DatCh() can lose the initial period

[request]

DatCh() can lose the initial period

When the First(s) of s is in the middle of a period of TimeSet t
DatCh(s, t, StatisticFunction) lose the initial value.

Check if is this the behaviour that we want for DatCh() but
there is a difference bewteen new TOL and old TOL

Serie con002 = CalInd(C,C)*2; Infinite serie constant value 2

Finite from 4/16 to 6/15
Serie aprjun = SubSer(con002, y2004m04d16, y2004m06d15);

Finite from 8/16 to 9/15
Serie augsep = SubSer(con002, y2004m08d16, y2004m09d15);

Finite with unknown values in all july and first days of august
Serie aprsep = aprjun >> augsep;

Change the timeset
Serie mthSer = DatCh(aprsep, Monthly, AvrS);

Set view = EvalSet(Dates(Monthly, y2004m4, y2004m9), Real(Date dte)
{

Text WriteLn(FormatDate(dte,"%c%Y/%m")+" "+

FormatReal(SerDat(mthSer,dte), "%.0lf"));

TRUE

});

Text WriteLn("the serie began at "+FormatDate(First(mthSer),"%c%Y/%m"));

/* Old TOL trace:
TOL (Lenguaje Orientado al Tiempo) Versión 2.0.1999.12.03 para WINDOWS
...
TOL ha sido inicializado
Tiempo de carga : 0:00:02
2004/04 2 ! April HAS its value
2004/05 2
2004/06 2
2004/07 ? ! correct in both TOLs
2004/08 2
2004/09 2
the serie began at 2004/04
Ha sido incluído el fichero E:/sw/tol/bugs/error200408-130.tol
*/

/* CVS TOL trace:
2004/04 ? ! April HAS NOT its value
2004/05 2
2004/06 2
2004/07 ? ! correct in both TOLs
2004/08 2
2004/09 2
the serie began at 2004/05
Included file E:/sw/tol/bugs/error200408-130.tol
*/

[danirus]

This bug has been resolved and has just been uploaded to Head branch of the Tol
CVS. In the near version of Tol, v1.1.2, this bug will be included.