* Copied qTESLA-p-I round2 (2019-08-19) code

* Code compiles, NIST-KAT works

* Included detached signature API

* Generated testvectors

* Fixed name in api.h

* code style

* Fixed error in Makefile

* Passing pytest

* Fixing types (uint8_t bytes and size_t indices)

* Replaced SHAKE with SHAKE128 where necessary

* Fixed bug: (signed) integer overflow

* Added qTESLA-p-III

* Code is now independent of machine endianness

* repaired Microsoft makefile

Add duplicate consistency files

* Integrate Kyber-AVX2 into PQClean

* Fix types and formatting in Kyber

* Workaround a valgrind crash

* Remove comment in shuffle.s

* Remove some extraneous truncations

* fixup! Fix types and formatting in Kyber

* Add support for specifying architecture and feature limits

* cpuinfo not supported on ppc

* Detect 32-bit python interpreter on 64-bit CPU

* Fix bugs in isolated tests

* Also support restricting operating system

* Enable valgrind exit errorcode

It seems we were not doing anything with the valgrind results... D:

* Use --exit-on-first-error if available

* Revert "Merge pull request #182 from PQClean/rainbow"

This reverts commit 8a008417fe988b48c8acc4aefc9c299a4d3c1645, reversing
changes made to d4b612fe9a.

Use Python's built-in ``diflib`` to compute diffs between the two versions.

* Reduce pytest xml log size
* store artifact to debug
* Fix test result path

This should drastically reduce the runtime of the valgrind tests

* Do tests with pytest to run them in parallel

* attempt to handle merge commits better for PR test path

Similar to how we solved this for travis

* Clean up imports

* don't run valgrind if not specified slow_test

* Fix functest after initializer rename

* upload tests results as junit

* Upload test-common files since #200 got merged

* Catch test results upload failure

* Move test support files into tests directory

* Fix common tests dir in test

* Fix Windows Makefile for common tests

* notrandombytes.obj should be in TEST_COMMON_DIR

This reverts commit def5bddb62.

There are schemes, like SABER (#192) that have more than one principal
submitter. Consistency warrants that we turn it into a list for all
schemes and don't do something with allowing either a str or a list:
that would just be very annoying to parse.

Closes #194

* fixes dynamic memory allocation test. previously a function called freeze() would trigger it

* this adds DilithiumII. Preprocessor conditionals still need to be removed

* fix ms Makefile

* fix MS compiler warnings

* clean-up

* remove preprocessor conditionals

* add dilithium3

* add dilithium4

* add duplicate consistency checks

* SHA2 state constants in common

* clean up symmetric.h

* Port SPHINCS+-SHA256 to sha256ctx struct

* Implement ctx struct for fips202

* Port Kyber{512,768,1024} to fips202 ctx struct

* Port NewHope to fips202 structs

* Port SPHINCS+-SHAKE256 to fips202 ctx structs

* Use opaque fips202 structs in MQDSS

* port dilithium to use fips202 ctx structs

* include -Wredundant-decls

* remove comment; format NTT constants

* reduce casts in power2round

Fixes https://github.com/PQClean/PQClean/issues/178.