Quick question -- OCAU please help

[yihfeng]

If I have a php script running an infinite loop (say

while (true)
{
    mysql_query("delete * from table");
}

and I load the script from a browser window. If I click on stop in the browser, does the script immediately terminate? Or will it still run in for a while before the server runs some clean up processes to clean it up?

Thanks!

[teegman]

The php script will keep on running till the maximum execution timelimit is hit, as defined in php.ini

(i.e. hitting stop ain't gonna stop it)

[yihfeng]

No wonder I can't find that goddamned bug in my application! And I suspect my University's sys admin is not amused by the amount of resources my "background" script is taking up.

I always thought that the PHP script will terminate once the HTTP request for it is terminated Guess I was wrong.

Thanks teegman! Is there a solution for this? Or should I just not run apps like this?

[teegman]

What are you trying to do?

Running a mysql query in an infinite loop is extremely bad practice.. the solution is thinking about what you are trying to achieve and deriving a way of programming it that isn't going to totally kill the system's resources

[yihfeng]

I'm trying to do this: http://forums.overclockers.com.au/sh...d.php?t=594906

What basically happens is that I run a loop checking if the database has been updated with a new message, and if it has, send it to the user... Not sure how I'll implement that though, maybe I'll think about it a little bit more.

Thanks though You've been most helpful

[yihfeng]

Oh yea, I do realize that how much easier this would be should it be written in Java rather than PHP since in Java you can have thread control, send one thread to sleep, and wake the other thread up as as you insert data into the database, but I'll see what can be done in PHP or if it even can be done

Cheers.

[teegman]

I wouldn't be thinking loops for something like this, what I'd be thinking is 'how do I know when the database will have new data that needs to be sent to the users?'. The simple answer is whenever the database is being updated. Presumably you have control of the code that is updating the database, so whenever anyone fires off an action that results in the database being updated, run whatever code you need to run to update the other users then. Did that make any sense?

oh and p.s., you trying the cometd stuff?

[yihfeng]

I wouldn't be thinking loops for something like this, what I'd be thinking is 'how do I know when the database will have new data that needs to be sent to the users?'. The simple answer is whenever the database is being updated. Presumably you have control of the code that is updating the database, so whenever anyone fires off an action that results in the database being updated, run whatever code you need to run to update the other users then. Did that make any sense?

Yes, and can that be done in PHP? How would one running script let another running script know that something has happened?

I know how to do that in Java, but looking through the PHP manual I couldnt' find how to do it. Unless I'm missing out something...

oh and p.s., you trying the cometd stuff?

Yes!

[mordy]

basically the php script runs from beginning to end. Usually a HTTP request will start it, and it will terminate as the HTTP connection is closed.

If the script has an infinite loop, it will just keep running continously in the apache handler thread/process and the HTTP connection wont be closed (it might time out on the user side)

I dont believe that the stop button on the browser sends a termination request, i dont believe the HTTP protocol has one of those. It just stops seeking a response.

[effekt26]

As i understand it, all php code is executes on the server, and once that has been executed, it returns the HTML markup to the browser. So, therefore, the script is continually running, and no html is being sent to the browser...

Well, thats my understanding of it, please correct me if i'm wrong