, and trying to deal with this input in C#, beware how you take in the compressed data. Compress2() does not use a preset dictionary.

Specifically, make sure that the first 2 bytes make sense. I shrugged off a strange second byte because the first byte was correct, but really read it and compare against the RFC 1950

There are many ways to properly read the bytes, but my mistake was trying to read them using Encoding.Ascii.GetBytes. Don't do that!