Log message:
security: Replace RMD160 checksums with BLAKE2s checksums

All checksums have been double-checked against existing RMD160 and
SHA512 hashes

Unfetchable distfiles (fetched conditionally?):
./security/cyrus-sasl/distinfo \ 
cyrus-sasl-dedad73e5e7a75d01a5f3d5a6702ab8ccd2ff40d.patch.v2

Log message:
security: Remove SHA1 hashes for distfiles

Log message:
py-m2crypto: Update to 0.36.0.

Based on joyent/pkgsrc#282 from sjorge.

0.36.0 - 2020-07-13
-------------------

- wrap SocketIO in io.Buffered* for makefile <lethliel>
- SSL.Connection.close accepts an argument to force the socket closing
  <Christophe Haen>
- SSL.Connection: make the clientPostConnectionCheck an instance
  attribute <Christophe Haen>
- Fixed bug with usage of unexisting method getreply at SSL_Transport
  <roman-nagaev>
- Add appveyor builds for python 3.7 and 3.8 <Daniel A. Wozniak>
- Fixed syntax warning on line 44. <randomfox>
- Update M2Crypto.six to 1.13.0 <Matěj Cepl>
- base64.decodestring() was finally removed in Python 3.8. <Matěj Cepl>
- wrap SocketIO in io.Buffered* for makefile <lethliel>
- NULL is legal argument for key and iv paramters of EVP_CipherInit(3)
  <Matěj Cepl>
- Expose X509_V_FLAG_ALLOW_PROXY_CERTS verification flag and
  X509_STORE_SET_FLAGS function <Christophe Haen>
- Stop testing for 2.6 and 3.4 on Travis. Start testing 3.8 <Matěj Cepl>
- Extend test cert validity to 2049 <Bernhard M. Wiedemann>
- Revert using typing module in 2.6. It is just not worthy. <Matěj Cepl>
- Update Debian/stable SSL as well <Matěj Cepl>
- Make tests pass again. <Matěj Cepl>
- Stop using string module, which has been deprecated. <Matěj Cepl>
- Tiny fixes to make pyls more happy <Matěj Cepl>
- CI: Rework Fedora CI configuration <Neal Gompa>

0.35.2 - 2019-06-10
-------------------

- tests.test_rsa: Fix typo to match for proper exception <Sebastian
  Andrzej Siewior>
- Expose CRLs verification flags <Christophe Haen>

0.35.1 - 2019-06-08
-------------------

- Actually, really fix compatibility with OpenSSL 1.1.1c. Thank you,
  Sebastian Andrzej Siewior from the Debian team for resolving it.

0.34.0 - 2019-05-30
-------------------

- Use more recent version of OpenSSL on Windows
- Be resilient against the situation when no erorr happened.
- Correct URL of https://www.schneier.com/academic/smime/
- Use shlex.split() for CPP

0.33.0 - 2019-04-26
-------------------

- eb4525c - Stop pretending to support Python 3.4. <Matěj Cepl>
- 6a89548 - Fix use of urlunsplit (25 hours ago) <Andreas Schwab>
- 0a5a356 - tests/test_ssl: use -ciphercuites for TLS1.3 cipher in
      openssl1.1 <Sebastian Andrzej Siewior>
- 8a0a3e3 - There are apparently multiword CPP variables. Taking that
      into account. <Matěj Cepl>

Log message:
security: Mark packages that fail with OpenSSL 1.1 BROKEN

Log message:
*: Recursive revision bump for openssl 1.1.1.

Log message:
Forget about Python 3.4

Log message:
py-m2crypto: updated to 0.32.0

0.32.0:
- setup.py: use ${CPP} as path to cpp
- Bump pipeline OpenSSL from 1.1.0i to 1.1.0j
- Stub wchar_t helpers and ignore unused WCHAR defs
- Add type comment to setup.py

Log message:
py-m2crypto: updated to 0.31.0

0.31.0:
- Compatibility with OpenSSL 1.1.1 (partly workaround, maybe requires
  further investigation)
- Fixes for Windows builds
- Fixes of installs on AWS Lambda
- Fixes of Mac OS X related failures
- Fix Python 2.6 compatibility issues

Log message:
py-m2crypto: updated to 0.30.1

0.30.1:
- Fix packaging (missed packaging testing file)

Log message:
py-m2crypto: updated to 0.30.0

0.30.0:
- Various small typos (Windows builds, Fix SSL.Connection.__del__)
- The project is now Linux-distribution agnostic
- Replace all old-style classes with the new ones (it shouldn't cause
  any problems, but feel free to file an issue, if it does)
- Do not by-pass a potential transfer decoding in m2urllib2
- Update M2Crypto.six with 1.11.0 and replace our local workarounds with
  new functions.
- SSLv3 just removed.
- Don't support Python 2.6 on Windows anymore. Windows users don't have
  python as a system package, so they are usually more likely to upgrade
  anyway.