InterNetNews/inn 2.6.3

INN 2.6.3

on GitHub

2019-02-07
Full changes and diff from previous release

Bug Fixes

• Fixed the selection of the elliptic curve to use with OpenSSL 1.1.0 or later; NIST P-256 was enforced instead of using the most secure curve.

• Fixed a regression since INN 2.6.1 that prevented articles with internationalized header fields (that is to say encoded in UTF-8) from being posted.

• nnrpd now properly logs the hostname of clients whose connection failed owing to an issue during the negotiation of a TLS session or high load average.

New Features

• A new inn.conf parameter has been added to fine-tune the cipher suites to use with TLS 1.3: the tlsciphers13 now permits configuring them. A separate cipher suite configuration parameter is needed for TLS 1.3 because TLS 1.3 cipher suites are not compatible with TLS 1.2, and vice-versa. In order to avoid issues where legacy TLS 1.2 cipher suite configuration configured in the tlsciphers parameter would inadvertently disable all TLS 1.3 cipher suites, the inn.conf configuration has been separated out.

• Support for Python 3 has been added to INN. Embedded Python filtering and authentication hooks for innd and nnrpd can now use version 3.3.0 or later of the Python interpreter. In the 2.x series, version 2.3.0 or later is still supported.
  When configuring INN with the --with-python flag, the PYTHON environment variable, when set, is used to select the interpreter to embed. Otherwise, it is searched in standard paths.
  In case you change the Python interpreter to embed, make sure that the Python scripts you use are written in the expected syntax for that version of the Python interpreter. Notably, buffer objects have been replaced with memoryview objects in Python 3, and UTF-8 encoding now really matters for string literals (Python 3 uses bytes and Unicode objects).
  INN documentation and samples of Python hooks have been updated to provide more examples.

• When a Python or Perl filter hook rejects an article, innd now mentions the reason in response to CHECK and TAKETHIS commands. Previously, the reason was given only for the IHAVE command.