On Sat, 05 Mar 2011, Andi Jahja wrote: Hi Andi,
> I proposed that Valtype( dDateTime ) returns "T" (currently "D") > ( dDateTime = date data type where time != 0 ), thus the corresponding > data type would be HB_IT_DATETIME instead of HB_IT_DATE | HB_IT_DATETIME. > Any objection or is there any problem foreseeable by this change? This seems to be reasonable but please remember that in xHarbour there is no separate type for DATE and DATETIME/TIMESTAMP values in HVM. HB_IT_TIMEFLAG/HB_IT_DATETIME was added to xHarbour with RDD code borrowed from Harbour and was never fully implemented so now neither HVM nor RDD supports this type in compatible way. To be honest it's even hard to say how it should work looking at current xHarbour code. So I think you should start defining expect behavior for such new type and then adopt existing code so it will work as you want. Such definition will also help other developers working on xHarbour HVM because they will know how this new type should work and can fix potential problems in implementation. The modification in VALTYPE() should be natural result of extensions in HVM not the source of fixes in HVM for problems you should find when you begin to analyze DATETIME/TIMESTAMP implementation deeply. If you are really interested then I suggest to read "DATETIME/TIMESTAMP VALUES" and "LITERAL DATE AND TIMESTAMP VALUES" in http://harbour-project.svn.sourceforge.net/svnroot/harbour-project/trunk/harbour/doc/xhb-diff.txt as start point. best regards, Przemek ------------------------------------------------------------------------------ What You Don't Know About Data Connectivity CAN Hurt You This paper provides an overview of data connectivity, details its effect on application quality, and explores various alternative solutions. http://p.sf.net/sfu/progress-d2d _______________________________________________ xHarbour-developers mailing list xHarbour-developers@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/xharbour-developers
