Masayuki Hatta pushed to branch master at Debian Java Maintainers / orchid


Commits:
19ac471e by Masayuki Hatta at 2026-11-15T11:36:19+09:00
Select entry guards as in guard-spec; add vanguards-lite

Replace the old EntryGuard logic with the guard selection of Tor's
guard-spec (proposal 271): a persistent sample of guards, confirmed
guards in order of first use, three primary guards, the retry
schedules, and directory requests through the primary guards.
Reachability is learnt from circuit builds; GuardProbeTask is gone.

Onion service circuits go through four layer-2 guards with lifetimes
of 1-12 days (vanguards-lite, proposal 333).

The state file now holds "Guard" lines in Tor's format and keeps
unknown fields.  Guards in the old EntryGuard format are migrated as
confirmed guards in file order, so clients keep their guards.

API: GuardEntry and the guard methods of Directory are replaced by
Directory.getGuardState()/setGuardState(); TorConfig.NumEntryGuards
is deprecated and unused.

- - - - -
02fde784 by Masayuki Hatta at 2026-11-15T11:59:01+09:00
Authenticate the Ed25519 identity of relays in the link handshake

Check the Ed25519 certificates of the CERTS cell as tor-spec says:
the identity to signing key certificate (type 4), the signing key to
TLS link certificate (type 5), which must certify the SHA-256 digest
of the TLS certificate of the connection, and the RSA to Ed25519
cross certificate (type 7), signed with the RSA identity key.

The authenticated Ed25519 identity must be the one in the router's
microdescriptor or descriptor, and a router listed with an Ed25519
identity must send the certificates; before, a guard or bridge was
authenticated by its 1024 bit RSA identity key alone.  Routers whose
Ed25519 identity is not known yet are authenticated as before.

- - - - -
c570b4ca by Masayuki Hatta at 2026-11-15T12:49:32+09:00
Close circuits on relay cells a client does not expect

RELAY_DROP and relay commands that Orchid does not handle used to be
dropped without a trace, which let a relay send a timing signal along
the circuit (DropMark).  They now close the circuit with a DESTROY,
as do relay cells in the client-to-relay direction and cells that no
hop recognizes, which used to end in an unhandled TorException.

An inbound RELAY_EARLY cell, which was ignored at the link layer,
closes its circuit at once, as in Tor.

- - - - -
c425ecc4 by Masayuki Hatta at 2026-11-15T12:55:08+09:00
DataDirectoryTest: do not use Files.getFileStore()

Files.getFileStore() looks the path up in /proc/mounts and throws
"Mount point not found" in chroots that do not list their mounts
there.  The two permission tests failed that way in the pbuilder
chroot of Debian's reproducible-builds infrastructure, before their
Assume could skip them, so the package failed to build there.

Ask the default file system whether it has POSIX permissions
instead.  Only the tests change.

- - - - -
83cdb475 by Masayuki Hatta at 2026-11-15T13:10:49+09:00
Report a TLS handshake as completed only after the host name check

SSLEngineSSLSocket told its HandshakeCompletedListeners that the
handshake had completed, and getSession() returned the server's
session, even when the certificate was then refused for the host
name.  Listeners are now told once, after the check has passed, and
outside the locks.  getSession() runs the handshake if needed, as
JSSE's sockets do, and after a failed handshake or check returns an
invalid session (SSL_NULL_WITH_NULL_NULL) without peer certificates.

Document that TorRequest.setEnforceSslCertificates(false) has still
checked the host name since 2.5.2, and say so in the error message
of a failed HTTPS request.

- - - - -
9e8521ce by Masayuki Hatta at 2026-11-15T17:22:00+09:00
Sample guards from the consensus, not from the descriptors at hand

Guards were sampled only among relays whose descriptor had already
been downloaded.  The sample is made right after the first start,
when few descriptors have arrived, and is kept for months, so the
directory server that answered first could pick a client's guards by
the descriptors it chose to serve.

Sample from the consensus whether or not the descriptor is there
(guard-spec 4.2), and let a circuit wait for the descriptor of its
primary guard instead of going through another guard (4.12).
Directory requests go to the primary guards from the start.

Also set the version in the dependency examples to 2.6.0 and add
that step to RELEASING.md.

- - - - -
b6efa0ef by Masayuki Hatta at 2026-11-15T17:44:51+09:00
New upstream version 2.6.0+ds
- - - - -
d6773244 by Masayuki Hatta at 2026-11-15T17:44:51+09:00
Update upstream source from tag 'upstream/2.6.0+ds'

Update to upstream version '2.6.0+ds'
with Debian dir 56417710182847054a41342660286f403e12ff3c
- - - - -
a257fd1d by Masayuki Hatta at 2026-11-15T17:50:08+09:00
d/control: Mark the binary package Multi-Arch: foreign

As suggested by the Multi-Arch hinter.  The package is Architecture: all and
provides only a Java library, so it can satisfy dependencies of packages of any 
architecture.

- - - - -
2af1ab64 by Masayuki Hatta at 2026-11-15T17:53:28+09:00
New upstream version 2.6.0+ds

- - - - -
f7e74901 by Masayuki Hatta at 2026-11-15T17:59:27+09:00
d/changelog: change distribution from UNRELEASED to unstable

- - - - -


46 changed files:

- .git-archival.txt
- README.md
- debian/changelog
- debian/control
- doc/RELEASING.md
- samples/android/app/build.gradle
- src/main/java/com/subgraph/orchid/ApplicationProperties.java
- src/main/java/com/subgraph/orchid/Directory.java
- − src/main/java/com/subgraph/orchid/GuardEntry.java
- src/main/java/com/subgraph/orchid/TorConfig.java
- src/main/java/com/subgraph/orchid/circuits/CircuitBuildTask.java
- src/main/java/com/subgraph/orchid/circuits/CircuitCreationRequest.java
- src/main/java/com/subgraph/orchid/circuits/CircuitIO.java
- src/main/java/com/subgraph/orchid/circuits/CircuitManagerImpl.java
- src/main/java/com/subgraph/orchid/circuits/guards/EntryGuards.java
- − src/main/java/com/subgraph/orchid/circuits/guards/GuardProbeTask.java
- + src/main/java/com/subgraph/orchid/circuits/guards/SampledGuard.java
- + src/main/java/com/subgraph/orchid/circuits/guards/VanguardsLite.java
- src/main/java/com/subgraph/orchid/circuits/path/CircuitNodeChooser.java
- src/main/java/com/subgraph/orchid/circuits/path/CircuitPathChooser.java
- src/main/java/com/subgraph/orchid/circuits/path/TorConfigNodeFilter.java
- src/main/java/com/subgraph/orchid/config/TorConfigImpl.java
- src/main/java/com/subgraph/orchid/connections/ConnectionHandshakeV3.java
- src/main/java/com/subgraph/orchid/connections/ConnectionImpl.java
- + src/main/java/com/subgraph/orchid/connections/Ed25519LinkAuthentication.java
- src/main/java/com/subgraph/orchid/directory/DirectoryImpl.java
- − src/main/java/com/subgraph/orchid/directory/GuardEntryImpl.java
- src/main/java/com/subgraph/orchid/directory/StateFile.java
- src/main/java/com/subgraph/orchid/http/TorRequest.java
- src/main/java/com/subgraph/orchid/http/TorSocketStream.java
- src/main/java/com/subgraph/orchid/http/get/TorSocketHttpsGet.java
- src/main/java/com/subgraph/orchid/http/post/TorSocketHttpsPost.java
- src/main/java/com/subgraph/orchid/http/ssl/IgnoreSslCertificates.java
- + src/main/java/com/subgraph/orchid/sockets/sslengine/NullSSLSession.java
- src/main/java/com/subgraph/orchid/sockets/sslengine/SSLEngineManager.java
- src/main/java/com/subgraph/orchid/sockets/sslengine/SSLEngineSSLSocket.java
- test/com/subgraph/orchid/circuits/UnsolicitedCellTest.java
- − test/com/subgraph/orchid/circuits/guards/EntryGuardsNetworkChangeTest.java
- + test/com/subgraph/orchid/circuits/guards/EntryGuardsTest.java
- + test/com/subgraph/orchid/circuits/guards/FakeNetwork.java
- + test/com/subgraph/orchid/circuits/guards/VanguardsLiteTest.java
- + test/com/subgraph/orchid/connections/Ed25519LinkAuthenticationTest.java
- test/com/subgraph/orchid/connections/LinkControlCellTest.java
- test/com/subgraph/orchid/directory/DataDirectoryTest.java
- + test/com/subgraph/orchid/directory/StateFileTest.java
- test/com/subgraph/orchid/sockets/sslengine/SSLEngineSSLSocketTest.java


The diff was not included because it is too large.


View it on GitLab: 
https://salsa.debian.org/java-team/orchid/-/compare/023dddade2b4f70e23aee3690daa79137cec407e...f7e7490108bc5c85a997f38702537a3f1cede9fe

-- 
View it on GitLab: 
https://salsa.debian.org/java-team/orchid/-/compare/023dddade2b4f70e23aee3690daa79137cec407e...f7e7490108bc5c85a997f38702537a3f1cede9fe
You're receiving this email because of your account on salsa.debian.org. Manage 
all notifications: https://salsa.debian.org/-/profile/notifications | Help: 
https://salsa.debian.org/help


_______________________________________________
pkg-java-commits mailing list
[email protected]
https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/pkg-java-commits

Reply via email to