Hi Ron, I have read your message and again, please attach self-contained program to demonstrate any error with regards to subject matter.
I'll fix it asap. Andi On Tue, 1 Nov 2011 10:29:01 -0430 Ron Pinkas <ron.pin...@xharbour.com> wrote: > Hi Andi, WCB :-) > > Please read my original message, i just reread it, and I beleive it perfectly > explains the issue in details, except that you apparntly fixed the DATETIME > support for calculations like hb_vmPlus(), though I didn't get a chance to > review it. Here is my original message: > > It appears that the definitions and usage of: > > HB_IT_DATE > HB_IT_TIMEFLAG > HB_IT_DATETIME > > HB_IS_DATE() > HB_IS_DATETIME() > HB_IS_TIMEFLAG() > > is inconsistent and not designed with system wide consideration. > > On quick read of all relevant sources it appears that the correct fix would > be to make sure that HB_IT_TIMEFLAG is used indeed as a FLAG where HB_IT_DATE > must always be be the base value. This means that all API which returns a > DATETIME value must be corrected to: > > pItem->type = HB_IT_DATETIME; > > instead of: > > pItem->type = HB_IT_TIMEFLAG; > > Once this is done then naturally HB_IS_DATE() will correctly returns TRUE for > either a DATE value or a DATETIME value, since both are indeed a form of a > DATE value. This in turn will fix currently missing support for DATETIME type > in calculations, etc., such as hb_vmPlus() and similar. > > Finally, all existing instances of HB_IT_TIMEFLAG / HB_IT_DATETIME, > HB_IS_DATETIME(), and HB_IS_TIMEFLAG() should be searched and verified and > corrected when needed. All "case" structure where HB_IT_TIMEFLAG has specific > logic separate from HB_IT_DATE must be verified to be ordered such that the > HB_IT_TIMEFLAG/HB_IS_DATETIME() "case" is ABOVE the HB_IT_DATE "case". > > Since I have not taken any active part in the addition of the HB_IT_DATETIME > support, and I therefore may not be familiar with all subtleties of the > implementation I will greatly appreciate the response, and appropriate fixing > to be addressed by the original contributor[s]. > > As it is right now the existing support for HB_IT_DATETIME is BROKEN and > incomplete and fixing this should be a PRIORITY IMO. > ----------------------------------------------------------------------- > > Will greatly appreciate your review. I would also strongly suggest > consultation with Przmek and synching datetime support with Harbour since a > language TYPE is a very low level core feature, which is a foundation issue > with regards to RDD as well. > > Ron > > > On Oct 16, 2011, at 2:23 AM, Andi Jahja <andi.ja...@yahoo.co.id> wrote: > > > Hello All, > > > > > > I am sorry for being too late to response as I intentionally kept > > anything relates to programming away for some times in order to have my > > mind refreshed :-) > > > > Regarding the issues as per subject, as usual, I would be very much > > obliged if anyone can post a self-contained program to demonstrate the > > errors. I am sorry, I am too old to be patient to read narrative > > description. > > > > Bese regards, > > > > Andi > > > > > > On Tue, 16 Aug 2011 12:37:44 -0700 > > Luis Krause Mantilla <lkrau...@shaw.ca> wrote: > > > >> Ron: > >> > >> I think it was Andi that brought all these changes from Harbour > >> so he needs to chime in.... > >> > >> On 16-08-2011 08:11, Ron Pinkas wrote: > >>> ? > >>> > >>> On Aug 10, 2011, at 11:48 AM, Ron Pinkas<ron.pin...@xharbour.com> wrote: > >>> > >>>> Would greatly appreciate input from whomever was involved in adding > >>>> DATETIME support. > >>>> > >>>> Sent from my iPad > >>>> > >>>> On Aug 8, 2011, at 7:02 PM, Ron Pinkas<ron.pin...@xharbour.com> wrote: > >>>> > >>>>> BTW, to avoid consequential incompatibility with Harbour, I suggest we > >>>>> use compatible definitions and API where possible. > >>>>> > >>>>> -------------------------------------------------- > >>>>> From: "Ron Pinkas"<ron.pin...@xharbour.com> > >>>>> Sent: Monday, August 08, 2011 4:27 PM > >>>>> To:<xharbour-developers@lists.sourceforge.net> > >>>>> Subject: HB_IT_DATE, HB_IT_TIMEFLAG, HB_IT_DATETIME > >>>>> > >>>>>> Hi Guys, > >>>>>> > >>>>>> It appears that the definitions and usage of: > >>>>>> > >>>>>> HB_IT_DATE > >>>>>> HB_IT_TIMEFLAG > >>>>>> HB_IT_DATETIME > >>>>>> > >>>>>> HB_IS_DATE() > >>>>>> HB_IS_DATETIME() > >>>>>> HB_IS_TIMEFLAG() > >>>>>> > >>>>>> is inconsistent and not designed with system wide consideration. > >>>>>> > >>>>>> On quick read of all relevant sources it appears that the correct fix > >>>>>> would be to make sure that HB_IT_TIMEFLAG is used indeed as a FLAG > >>>>>> where HB_IT_DATE must always be be the base value. This means that all > >>>>>> API which returns a DATETIME value must be corrected to: > >>>>>> > >>>>>> pItem->type = HB_IT_DATETIME; > >>>>>> > >>>>>> instead of: > >>>>>> > >>>>>> pItem->type = HB_IT_TIMEFLAG; > >>>>>> > >>>>>> Once this is done then naturally HB_IS_DATE() will correctly returns > >>>>>> TRUE for either a DATE value or a DATETIME value, since both are > >>>>>> indeed a form of a DATE value. This in turn will fix currently missing > >>>>>> support for DATETIME type in calculations, etc., such as hb_vmPlus() > >>>>>> and similar. > >>>>>> > >>>>>> Finally, all existing instances of HB_IT_TIMEFLAG / HB_IT_DATETIME, > >>>>>> HB_IS_DATETIME(), and HB_IS_TIMEFLAG() should be searched and verified > >>>>>> and corrected when needed. All "case" structure where HB_IT_TIMEFLAG > >>>>>> has specific logic separate from HB_IT_DATE must be verified to be > >>>>>> ordered such that the HB_IT_TIMEFLAG/HB_IS_DATETIME() "case" is ABOVE > >>>>>> the HB_IT_DATE "case". > >>>>>> > >>>>>> Since I have not taken any active part in the addition of the > >>>>>> HB_IT_DATETIME support, and I therefore may not be familiar with all > >>>>>> subtleties of the implementation I will greatly appreciate the > >>>>>> response, and appropriate fixing to be addressed by the original > >>>>>> contributor[s]. > >>>>>> > >>>>>> As it is right now the existing support for HB_IT_DATETIME is BROKEN > >>>>>> and incomplete and fixing this should be a PRIORITY IMO. > >>>>>> > >>>>>> Ron > > > > > > ------------------------------------------------------------------------------ > > All the data continuously generated in your IT infrastructure contains a > > definitive record of customers, application performance, security > > threats, fraudulent activity and more. Splunk takes this data and makes > > sense of it. Business sense. IT sense. Common sense. > > http://p.sf.net/sfu/splunk-d2d-oct > > _______________________________________________ > > xHarbour-developers mailing list > > xHarbour-developers@lists.sourceforge.net > > https://lists.sourceforge.net/lists/listinfo/xharbour-developers ------------------------------------------------------------------------------ RSA® Conference 2012 Save $700 by Nov 18 Register now http://p.sf.net/sfu/rsa-sfdev2dev1 _______________________________________________ xHarbour-developers mailing list xHarbour-developers@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/xharbour-developers
