Test runner refactor (#508)

* Port test runner script to Python.

This allows us to keep the test output in sorted order while still running the
tests in parallel.  It also now defaults to using the number of available CPU
threads for parallel execution, rather than the previously hard-coded default.

* Also port decompyle_test.sh script to python within run_tests.py

* Fix cmake check target for multi-config generators.

Adds testing of release builds on both MSVC and GCC.

* Fix diff comparisons on Windows

* Ubuntu runners don't have ninja by default

.github/workflows/linux-ci.yml

@@ -11,10 +11,19 @@ jobs:
       - uses: actions/checkout@v1
       - name: Configure and Build
         run: |
-          mkdir build && cd build
-          cmake -DCMAKE_BUILD_TYPE=Debug ..
-          make -j4
+          (
+            mkdir build-debug && cd build-debug
+            cmake -DCMAKE_BUILD_TYPE=Debug ..
+            make -j4
+          )
+
+          (
+            mkdir build-release && cd build-release
+            cmake -DCMAKE_BUILD_TYPE=Debug ..
+            make -j4
+          )
+
       - name: Test
         run: |
-          cd build
-          make check JOBS=4
+          cmake --build build-debug --target check
+          cmake --build build-release --target check