Array bound checks removal increasing importance

[Andrei Alexandrescu via Digitalmars-d]

On 5/31/14, 6:47 PM, Meta wrote:
> On Saturday, 31 May 2014 at 23:30:41 UTC, Walter Bright wrote:
>> There isn't a suitable place. To make it work, data flow analysis
>> would have to be added to the front end. While doable, this is not a
>> simple addition. Eventually, we'll have to do it as a lot of things
>> become possible & better with data flow analysis, not just bounds
>> check elimination.
>
> I've always wondered if VRP can be leveraged in certain situations. I
> can't remember exactly how it's supposed to work, but very basically,
> isn't it just numeric variables (and expressions?) having an associated
> range that they carry around with them at compile time, so something
> like this is possible:
>
> long n1 = long.max;
> int n2 = n1 % 3; //No cast needed due to VRP
>
> Couldn't this be used for other things as well, such as detecting
> numeric overflow at compile time, or like Nordlow suggested, figuring
> out when it's safe to elide an array bounds check?

For $/n VRP can't be used because it's geared toward constants, not 
relative to variables. So VRP could inform of something like "this 
number is between 0 and 5", not "this number is between 0 and a.length". 
-- Andrei