Inconsistent behavior of regression tests due to derivation check

[kevinmorio]

The default timeout of the derivation checks introduced in PR ?? causes inconsistent behavior when running the regression tests.
Since the timeout defaults to 5 seconds, the result of the derivation checks (success or timeout) depends on the performance of the system they are executed on.

The regression tests contain the following case studies with timed out derivation checks:

• case-studies-regression/fast-tests/related_work/YubiSecure_KS_STM12/Yubikey_and_YubiHSM_multiset_analyzed.spthy
• case-studies-regression/sapic/fast/feature-secret-channel/U2F_analyzed.spthy
• case-studies-regression/sapic/slow/PKCS11/pkcs11-templates_analyzed.spthy
• case-studies-regression/ake/dh/DHKEA_NAXOS_C_eCK_PFS_partially_matching_analyzed.spthy
• case-studies-regression/ake/dh/DHKEA_NAXOS_C_eCK_PFS_keyreg_partially_matching_analyzed.spthy

The first of these is the reason why the current regression tests fail. That is, the GitHub runner is fast enough such that the derivation check succeeds. The four other case studies are not part of the default regression tests.

[jdreier]

Hmm I guess to avoid this issue we need to either deactivate the derivation checks for most regression tests (except for those testing the derivation checks) to have stable results (in most cases), or to increment the timeout (which might cause longer run times, and might not always guarantee stable results). Any opinions?

[rsasse]

I would be fine in turning them off for the regression, except where they are tested themselves. This is the cleaner (and faster running) solution.