Regression in pyk integration tests

[tothtamas28]

Analysis: runtimeverification/pyk#649 (comment)

[ehildenb]

If we can't make progress on the issue directly, we can consider reverting the PR which caused this problem (#3650). That revert should be tested downstream before committing to it though, because if reverting it causes other regressions, we need to make sure we have workarounds for those other regressions.

[geo2a]

The problem was in the fact we've returned a "branching" result with only one branch, which pyk interpreted as "stuck".

It turns out, Sam ahs already made a subsequent patch (#3653) to HB, and I've completely forgotten about it. I've approved the dependency update to K (runtimeverification/k#3633) which would merge it. Once this change propagates here, the test will be green again.

I've verified locally by doing:

kup install k --version 06734ea77b9ed0037619d81cf193cb1198701f2a
cd pyk
make test-integration TEST_ARGS='-k imp-simple-sum-loop'

@tothtamas28 we can keep this open until the update propagates from K to pyk.