Fix ReplSpec tests in GH action

[rsoeldner]

Previously, we noticed failing tests of the ReplSpec which rely on a pseudo-terminal. The major reason is the parallel execution of the test-suite (using cabal tests).

This PR (should) fix the behaviour. Test PR #1215

[imalsogreg]

Let's try.

[emilypi]

alternatively, since we have some time to look at this deeply, perhaps we should take a look at https://hackage.haskell.org/package/hspec-core-2.11.0.1/docs/Test-Hspec-Core-Spec.html#v:sequential and https://hackage.haskell.org/package/hspec-core-2.11.0.1/docs/Test-Hspec-Core-Spec.html#v:parallel when running these tests in the actual PactTests.hs file, rather than relying on CI to do "the right thing". I'll take a stab at this today and if nothing turns up, let's merge as is, but ideally we'd have these be reproducible at the local level via cabal test.

[jwiegley]

How did that analysis turn out, Emily?

[emilypi]

Hard to say - i can't get anything to trigger locally. I'm tempted to just say "merge this" and then we can re-evaluate later. I don't want to block.

[imalsogreg]

@rsoeldner Let's also make note of the time it takes to complete the tests on GitHub, making sure that running them serially doesn't slow CI down too much.

edit:
It appears to take ~3 minutes to run the hspec suite both in your PR and on master.

CI does not pass

[rsoeldner]

@emilypi , It looks like this branch passes CI,

We run tests against master and the branch of interest, for each PR. This PR runs the failing tests against master https://github.com/kadena-io/pact/actions/runs/4863641863 (If you click at the action report, checkout repository step, you see master being checked out) and then for my PR https://github.com/kadena-io/pact/actions/runs/4863600190.