Re: [boost] [safe_numerics] Review Part 1 (Documentation)

Subject: Re: [boost] [safe_numerics] Review Part 1 (Documentation)
From: Robert Ramey (ramey_at_[hidden])
Date: 2017-03-03 04:17:57

On 3/2/17 4:49 PM, Steven Watanabe via Boost wrote:
> AMDG

>> ....  I guess this illustrates the impossibility
>> for normal people to actually write demonstrably correct code.
... without the help of something like this.

> Tell me about it:
> https://github.com/boostorg/random/blob/develop/include/boost/random/uniform_int_distribution.hpp#L93

I took a look at this, and it looks good to me.

>
>> Ahhh - finally I see your point.  assignment using d as
>> an accumuator loses the range calculated at compile time
>> so subsequent operations can't be guaranteed to not
>> overflow.
>>
>
> Yeah.  This makes it a bit inconvenient to use
> trap_exception with anything other than a strict
> functional style.

Actually this turns out to be a very quite interesting point.

I started an example of taking a small micro computer program
for controlling a stepper motor and using the compile time
trapping facility to tweak the program so that it would guaranteed
to not produce an invalid result without invoking any runtime overhead.

Things went pretty well until I had to actually update the to position.
At this point I had to more than "tweak" the program or give up
on my goal of avoiding runtime overhead to guarantee no incorrect
behavior.  At that point I suspended work on the example because
it failed to illustrate my hope that I could take a legacy program
for an foreign processor and make minimal changes to guarantee
correctness w/o runtime overhead.  But the experiment was very
interesting and useful and I hope to get back to it when I
understand the science of computer programming better.

Robert Ramey

>
> In Christ,
> Steven Watanabe
>
>
> _______________________________________________
> Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
>