Get file size when in buffer

I gota BYTE* buffer, which contains the contents of a file.

Problem is when I want to output the contents of the buffer, into a new file (copy).
I need to retrieve the size of the BYTE* buffer, and casting it to char* and using strlen() doesn't work since it breaks on EOF-bytes, which there sometimes are in files.

Any solutions?

None. You need to know the size beforehand.
Although if it's a text file, then strlen might work to get the length of the text inside the file.

Consider using std::vector<BYTE> instead of BYTE* as it contains more features (like buffer size and automatic memory deallocation).

Quote:

---

Originally Posted by Magos

Consider using std::vector<BYTE> instead of BYTE* as it contains more features (like buffer size and automatic memory deallocation).

---

and I'm pretty sure vector doesn't care about NULLs or EOF bytes, the size it reports is simply the size of the data you put in it.

Yes, I agree with that. The problem, however, with using vector<BYTE> and file-reading functions is that you either need to use another buffer to store temporary results, or read one byte at a time. Reading one byte at a time is inefficient, but would technically work. Reading multiple bytes still presents the problem of knowing how many byte was read, which is what the original post (sort of) is asking for.

--
Mats

Quote:

---

Originally Posted by cpjust

and I'm pretty sure vector doesn't care about NULLs or EOF bytes, the size it reports is simply the size of the data you put in it.

---

If that size isn't what you need, then what size do you need?
A vector of size 0 is equal to a NULL pointer.

Quote:

---

Originally Posted by matsp

Yes, I agree with that. The problem, however, with using vector<BYTE> and file-reading functions is that you either need to use another buffer to store temporary results, or read one byte at a time.

---

I'm not sure the impact would be that great, since C++ iostreams are buffered. It would be easy to do some time trials to be sure. By default, I wouldn't worry about it. Maybe something like this:

Code:

---

std::ifstream input_file(...);
input_file >> std::noskipws;
std::vector<BYTE> file_data(std::istream_iterator<char>(input_file), std::istream_iterator<char>());

---

Quote:

---

Originally Posted by matsp

Yes, I agree with that. The problem, however, with using vector<BYTE> and file-reading functions is that you either need to use another buffer to store temporary results, or read one byte at a time. Reading one byte at a time is inefficient, but would technically work. Reading multiple bytes still presents the problem of knowing how many byte was read, which is what the original post (sort of) is asking for.

--
Mats

---

What is wrong with

Code:

---

std::stringstream ss;
// this copies the entire contents of the file into the string stream
ss << fin.rdbuf();
// get the string out of the string stream
std::string contents = ss.str();
// construct the vector from the string.
std::vector<char> buff(contents.begin(), contents.end());

---

??

What is wrogn with keepign track of how many bytes you read form th source file, you need to do this anyway, sicne a read request for X bytes is not guaranteed to return X bytes even if there are more than X bytes left in the file.

Would this be more efficient than vart's example?

Code:

---

std::vector<char> buff;
std::copy(std::istreambuf_iterator<char>(fin),
          std::istreambuf_iterator<char>(),
          std::back_inserter(buff));

---

rdbuf() is very efficient, but there's a temporary string involved.

If you look back at a thead I posted in a few months ago, this method was significantly faster:

Code:

---

void Test4( ifstream&  file )
{
        char block[BUFSIZ + 1];
        string fileData;
        while ( file.read( block, BUFSIZ ) )
        {
                fileData += block;
        }
        cout << "Test4 Bytes Read: " << fileData.size() << endl;
}

---