Below is the list of changes that have just been committed into a local
5.0 repository of iggy. When iggy does a push these changes will
be propagated to the main repository and, within 24 hours after the
push, to the public repository.
For information on how to access the public repository
see http://dev.mysql.com/doc/mysql/en/installing-source-tree.html
ChangeSet@stripped, 2007-09-21 11:38:23-04:00, iggy@alf.(none) +1 -0
Bug #15327: configure: --with-tcp-port option being partially ignored
make sure that if builder configured with a non-standard (!= 3306)
default TCP port that value actually gets used throughout.
win/configure.js@stripped, 2007-09-21 11:38:19-04:00, iggy@alf.(none) +35 -2
Bug #15327: configure: --with-tcp-port option being partially ignored
If MYSQL_TCP_PORT defaulted in configure (factory default 3306
at the time of this writing), set MYSQL_TCP_PORT to factory
default, then clear factory default after. That way, we lose no
information, and we can distinguish between "defaulted" and the
pathological case "builder specifically configured a port that
coincides with factory default." This can in theory happen if
builder configures and builds several servers from a script
(--with-tcp-port=3306, --with-tcp-port=3316, --with-tcp-port=3326).
Not all that probable, but much preferable to having more "magic"
happen in the server when we can solve this without any guesswork.
diff -Nrup a/win/configure.js b/win/configure.js
--- a/win/configure.js 2007-04-23 15:41:22 -04:00
+++ b/win/configure.js 2007-09-21 11:38:19 -04:00
@@ -31,6 +31,7 @@ try
configureInTS.Close();
var default_comment = "Source distribution";
var default_port = GetValue(configureIn, "MYSQL_TCP_PORT_DEFAULT");
+ var actual_port = 0;
var configfile = fso.CreateTextFile("win\\configure.data", true);
for (i=0; i < args.Count(); i++)
@@ -58,10 +59,41 @@ try
default_comment = parts[1];
break;
case "MYSQL_TCP_PORT":
- default_port = parts[1];
+ actual_port = parts[1];
break;
}
}
+ if (actual_port == 0)
+ {
+ // if we actually defaulted (as opposed to the pathological case of
+ // --with-tcp-port=<MYSQL_TCP_PORT_DEFAULT> which might in theory
+ // happen if whole batch of servers was built from a script), set
+ // the default to zero to indicate that; we don't lose information
+ // that way, because 0 obviously indicates that we can get the
+ // default value from MYSQL_TCP_PORT. this seems really evil, but
+ // testing for MYSQL_TCP_PORT==MYSQL_TCP_PORT_DEFAULT would make a
+ // a port of MYSQL_TCP_PORT_DEFAULT magic even if the builder did not
+ // intend it to mean "use the default, in fact, look up a good default
+ // from /etc/services if you can", but really, really meant 3306 when
+ // they passed in 3306. When they pass in a specific value, let them
+ // have it; don't second guess user and think we know better, this will
+ // just make people cross. this makes the the logic work like this
+ // (which is complicated enough):
+ //
+ // - if a port was set during build, use that as a default.
+ //
+ // - otherwise, try to look up a port in /etc/services; if that fails,
+ // use MYSQL_TCP_PORT_DEFAULT (at the time of this writing 3306)
+ //
+ // - allow the MYSQL_TCP_PORT environment variable to override that.
+ //
+ // - allow command-line parameters to override all of the above.
+ //
+ // the top-most MYSQL_TCP_PORT_DEFAULT is read from win/configure.js,
+ // so don't mess with that.
+ actual_port = default_port;
+ default_port = 0;
+ }
configfile.WriteLine("SET (COMPILATION_COMMENT \"" +
default_comment + "\")");
@@ -70,7 +102,8 @@ try
GetValue(configureIn, "PROTOCOL_VERSION") + "\")");
configfile.WriteLine("SET (DOT_FRM_VERSION \"" +
GetValue(configureIn, "DOT_FRM_VERSION") + "\")");
- configfile.WriteLine("SET (MYSQL_TCP_PORT \"" + default_port + "\")");
+ configfile.WriteLine("SET (MYSQL_TCP_PORT_DEFAULT \"" + default_port + "\")");
+ configfile.WriteLine("SET (MYSQL_TCP_PORT \"" + actual_port + "\")");
configfile.WriteLine("SET (MYSQL_UNIX_ADDR \"" +
GetValue(configureIn, "MYSQL_UNIX_ADDR_DEFAULT") + "\")");
var version = GetVersion(configureIn);
| Thread |
|---|
| • bk commit into 5.0 tree (iggy:1.2477) BUG#15327 | Ignacio Galarza | 21 Sep |