--- In [email protected], "Sheri" <sheri...@...> wrote:
> function: Objects* Search(IN VARIANT for (optional), IN VARIANT wholeWord > (optional), IN VARIANT caseSensitive (optional), IN VARIANT replacingWith > (optional)) > The values in the failing row are > chgVec[0,1]:123 > chgVec[0,2]:abc > I now think maybe the reason it works when I use unicode.new(chgVec[..]) in > arguments is that your argument type-setting process is somehow seeing the > unicode HANDLE LABEL "u\x5.." (which is the STRING seen when such a variable > is debugged), and therefore setting the type as a string. Despite this, the > content of what is subsequently processed is unicode encodes of the actual > values (not the handle labels previously noted). Yeah, I gotta gind a service that takea a variant and step my code when I get back home. BTW co.double, com.real: I think not, because of course they have to return something, and they have to return strings, and the strings have to stand in for VARIANTs, which means more handles. Much easier to just stick with com.*_typed services.
