hmm, I didn't know about the Portlet issues. Sounds like Portlets in general a broken - but of course a statement like that doesn't solve anyone's problem.
I know squat about Ptlets so I'm afraid I can't be of much help. Geoff On 6/15/06, John Singleton <[EMAIL PROTECTED]> wrote:
This initial hit isn't so much of a problem for Servlet based applications, but for Portlet applications it is, as it's per portlet, and even per portlet instance for some classes of Portlet. It's bad enough for us to have given up trying to use Tapestry for our portlet applications. John ----- Original Message ---- From: Geoff Longman <[EMAIL PROTECTED]> To: Tapestry users <users@tapestry.apache.org> Sent: Thursday, 15 June, 2006 3:30:50 PM Subject: Re: To Generate Tapestry Page Class Before 1st Access Trying to do this is a wasted effort regardless of the tool chosen to do it. I mean that even if you hit every page in the application, you are only injecting one instance of each page into the pool. All you buy is a faster response for the first access. Subsequent, concurrent, access of any page may cause another instance to be created and pooled. So, until the pool contains enough instances to serve the load, you end up paying the price over and over anyways. The cost is minimal, no customer of ours has ever complained about this. Geoff On 6/15/06, Bryan Lewis <[EMAIL PROTECTED]> wrote: > I believe the slowness you're seeing at initial page load isn't so much > the class enhancement (and not the compilation, of course, since that's > already done); it's the parsing and initialization of the component > graph. There is a way to trick Tapestry into preloading all the pages > on first access... check a couple of old threads here: > http://marc.theaimsgroup.com/?l=tapestry-user&w=2&r=1&s=preload&q=b > > But it's cumbersome and imperfect and it turns out not to be a big > benefit. It makes the first access of the app a lot slower, in exchange > for a small speed-up on the first visit to each page. We did it that way > for a long time and stopped it. The one nice thing about it was the > syntax checking of all the pages at startup, which is handy in > development when you make a global change and don't want to visit all > the pages manually. You could achieve the same benefit (and more) by > using an http tester tool (JMeter, for example) to visit all your pages. > > > seng kim khong wrote: > > > Hi All, > > > > If I not wrong, the page class in tapestry is an abstract class and it > > will > > be extends in run time and generate instance based on that class on > > the fly. > > But this will let the 1st time loading of tat page become slower. > > > > So, I would like to check out is it tapestry got provide any script to > > compile (maybe I use the wrong keyword) the page class 1st before the > > user > > request to that particular page to make it can be run faster. This > > concept > > is like the jspc which can turn the jsp page to be servlet 1st before > > user > > request to that paricular jsp page. > > > > Thank you in advance for any info. > > > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [EMAIL PROTECTED] > For additional commands, e-mail: [EMAIL PROTECTED] > > -- The Spindle guy. http://spindle.sf.net Blog: http://jroller.com/page/glongman Other interests: http://www.squidoo.com/spaceelevator/ --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
-- The Spindle guy. http://spindle.sf.net Blog: http://jroller.com/page/glongman Other interests: http://www.squidoo.com/spaceelevator/ --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
