Blocking points for further D adoption

[Jacob Carlborg via Digitalmars-d]

On 2016-06-02 20:14, Jack Stouffer wrote:

> Just to be clear, it's not a good idea to have a full blown server in
> your stdlib. Non-toy web servers are complicated pieces of software
> involving > 10KLOC. Not only that, but there are many ways to skin a cat
> in this field. Different products need varying, sometimes mutually
> exclusive, features from their servers.
>
> Therefore, I don't web servers are good candidates for standardization.

Don't you have that issue with most stuff. Not everything can fit 
everyone's need. I have never used std.bigint but it still present in 
Phobos because it's useful from someone.

I agree with the complexity of web servers but they don't need to handle 
all the gory details off clients not following the protocol. I would 
think it works perfectly fine for non-public facing servers. For public 
facing servers it should sit behind a well test well understood 
implementation like Apache or nginx, regardless if the implementation is 
in Go, Node.js or D.

-- 
/Jacob Carlborg