Thanks, btw. Den onsdag 23 mars 2016 kl. 13:02:28 UTC+1 skrev Alex Hultman: > > Ah, okay. Nice to know we are doing basically the same optimization then. > I should probably use uv_tcp_t then. > > Den onsdag 23 mars 2016 kl. 12:54:15 UTC+1 skrev Ben Noordhuis: >> >> On Wed, Mar 23, 2016 at 12:33 PM, Alex Hultman <[email protected]> >> wrote: >> > Hi. >> > >> > I've briefly looked at the code but I didn't directly find what I was >> > looking for. >> > >> > I'm wondering if uv_tcp_t tries to write to the socket before it >> decides to >> > queue the data for next writable event? I do this in my use of >> uv_poll_t and >> > it is very fast because I do not need to copy the data and queue it if >> it >> > turns out the socket was writable at the very function call. >> > >> > So if I have a char *data I can directly write it to the kernel buffer >> and >> > skip the queueing. Is this empoyed in uv_tcp_t, or do you have any >> > opposition to this technique? >> >> uv_write() tries to write out the data immediately but will still >> queue the 'write complete' callback. >> >> uv_try_write() is like uv_write() but it skips the callback when it >> can, details in the documentation. Does that answer your question? >> >
-- You received this message because you are subscribed to the Google Groups "libuv" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at https://groups.google.com/group/libuv. For more options, visit https://groups.google.com/d/optout.
