i am not sure, but was/is there any reason why $sql_query is not always
properly initilized? or is it just missing?
i think it should happen in common.inc.php on line #429
with $GLOBALS['sql_query'] = '';
like it is done with $GLOBALS['db'] and $GLOBALS['table'] some lines above
sometimes $file['error'] is a numeric (this is expected) but sometimes
it's an array.
only reaches the default case.
To reproduce, use
CREATE TABLE `audios` (
`id` smallint(5) unsigned NOT NULL auto_increment,
`title` varchar(50) NOT NULL,
`audio` longblob NOT NULL,
PRIMARY KEY (`id`)
) ENGINE=MyISAM DEFAULT CHARSET=latin1 AUTO_INCREMENT=78 ;
insert a row then try to edit it. I get
Unknown error in file upload
for every edit.
how about fall back to cookie or http auth if config auth fails?
would make it more easy to run phpMyAdmin out of the box (at least for
but only if config is set to root without password
if config_auth_fail, user == 'root', pw == ''
than switch to cookie auth
and display message about it
On 3/22/07, Marc Delisle <Marc.Delisle(a)cegepsherbrooke.qc.ca> wrote:
> I would prefer to remove "config" auth.
While I don't object, some users may feel alienated if this change
isn't handled carefully. There are no doubt plenty of users who enjoy
the lack of authentication (many users are home users on a secured
intranet) that comes with config, and some who use .htaccess with
'config' to secure their systems (don't ask me why, but they like it
that way). I'm just saying that some users may be offended if the
feature silently disappears. Perhaps a poll on phpmyadmin.net or at
least a comment soliciting email feedback is warranted (perhaps not).
Just my thoughts, hope you all have a great day!
any odds against splitting common.lib.php into .lib.php and .inc.php?
holding all functions
holding all procedural code
(easier testing of function)
seems, somehow the BOM from changelog in trunk is missing ... i think this
was my fault ... i messed it up somehow ...
can someone fix this? or how can this be fixed ...?