6 Jan
2016
6 Jan
'16
10:28 p.m.
Le 06/01/16 21:07, Benedek Thaler a écrit :
Can you dig a bit further?
I drilled down the code with gdb, here is the detailed listing: http://pastebin.com/pUjxEsiW
As far as I understand, the binary_expr gets constructed, holding the right value (3) in m_rhs. The address of this member is watched, which gets modified twice, by a lazy_ostream object, which happens to be at the same address. The latest value is the same as the one Boost.Test prints when reports the failed test.
Thanks, let's continue in the ticket :) Raffi