Inconsistencies when using C-style code and the WPDS

[maximiliankaul]

See https://github.com/Fraunhofer-AISEC/codyze/tree/mk/wpdscstyle for a test case. This branch just takes the Botan cpp testcases and mimics the behavior for an imaginary C - API. The results should not differ, but they do.

[maximiliankaul]

There is a general problem with WPDS not correctly checking for an order rule to correct properly.
Neither the

codyze/src/test/java/de/fraunhofer/aisec/codyze/crymlin/WpdsTest.kt

Line 530 in 08d6ff7

fun testWPDSCStyle2() {

.c test nor the

codyze/src/test/java/de/fraunhofer/aisec/codyze/crymlin/WpdsTest.kt

Line 538 in 08d6ff7

fun testWPDSCppStyle2() {

.cpp test fail besides missing the required finish() call

codyze/src/test/resources/unittests/WPDS_c_cpp_style.mark

Line 28 in 08d6ff7

foo.finish()

Note: NFA works correctly.

[maximiliankaul]

@oxisto I checked 06320a3 with the cpp code as discussed and it's broken, too :(

[maximiliankaul]

@oxisto same with ba136fe

[maximiliankaul]

Possible fix: TypestateAnalysis.kt:277 (oxisto)