[Bug 3271] Rsync instances stay in memory when using in daemon mode

bugzilla-daemon at dp3.samba.org bugzilla-daemon at dp3.samba.org
Sat Dec 3 04:23:59 GMT 2005


------- Comment #9 from euclid80 at users.sf.net  2005-12-02 21:23 MST -------
(In reply to comment #8)
> Thanks for the suggested patches.  I agree that those continue statements after
> select() returns -1 look like they should be improved, but I'm not sure why you
> allowed the code in the read_timeout() function to fall through after an error:

I didn't change read_timeout().

> In the writefd_unbuffered() function, what errno is being returned when your
> new code is being triggered?

Good point.  When select() returns an error, we ought not to write.  So, my
code could be put right after check_timeout().  The bug is caused by the fact
that select() doesn't return an error.  It keeps returning 0.  In a perfect
world, select() would detect that the socket is in an error state, but it
doesn't.  It sees a full transmit buffer and says that the socket is not ready.

How about this:

if (count == 0) {
  --- my patch ---

if (count <= 0) {
   --- leave this as is ---

Configure bugmail: https://bugzilla.samba.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug, or are watching the QA contact.

More information about the rsync mailing list