Apply group filters to the samples statistics in the "find zone" window

reduce memory by storing the thread ids next to the zones intead of
making zone lists per thread.

speed-up by reading the zones in linear order rather than bisecting
the whole list for each sample.

server/TracyView.hpp

@@ -516,6 +516,7 @@ private:
         {
             uint16_t id;
             Vector<short_ptr<ZoneEvent>> zones;
+            Vector<uint16_t> zonesTids;
             int64_t time = 0;
         };
 
@@ -563,12 +564,10 @@ private:
         } binCache;
 
         struct {
-            unordered_flat_map<uint16_t, Vector<short_ptr<ZoneEvent>> > threads;
             Vector<SymList> counts;
-            int64_t timeRange = 0;
             bool scheduleUpdate = false;
-            bool needZonesPerThread = false;
-        } samplesCache;
+            bool enabled = false;
+        } samples;
 
         void Reset()
         {
@@ -595,8 +594,6 @@ private:
         {
             ResetSelection();
             groups.clear();
-            samplesCache.threads.clear();
-            samplesCache.counts.clear();
             processed = 0;
             groupId = 0;
             selCs = 0;
@@ -613,6 +610,8 @@ private:
             selTotal = 0;
             selTime = 0;
             binCache.numBins = -1;
+            samples.counts.clear();
+            samples.scheduleUpdate = true;
         }
 
         void ShowZone( int16_t srcloc, const char* name )