[clang-tidy] Apply performance-inefficient-vector-operation fixes.

This CL applies clang-tidy's performance-inefficient-vector-operation
[1] on the WebRTC codebase.

All changes in this CL are automatically generated by both clang-tidy
and 'git cl format'.

[1] - https://clang.llvm.org/extra/clang-tidy/checks/performance-inefficient-vector-operation.html

Bug: webrtc:10252
Change-Id: I824caab2a5746036852e00d714b89aa5ec030ee3
Reviewed-on: https://webrtc-review.googlesource.com/c/120052
Commit-Queue: Mirko Bonadei <mbonadei@webrtc.org>
Reviewed-by: Karl Wiberg <kwiberg@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#26442}

pc/rtc_stats_traversal_unittest.cc

@@ -41,6 +41,7 @@ class RTCStatsTraversalTest : public testing::Test {
 
   void TakeReferencedStats(std::vector<const RTCStats*> start_nodes) {
     std::vector<std::string> start_ids;
+    start_ids.reserve(start_nodes.size());
     for (const RTCStats* start_node : start_nodes) {
       start_ids.push_back(start_node->id());
     }