Use previous target path when resuming a download after crash

content/browser/download/download_item_impl_delegate.cc

Index: content/browser/download/download_item_impl_delegate.cc
diff --git a/content/browser/download/download_item_impl_delegate.cc b/content/browser/download/download_item_impl_delegate.cc
index 96be706eee0373c3021b95ce2b80db0f08bea110..272d241f1af959fbc33f5a4d6c2cd1d5d6ea053e 100644
--- a/content/browser/download/download_item_impl_delegate.cc
+++ b/content/browser/download/download_item_impl_delegate.cc
@@ -29,8 +29,11 @@ void DownloadItemImplDelegate::Detach() {
 
 void DownloadItemImplDelegate::DetermineDownloadTarget(
     DownloadItemImpl* download, const DownloadTargetCallback& callback) {
-  // TODO(rdsmith/asanka): Do something useful if forced file path is null.
-  base::FilePath target_path(download->GetForcedFilePath());
+  // TODO(qinmin): Do something useful if both forced file path and
+  // target file path are empty. This code path is unreachable. It's
+  // here because we cannot rely on embedder to replace the implementation.
+  base::FilePath target_path = download->GetForcedFilePath().empty() ?

[asanka, 2017/01/10 22:24:03]

The changes in this file are not necessary. The last-target-path logic is
//chrome specific.

[qinmin, 2017/01/10 23:04:42]

ok, reverted the change

+      download->GetTargetFilePath() : download->GetForcedFilePath();
   callback.Run(target_path,
                DownloadItem::TARGET_DISPOSITION_OVERWRITE,
                DOWNLOAD_DANGER_TYPE_NOT_DANGEROUS,