--with-port=6667 should trump

From: Eric Prud'hommeaux <eric(at)w3.org>
Date: Sun Sep 02 2007 - 18:30:48 EDT

sql/mysqld.cc:1361 (5.1.21-beta tarball) allows /etc/services to trump $(configure --with-tcp-port=3307) by setting mysqld_port to the configure option first, and then overriding it with the /etc/services entry. Why?

I expect that very few folks control that from /etc/services, and we can allow them to continue by compiling with --with-tcp-port=0 checking first if mysqld_port is 0 (patch attached).
  if (!mysqld_port && (serv_ptr= getservbyname("mysql", "tcp")))

related discussion on #mysql-dev at 2007-09-02T21:28:41Z:

21:28:41Z  set_ports() seems a little twisted:
21:28:43Z      mysqld_port= MYSQL_PORT;
21:28:43Z      if ((serv_ptr= getservbyname("mysql", "tcp")))
21:28:43Z        mysqld_port= ntohs((u_short) serv_ptr->s_port); /* purecov: inspected */
21:29:39Z  so if you $(configure --withe-port=3307), you'll get 3307 
	  	  *unless* you have a /etc/services

21:30:02Z <Toad> yea does seem backwards 21:30:15Z <ericP> it's clever code, but i think the config option should

                    always trump
21:30:24Z <ericP> i.e. no reason to consult /etc/services 21:30:55Z <Toad> I wonder if anyone ever uses /etc/services to try to set

                   their service ports

21:30:56Z  *maybe* if the user did --with-port=0...
21:30:58Z  I know I don't expect that to happen
21:31:03Z -!- weigon__ is now known as weigon
21:31:59Z  just in case, i guess i'll motion for 
	  [[if (!mysql_port && (serv_ptr= getservbyname("mysql", "tcp")))...]]

-- 
-eric

office: +1.617.258.5741 NE43-344, MIT, Cambridge, MA 02144 USA
mobile: +1.617.599.3509

(eric@w3.org)
Feel free to forward this message to any list for any purpose other than
email address distribution.

Received on Sun Sep 2 18:31:15 2007