Re: read() call in timer_cb

On Mon, Jun 8, 2020 at 10:35 PM Daniel Stenberg <daniel_at_haxx.se> wrote:

> On Mon, 8 Jun 2020, James Read via curl-library wrote:
>
> > This is because my web crawler uses this function and the read() call
> seems
> > to be causing problems with my web crawler.
> >
> > What is the read() call for?
>
> Obviously it reads from the file descriptor created by timerfd_create().
> Which
> is a Linux timer API.
>
> > It has been suggested to me elsewhere that I could use select instead. Is
> > that right?
>
> Instead of epoll you mean? If you intend to use the multi_socket API then
> using select is not advisable, as it really is not suitable for
> event-based
> systems.
>
>
It was suggested to me to use select instead of read. I know. That left me
scratching my head as well. As far as I can see the read call is made so
that the special case of receiving an errno = EAGAIN can be returned
without calling curl_multi_socket_action(g->multi, CURL_SOCKET_TIMEOUT, 0,
&g->still_running) How select is an alternative to this is beyond my very
limited knowledge.

It also escapes me why calling read would interfere
with mysql_store_result_nonblocking(mysql_conn, &result) which is what
seems to be happening in my program.

James Read

> > What would the behaviour of the program be if the read call was simply
> > taken out?
>
> I can't answer that, I really don't know that interface.
>
> --
>
> / daniel.haxx.se | Commercial curl support up to 24x7 is available!
> | Private help, bug fixes, support, ports, new features
> | https://www.wolfssl.com/contact/
>

-------------------------------------------------------------------
Unsubscribe: https://cool.haxx.se/list/listinfo/curl-library
Etiquette: https://curl.haxx.se/mail/etiquette.html
Received on 2020-06-09