Open Hub
Open Hub

 Settings

1
I Use This!
Very High Activity

Commits : Listings

Analyzed about 19 hours ago. based on code collected 2 days ago.
Nov 06, 2025 — Nov 06, 2026
Commit Message Contributor Files Modified Lines Added Lines Removed Code Location Date
Add an end-to-end DTLS 1.3 handshake test, relates to github #1468.
Paul Gregoire
More... about 2 months ago
Fail loudly if a DTLS handshake transcript is read while undecided, relates to github #1468.
Paul Gregoire
More... about 2 months ago
Complete the server half of a DTLS 1.3 handshake, relates to github #1468.
Paul Gregoire
More... about 2 months ago
Hash DTLS 1.3 handshake transcripts without the DTLS message header, relates to github #1468.
Paul Gregoire
More... about 2 months ago
Complete the client half of a DTLS 1.3 handshake, relates to github #1468.
Paul Gregoire
More... about 2 months ago
Offer DTLS 1.3 from the DTLS client, relates to github #1468.
Paul Gregoire
More... about 2 months ago
Add the DTLS 1.3 reliable handshake and pack handshake flights into datagrams
Paul Gregoire
More... about 2 months ago
Add DTLS 1.3 record layer send/receive paths and epoch switching (RFC 9147 4), relates to github #1468.
Paul Gregoire
More... about 2 months ago
Add DTLS 1.3 record number encryption mask primitive for AES and ChaCha20 (RFC 9147 4.2.3), relates to github #1468.
Paul Gregoire
More... about 2 months ago
Add the DTLS 1.3 ack content type (RFC 9147 7.1), relates to github #1468.
Paul Gregoire
More... about 2 months ago
Add DTLS 1.3 record protection to TlsAEADCipher via a new TlsDTLS13Cipher interface, with the unified header codec and sequence number reconstruction (RFC 9147 4.2.2), relates to github #1468.
Paul Gregoire
More... about 2 months ago
Added the 1.86 security advisories section listing the thirteen CVEs the release covers and cross-referenced each with its Defects Fixed entry, added entries for the NTRU and HQC constant-time work, the legacy PKCS#12 keystore overlay reporting and the Ant test-material packaging changes, and filled in the release date.
David Hook
More... about 2 months ago
Calling the key's deriveChildKey avoids future problems
Peter Dettman
More... about 2 months ago
The legacy jdk1.3 and jdk1.4 PKCS#12 keystore overlays report an unrecognised bag through java.util.logging where the base class does, rather than writing it and an ASN.1 dump to stdout, the Ant bcpg copy no longer sweeps in the OpenPGPTestKeys fixture that sits in the main-namespace openpgp package and so matches none of its *Test.java excludes, the Ant bcpkix source and javadoc copies exclude test directories at any depth so the est/test/san certificate fixtures stop reaching the sources jar, and jdk15+.xml gains the test-signed target jdk14.xml already carries.
David Hook
More... about 2 months ago
The three argument doFinal on TupleHash and ParallelHash now binds the requested output length as the L parameter of NIST SP 800-185 sec. 5.3 and sec. 6.3, as KMAC already did, rather than encoding the object's configured output size.
David Hook
More... about 2 months ago
OpenPGPDocumentSignature.isValidAt now folds in the signature's own effectiveness, so a data signature past its Signature Expiration Time (RFC 9580 sec. 5.2.3.18) is no longer reported valid at that time and the method agrees with isEffectiveAt on the same object.
David Hook
More... about 2 months ago
An expired primary key now takes its subkeys with it: the binding check applies the primary key's Key Expiration Time to the whole certificate, and a subkey no longer inherits that validity period, which RFC 9580 sec. 5.2.3.13 counts from the creation time of the key the carrying self-signature is made on.
David Hook
More... about 2 months ago
The high-level OpenPGP message reader now sanitizes document signatures against the policy its processor was configured with rather than the implementation's own default, so a policy passed to the documented OpenPGPApi(implementation, policy) constructor governs signature acceptance on both the one-pass and the prefixed-signature path as it already governs the decompressed-size bound.
David Hook
More... about 2 months ago
The twelve private key classes that define destroy() but left Destroyable to be inherited from PrivateKey, which extends it only from Java 8, now declare the interface themselves - it dates to Java 1.4 - so destroy() is reachable and instanceof Destroyable answers true on the legacy jdk15to18 distribution as well, where the four key destruction tests now run from the jdk1.5 suite overlay, relates to github #2432.
David Hook
More... about 2 months ago
The key destruction tests now compile under the genuine 1.5 javac of the jdk15to18 build, reaching destroy() and isDestroyed() through Destroyable rather than off PrivateKey, which extends it only from Java 8, and making final the locals their anonymous Callables capture, relates to github #2432.
David Hook
More... about 2 months ago
Merge branch 'xmss-promote-crypto-signers-pick', promoting XMSS and XMSS^MT into org.bouncycastle.crypto.
David Hook
More... about 2 months ago
Merge branch 'main' of gitlab.cryptoworkshop.com:root/bc-java
David Hook
More... about 2 months ago
exclude package based lms/params tests
David Hook
More... about 2 months ago
Support Destroyable for the ten remaining private key types, relates to github #2432.
royb
More... about 2 months ago
Added TODO[ecgost] notes at the four hand-rolled ECGOST3410 encoders and decoders in the ecgost and ecgost12 provider key classes and at the lightweight PublicKeyFactory, PrivateKeyFactory and ECGOST3410Util they duplicate, recording the divergences found in review - bitLength()-based 256/512 detection where the lightweight side sizes from the curve field, three variants of the PKCS#8 parameter-sequence guard, the 2001 public key's bare-OID parameter leniency, the X9.62 ECPrivateKey fallback form emitted for keys built from an ECPrivateKeySpec that the lightweight PrivateKeyFactory cannot read because its own fallback branch is unreachable, and decoded provider keys losing their GOST parameter sets - together with the proposed consolidation onto the lightweight factories.
Peter Dettman
More... about 2 months ago
MTCContentSigner and MTCSignatureVerifierProvider now follow draft-ietf-plants-merkle-tree-certs sec. 7.2 for any subtree, taking the entry index from the serial, evaluating the full inclusion proof and hashing the MTCProof extensions, rather than assuming a two-entry subtree with the entry at index 0; MerkleTreeCertificateValidator rejects a signatureValue with unused bits; the Appendix C accumulated subtree test vectors are added, and the remaining section citations that moved in draft-05 are corrected.
David Hook
More... about 2 months ago
Merkle Tree Certificate javadoc now cites draft-ietf-plants-merkle-tree-certs-05's section numbers, fixing stale Section 6.1 references (now Section 6.2) and other drifted citations left over from draft-04.
David Hook
More... about 2 months ago
The PKCS#12 PBMAC1 MAC key length is now bounded from below as well as above, the new PKCS12Util.validateMacKeyLength applying the 20-octet floor of RFC 9579 sec. 9 on the five paths that derive a MAC key while validateKeyLength keeps its old bounds for the PBES2 content-encryption keyLength, where 16 octets is AES-128 and BC writes it itself; JcePBMac1CalculatorBuilder now names id-PBKDF2 rather than id-PBES2 as the PBMAC1 keyDerivationFunc per sec. 4, which BcPKCS12PBMac1CalculatorBuilder had refused as an unrecognised PBKDF; and the PKCS12-PBMAC1 keystore asks for a 64-octet rather than 256-octet key to match its HMAC-SHA-512 output size per sec. 5, as does the PKCS12StoreParameter builder default, whose keySize parameter is renamed from keySizeinBits to keySizeInOctets and which is now documented, relates to github #2431.
David Hook
More... about 2 months ago
Composite ML-KEM now gives every component a key generation spec so the caller's SecureRandom reaches all of them from initialize(null, random) - it reached neither half of the X25519 and X448 sets and never the ML-KEM half of any - publishes the SupportedKeyClasses/SupportedKeyFormats attributes on its Cipher and KeyGenerator services rather than building and dropping them, and rejects a wrapped key shorter than the composite encapsulation before decapsulating a zero-padded one; the never-read second column of the component size table and a getBaseName branch no composite KEM pairing can reach are removed, and the key generator takes the OID from the key rather than the one the service is locked to, which is null when it is not locked to a parameter set.
David Hook
More... about 2 months ago
DefaultAlgorithmNameFinder now names all 30 composite algorithms - the 18 Composite ML-DSA combinations and the 12 Composite ML-KEM ones on the IANA arc 1.3.6.1.5.5.7.6.37-66 - so getAlgorithmName() on a composite certificate returns the name the provider registers the service under rather than the dotted OID string, with the matching rows added to the entry-count-locked test.
David Hook
More... about 2 months ago