Re: [Evolution] IMAP scanning gets stuck

["Ujwal S. Sathyam" <ujwal netbrowser com>]

On Fri, 2001-11-23 at 17:34, Jeffrey Stedfast wrote:
> On Fri, 2001-11-23 at 20:10, Ujwal S. Sathyam wrote:
> [snip]
> > > Anyways, we already set the socket to O_NONBLOCK before we connect() and
> > > do a select() on the socket after afaik, so this doesn't seem to work on
> > > linux systems.
> >
> > The timeout on a blocking connect() in Linux can vary depending on the
> > destination IP address. I have seen timouts of 14 minutes for
> > destination outside the LAN. In my code, I usually set the socket to
> > O_NONBLOCK with fcntl() before doing the connect. I then use poll with a
> > 10 second timeout (I have a pretty busy server), and that has always
> > worked well for me. 
>
> This is basically what we are currently doing, but according to you it's
> not working. We *do* set the socket to O_NONBLOCK with an fcntl() call
> and then call connect() and then select() on the socket and a cancel_fd
> which is basically the same as polling on the socket while also allowing
> user-intervention.
>
> as seen here:
>
>       flags = fcntl (fd, F_GETFL);
>       fcntl (fd, F_SETFL, flags | O_NONBLOCK);
>       
>       ret = connect (fd, (struct sockaddr *)&sin, sizeof (sin));
>       if (ret == 0) {
>               fcntl (fd, F_SETFL, flags);
>               return fd;
>       }
>       
>       if (errno != EINPROGRESS) {
>               close (fd);
>               return -1;
>       }
>       
>       FD_ZERO (&rdset);
>       FD_ZERO (&wrset);
>       FD_SET (fd, &wrset);
>       FD_SET (cancel_fd, &rdset);
>       fdmax = MAX (fd, cancel_fd) + 1;
>       tv.tv_usec = 0;
>       tv.tv_sec = 60 * 4;
>       
>       if (select (fdmax, &rdset, &wrset, 0, &tv) <= 0) {
>               close (fd);
>               errno = ETIMEDOUT;
>               return -1;
>       }

Yes, that looks correct indeed, though I prefer to use poll. Which file
is this so that I can put in my debugging printfs to see if it is really
doing that? 4 minutes is pretty long. How about 30 seconds or 1 minute?
If I want to hack the source, do I get it from CVS HEAD or the 1.0
brach? If it is the branch, what is the exact tag?

Also, do you have the linger socket option set? I have found that the
option is set, close() sometimes blocks until the kernel decides to
actually close the connection.

Also, you were mentioning a Stop/Cancel button? Where is that? Just so
that you know, this is in the periodic mail check, not when I hit
Send/Receive. Even when I hit Send/Receive, sometimes hitting the cancel
button does not work.

Ujwal