private static readonly CompressionUtils _instance = new CompressionUtils(); // only for logging
// uncompress gzip byte array
public static byte[] uncompress(byte[] compressedBytes)
{
// Semi-arbitrary choice of 16 KB for the stream window to process. The goal here is to pick
// a number that's large enough to reduce the number of processing iterations, but small enough
// to avoid being a huge memory / garbage collection hit.
const uint STREAM_WINDOW_SIZE = 16 * 1024;
byte[] uncompressBytes = null;
try {
uint uncompressedSize = 0;
// Pull the uncompressed size out of the raw GZip bytes so that we can size our output stream appropriately.
// Without this, on larger streams the MemoryStream will get resized multiple times which thrashes the garbage collector.
if (compressedBytes.Length > 4)
{
// In a gzip file, the last four bytes are the uncompressed size, assuming the
// total uncompressed size is < 4 GB, which is a safe assumption for our data.
// https://tools.ietf.org/html/rfc1952#page-5
uncompressedSize =
(((uint)compressedBytes[compressedBytes.Length - 4]) << 0) |
(((uint)compressedBytes[compressedBytes.Length - 3]) << 8) |
(((uint)compressedBytes[compressedBytes.Length - 2]) << 16) |
(((uint)compressedBytes[compressedBytes.Length - 1]) << 24);
}
else
{
_instance.LogError("Compressed data too small to have length field: " + compressedBytes.Length);
}
uncompressBytes = new byte[uncompressedSize];
// Uncompress the GZipped memory stream.
using (MemoryStream stream = new MemoryStream(compressedBytes))
using (GZipInputStream zip = new GZipInputStream(stream))
{
MemoryStream outputStream = new MemoryStream(uncompressBytes);
StreamUtils.Copy(zip, outputStream, new byte[STREAM_WINDOW_SIZE]);
}
} catch (Exception e) {
_instance.LogError("Unable to uncompress bytes: " + e.ToString());
}
return(uncompressBytes);
}
