It's done by the ant target 'prefetch' in build.xml, calls build- tools/prefetch.sh
On 2006-04-05 20:40 PDT, Andy Chandler wrote: > I was referring to a stop /restart - we were speculating that the > cache was taking time to reload into memory. However the primary > developer on that project is getting pulled in too many directions > so I'm going to have to muddle through testing it myself. I > appreciate the tips received so far - just out of curiosity where > should I be looking to see how the OL distribution is doing the wget? > > > -----Original Message----- > From: P T Withington [mailto:[EMAIL PROTECTED] > Sent: Wednesday, April 05, 2006 4:09 PM > To: Andy Chandler > Cc: Scott Evans; [email protected] > Subject: Re: [Laszlo-user] pre-compile? > > When you 'bounce' tomcat, what exactly do you mean? The Openlaszlo > servlet definitely has a cache that should be preserved across tomcat > stop/start. The OL distribution is built using a script that wget's > the 'standard' demos/examples, etc. to pre-fill the cache. If you > are only seeing a delay on the first fetch by any browser of your > app, I suggest that you use a similar technique. This is the > simplest way to cause the compilation of a non-SOLO app. The server > will recompile your app if the source files for your app are newer > than the cached compiled swf. > > On 2006-04-04 00:37 EDT, Andy Chandler wrote: > > > I'll try to get you an answer on that - The app itself is fairly > > decent > > sized but I'm not sure its not broken into different pages. What > > I can > > tell you is that it is slow the first time you view the page after > > tomcat is > > bounced, however all subsequent views in the same browser are fast > > OR if you > > open a different browser on the same client it is also fast. > > Unfortunately > > he couldn't yet tell me if views from other computers are also fast > > after > > that one slow one. I'll try to do some more tests over the next > > day or > > two. Can anyone tell me though if there is a difference between > > passing > > the lzx the don't compile flag versus creating our own wrapper html > > that > > references the compiled swf? > > > > > > > > -----Original Message----- > > From: Scott Evans [mailto:[EMAIL PROTECTED] > > Sent: Monday, April 03, 2006 10:53 PM > > To: P T Withington > > Cc: Andy Chandler; [email protected] > > Subject: Re: [Laszlo-user] pre-compile? > > > > Or is this just network latency (and then browser caching)? How > > big is > > the swf? > > > > > > gse > > > > > > On Mon, 3 Apr 2006, P T Withington wrote: > > > >> There is a cache. You can pre-load it by fetching the app. You > can > >> use wget or curl to do that in a script. > >> > >> Only the very first user of your app would see a delay due to > loading > >> the cache if your app has not changed. The browser will also cache > >> your app -- you may be seeing the effect of the browser fetching > the > >> app locally for subsequent access, if you see a speed up for all > >> users on the second access. > >> > >> On 2006-04-03 17:04 EDT, Andy Chandler wrote: > >> > >>> (I keep forgetting to use the correct account to send to the > list - > >>> apologies if this winds up coming through twice) > >>> > >>> > >>> > >>> I was asked to pass this question along to the list : > >>> > >>> > >>> > >>> We have a fairly large/complicated application and it > >>> has been > >>> observed that whenever a user visits a page for the first time it > >>> can take > >>> around 15 seconds for the page to load - thereafter it loads much > >>> quicker. > >>> We believe it is due to compilation or perhaps loading the swf > into > >>> a cache > >>> on the tomcat side of the equation and we were looking to see how > >>> we could > >>> avoid that step if the resource hasn't been updated since the last > >>> compilation or if its caching is there any way to pre-load it? > >>> > >>> _______________________________________________ > >>> Laszlo-user mailing list > >>> [email protected] > >>> http://www.openlaszlo.org/mailman/listinfo/laszlo-user > >> > >> _______________________________________________ > >> Laszlo-user mailing list > >> [email protected] > >> http://www.openlaszlo.org/mailman/listinfo/laszlo-user > >> > _______________________________________________ Laszlo-user mailing list [email protected] http://www.openlaszlo.org/mailman/listinfo/laszlo-user
