[bugfix] remainder_in_ring: using exact division is plain wrong.
authorAlexei Sheplyakov <varg@theor.jinr.ru>
Mon, 29 Sep 2008 05:58:35 +0000 (09:58 +0400)
committerAlexei Sheplyakov <varg@theor.jinr.ru>
Tue, 30 Sep 2008 20:04:24 +0000 (00:04 +0400)
ginac/polynomial/remainder.tcc

index 4db2843..60767dc 100644 (file)
@@ -90,7 +90,7 @@ bool remainder_in_ring(T& r, const T& a, const T& b)
                if (zerop(r[k]))
                        continue;
 
-               const ring_t qk = div(r[k], b_lcoeff);
+               const ring_t qk = truncate1(r[k], b_lcoeff);
 
                // Why C++ is so off-by-one prone?
                for (std::size_t j = k, i = b.size(); i-- != 0; --j) {