This is my code

var fs = require('fs');
var fp = fs.openSync('binary.txt', "w");

var byte = '\0';
fs.writeSync(fp, byte, null, 'ascii');

After executing it when I open the binary.txt file it contains 0x20 and not the null byte as expected.

Now when I use

fs.writeSync(fp, byte, null, 'utf-8');

I get the wanted null byte in the file.

This is my code

var fs = require('fs');
var fp = fs.openSync('binary.txt', "w");

var byte = '\0';
fs.writeSync(fp, byte, null, 'ascii');

After executing it when I open the binary.txt file it contains 0x20 and not the null byte as expected.

Now when I use

fs.writeSync(fp, byte, null, 'utf-8');

I get the wanted null byte in the file.

• javascript
• node.js

1 Answer 1

This isn't because of the file specifically, but rather then way Node converts ASCII into bytes to write. You'll see the same behavior in this:

new Buffer('\0', 'ascii')[0]
// 32

If you want to write a NULL byte to a file, don't use a string, just write the byte you want.

fs.writeSync(fp, new Buffer([0x00]));

Generally when doing file IO, I would remend only using strings when the content is explicitly text content. If you are doing anything beyond that, stick with Buffers.

Specifics

It is actually V8, not Node that performs this conversion. Node exposes the ascii encoding as a faster method of converting to binary. To achieve this, it uses V8's String::WriteOneByte method and unless explicitly instructed not to, this function automatically converts '\0' into ' '.