Relaxing the definition of isSomeString and isNarrowString

[Andrei Alexandrescu via Digitalmars-d]

On 8/25/14, 1:21 PM, monarch_dodra wrote:
> On Sunday, 24 August 2014 at 12:24:03 UTC, Andrei Alexandrescu wrote:
>> Look e.g. at
>> https://github.com/D-Programming-Language/phobos/blob/master/std/utf.d#L1074.
>> That's a memory allocation each and every time decodeImpl is called.
>> I'm not kidding. Take a look at http://goo.gl/p5pl3D vs.
>> http://goo.gl/YL2iFN. It's egregious.
>
> Dmitry already replied, but I want to stress the reply. That's complete
> BS. decodeImpl does *not* allocate. That line of code has been commented
> on, tested and benched. The enum is only ever used for CT indexing,
> which will *not* allocate.
>
> I'm repeatedly seeing people complain that decoding is slow. Maybe. And
> that it "allocates". But that's BS. A lot of people have been repeating
> this false information now, and I find it worrysome that both you and
> Walter have made this claim now.

That escalated quickly. Chillax.

The use of that enum array is fine but somewhat fragile - a small 
refactoring may trigger an allocation. Trying a static immutable array 
may be in order.

The larger point is there's significant optimization work that can be 
done on decoding before looking at other changes.


Andrei