fix: only increment the current_concurrency counter if the job max …

…`max_concurrency` set

Seems to have been a small oversight when popping idempotent jobs
from the queue on Redis.

Fixes: Issue NicolasLM#15

spinach/brokers/redis_scripts/get_jobs_from_queue.lua

@@ -34,7 +34,7 @@ repeat
             -- job is idempotent, must track if it's running
             redis.call('hset', running_jobs_key, job["id"], job_json)
             -- If tracking concurrency, bump the current value.
-            if max_concurrency ~= -1 then
+            if max_concurrency ~= nil and max_concurrency ~= -1 then
                 redis.call('hincrby', current_concurrency_key, job['task_name'], 1)
             end
         end

tests/test_redis_brokers.py

@@ -182,6 +182,20 @@ def test_cant_exceed_max_concurrency(broker):
     queued = broker._r.lpop(broker._to_namespaced('foo_queue'))
     assert json.loads(queued.decode())['id'] == str(job2.id)
 
+def test_does_not_set_concurrency_key_when_no_max_concurrency(broker):
+    job = Job(
+        CONCURRENT_TASK_NAME, 'foo_queue', datetime.now(timezone.utc), 1,
+        # kwargs help with debugging but are not part of the test.
+        task_kwargs=dict(name='job'),
+    )
+    broker.enqueue_jobs([job])
+    returned_jobs = broker.get_jobs_from_queue('foo_queue', 2)
+    assert returned_jobs[0].task_kwargs == dict(name='job')
+    current = broker._r.hget(
+        broker._to_namespaced(CURRENT_CONCURRENCY_KEY), CONCURRENT_TASK_NAME
+    )
+    assert current is None
+
 
 def test_get_jobs_from_queue_returns_all_requested(broker):
     # If a job is not returned because it was over concurrency limits,