Why does MySQL’s version comment change when logging is enabled?
I wonder if the MySQL archaeologists out there would be willing to unearth some (presumably ancient) history for me. Why does the logging configuration merit special mention in the version_comment variable?
The more I think about this, the more bizarre it seems. I enabled logging. The version reported by the server changed. No, really, is my server somehow a different version of MySQL now?
130203 15:39:55 [Note] ./bin/mysqld: ready for connections.
Version: '5.6.7-rc-log'
I assume there’s a good story behind this somewhere. I’m thinking a priest, a rabbi, and Monty walk into a bar, and black vodka is probably involved at some point too :-)



I thought just the same when I saw that!
golimar
3 Feb 13 at 5:56 pm
mysqld.cc has set_server_version() which has
if (opt_log || opt_update_log || opt_slow_log || opt_bin_log)
strmov(end, “-log”); // This may slow down system
The “// This may slow down system” comment is priceless IMHO.
Bzr archeology traces this comment-but not the “-log” itself-to Monty:
revno: 1346.620.1
committer: monty@mysql.com
timestamp: Wed 2004-05-19 16:38:12 +0300
message:
Added support for projects ‘classic’, ‘classic nt’, ‘pro’ and ‘pro nt’
Ensured that all projects compile
Removed compiler warnings
Better setting of server_version variable.
Fix that make_win_src_distribution creates the privilege tables.
Laurynas
4 Feb 13 at 2:35 am
My wager is it’s a quick hack to know if they can just ask for the logs that they know exist or if they have to get the reporter to turn on logging and reproduce the bug before they can get logged information.
Rob Smith
4 Feb 13 at 3:38 am
Laurynas,
you need to go deeper :)
The previous version of this line – before that revno 1346.620.1 – looked like
if (opt_log || opt_update_log || opt_slow_log || opt_bin_log)
strcat(server_version,”-log”);
it has no comment yet, and it is from the revision
revno: 2
revision-id: sp1r-bk@work.mysql.com-20000731192914-08846
parent: sp1r-bk@work.mysql.com-20000731191005-11438
committer: bk@work.mysql.com
timestamp: Mon 2000-07-31 21:29:14 +0200
message:
Import changeset
So, here you are. It’s from before-bk days, if Monty has the old CVS repository, he might be able to look this line up. Otherwise nobody can/
Sergei
4 Feb 13 at 3:58 am
http://bugs.mysql.com/bug.php?id=57628
(remove the useless -log suffix from server version)
sbester
4 Feb 13 at 3:13 pm
Some interesting MySQL archaeology in the above comments!
Raghavendra Prabhu
17 Feb 13 at 5:49 pm