27 Nov
2018
27 Nov
'18
7:34 p.m.
AMDG On 11/27/2018 12:14 PM, Daniela Engert via Boost wrote:
Am 27.11.2018 um 19:58 schrieb Andrey Semashev via Boost:
<snip> Exactly my position in this regard. If only there were a signed integral type with modulo arithmetic in C++. In other words: the raw underlying hardware implementation without the UB semantics demanded by C++.
That would be useful if two's complement arithmetic were really the correct behavior. Since that's rarely the case, it would only serve to mask the problem. In Christ, Steven Watanabe