Encode

$encode(text) is a compatibility function supplied by builtins, not a registered C built-in. It returns $xform(+enc text). The defining script must be loaded; default startup normally loads it. decode supplies the inverse wrapper.

The current enc transformation converts each input byte into two uppercase letters A through P, representing its high and low four-bit nibbles. It is deterministic, preserves case distinctions in the input, and produces twice the input byte length. Empty input returns empty. This is not ordinary hexadecimal, Base64, encryption, or a hash.

$encode(hello there)
$decode($encode(hello there))

The first returns GIGFGMGMGPCAHEGIGFHCGF; the second returns hello there. Prefer generating encoded keys at runtime rather than unnecessarily coupling scripts to hardcoded representations.

Structure keys

The output alphabet is suitable for variable-name components. For example, in an EPIC script:

@ key = encode([some nickname])
@ users[$key] = [cached value]

This avoids inserting spaces or punctuation from arbitrary text into the variable name. Different input byte strings produce distinct encoded keys; if an application wants case-insensitive identity, it must normalize the input consistently before encoding. Empty input produces no component, so handle empty keys explicitly rather than assuming a nonempty identifier.

Ordinary script arguments are C strings and cannot carry embedded NUL bytes intact. For transformations that produce binary data, use one xform chain ending in a text-safe encoding; intermediate lengths are retained there. Encoding does not validate the meaning or freshness of cached user, channel, or server data.

Source references

• script/builtins: encode and decode wrappers.
• source/ircaux.c: nibble encoder, decoder, and transform registry.
• source/functions.c: xform interface and current built-in registry.
• source/alias.c: bracketed variable-name expansion and structure handling.