Hi Tamás, Zolnai Tamás píše v St 20. 03. 2013 v 22:33 +0100:
> Yes, I give it a try. Great, thanks a lot! :-) > I think the mentioned multiplication is in > MetricField::ConvertDoubleValue(), which not need to change, as I see. So whenever we are converting to and from double, there is the risk of losing precision. I guess the best would be to change ConvertDoubleValue() to sal_Int64 too [and change the name accordingly ;-)], and use BigInt (see tools/inc/tools/bigint.hxx) for the computations not to overflow. Though - looking at BigInt - I am afraid you will have to add BigInt( long long nVal ) constructor in order to make 32bit architectures happy when you provide them with sal_Int64. Similarly, you will have to create an operator long long() const; that would return the long long - and would do that even in the bIsBig case when the value is larger than long, but still fits sal_Int64. I am sorry this is getting complicated - but I believe this is a cleanup that is still worth doing :-) All the best, Kendy _______________________________________________ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
