On 4/13/2014 9:26 AM, Dicebot wrote:
C++ fault is that it makes very easy even for experienced programmer to write
faulty code and consequences of a mistake can be rather dire. C has similar
issues but C is much smaller and simpler language which allows to keep all
possible danger points in mind. I have yet to meet a single C++ programmer who
can remember about all corner cases at the same time. I am not sure even Bjorne
himself can.
D makes situation much better by refining base semantics to adhere "safe by
default, fast when explicitly asked" principle (array bounds checks, default
initialization of locals). It is still not 100% consistent but greatly reduces
stress of the context comparing to C++.
I'd still prefer C++ over C for most projects but only if using very small
well-defined subset of C++ verified by static analysis tool continuously.
Interestingly, I've been seeing that what makes D code much more robust is a
consequence of 'unittest' coupled with -cov.
