Re: [Dovecot] Dovecot 2.0.5 and BSDI 4.3 again

2010-10-16 Thread The Doctor
On Sat, Oct 16, 2010 at 11:43:05AM +0100, Timo Sirainen wrote: > On 16.10.2010, at 3.41, The Doctor wrote: > > > The logs still show > > > > Oct 15 20:36:24 doctor dovecot: master: Dovecot v2.0.5 starting up (core > > dumps disabled) > > Oct 15 20:36:24 doctor dovecot: log: Error: conn.fd: -2 >

Re: [Dovecot] Dovecot 2.0.5 and BSDI 4.3 again

2010-10-16 Thread Odhiambo Washington
On Sat, Oct 16, 2010 at 6:21 PM, The Doctor wrote: > On Sat, Oct 16, 2010 at 11:43:05AM +0100, Timo Sirainen wrote: > > On 16.10.2010, at 3.41, The Doctor wrote: > > > > > The logs still show > > > > > > Oct 15 20:36:24 doctor dovecot: master: Dovecot v2.0.5 starting up > (core dumps disabled) > >

Re: [Dovecot] Dovecot 2.0.5 and BSDI 4.3 again

2010-10-16 Thread The Doctor
On Sat, Oct 16, 2010 at 11:43:05AM +0100, Timo Sirainen wrote: > On 16.10.2010, at 3.41, The Doctor wrote: > > > The logs still show > > > > Oct 15 20:36:24 doctor dovecot: master: Dovecot v2.0.5 starting up (core > > dumps disabled) > > Oct 15 20:36:24 doctor dovecot: log: Error: conn.fd: -2 >

Re: [Dovecot] Dovecot 2.0.5 and BSDI 4.3 again

2010-10-16 Thread Timo Sirainen
On 16.10.2010, at 3.41, The Doctor wrote: > The logs still show > > Oct 15 20:36:24 doctor dovecot: master: Dovecot v2.0.5 starting up (core > dumps disabled) > Oct 15 20:36:24 doctor dovecot: log: Error: conn.fd: -2 > Oct 15 20:36:24 doctor dovecot: log: Error: l->fd: 20 > Oct 15 20:36:24 docto

Re: [Dovecot] Dovecot 2.0.5 and BSDI 4.3 again

2010-10-15 Thread The Doctor
On Fri, Oct 15, 2010 at 10:34:10AM -0600, The Doctor wrote: > On Fri, Oct 15, 2010 at 10:19:37AM -0600, The Doctor wrote: > > On Fri, Oct 15, 2010 at 04:17:25PM +0100, Timo Sirainen wrote: > > > On Thu, 2010-10-14 at 22:43 -0600, The Doctor wrote: > > > > > > } else if (errno == EIN

Re: [Dovecot] Dovecot 2.0.5 and BSDI 4.3 again

2010-10-15 Thread The Doctor
On Fri, Oct 15, 2010 at 10:19:37AM -0600, The Doctor wrote: > On Fri, Oct 15, 2010 at 04:17:25PM +0100, Timo Sirainen wrote: > > On Thu, 2010-10-14 at 22:43 -0600, The Doctor wrote: > > > > > } else if (errno == EINVAL && > > > > >(fstat(l->fd, &st) < 0 |

Re: [Dovecot] Dovecot 2.0.5 and BSDI 4.3 again

2010-10-15 Thread The Doctor
On Fri, Oct 15, 2010 at 04:17:25PM +0100, Timo Sirainen wrote: > On Thu, 2010-10-14 at 22:43 -0600, The Doctor wrote: > > > > } else if (errno == EINVAL && > > > >(fstat(l->fd, &st) < 0 || > > > > !S_ISFIFO(st.st_mode))) { > > > >

Re: [Dovecot] Dovecot 2.0.5 and BSDI 4.3 again

2010-10-15 Thread Timo Sirainen
On Thu, 2010-10-14 at 22:43 -0600, The Doctor wrote: > > > } else if (errno == EINVAL && > > >(fstat(l->fd, &st) < 0 || > > > !S_ISFIFO(st.st_mode))) { > > > /* BSDI fails accept(fifo) with EINVAL. */ Oh, the checks in this code

Re: [Dovecot] Dovecot 2.0.5 and BSDI 4.3 again

2010-10-14 Thread The Doctor
On Thu, Oct 14, 2010 at 04:45:00PM +0100, Timo Sirainen wrote: > On Sun, 2010-10-10 at 17:02 -0600, The Doctor wrote: > > > Oct 10 16:30:45 doctor dovecot: log: Error: net_accept() failed: Invalid > > argument > > So my check below doesn't work. Add here: > > if (errno != EINVAL) > i_error("e

Re: [Dovecot] Dovecot 2.0.5 and BSDI 4.3 again

2010-10-14 Thread Timo Sirainen
On Sun, 2010-10-10 at 17:02 -0600, The Doctor wrote: > Oct 10 16:30:45 doctor dovecot: log: Error: net_accept() failed: Invalid > argument So my check below doesn't work. Add here: if (errno != EINVAL) i_error("errno = %d %m"); else if (fstat(l->fd, &st) < 0) i_error("fstat() failed: %m");

Re: [Dovecot] Dovecot 2.0.5 and BSDI 4.3 again

2010-10-11 Thread Charles Marcus
On 2010-10-10 7:02 PM, The Doctor wrote: > Configuration being used > > > # 2.0.0: dovecot.conf > # OS: BSD/OS 4.3 i386 This says you're using 2.0.0, not 2.0.5? -- Best regards, Charles

[Dovecot] Dovecot 2.0.5 and BSDI 4.3 again

2010-10-10 Thread The Doctor
Right, Changes made to dovecot-2.0.5/src/lib-master/master-service.c conn.listen_fd = l->fd; conn.fd = net_accept(l->fd, &conn.remote_ip, &conn.remote_port); i_error("conn.fd: %d",conn.fd); i_error("l->fd: %d",l->fd);