On Nov 11, 2013, at 5:40 PM, Sébastien Durand <[email protected]> wrote:

> It's just a dead simple draft, it works, but what are the drawbacks ? Do I
> put myself in trouble ? (I don't have a deep tcp/http background and I fear
> it's not that simple.)

It’s not that simple. HTTP isn’t rocket science but it has a lot of details and 
some subtleties. You do NOT want to go about re-implementing an HTTP client on 
your own. It’s going to be a big pain in the ass and get in the way of what you 
really want to do. And Ruby already has perfectly good HTTP support.

Just off the top of my head, your code will break with redirects, chunked-mode 
responses or gzip encoding, and it doesn’t support keep-alive. If you added 
code to parse the response, I could probably find six or seven other things 
that didn’t do. I’m not trying to be harsh, but I've spent a decade working 
closely with HTTP and I know the pitfalls.

—Jens

Reply via email to