renmengye / base62-csharp Goto Github PK

We found that sometimes we were getting an extra 0 byte back from FromBase62. The following change to EncodingExtensions.cs fixed it:

75c75,79
<                                       stream.Write(new byte[] { (byte)(index << (mod)) }, 0, 8 - mod);

---
>                                       if (mod > 0)
>                                       {
>                                               // only pad with more bits if it didn't come out at an exact byte boundary
>                                               stream.Write(new byte[] {(byte)(index << (mod))}, 0, 8 - mod);
>                                       }

You've published this with an MIT license so I understand I can just copy the code but it would also be nice if I could just include your code in mine thru NuGet.

I'm willing to help out if you're interested and want any help.

I recognize this project is a number of years old, but I wanted to leave a remark for anyone else coming across this just in case. In some cases, when encoding a Guid, the algorithm will pad an erroneous zero to the end. For example:

Guid: b45a57cd-440a-41f4-b57f-fe87c4fca78e
Expected result: pLTQj0f4yY3QlzzGz4zbF7
Actual result: pLTQj0f4yY3QlzzGz4zbF70

When running a unit test with one million iterations, the algorithm failed a number of times, returning an encoded string with a length of 23 rather than the expected length of 22.

I think that it would be good to have this validity checks when decoding:

else
{
    // Check if the ending is good
    int mod = (int)(stream.Position % 8);
    if (mod == 0)
        throw new InvalidDataException("an extra character was found");

    if ((index >> (8 - mod)) > 0)
        throw new InvalidDataException("invalid ending character was found");

    stream.Write(new byte[] { (byte)(index << mod) }, 0, 8 - mod);
}

I would like to use this code in a project, but I need a non-GNU license to do so. Can you please add one? MIT would be ideal. Many thanks!