core: reserve vector sizes as much as we can (#9118)

avoid reallocations as much as possible with a few edge cases where the
reservation overshoots a tiny bit. but a few bytes of memory short term
is better used then the overhead of potential reallocation.

src/protocols/ToplevelExport.cpp

@@ -393,6 +393,8 @@ void CToplevelExportProtocol::onOutputCommit(PHLMONITOR pMonitor) {
         return; // nothing to share
 
     std::vector<WP<CToplevelExportFrame>> framesToRemove;
+    // reserve number of elements to avoid reallocations
+    framesToRemove.reserve(m_vFramesAwaitingWrite.size());
 
     // share frame if correct output
     for (auto const& f : m_vFramesAwaitingWrite) {