Fix for bug #8068 "TIMEDIFF with first negative argument gives wrong
result" (and similar bug in ADDTIME/SUBTIME). Both Item_func_add_time/Item_func_timediff::val_str() now use calc_time_diff() function which was backported from 5.0 (and which was was fixed to properly handle microseconds part of its second argument). Also now we correctly set sign of result in case when first argument is negative and second is positive.
Showing
Please register or sign in to comment