Hash large files with crypto.subtle.digest("SHA-256", buffer)

Question

i have developed a web application where a user can select multiple files via a input field. Then the sha-256 checksums are calculated by the following code. The code (taken from developer.mozilla.org) only works for small files. What do I have to change to handle large files (e.g. 1GB+) too?

function sha256(buffer){
  return crypto.subtle.digest("SHA-256", buffer).then(function (hash) {
    return hex(hash);
  });
}

function hex(buffer) {
  var hexCodes = [];
  var view = new DataView(buffer);
  for (var i = 0; i < view.byteLength; i += 4) {
    // Using getUint32 reduces the number of iterations needed (we process 4 bytes each time)
    var value = view.getUint32(i)
    // toString(16) will give the hex representation of the number without padding
    var stringValue = value.toString(16)
    // We use concatenation and slice for padding
    var padding = '00000000'
    var paddedValue = (padding + stringValue).slice(-padding.length)
    hexCodes.push(paddedValue);
  }

  // Join all the hex strings into one
  return hexCodes.join("");
}

Cabezon

does this internal digest to your application or some other parties will test it? – Trustful

the function sha256 takes an ArrayBuffer as argument and returns a string (sha256-hash-string). In my application i read a file as ArrayBuffer then i call the sha256 from that and alert the resulting hash. with small files a alert box opens and shows the correct hash but with large files there is no alert box. – Adjectival

I look for the limitation but couldn't find. One must have looked a the implementation. If the digest is not going to be used outside of your application than I can give a simple solution to you. If you aim to send the digest to third parties to be checked no solution yet. – Trustful

what is the simple solution? – Adjectival

well, divide the file into, smaller size that is acceptable by crypto.subtle.digest. get the digest and prepend it to the next one,... The problem; this will not be the same digest as the whole file. Since the SHA-256 adds some padding at the end. This will be the non-standard usage of the SHA-256 but will have same collusion resistance, pre-image and second pre-image attack resistance. – Trustful

Your code works for any size of buffer. So, what doesn't work for large files? – Dirichlet

#8974875 – Trustful

Oh, I see. Yes, it looks like that particular API is useless for hashing gigabytes of data. – Dirichlet