Question How do I convert IPv6 into a 128bit number.

[JamesBowen]

I have a requirement to convert an IPv6 address into a number (128bit number).
Because it's a 128bit number I will have to store it in the database as a character field.

I found some javascript code which does what I need, but for something so simple I can't figure out what need to do to convert it to ABL Code.

var ip = '2001:0db8:0:0:8d3:0:0:0';


// simulate your address.binaryZeroPad(); method
var parts = [];
ip.split(":").forEach(function(it) {
    var bin = parseInt(it, 16).toString(2);
    while (bin.length < 16) {
        bin = "0" + bin;
    }
    parts.push(bin);
})
var bin = parts.join("");

// Use BigInteger library
var dec = bigInt(bin, 2).toString();
console.log(dec);

This is what I have but it's not working as expected.

def var ipAddress as char.
def var part as raw  extent .
def var hexPart as character .
def var hexPartTemp as character .
def var partLoop as Integer.
def var base10 as character   .
def var ipv6length as Integer.

ipAddress =  '2001:0db8:0:0:8d3:0:0:0'.

ipv6length = num-entries(ipAddress, ":").

extent(part) = ipv6length.

do partLoop = 1 to ipv6length:

    hexPartTemp = entry(partLoop, ipAddress, ':' ).

    hexPart = "0000".
    OVERLAY(hexPart, 5 - length(hexPartTemp) ) = hexPartTemp.
 
    message hexPart.
    part[partLoop] = hex-decode( hexPart ).
    
    base10 = base10 + STRING( GET-UNSIGNED-SHORT( part[partLoop], 1) ).

end.

message base10.

 

[WinningJr]

I'm thinking you would be better off converting a byte array from System.Net.IPAddress to a number.

20-01-0D-B8-00-00-00-00-08-D3-00-00-00-00-00-00

I base64string-er-ated it in the example below just to show it being turned back into an IP address from bypte array that came from a string:

DEFINE VARIABLE strIP AS CHARACTER NO-UNDO.
DEFINE VARIABLE ipAddress AS System.Net.IPAddress NO-UNDO.
DEFINE VARIABLE ipAddress2 AS System.Net.IPAddress NO-UNDO.
DEFINE VARIABLE bBytes AS "System.Byte[]" NO-UNDO.
DEFINE VARIABLE bBytes2 AS "System.Byte[]" NO-UNDO.
DEFINE VARIABLE strBytes AS CHARACTER NO-UNDO.

strIP = '2001:0db8:0:0:8d3:0:0:0'.
ipAddress = System.Net.IPAddressarse(strIP).
bBytes = ipAddress:GetAddressBytes().
MESSAGE
System.BitConverter:ToString(bBytes)
VIEW-AS ALERT-BOX.

strBytes = System.Convert:ToBase64String(bBytes).
bBytes2 = System.Convert:FromBase64String(strBytes).
ipAddress2 = NEW System.Net.IPAddress (bBytes2).

MESSAGE ipAddress2:ToString()
VIEW-AS ALERT-BOX.

 

[Stefan]

The problem is that ABL decimals are not accurate to 128 bits, try:

message exp( 2, 127 ).

If you want to go down this road you will need two 64 bit integers.

 

[JamesBowen]

There is possiblitly

The problem is that ABL decimals are not accurate to 128 bits, try:

message exp( 2, 127 ).

If you want to go down this road you will need two 64 bit integers.

170141183460469273745278895279919563358.9442406122

 

[Stefan]

Hmm... it's a bit odd. ABL decimals are supposed to support 50 digits. If I just keep multiplying by two, the decimal behaves:

def var i as int.
def var de as decimal format '>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>.9' initial 1.

repeat:

    display
        i
        de
        exp( 2, i ) format '>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>.9'
        .
    assign
        i = i + 1
        de = de * 2       
        .   
end.

Not sure why the exp function starts misbehaving at exponent 70.

Update: And I now see that you are not even using a decimal but a character... oops.

 

[Stefan]

As an easy example use :1 - your code returns 0256 which seems to indicate that your high and low bytes are swapped.

If you add:

    hexPart = substring( hexPart, 3, 2 ) + substring( hexpart, 1, 2 ).

after the overlay you should be ok.