service worker: Remove a DCHECK from ServiceWorkerSubresourceLoader
Before this CL we had a DCHECK in OnFallback() to make sure
that the loader isn't the only one who has a reference to
|fallback_factory_|. The intent was to make sure that the
fallback factory lives long enough to handle the response,
but it turned out we wouldn't need to ensure that because
we don't need the request in that case.
Background context:
I hit this DCHECK while implementing PoC CL of generating
code cache in idle tasks.
Bug: 915081
Change-Id: Ia636edf31c6c1cf9b95ae0a3db9f913d578bd733
Reviewed-on: https://chromium-review.googlesource.com/c/1377487
Reviewed-by: Matt Falkenhagen <falken@chromium.org>
Commit-Queue: Kenichi Ishibashi <bashi@chromium.org>
Cr-Commit-Position: refs/heads/master@{#616596}diff --git a/content/renderer/service_worker/service_worker_subresource_loader.cc b/content/renderer/service_worker/service_worker_subresource_loader.cc
index e20f3cb..b9f55a3 100644
--- a/content/renderer/service_worker/service_worker_subresource_loader.cc
+++ b/content/renderer/service_worker/service_worker_subresource_loader.cc
@@ -448,15 +448,13 @@
fallback_factory_->CreateLoaderAndStart(
url_loader_binding_.Unbind(), routing_id_, request_id_, options_,
resource_request_, std::move(client), traffic_annotation_);
+
// Per spec, redirects after this point are not intercepted by the service
// worker again (https://crbug.com/517364). So this loader is done.
//
- // Assume ServiceWorkerSubresourceLoaderFactory is still alive and also
- // has a ref to fallback_factory_, so it's OK to destruct here. If that
- // factory dies, the web context that made the request is dead so the request
- // is moot.
- DCHECK(!fallback_factory_->HasOneRef());
-
+ // It's OK to destruct this loader here. This loader may be the only one who
+ // has a ref to fallback_factory_ but in that case the web context that made
+ // the request is dead so the request is moot.
RecordTimingMetrics(false /* handled */);
delete this;
}
