Hi Davi,
Thanks for your input!
Davi Arnaut wrote:
> Hi Tatiana,
>
> On 7/21/09 11:34 AM, Tatiana A. Nurnberg wrote:
>> #At file:///Users/tnurnberg/forest/35132/50-35132/ based on
>> revid:timothy.smith@stripped
>
>> @@ -387,7 +388,14 @@ int main(int argc,char *argv[])
>> {
>> if (option_wait&& !interrupted)
>> {
>> - mysql_close(&mysql);
>> + if (mysql.net.vio != 0)
>> + {
>> + free_old_query(&mysql);
>> + mysql.status=MYSQL_STATUS_READY; /* Force command */
>> + mysql.reconnect=0;
>> + simple_command(&mysql,COM_QUIT,NullS,0,1);
>> + end_server(&mysql); /* Sets mysql->net.vio= 0 */
>> + }
>
> Why this needed? It seems we should only attempt to reconnect if the
> connection is broken. Otherwise, the same connection could be used..
That, in all honesty, is some seriously surprising code. :)
Not the new stuff -- that just replaces the destructive mysql_close()
with a less destructive variant thereof --, but the surrounding logic.
Note that if --force is set, we do NOT reopen, ever. I have a case
open with monty (who wrote that bit) to clear whether it's a bug or
something arcane. That said, I have a version here
http://lists.mysql.com/commits/80902 that rectifies all that (and adds
a shitload of comments while at it, I think we both coulda saved time
trying to understand the code had it been properly documented -- though
I may have error on the side of "too much" this time). Anyway, let's
hear monty first, he'll know.
regards,
Tatiana
--
Tatiana Azundris, Bugs Bunny * IRC/Skype: Azundris * azundris@stripped
Sun Microsystems GmbH, Sonnenallee 1, 85551 Kirchheim-Heimstetten
Geschaeftsfuehrer: Thomas Schroeder, Wolfang Engels, Dr. Roland Boemer
Vorsitz des Aufsichtsrats: Martin Haering HRB MUC 161028 49.011, 8.376
