OnDownloadStarted takes DownloadItem* instead of DownloadId

content/browser/download/download_manager_impl.cc

Index: content/browser/download/download_manager_impl.cc
diff --git a/content/browser/download/download_manager_impl.cc b/content/browser/download/download_manager_impl.cc
index bfd5ac0be555b9f4c8c8360750e5d13c4c95b338..74c15d109b1c8ca55ad0295616a7eab278ee60ab 100644
--- a/content/browser/download/download_manager_impl.cc
+++ b/content/browser/download/download_manager_impl.cc
@@ -344,7 +344,7 @@ bool DownloadManagerImpl::Init(content::BrowserContext* browser_context) {
 }
 
 // We have received a message from DownloadFileManager about a new download.
-content::DownloadId DownloadManagerImpl::StartDownload(
+DownloadItem* DownloadManagerImpl::StartDownload(
     scoped_ptr<DownloadCreateInfo> info,
     scoped_ptr<content::ByteStreamReader> stream) {
   DCHECK(BrowserThread::CurrentlyOn(BrowserThread::UI));
@@ -376,7 +376,7 @@ content::DownloadId DownloadManagerImpl::StartDownload(
                  (delegate_ && delegate_->GenerateFileHash()), bound_net_log,
                  callback));
 
-  return download_id;
+  return GetDownload(download_id.local());

[Randy Smith (Not in Mondays), 2012/10/08 19:20:52]

I think that it would be a better design to refactor this so that
CreateDownloadItem() returns the DownloadItem so you don't need to do
GetDownload().  OTOH, I'm doing exactly that in the DFM nuke CL, and there's
really no point in duplicating the work.  So this is fine, but I wanted to call
out the code preference.

[benjhayden, 2012/10/08 19:49:45]

Ya, I thought about making CreateDownloadItem return the item, but noticed that
that would require some surgery around bound_net_log, and decided to let you
handle it. :-)

 }
 
 void DownloadManagerImpl::OnDownloadFileCreated(