Calculating MD5 Hash (RFC 1321 conform) in Matlab via Java

Question

I want to calculate MD5 (or other) file hashes (RFC 1321 conform) within MATLAB using the Java-Security-Implementations. Thus I coded

mddigest=java.security.MessageDigest.getInstance('MD5');
filestream=java.io.FileInputStream(java.io.File(filename));
digestream=java.security.DigestInputStream(filestream,mddigest);
md5hash=reshape(dec2hex(typecast(mddigest.digest,'uint8')),1,[])

and the routine is working fine. Somehow, the result differs from given tools.
Maybe there are problems with the file-encoding? Should't MATLAB solve that internally?
I'd like to reproduce the results, one gets by md5sum (on linux), which equal those from HashCalc (Windows).

Answer 1

There is two problems:

1. You don't read the file.
2. You have to transpose the matrix before reshape it.

This code works:

mddigest   = java.security.MessageDigest.getInstance('MD5'); 
filestream = java.io.FileInputStream(java.io.File(filename)); 
digestream = java.security.DigestInputStream(filestream,mddigest);

while(digestream.read() ~= -1) end

md5hash=reshape(dec2hex(typecast(mddigest.digest(),'uint8'))',1,[]);

/!\ Edit : p.vitzliputzli answered a very much faster solution which should be used instead of this one.

Answer 2

Stephane's solution works but is quite slow due to MATLAB's limitation of not being able to supply a JAVA byte[] array to the read method of the DigestInputStream (or any other InputStream).

However, we can adapt Thomas Pornin's solution (discarding the FileInputStream) in order to arrive at:

mddigest   = java.security.MessageDigest.getInstance('MD5'); 

bufsize = 8192;

fid = fopen(filename);

while ~feof(fid)
    [currData,len] = fread(fid, bufsize, '*uint8');       
    if ~isempty(currData)
        mddigest.update(currData, 0, len);
    end
end

fclose(fid);

hash = reshape(dec2hex(typecast(mddigest.digest(),'uint8'))',1,[]);

This solution takes about 0.018s to compute the hash of a 713kB file whereas the other solution takes about 31s.