I'm using lz4 on mac and doing an experiment to compress a string (named str) in my program.
#include <fstream>
#include <iostream>
#include "lz4.h"
using namespace std;
int main(){
char str[] = "10100100010000100000100000010000000100000000100000000010000000000";
size_t len = sizeof(str);
char* target = new char[len];
int nCompressedSize = LZ4_compress_default((const char *)(&str), target, len, len);
ofstream os("lz4.dat",ofstream::binary);
os.write(target, nCompressedSize);
os.close();
delete[] target;
target = 0;
ifstream is( "lz4.dat", ifstream::binary );
is.seekg (0,is.end);
size_t nCompressedInputSize = is.tellg();
is.clear();
is.seekg(0,ios::beg);
//Read file into buffer
char* in = new char[nCompressedInputSize];
int32_t n=is.read(in,nCompressedSize);
cout<<"Byte number:"<<nCompressedSize<<",file size:"<<n<<",bytes read:"<<in<<endl;
is.close();
return 0;
}
Run this program, I checked the "lz4.dat" file:
$ls -lrt lz4.dat
-rw-r--r-- 1 x staff 34 7 15 14:50 lz4.dat
It's 34 bytes, OK, but the program output is:
Byte number:34,file size:1,bytes read:@1010
Very strange, seems the file size received is 1 byte, and I actually output some randome @1010. Why my "is.tellg()" didn't get correct file length?
Thanks.
ifstream::read()
doesn't return the bytes read. It returns a reference to *this
, which has operator bool()
, which is used in case, I think. So you in n
, you get whether the operation was succeeded.
Output seems to be completely fine, it is the beginning of the compressed data. I think there is only several bytes printed, because it contains a terminating zero. And it resembles your input, because lz4 puts literals into the stream verbatim (lz4 doesn't have an entropy encoding)