Fix span reporting when actor aborts with pending waitUntil tasks

The span attached to a subrequest is held by the fetch promise, which can be passed to waitUntilTasks. When an actor aborts while another request's waitUntil task is still pending, span reporting can partially fail. Span reporting relies on timing from the current request.

If there are reasons not to reorder ~IncomingRequest to cancel pending operations before removing from the list, we can reconsider the assumption that an active request is available during that time.

Author: mar-cf

src/workerd/io/io-context.c++

@@ -310,8 +310,6 @@ IoContext::IncomingRequest::~IoContext_IncomingRequest() noexcept(false) {
     }
   }
 
-  context->incomingRequests.remove(*this);
-
   KJ_IF_SOME(a, context->actor) {
     a.getMetrics().endRequest();
   }
@@ -338,6 +336,10 @@ IoContext::IncomingRequest::~IoContext_IncomingRequest() noexcept(false) {
       context->waitUntilTasks.clear();
     }
   }
+
+  // Remove incoming request after canceling waitUntil tasks, which may have spans attached that
+  // require accessing a timer from the active request.
+  context->incomingRequests.remove(*this);
 }
 
 InputGate::Lock IoContext::getInputLock() {

src/workerd/io/tracer.c++

@@ -411,7 +411,7 @@ void BaseTracer::adjustSpanTime(tracing::CompleteSpan& span) {
           KJ_FAIL_ASSERT(
               "reported span without current request", span.operationName, hasCompleteTime);
         } else {
-          LOG_NOSENTRY(WARNING, "reported span without current request", span.operationName,
+          KJ_LOG(WARNING, "reported span without current request", span.operationName,
               hasCompleteTime);
         }
       }