Re: Asynchronous DNS

[Alain Schneble]

Lars Ingebrigtsen <address@hidden> writes:

> I had assumed that getaddrinfo_a would have the normal timeouts that
> getaddrinfo has -- that it would say that it's failed after a certain
> number of seconds.  Some testing here (switching off the wifi after
> issuing a DNS request) seems to indicate that that may not be the case?

I was just looking into the glibc implementation of getaddrinfo_a again.
It invokes getaddrinfo from its worker threads and it seems like it
should not behave differently regarding timeouts.  The only timeout I
found was to make the worker threads sleep when there is nothing to do.
But they should get woken up if a new request arrives or they will
eventually die when the timeout has elapsed.  But all this should not be
related to the timeout you are referring to.

> I couldn't find anything in the documentation about this...  am I
> overlooking something?  If not, I have to implement some timeouts myself
> in that code.

But what seems strange to me right now is this code in check_for_dns:

  /* The DNS lookup failed. */
  else if (!EQ (p->status, Qconnect))
    {
      ...

Actually, I tried it out and DNS lookups seem to timeout quite fast on
my machine if there is no network connection.  But they do not enter the
if-branch above.  That seems like an error.  Am I day-dreaming or
shouldn't it be rewritten to look like this?

  /* The DNS lookup failed. */
  else if (EQ (p->status, Qconnect))
    {
      ...