memory instrumentation: soften DCHECK on duplicated PIDs

Some developers seem to be hitting a DHCECK that suggests that
two clients with the same PID reply to memory dumps. Temporarily
removing the DCHECK while investigating as this seem to be hit
at startup by some developers.

BUG=744722

Change-Id: I82c06b424661110c31c26aea56420f177a23b237
Reviewed-on: https://chromium-review.googlesource.com/574854Reviewed-by: Hector Dearman <hjd@chromium.org>
Commit-Queue: Primiano Tucci <primiano@chromium.org>
Cr-Commit-Position: refs/heads/master@{#487466}

services/resource_coordinator/memory_instrumentation/coordinator_impl.cc

@@ -416,7 +416,10 @@ void CoordinatorImpl::FinalizeGlobalMemoryDumpIfAllManagersReplied() {
   std::map<base::ProcessId, mojom::ProcessMemoryDumpPtr> finalized_pmds;
   for (auto& response : request->responses) {
     const base::ProcessId pid = response.second.process_id;
-    DCHECK(!finalized_pmds.count(pid));
+    // TODO(hjd): this shouldn't really happen, but seem to be hitting this in
+    // crbug.com/744722 . Temporarily softening the DCHECK while investigating.
+    if (finalized_pmds.count(pid) > 0)
+      continue;
 
     // The dump might be nullptr if the client crashed / disconnected before
     // replying.