Topic: GetCharABCWidths + cyrillic

Colleagues, Are difficult algorithm which works approximately so: Receives width of characters the specified function and then   it is symbol-by-symbol,  the received width for  offsets of "the printing carriage". With Latin characters all works perfectly, and here with cyrillic such sensation that for some characters the wrong width is received. All head broke. Roofing felts the problem in algorithm, roofing felts GetCharABCWidths incorrectly works with cyrillic. Somebody faced a problem at usage  sheaves?