[WIP] 0.15.0 release candidate 1

[xuganyu96]

This PR tracks the progress of 0.15.0-rc1. Currently a work in progress.

[coveralls]

coverage: 83.543% (-0.004%) from 83.547%
when pulling 0e7a295 on gyx-0.15.0-release-notes
into 9a61d90 on main.

[baentsch]

coverage: 83.543% (-0.004%) from 83.547% when pulling 0e7a295 on gyx-0.15.0-release-notes into 9a61d90 on main.

Am I the only one wondering about this? How can code coverage change for a PR that basically just changes constants in preparation for a release? If there is any point in time that code coverage should not change, it's now, no? This doesn't give me great confidence in its actual value/reliability: and running for more than 3 hours to detect this "change in code coverage" adds insult to injury.

[xuganyu96]

Constant time tests

I ran the constant time tests on an AWS c7i.large instance with the following build instruction:

# skip SLH-DSA pre-hashed variants
export SKIP_ALGS='SPHINCS\+-SHA(.)*s-simple,SPHINCS\+-SHAKE-(.)*,SLH_DSA_(SHA2|SHA3|SHAKE)(.)*'
cmake -GNinja -DOQS_DIST_BUILD=OFF -DOQS_OPT_TARGET=generic -DCMAKE_BUILD_TYPE=Debug -DOQS_ENABLE_TEST_CONSTANT_TIME=ON ..
cmake -GNinja -DOQS_DIST_BUILD=OFF -DOQS_OPT_TARGET=haswell -DCMAKE_BUILD_TYPE=Debug -DOQS_ENABLE_TEST_CONSTANT_TIME=ON ..
python3 -m pytest --verbose --numprocesses=auto tests/test_constant_time.py

All tests passed.