Unexpected-start-end-date-range-in-a-Wealth-Data-DataSet

When I look at the DataManger for Wealth-Data->Dow30 I see this:



If I try to acces this data from a Strategy, there are many incomplete series.

To demonstrate I created this strategy:

CODE:
using System;
using System.Collections.Generic;
using WealthLab.Core;
using WealthLab.Backtest;

namespace WealthScript1 
{
    public class CheckDataStrategy : UserStrategyBase
    {
      static SortedDictionary<string, string> symbolMessages = new SortedDictionary<string, string>();
        public override void Initialize(BarHistory bars)
        {
         DateTime firstValidDay = DateTime.MaxValue;
         DateTime lastValidDay = DateTime.MinValue;
         // find first and last day with valid data (Close != NaN)
         for (int i = 0; i < bars.Count; i++)
         {
            if(!double.IsNaN(bars.Close[i]))
            {
               if(firstValidDay == DateTime.MaxValue) firstValidDay = bars.DateTimes[i];
               lastValidDay = bars.DateTimes[i];
            }
         }         
         symbolMessages.Add(bars.Symbol, 
   $"{bars.Symbol:-6} First: {firstValidDay.Date:yyyy-MM-dd}, Last: {lastValidDay.Date:yyyy-MM-dd}");         
        }

      public override void BacktestComplete()
      {
         foreach(string msg in symbolMessages.Values)
            WLHost.Instance.AddLogItem("CheckData", msg, WLColor.Black);
      }
      
      // not used
        public override void Execute(BarHistory bars, int idx) {}
    }
}

When run as Portfolio Backtest with DataSet Dow 30;WealthData
it produces a LogViewer output like this:


Please note the different "First Date" and "Last Date" for symbols AIC, BAC, C and so on.

It is possible to see the same effect by just clicking the "bad" symbols in DataSets->WealthData->DOW 30

I have Data Range: All Data.

It makes no difference if Strategy Settings->BacktestData->Obtain data from selected... is checked or not.

Happens with Build 89 and Build 90. I did not check older builds.
Happens on two (very different) PCs.

This problem is hard to see unless you know what you are looking for. On the other hand it will disturb backtest results heavily because much less data is used during the backtest than expected.