Comments for Javascriptguy's blog http://jsguy.com So much javascript, so little time Fri, 28 May 2010 23:28:43 +0000 http://wordpress.org/?v=2.9.2 hourly 1 Comment on Long poll jsonp by admin http://jsguy.com/?p=103&cpage=1#comment-1009 admin Fri, 28 May 2010 23:28:43 +0000 http://jsguy.com/?p=103#comment-1009 Are you suggesting that php automatically opens a mysql connection each time? Are you suggesting that php automatically opens a mysql connection each time?

]]> Comment on Long poll jsonp by chx http://jsguy.com/?p=103&cpage=1#comment-968 chx Fri, 28 May 2010 02:31:26 +0000 http://jsguy.com/?p=103#comment-968 Be careful with this -- you might pile up a ton of MySQL connections and each connection uses some MySQL buffers... you might want to close the MySQL connection before sleeping. Be careful with this — you might pile up a ton of MySQL connections and each connection uses some MySQL buffers… you might want to close the MySQL connection before sleeping.

]]> Comment on Long poll jsonp by Julian http://jsguy.com/?p=103&cpage=1#comment-726 Julian Sat, 22 May 2010 19:37:40 +0000 http://jsguy.com/?p=103#comment-726 Glad I could help :) Glad I could help :)

]]> Comment on Long poll jsonp by admin http://jsguy.com/?p=103&cpage=1#comment-696 admin Sat, 22 May 2010 04:22:08 +0000 http://jsguy.com/?p=103#comment-696 Excellent - the loading spinner is gone, thanks very much for updating your script! :) Excellent – the loading spinner is gone, thanks very much for updating your script! :)

]]> Comment on Long poll jsonp by admin http://jsguy.com/?p=103&cpage=1#comment-695 admin Sat, 22 May 2010 04:18:48 +0000 http://jsguy.com/?p=103#comment-695 Great idea Akeru! I've updated the chat.php file, it now has a configurable timeout - I've set it to a low value, (25sec), as that seems the most common value on shared hosts, you can of course set it much higher on your own host. Great idea Akeru! I’ve updated the chat.php file, it now has a configurable timeout – I’ve set it to a low value, (25sec), as that seems the most common value on shared hosts, you can of course set it much higher on your own host.

]]> Comment on Long poll jsonp by Julian http://jsguy.com/?p=103&cpage=1#comment-671 Julian Fri, 21 May 2010 16:41:09 +0000 http://jsguy.com/?p=103#comment-671 Just popping in to point out jQuery-JSONP 2.x has been released. The new implementation should limit the loading indicator problem. Nice to see the plugin in use :) Just popping in to point out jQuery-JSONP 2.x has been released. The new implementation should limit the loading indicator problem.

Nice to see the plugin in use :)

]]> Comment on Long poll jsonp by Akeru http://jsguy.com/?p=103&cpage=1#comment-416 Akeru Fri, 14 May 2010 20:15:29 +0000 http://jsguy.com/?p=103#comment-416 Thanks ! I played a bit with it today. Had no problem running the demo but ran into some issues. The first one is that the mtime is not accurate enough (2secs) so I sometimes got 3,4 times the same message until I changed the usleep to a longer value. By playing a bit with the ts parameter before is was received (using TamperData, or such) I was able to create an infinite loop of refresh, so the ts field might not be a good idea. Use maybe session for that. I also think it is a better idea to use another domain for longpolling only and send the message to the main domain. Sending should be fast and should not be hindered by waiting polls. There should also be a way to end nicely the loop after some times (less than the specified JS timeout) because on your demo, my browsers (safari and firefox) both ends up with more than one long polling running. So, adding something like $started = time(); while (... && (time() - $started) < 60) { ... } echo json_encode(....); Would prevent memory leaks and ends nicely the query before starting a new one. Anyway, thanks for the proof of concept ! Thanks ! I played a bit with it today. Had no problem running the demo but ran into some issues.

The first one is that the mtime is not accurate enough (2secs) so I sometimes got 3,4 times the same message until I changed the usleep to a longer value.

By playing a bit with the ts parameter before is was received (using TamperData, or such) I was able to create an infinite loop of refresh, so the ts field might not be a good idea. Use maybe session for that.

I also think it is a better idea to use another domain for longpolling only and send the message to the main domain. Sending should be fast and should not be hindered by waiting polls.

There should also be a way to end nicely the loop after some times (less than the specified JS timeout) because on your demo, my browsers (safari and firefox) both ends up with more than one long polling running. So, adding something like

$started = time();
while (… && (time() – $started) < 60) { … }
echo json_encode(….);

Would prevent memory leaks and ends nicely the query before starting a new one.

Anyway, thanks for the proof of concept !

]]> Comment on Long poll jsonp by admin http://jsguy.com/?p=103&cpage=1#comment-408 admin Fri, 14 May 2010 07:15:24 +0000 http://jsguy.com/?p=103#comment-408 You bet, <a href="http://jsguy.com/labs/longpolljsonp.zip" rel="nofollow">uploaded it here</a>, let me know if you have problems, or any good ideas! You bet, uploaded it here, let me know if you have problems, or any good ideas!

]]> Comment on Long poll jsonp by Akeru http://jsguy.com/?p=103&cpage=1#comment-404 Akeru Thu, 13 May 2010 19:10:26 +0000 http://jsguy.com/?p=103#comment-404 Nice article ! Could you provide a zip file with the JS and PHP code to download, so I could play a bit with the concept ? Nice article ! Could you provide a zip file with the JS and PHP code to download, so I could play a bit with the concept ?

]]>