← All modules

Base64 in pure Raku with the two alphabets that matter — standard and URL-safe — from the author of zef, with no dependencies to drag in.

Version

0.1.0 github:ugexe

Depends

nothing outside the core

License

Artistic-2.0

Its own test suite

2 files, green

Checked

2026-08-28 against Raku++ 3.20.1 and Rakudo 2026.08

Where it lives

raku.land

Install it #

$ rakupp install Base64

zef install Base64 writes the same store; either installer leaves the module usable by both engines.

What it is for #

Base64 is how binary rides in places that only accept text — JSON fields, HTTP headers, URLs, data: URIs. This is ugexe's implementation: two subs, encode-base64 and decode-base64, pure Raku, depending on nothing — the property that made it the encoding layer under fourteen other distributions, where a dependency of a dependency is exactly what you do not want:

use Base64;

say encode-base64('Raku is fun!', :str);
say decode-base64('UmFrdSBpcyBmdW4h', :bin).decode;

UmFrdSBpcyBmdW4h
Raku is fun!

The adverbs are doing quiet work there. Without :str, encode-base64 returns the encoded characters as a sequence — a design that lets you stream them — so :str is the spelling for the common case of wanting one string. On the way back, :bin makes decode-base64 hand you a Buf of the original bytes; .decode then turns bytes into text when the payload was text, and stays un-called when it was an image.

The URL-safe alphabet #

Standard Base64 uses + and / — two characters with jobs of their own inside URLs and filenames. RFC 4648's URL-safe variant swaps them for - and _, and :uri selects it on both subs:

use Base64;

my $bytes = Blob.new(251, 255, 190);
say encode-base64($bytes, :str);
say encode-base64($bytes, :str, :uri);
say decode-base64('-_--', :uri, :bin).list;

+/++
-_--
(251 255 190)

Those three bytes are chosen to hit the difference: encode them one way and you get +/++, the other way -_--. If you have ever seen a JWT — three dot-separated blocks of exactly this alphabet — you have seen :uri Base64 in the wild, and this is the sub pair that reads and writes it.

Base64 or MIME::Base64? #

The ecosystem has two Base64 distributions with a page here, and the choice is simpler than it looks. MIME::Base64 matches Perl's module of the same name — it line-wraps its output at 76 columns for the email tradition that named it, and speaks Str-in/Str-out by default. This one is bytes-first, never wraps, and has the URL-safe alphabet built in. For a data: URI, an HTTP Authorization: Basic header, or anything JWT-shaped, this module's defaults are the right ones; for MIME mail bodies, use the one named after the job.

Where the two engines differ #

Nothing on this page. Both alphabets, the sequence-vs-string return shapes, and the byte-exact round trips answer identically under Raku++ and Rakudo.

What was run to put this page here #

1. Parse — every file of the distribution is parsed by Raku++ itself.
2. Install — rakupp install Base64. This distribution is not in the zef index, so the installer resolved it from the REA archive; it depends on nothing.
3. Test — the distribution's own suite: 2 files, green.
4. Run — every example on this page, twice under each engine, as the site is built.

Nothing had to be fixed anywhere — an implementation built from rotor, Slip and bit-shifts over blobs, running byte-identically on both engines, is the sort of quiet conformance statement this site collects.