ss: Fix a bug where history tree node utilization is inaccurate

When a node is read from a file, the size of interval section is left at 0,
so using that metric to monitor the performance of the state system was
misleading.

Change-Id: I885a41ae744ea7f38dd25f1adce42268fe2fbd6f
Signed-off-by: Geneviève Bastien <gbastien+lttng@versatic.net>
Reviewed-on: https://git.eclipse.org/r/52681
Reviewed-by: Alexandre Montplaisir <alexmonthy@voxpopuli.im>
Reviewed-by: Hudson CI

diff --git a/statesystem/org.eclipse.tracecompass.statesystem.core/src/org/eclipse/tracecompass/internal/statesystem/core/backend/historytree/HTNode.java b/statesystem/org.eclipse.tracecompass.statesystem.core/src/org/eclipse/tracecompass/internal/statesystem/core/backend/historytree/HTNode.java
index 688c7212b8d89b8946906affed9868112b22c32a..2f91b5182c45351437f655455faf30f41d93d005 100644 (file)
--- a/statesystem/org.eclipse.tracecompass.statesystem.core/src/org/eclipse/tracecompass/internal/statesystem/core/backend/historytree/HTNode.java
+++ b/statesystem/org.eclipse.tracecompass.statesystem.core/src/org/eclipse/tracecompass/internal/statesystem/core/backend/historytree/HTNode.java
@@ -206,7 +206,9 @@ public abstract class HTNode {
          * should only have the intervals left
          */
         for (i = 0; i < intervalCount; i++) {
-            newNode.intervals.add(HTInterval.readFrom(buffer));
+            HTInterval interval = HTInterval.readFrom(buffer);
+            newNode.intervals.add(interval);
+            newNode.sizeOfIntervalSection += interval.getIntervalSize();
         }
 
         /* Assign the node's other information we have read previously */