On Tue, 18 May 2021 11:30:32 -0700 (MST) fredvs <fi...@hotmail.com> wrote:
> (But I like your first peace of code too.) Well, let that second piece rest in peace, rather... > For the new code, maybe using a variant variable and create a new > sumfield() procedure: Yes, that might be a more general solution. It's maybe a bit clumsy because of the neccessary _explicit_ coercion conversions, but it will probabely work in all cases, or can be made to do so. > procedure tmsebufdataset.sumfield(const afield: tfield; out asum: > variant); > > But for this, we need a variant-guru (I am beginner in variant type). Well, I'm definitively _not_ one - I even depise variants and avoid them where possible. Variants are really not a data type, at least not a _simple_ one. They are rather complex records containing a variant part for the real data, preceded by a description of what it entails. And this description has to be provided by the issuer, when a variant is built. Although this can be done implicitely by the compiler at times, i.e. if it can predetermine what the variant is meant to contain. -- (Weitergabe von Adressdaten, Telefonnummern u.ä. ohne Zustimmung nicht gestattet, ebenso Zusendung von Werbung oder ähnlichem) ----------------------------------------------------------- Mit freundlichen Grüßen, S. Schicktanz ----------------------------------------------------------- _______________________________________________ mseide-msegui-talk mailing list mseide-msegui-talk@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/mseide-msegui-talk