Daniel Stenberg <daniel_at_haxx.se> wrote:
> On Sun, 7 Jun 2009, Eric Wong wrote:
>> On a side note, would a patch to change the nobuffer code path to use
>> plain write(2) be accepted? It would save some data copies to provide
>> a small speedup.
>
> You mean it avoids the copying fwrite() does internally, or what copying
> are you referring to?

Yup, the copy to the stdio internal buffer.

> But sure, I could accept such a patch. I'm not actually sure why we use
> the buffered output approach in the first place, but perhaps we gain
> something at some points...

OK. Though it may just be easier/less code to support with just the
current approach. *shrug*

-- 
Eric Wong
-------------------------------------------------------------------
List admin: http://cool.haxx.se/cgi-bin/mailman/listinfo/curl-users
FAQ:        http://curl.haxx.se/docs/faq.html
Etiquette:  http://curl.haxx.se/mail/etiquette.html