NAME¶

VarNumFromParseNum (OLEAUT32.47)

SYNOPSIS¶

HRESULT VarNumFromParseNum
(
NUMPARSE* pNumprs,
BYTE* rgbDig,
ULONG dwVtBits,
VARIANT* pVarDst
)

DESCRIPTION¶

Convert a NUMPARSE structure into a numeric Variant type.

PARAMS¶

pNumprs [In] Source for parsed number. cDig must be set to the size of rgbDig. rgbDig [In] Source for the numbers digits. dwVtBits [In] VTBIT_ flags from "oleauto.h" indicating the acceptable dest types. pVarDst [Out] Destination for the converted Variant value.

RETURNS¶

Success: S_OK. pVarDst contains the converted value. Failure: E_INVALIDARG, if any parameter is invalid. DISP_E_OVERFLOW, if the number is too big for the types set in dwVtBits.

NOTES¶

- The smallest favoured type present in dwVtBits that can represent the number in pNumprs without losing precision is used. - Signed types are preferred over unsigned types of the same size. - Preferred types in order are: integer, float, double, currency then decimal. - Rounding (dropping of decimal points) occurs without error. See VarI8FromR8(3w) for details of the rounding method. - pVarDst is not cleared before the result is stored in it. - WinXP and Win2003 support VTBIT_I8, VTBIT_UI8 but that's buggy (by design?): If some other VTBIT's for integers are specified together with VTBIT_I8 and the number will fit only in a VT_I8 Windows(tm) will "cast" the number to the smallest requested integer truncating this way the number. Wine doesn't implement this "feature" (yet?).

IMPLEMENTATION¶

Declared in "oleauto.h". Implemented in "dlls/oleaut32/variant.c". Debug channel "variant".