File Coverage

blib/lib/CryptX.pm

Criterion	Covered	Total	%
statement	16	21	76.1
branch	3	10	30.0
condition	1	3	33.3
subroutine	5	7	71.4
pod			n/a
total	25	41	60.9

package CryptX;

use strict;
use warnings ;
our $VERSION = '0.089_001';

require XSLoader;
XSLoader::load('CryptX', $VERSION);

use Carp;
my $has_json;

BEGIN {
  $has_json = 1 if eval { require JSON; 1 };
}

sub _croak {
  die @_ if ref $_[0] || !$_[-1];
  if ($_[-1] =~ /^(.*)( at .+ line .+\n$)/s) {
    pop @_;
    push @_, $1;
  }
  die Carp::shortmess @_;
}

sub _decode_json {
  croak "FATAL: cannot find JSON module" if !$has_json;
  return JSON->new->utf8->decode(shift);
}

sub _encode_json {
  croak "FATAL: cannot find JSON module" if !$has_json;
  return JSON->new->utf8->canonical->encode(shift);
}

1;

=pod

=head1 NAME

CryptX - Cryptographic toolkit

=head1 SYNOPSIS

CryptX is the distribution entry point. In normal code, load one of the concrete modules listed below.

 ## one-shot hashing
 use Crypt::Digest qw(digest_data_hex);
 my $sha256 = digest_data_hex('SHA256', 'hello world');

 ## classic AES-CBC encryption with padding
 use Crypt::Mode::CBC;
 my $cbc = Crypt::Mode::CBC->new('AES');
 my $iv = random_bytes(16); # 16-byte AES block-size IV
 my $cbc_ciphertext = $cbc->encrypt('hello world', $key, $iv);

 ## authenticated encryption (AEAD) with AES
 use Crypt::AuthEnc::GCM qw(gcm_encrypt_authenticate);
 my $key = random_bytes(32);   # 32-byte AES-256 key
 my $nonce = random_bytes(12); # 12-byte unique nonce
 my ($ciphertext, $tag) = gcm_encrypt_authenticate('AES', $key, $nonce, 'header', 'hello world');

 ## message authentication
 use Crypt::Mac::HMAC qw(hmac_hex);
 my $mac = hmac_hex('SHA256', $key, 'hello world');

 ## secure random data + UUID helpers
 use Crypt::PRNG qw(random_bytes random_string);
 use Crypt::Misc qw(random_v4uuid random_v7uuid);
 my $salt = random_bytes(16);
 my $token = random_string(24);
 my $uuid4 = random_v4uuid();
 my $uuid7 = random_v7uuid();

 ## classic password-based key derivation
 use Crypt::KeyDerivation qw(pbkdf2);
 my $dk = pbkdf2('password', $salt, 100_000, 'SHA256', 32);

 ## bare stream cipher (authenticate separately)
 use Crypt::Stream::ChaCha;
 my $stream = Crypt::Stream::ChaCha->new($key, $nonce);
 my $stream_ciphertext = $stream->crypt('hello world');

 ## modern signatures
 use Crypt::PK::Ed25519;
 my $signer = Crypt::PK::Ed25519->new->generate_key;
 my $sig = $signer->sign_message('hello world');
 my $ok = $signer->verify_message($sig, 'hello world');

 ## key agreement
 use Crypt::PK::X25519;
 my $alice = Crypt::PK::X25519->new->generate_key;
 my $bob = Crypt::PK::X25519->new->generate_key;
 my $shared_secret = $alice->shared_secret($bob);

=head1 DESCRIPTION

Perl cryptographic modules built on the bundled L library.
The distribution also includes L, a L backend
built on the bundled L library
used internally by LibTomCrypt.

This module mainly serves as the top-level distribution/documentation page. For actual work,
use one of the concrete modules listed below.

=head2 Algorithm Selection Guide

=head3 Authenticated Encryption (AEAD)

For new designs, prefer authenticated encryption (AEAD) over bare cipher modes:

=over

=item * B (L) - Fast, constant-time,
widely deployed (TLS 1.3, WireGuard, SSH). Use this as the default AEAD choice.

=item * B (L) - Extended 24-byte nonce
variant. Prefer over ChaCha20-Poly1305 when nonces are generated randomly.

=item * B (L) - The standard AEAD mode for AES. Hardware-accelerated
on modern CPUs. Requires unique nonces; nonce reuse breaks the security entirely.

=item * B (L) - Deterministic AEAD, nonce-misuse resistant.
Slightly slower but safer when nonce uniqueness cannot be guaranteed.

=item * B (L) - Nonce-misuse-resistant AEAD (RFC 8452).
Faster than AES-SIV; pick this when you need GCM-like performance with nonce-reuse safety.

=item * B (L) - Very fast single-pass AEAD. Check patent status
for your jurisdiction.

=item * B (L) - Two-pass AEAD based on CTR+OMAC. No patents,
no nonce-length restrictions.

=item * B (L) - Used in WiFi (WPA2) and Bluetooth. Requires
knowing the plaintext length in advance.

=back

=head3 Cryptographically Secure Randomness and UUIDs

=over

=item * B (L) - Use this for salts, keys, nonces, tokens,
and any other secret random values. The functional helpers
C, C, C, C,
C, and C cover most use cases. The OO API
and the algorithm-specific wrappers (L, L, etc.)
are mainly for deterministic streams or interoperability with a specific PRNG.

=item * B (L, L) - Use C
for opaque random identifiers. Use C when you want roughly time-ordered identifiers
that sort by creation time at millisecond granularity. UUIDs are identifiers, not replacements for
secret random bytes.

=back

=head3 Stream Ciphers

Stream ciphers encrypt data byte-by-byte without block padding. For most
applications prefer an AEAD mode (see above) which bundles encryption with
authentication. Use bare stream ciphers only when you handle authentication
separately.

=over

=item * B (L) - The default stream cipher choice.
Same core as ChaCha20-Poly1305 without the built-in MAC.

=item * B (L) - Extended 24-byte nonce variant of ChaCha.
Prefer when nonces are generated randomly.

=item * B / B (L, L) -
Predecessor of ChaCha. Prefer ChaCha for new designs; Salsa20 only for
interoperability (e.g. NaCl/libsodium).

=item * B (L) - B Provided for
legacy interoperability only.

=item * B, B, B (L, L, L) -
Niche ciphers from the eSTREAM portfolio. Use ChaCha unless a specific protocol requires one of these.

=back

=head3 Block Cipher Modes (without authentication)

Use these only when authentication is handled separately or not needed:

=over

=item * B (L) - Turns a block cipher into a stream cipher. Parallelizable.

=item * B (L) - Classic mode, needs padding. Prefer CTR or an AEAD mode.

=item * B (L) - B Each block encrypted independently.

=back

The individual L, L, etc. modules implement
raw single-block encryption and are rarely used directly. In almost all cases you should
use them through an AEAD mode (L, L) or a block
cipher mode (L, L) instead. When choosing a cipher,
B is the default; it is hardware-accelerated on most modern CPUs.

=head3 Hash Functions

=over

=item * B / B / B (L, etc.) - The default
choice for general hashing. Widely supported and well analyzed.

=item * B / B (L, etc.) - Alternative to SHA-2
with a completely different construction (Keccak sponge).

=item * B / B (L, etc.) - Very fast, especially
in software. BLAKE2b for 64-bit platforms, BLAKE2s for 32-bit.

=item * B / B / B - Extendable-output functions (XOFs).
Use when you need variable-length output.

=back

=head3 Checksums

Use L and L only for
non-adversarial integrity checks such as accidental corruption detection.
They are not cryptographic integrity or authenticity mechanisms. For
cryptographic use, prefer L, L, or an AEAD mode
from L.

=head3 Message Authentication Codes

=over

=item * B (L) - The standard MAC construction. Works with any hash.
Use HMAC-SHA256 as the default.

=item * B (L) - One-time MAC, very fast. Used as part of
ChaCha20-Poly1305. Requires a unique key per message.

=item * B (L) - Keyed BLAKE2. Faster than HMAC-SHA256 in
software.

=item * B (L) - Block-cipher-based MAC. Use when you already
have AES but not a hash function.

=back

=head3 Public-Key Cryptography

=over

=item * B (L) - Modern digital signatures. Fast, constant-time,
small keys/signatures. The default choice for new signature schemes.

=item * B (L) - Higher security margin than Ed25519 (~224-bit vs ~128-bit).

=item * B (L) - Elliptic-curve Diffie-Hellman key agreement. The
default choice for key exchange.

=item * B (L) - Higher security margin than X25519.

=item * B (L) - Widely used (TLS, Bitcoin). Prefer Ed25519 for new
designs unless ECDSA is required for interoperability.

=item * B (L) - Legacy but very widely used. Use 2048-bit keys minimum, 4096-bit
preferred. Prefer OAEP for encryption and PSS for signatures.

=item * B (L) - Legacy. Prefer Ed25519 or ECDSA.

=item * B (L) - Classic Diffie-Hellman. Prefer X25519 for new designs.

=back

=head3 Key Derivation / Password hashing

=over

=item * B (L) - Extract-then-expand KDF. Use for deriving
keys from shared secrets (e.g. after ECDH).

=item * B (L) - Memory-hard password hashing. The
recommended choice for password storage.

=item * B (L) - Use mainly for compatibility
with formats and protocols that specifically require bcrypt-based key derivation (for example
some OpenSSH workflows). Prefer Argon2 for new password-storage designs.

=item * B (L) - Memory-hard KDF. Use Argon2 if available.

=item * B (L) - Widely supported but CPU-only hardness.
Use Argon2 or Scrypt when possible.

=item * B (L, L) - Legacy
derivation only. Keep this for interoperability with older formats; do not use it for new designs.

=back

=head2 Error Handling

Most CryptX modules report errors by calling C (from L).
Invalid parameters, unsupported algorithms, wrong key sizes, malformed
input, and internal library failures usually croak with a descriptive
message. Catch exceptions with C or L.

Some validation-style helpers use a return value instead of croaking. The
most important examples are the C<*_decrypt_verify> functions in the
authenticated encryption modules C.
These return C when authentication fails, indicating the ciphertext was tampered with
or the wrong key/nonce was used. Some parser/decoder helpers in other modules
also return C or false for malformed input, so check the concrete
module POD when you need exact failure semantics.

=head2 Module Map

=over

=item * Top-level family modules

L, L, L,
L, L, L, L,
L, L, L, L

=item * Symmetric ciphers

L, L, L, L, L, L, L,
L, L, L, L, L, L,
L, L, L, L, L, L,
L, L, L, L, L, L,
L, L

=item * Block cipher modes

L, L, L, L, L

=item * Stream ciphers

L, L, L, L, L,
L, L, L

=item * Authenticated encryption modes

L, L, L, L, L, L, L, L

=item * Hash functions

L, L, L, L,
L, L, L, L,
L, L, L, L, L, L,
L, L, L, L, L, L,
L, L, L, L, L,
L, L, L, L,
L, L, L, L, L,
L, L, L

=item * Checksums

L, L

=item * Message authentication codes

L, L, L, L, L,
L, L, L, L, L

=item * Public-key cryptography

L, L, L, L, L, L, L, L

=item * Cryptographically secure random number generators

L, L, L, L, L

=item * Key derivation functions

L

=item * ASN.1 parser

L

Use C only when you need custom ASN.1 / DER parsing or encoding.
Most common key and certificate formats are already handled by the PK modules.

=item * Miscellaneous helpers

L (base64/base32/base58 codecs, PEM helpers, constant-time compare, UUID generation, octet increment helpers, and related utility functions)

=back

=head2 Diagnostic Functions

These low-level functions expose details of the bundled LibTomCrypt build.
They are intended for troubleshooting and bug reports, not for regular use.

=head3 ltc_build_settings

 my $str = CryptX::ltc_build_settings();

Returns a multi-line string describing every compile-time option that was
enabled when the bundled LibTomCrypt library was built (ciphers, hashes,
MACs, PK algorithms, compiler flags, etc.).

=head3 ltc_mp_name

 my $name = CryptX::ltc_mp_name();
 # e.g. "LTM" (LibTomMath)

Returns the name of the math provider back-end in use.

=head3 ltc_mp_bits_per_digit

 my $bits = CryptX::ltc_mp_bits_per_digit();
 # e.g. 60

Returns the number of bits per digit used by the math provider.

=head2 Math::BigInt backend

Part of CryptX is L, a L backend based on
the bundled LibTomMath library. It is separate from the cryptographic APIs
above, but it ships in the same distribution and uses the same big-integer
engine that LibTomCrypt relies on.

=head1 LICENSE

This program is free software; you can redistribute it and/or modify it under the same terms as Perl itself.

=head1 COPYRIGHT

Copyright (c) 2013-2026 DCIT, a.s. L / Karel Miko

=cut