I just noticed that since upgrading (from v3.5) to 4.0.4, no SQL history had been recorded using the configuration storage settings. It was working fine before.
So I logged out and checked logging in with the control setting but it said the details were wrong. I logged in as root again and edited the user to set the password it has in the config (which should have been working and I haven't changed it).
Then it would let me log in with the control user account, but now it has an error, "Connection for controluser as defined in your configuration failed", even though it definitely works fine now! Something strange is going on here...
PHP: 5.4.9
MySQL: 5.5.31
phpMyAdmin: 4.0.4 (latest)
I should also add:
Sorry, I just realised my mistake with the controluser setting. The original password had a 1 (one) instead of l (L), which was in the config, then when I changed the password I used l instead of 1. D'oh! So ignore that part of the bug, the control user is working fine now.
But the issue with SQL history still stands - SQL history is not being saved.
With version 4.0.4.1 and setting
$cfg['QueryHistoryDB'] = true;
I cannot reproduce this bug. I type a query in query window and it gets saved in the history table; it's also displayed in the SQL history tab.
Ah OK, I was missing
$cfg['QueryHistoryDB']
. Why is this an option when I've already set the$cfg['Servers'][$i]['history']
value?Good question. This is an option because, according to the doc, when it's set to false, a non-permanent history is supposed to be built. It looks like this feature no longer works (or it was decided to be removed and the doc was not changed) so I'm reopening your ticket and will change the subject.
The non-permanent history (session-based) problem is fixed for 4.0.5.
Hi, how was this latest change you made supposed to work? PMA is still not recording the query history (in v4.0.8) unless I have the
$cfg['QueryHistoryDB']
option set. If that's how it's supposed to work then that's fine, but it sounded like you were removing the need to set that variable.Yes, that's how it's documented to work. I fixed the non-permanent history mechanism.