Fix regression tests coming back from the dead.

If it had been more than 15 runs (approx 15 days) since the last succesful
baseline reference job, then all regression tests would fail. Furthermore, even
the baseline reference job, would fail, making it impossible to get started again.

Now, it checks the last 60 runs instead of the last 15, and it doesn't bother checking
if you're running a reference job - so we should be able to resume even if there
are no succesful results in the past.

Author: rwest

.github/workflows/CI.yml

@@ -239,13 +239,14 @@ jobs:
 
       # Retrieve Stable Results for reference
       - name: Find ID of Reference Results
+        if: ${{ env.REFERENCE_JOB == 'false' }}
         env:
           GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
         # this will search for the last successful execution of CI on main
         run: |
-          run_id=$(gh run list -R ReactionMechanismGenerator/RMG-Py --workflow="Continuous Integration" --branch main --limit 15 --json databaseId,conclusion --jq 'map(select(.conclusion == "success")) | .[0].databaseId')
+          run_id=$(gh run list -R ReactionMechanismGenerator/RMG-Py --workflow="Continuous Integration" --branch main --limit 60 --json databaseId,conclusion --jq 'map(select(.conclusion == "success")) | .[0].databaseId')
           if [ -z "$run_id" ] || [ "$run_id" = "null" ]; then
-            echo "::error::No successful reference run found"
+            echo "::error::No successful reference run found in last 60 runs."
             exit 1
           fi
           echo "CI_RUN_ID=$run_id" >> $GITHUB_ENV