Guard against nil execution_output in TestRun::Process

Fixes NoMethodError when tooling_job.execution_output is nil (e.g. ops
error with no output). Matches the existing guard in Analysis::Process
and Representation::Process added in PR #6696.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Author: iHiD

app/commands/submission/test_run/process.rb

@@ -112,6 +112,8 @@ def broadcast!(test_run)
 
   memoize
   def results
+    return {} if tooling_job.execution_output.nil?
+
     res = JSON.parse(tooling_job.execution_output['results.json'], allow_invalid_unicode: true)
     res.is_a?(Hash) ? res.symbolize_keys : {}
   rescue StandardError => e

test/commands/submission/test_run/process_test.rb

@@ -43,6 +43,15 @@ class Submission::TestRun::ProcessTest < ActiveSupport::TestCase
     assert submission.reload.tests_exceptioned?
   end
 
+  test "handle ops error with no execution output" do
+    submission = create :submission
+    job = create_test_runner_job!(submission, execution_status: 500)
+
+    Submission::TestRun::Process.(job)
+
+    assert submission.reload.tests_exceptioned?
+  end
+
   test "handle tests pass" do
     submission = create :submission
     results = { 'status' => 'pass', 'message' => "", 'tests' => [] }