Erlang and JavaScript MD5 Digest match

问题

Testing the Javascript Implementation of MD5 here: http://www.webtoolkit.info/javascript-md5.html gives the following output:

MD5("muzaaya") = "00e081abefbbbf72b2d5258196a9b6d0"

Going to my erlang shell, and calculating the MD5 of the same value i get this:

Eshell V5.8.4  (abort with ^G)
1> erlang:md5("muzaaya").
<<0,224,129,171,239,187,191,114,178,213,37,129,150,169,
  182,208>>
2>

How can i compare the two? If the MD5 result from the JavaScript front end app comes to my Erlang backend, i would like to be able to compare the two Digests. How can i match the Javascript MD5 digest to that of Erlang?

回答1:

An MD5 hash is in essence a 128-bit number.

You receive the MD5 value in Erlang as a binary of 16 bytes (16 * 8 = 128 bits). Each byte in that binary has to be converted into hexadecimal representation to be comparable to JavaScript's MD5 output (which is a hexadecimal string with two characters per byte):

2> lists:flatten([io_lib:format("~2.16.0b", [B]) || <<B>> <= MD5]).
"00e081abefbbbf72b2d5258196a9b6d0"

First, we take each byte from the binary and use the io_lib module to format it to a hexadecimal string. Then we use the flatten function to display it as a readable string (although this isn't necessary if you're going to write the value to a file or a socket since they are able to handle deep io lists, nested lists of characters or binaries).

The format string used, ~2.16.0b means format an integer (b) using base 16 and padding to width 2 with the padding character 0 (see io:format/3 for a full guide).

If you want a binary, you could use the following binary comprehension instead:

3> << << (list_to_binary(io_lib:format("~2.16.0b", [C])))/binary >>
     || <<C>> <= MD5 >>.
<<"00e081abefbbbf72b2d5258196a9b6d0">>

(Instead of io_lib:format/2 there is also http_util:integer_to_hexlist/1, although I don't know if it is faster)

回答2:

Yet another and faster version:

hstr(B) when is_binary(B) ->
  << <<(hex(A)), (hex(B))>> || <<A:4,B:4>> <= B >>.

-compile({inline, [hex/1]}).

hex(0)  -> $0;
hex(1)  -> $1;
hex(2)  -> $2;
hex(3)  -> $3;
hex(4)  -> $4;
hex(5)  -> $5;
hex(6)  -> $6;
hex(7)  -> $7;
hex(8)  -> $8;
hex(9)  -> $9;
hex(10) -> $a;
hex(11) -> $b;
hex(12) -> $c;
hex(13) -> $d;
hex(14) -> $e;
hex(15) -> $f.

but the fastest version will be

md5_hex(L) ->
  << A1:4, A2:4,  A3:4,  A4:4,  A5:4,  A6:4,  A7:4,  A8:4,
    A9:4,  A10:4, A11:4, A12:4, A13:4, A14:4, A15:4, A16:4,
    A17:4, A18:4, A19:4, A20:4, A21:4, A22:4, A23:4, A24:4,
    A25:4, A26:4, A27:4, A28:4, A29:4, A30:4, A31:4, A32:4
    >> = erlang:md5(L),
  << (hex(A1)), (hex(A2)),  (hex(A3)),  (hex(A4)),
    (hex(A5)),  (hex(A6)),  (hex(A7)),  (hex(A8)),
    (hex(A9)),  (hex(A10)), (hex(A11)), (hex(A12)),
    (hex(A13)), (hex(A14)), (hex(A15)), (hex(A16)),
    (hex(A17)), (hex(A18)), (hex(A19)), (hex(A20)),
    (hex(A21)), (hex(A22)), (hex(A23)), (hex(A24)),
    (hex(A25)), (hex(A26)), (hex(A27)), (hex(A28)),
    (hex(A29)), (hex(A30)), (hex(A31)), (hex(A32)) >>.

but you should not have to do this optimization.

EDIT: This version of hex/1 is even faster:

hex(X) ->
  element(X+1, {$0, $1, $2, $3, $4, $5, $6, $7, $8, $9, $a, $b, $c, $d, $e, $f}).

EDIT2: Yet another approach:

md5_hex(L) ->
    <<X:128>> = erlang:md5(L),
    B = integer_to_binary(X,16),
    list_to_binary([lists:duplicate(32-byte_size(B),$0)|B]).

回答3:

If you need an one-liner it can be something like this:

1> B = erlang:md5("muzaaya").
<<0,224,129,171,239,187,191,114,178,213,37,129,150,169,
  182,208>>
2> lists:flatten([io_lib:format("~2.16.0b", [C]) || <<C>> <= B]).
"00e081abefbbbf72b2d5258196a9b6d0"

回答4:

Here's bitstring comprehension version, probably the fastest and most memory efficient:

hstr(B) when is_binary(B) ->
    T = {$0,$1,$2,$3,$4,$5,$6,$7,$8,$9,$a,$b,$c,$d,$e,$f},
    << <<(element(X bsr 4 + 1, T)), (element(X band 16#0F + 1, T))>>
    || <<X:8>> <= B >>.

3> M:hstr(erlang:md5("muzaaya")).

4> <<"00e081abefbbbf72b2d5258196a9b6d0">>

回答5:

If you want to do it on the JavaScript side you can use this

function md5HexToArray ( hexStr ) {  
  var i, arr = [], arraylength = hexStr.length/2;

  for( i = 0; i < arraylength ; i++ ) {
     arr[i] = parseInt( hexStr.substr(i*2,2), 16) ;
  }

  return arr;
};

But @Wrikken's comment looks like it should work just fine too.

来源：https://stackoverflow.com/questions/6818214/erlang-and-javascript-md5-digest-match