Revert "POSIX: Don't spawn zombies." (r14488)

Something else is trying to reap children in the ui_tests and causing
a mess. Reverting since it's a Friday night.


git-svn-id: svn://svn.chromium.org/chrome/trunk/src@14499 0039d316-1c4b-4281-b951-d872f2087c98

1 files changed, 2 insertions, 13 deletions

diff --git a/chrome/browser/renderer_host/browser_render_process_host.cc b/chrome/browser/renderer_host/browser_render_process_host.cc
index 462b382..a75d9de 100644
--- a/chrome/browser/renderer_host/browser_render_process_host.cc
+++ b/chrome/browser/renderer_host/browser_render_process_host.cc
@@ -661,24 +661,13 @@ void BrowserRenderProcessHost::OnChannelError() {
   DCHECK(process_.handle());
   DCHECK(channel_.get());
 
-  bool child_exited;
-  if (base::DidProcessCrash(&child_exited, process_.handle())) {
+  if (base::DidProcessCrash(process_.handle())) {
     NotificationService::current()->Notify(
         NotificationType::RENDERER_PROCESS_CRASHED,
         Source<RenderProcessHost>(this), NotificationService::NoDetails());
   }
 
-  // If the process crashed, then the kernel closed the socket for it and so
-  // the child has already died by the time we get here. Since DidProcessCrash
-  // called waitpid with WNOHANG, it'll reap the process. However, if
-  // DidProcessCrash didn't reap the child, we'll need to in
-  // ~BrowserRenderProcessHost via ProcessWatcher. So we can't close the handle
-  // here.
-  //
-  // This is moot on Windows where |child_exited| will always be true.
-  if (child_exited)
-    process_.Close();
-
+  process_.Close();
   channel_.reset();
 
   // This process should detach all the listeners, causing the object to be