Hi!
>>>>> "Benjamin" == Benjamin Pflugmann <[EMAIL PROTECTED]> writes:
Benjamin> Hi.
Benjamin> On Mon, Jan 22, 2001 at 11:37:23PM +0200, [EMAIL PROTECTED] wrote:
>> >>>>> "Huang-Ming" == Huang-Ming <[EMAIL PROTECTED]> writes:
>>
Huang-Ming> hi Tim,
>>
Huang-Ming> thanks for the prompt response which is hard to be seen with Oracle and
Microsoft.
>>
Huang-Ming> sorry, i should have had studied hard the manual a year ago. regarding the
error, however, would the error message be better to tell users to adjust relevant
configuration than only to say "MySQL gone away"? it's misleading to some serious
fault in MySQL which actually is not therein, isn't it?
>>
>> The reason is that to guard against wrong usage and faulty
>> connections, MySQL will abort any transactions that doesn't follow the
>> protocol it enforces. In this case MySQL notices a packet that is
>> much bigger than the maximum allowed packet size and it can't know if
>> the packet is valid or is from someone that is trying to use mysqld in
>> a way that the MySQL maintainer haven't approved of. Just to be safe,
>> mysqld will in this case abort the connections and the client will get
>> the 'MySQL gone away' error.
Benjamin> Given that, a cleaner(?) way would IMHO be, that the server sends
Benjamin> (hm. within connection approvement, if there already exists one?) the
Benjamin> status of that variable to the client.
Benjamin> Then the client library could check the size itself and issue itself
Benjamin> the appropriate error. Or at least add a note that you might have hit
Benjamin> the limit and what the current limit is.
Benjamin> I assume there are reasons why the current implementation differs.
Benjamin> But anyhow, the fact that even an developer *grin to Tim* considered
Benjamin> this behaviour a possible bug, shows that it is too non-intuitive.
Benjamin> Sorry, but my knowledge about the internals are too bad in order to
Benjamin> make a more constructive proposal what to change.
Benjamin> Bye,
Benjamin> Benjamin.
The above should be easy to do; It will howver require a protocol change,
so I will postpone this to 4.0; In 4.0 we anyway have to do a protocol
change to fix a lot of small issues that has come up with MySQL 3.23.
Regards,
Monty
---------------------------------------------------------------------
Before posting, please check:
http://www.mysql.com/manual.php (the manual)
http://lists.mysql.com/ (the list archive)
To request this thread, e-mail <[EMAIL PROTECTED]>
To unsubscribe, e-mail <[EMAIL PROTECTED]>
Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php