How do you store bytes to a var from a loop so that web response can be decompressed?

Question

I have been building my own HTTP/1.0 proxy server using HTTPWebRequest/HTTPWebResponse. Part of the original code streamed the response from the remote server directly to the client like so:

if (response != null)
{
    List<Tuple<String, String>> responseHeaders = ProcessResponse(response);
    StreamWriter myResponseWriter = new StreamWriter(outStream);
    Stream responseStream = response.GetResponseStream();
    HttpStatusCode statusCode = response.StatusCode;
    string statusDesc = response.StatusDescription;

    Byte[] buffer;
    if (response.ContentLength > 0)
    {
        buffer = new Byte[response.ContentLength];
    }
    else
    {
        buffer = new Byte[BUFFER_SIZE];
    }

    int bytesRead;

    //send the response status and response headers
    WriteResponseStatus(statusCode, statusDesc, myResponseWriter);
    WriteResponseHeaders(myResponseWriter, responseHeaders);

    while ((bytesRead = responseStream.Read(buffer, 0, buffer.Length)) > 0)
    {
        // this is the response body being streamed directly to the client browser
        outStream.Write(buffer, 0, bytesRead);
    }
}

I have been trying to intercept the response body so that I can modify the contents, but before I can do this I need to decompress the contents if it has been gzip/br/deflate by the remote server. This is what I have come up with so far, but as you can see from my comments I just can't work out how to store the byte stream into one var so that I can then send it for decompression:

Byte[] buffer;
if (response.ContentLength > 0)
    buffer = new Byte[response.ContentLength];
else
    buffer = new Byte[BUFFER_SIZE];

int bytesRead;
var res = "";

// if the url and content type matches the criteria, then we want to edit it
if (hostPathMatch.Count > 0 && contentTypeMatch.Count > 0)
{
    while ((bytesRead = responseStream.Read(buffer, 0, buffer.Length)) > 0)
    {   
        // how to we send this response stream to a var so that the entire contents can be sent to decompress

        //res += UTF8Encoding.UTF8.GetString(buffer, 0, bytesRead); // this doesnt work as it mangles gzipped contents
    }

    //was the page compressed? check the content-encoding header.
    if (responseHeaders.Any(p => p.Item1.ToLower() == "content-encoding" && p.Item2.ToLower() == "gzip"))
    {
        Output._Log.Output_Log("CONTENT IS GZIPPED");
        res = Tools.Unzip(res); // expects byte[], returns UTF8
    }

    // THIS IS WHERE WE WILL MODIFY THE BODY CONTENTS
    res = res.Replace("Foo", "Bar Bar");

    // then we will re-compress

    // update the response headers with the correct content length after modification
    responseHeaders.RemoveAll(p => p.Item1 == "Content-Length");
    responseHeaders.Add(new Tuple<string, string>("Content-Length", res.Length));

    //send the response status and response headers
    WriteResponseStatus(statusCode, statusDesc, myResponseWriter); 
    WriteResponseHeaders(myResponseWriter, responseHeaders);
}
else // we didnt want to modify this file, so just stream it out directly to the browser
{
    //send the response status and response headers
    WriteResponseStatus(statusCode, statusDesc, myResponseWriter);
    WriteResponseHeaders(myResponseWriter, responseHeaders);

    while ((bytesRead = responseStream.Read(buffer, 0, buffer.Length)) > 0)
    {
        outStream.Write(buffer, 0, bytesRead);
    }
}

Answer 1

For the case where you don't know the total size in advance, you could do something like following

[...]

List<byte> data = new List<byte>();
while ((bytesRead = responseStream.Read(buffer, 0, buffer.Length)) > 0)
{
    for(int i = 0; i < bytesRead; ++i)
        data.Add(buffer[i]);
}

var bytes = data.ToArray();

[...]

Should you know it in advance you could use something like

[...]

int offset = 0;
byte[] data = new byte[TOTAL_SIZE];
while ((bytesRead = responseStream.Read(buffer, 0, buffer.Length)) > 0)
{
    buffer.CopyTo(data, offset);
    offset += bytesRead;
}

[...]

Note that nothing of the snippets above have been tested and some checks would have to be added (e.g. out-of-range, etc.). Furthermore, depending on the use cases it might not fit your needs (e.g. for huge data).

PS: DO NOT FORGET to dispose/cleanup all of the IDisposables.