mysql_insert_id and mysql_pconnect

Anon

Hi,
I am using mysql_pconnect and was just wondering if it is safe to use this along with mysql_insert_id. From the manual it seems like it probably is but it doesn't deal with persistant connections explicitly anywhere so I was wondering if anyone had more info...
Then again, I just reread the manual on pconnect and it seems like most people are advising to use mysql_connect instead so maybe I'll just go through all my pages and change them to connect instead of pconnect... Any comments?
tia,
kelvin.

gid__

We personally use only mysql_connect. We tried pconnect awhile back but ran into problems. Sorry, but I can't seem remeber what the problems were 🙁 mysql_connect's are rather fast, if mysql is running on localhost use ":/tmp/mysql.sock " (I think) for the host to connect to the local socket for faster connect times.

One thing with pconnect, is that since the connection stays up, if there are any memory leaks in anything, then nothing really gets reset. Also apache threads cannot share a mysql connection (as far as I know), each thread has to have it's own connection to your mysqld.

Anon

thanks for the info... it was kinda looking like it might be a good idea to change to mysql_connect from what i read earlier today, I'm just sure that I looked into it when I started this project (months ago!) it seemed like pconnect was a better idea...
At least all it will take is a global find and replace to change it!
thanks,
kelvin.

gid__

just as a tip, I usually make a file called "_mysql.php" and put the mysql connection and datasource selection in there.

And then at the top of each file that needs the database, I can just do a require_once('_mysql.php');
then I'll be connected. This way I only have to write code to connect once. And more imporantly only have to change it once. 🙂

I use lots of these types of files for common tasks in web sites I make, it means less code which in turn means it's easier to maintain/develop.