UrlBase64

UrlBase64 is a package for all versions of the .NET Framework and .NET Core, which supports the creation of url-safe base64 encoding. UrlBase64 provides symmetric routines to both encode and later decode data in a fashion that makes it safe for use on the general web and as input to ASP.NET web applications (under IIS or similar).

Installation

UrlBase64 is available via nuget and the Visual Studio package manager. To install, simply run the following in the package manager console:

Install-Package UrlBase64

Or search for UrlBase64 in the Visual Studio package manager.

Basic Usage

All functions are contained in the static UrlBase64 class, found in the NeoSmart.Utils namespace.

using NeoSmart.Utils;

void UsageSample()
{
    var foo = Encoding.UTF8.GetBytes("foo");
    var encoded = UrlBase64.Encode(foo);
    var decoded = UrlBase64.Decode(encoded);

    var bar = Encoding.UTF8.GetString(decoded);

    Assert.AreEqual("foo", bar);
}

Advanced Usage

The UrlBase64 library comes with several variants of the Encode() and Decode() routines that can be used for better performance. In particular:

It is possible (and, assuming it works for you, preferable) to use EncodeUtf8() or one of the other Encode() variants that return a span or array of ASCII/UTF-8 bytes (being the URL-safe base64-encoded result) instead of a string for better performance and lower memory usage.
Similarly, it is preferred to use the Decode(byte[]) or Decode(ReadOnlySpan<byte>) overloads instead of Decode(string).
There are overloads of both Encode() and Decode() which can complete in an allocation-free manner. These take a second Span<byte> buffer parameter that provides the scratch space to decode into and return a Span<byte> instead of a byte[] that represents the result of the encode or decode operation. This permits you to allocate a buffer once and reuse it for all subsequent calls (or use something like ArrayPool<byte>, etc) instead of having the Encode() or Decode() operation allocate a new buffer for the result each time.
To ensure you are passing in a buffer of sufficient size (otherwise the Encode() or Decode() operations will throw!), use UrlBase64.GetMaxEncodedLength() or UrlBase64.GetMaxDecodedLength() to obtain an adequately sized buffer to encode/decode into.

Padding Options (Encoding)

UrlBase64 supports two different padding modes (internally: PaddingPolicy) for generating base64 content:

public enum PaddingPolicy
{
    Discard,
    Preserve,
}

The PaddingPolicy option controls the behavior of UrlBase64 when encoding content that does not fall on a 4-character (output) boundary. Per RFC 4648, base64 specifies that an "optional depending on the circumstances" trailing = sign is used to pad the output to be a multiple of 4 characters long. UrlBase64 supports both padded and unpadded output via an optional PaddingPolicy parameter to UrlBase64.Encode controlling this behavior. The default behavior at this time is to omit the trailing padding given that it a) can (and usually is) be inferred automatically when dropped from the encoded output, and b) utilizes a symbol that requires encoding when used in URLs.

var bytes = Encoding.UTF8.GetBytes("Mary had a little lamb");
Console.WriteLine(bytes, UrlBase64.Encode(bytes, PaddingPolicy.Discard));
Console.WriteLine(bytes, UrlBase64.Encode(bytes, PaddingPolicy.Preserve));

which emits the following:

TWFyeSBoYWQgYSBsaXR0bGUgbGFtYg

TWFyeSBoYWQgYSBsaXR0bGUgbGFtYg==

License and Copyright

UrlBase64 is developed by Mahmoud Al-Qudsi of NeoSmart Technologies. UrlBase64 is released under the terms of the MIT Public License. Copyright NeoSmart Technologies 2017-2023.

neosmart / UrlBase64