Woo-hoo, I found the culprit after you got me thinking about the classpath. The slow web-app project's .classpath had:
<classpathentry kind="src" output="target/classes" path="src/main/resources"/> <classpathentry kind="src" output="target/test-classes" path="src/test/resources"/> That causes eclipse to do too much file-system thrashing during the build. For some reason (translation: too embarrassing to describe), several months ago I'd removed the attribute: excluding="**" Re-adding the 'excluding' reduces my reload time to 0.6 second. (Now if I get myself a new laptop with a solid-state drive like the other guys, I should get down to their 0.3 seconds. Drool.) To answer your questions (for the sake of future searches): * I was testing on simple pages with only a few components. * I was talking about the server-side request duration, as measured by AppModule's TimingFilter. On Wed, Apr 6, 2011 at 8:39 PM, Howard Lewis Ship <hls...@gmail.com> wrote: > How complicated are your pages? I've heard of these problems when > pages have deeply, deeply nested component structures (as in, loading > a page instantiates 15000 components). > > The other thing that can slow things down is all the class path > searching involved in locating the page and component classes. That > may be more efficient when the application classes are packaged into a > JAR stored under WEB-INF/lib, rather than being base classes under > WEB-INF/classes. > > Lastly, is the slowness in how long the request takes, or the time > until the changes are visible ... that is, does browser hang waiting > for content, or is the content up immediately, but doesn't reflect the > new code & resources? > > On Wed, Apr 6, 2011 at 5:30 PM, Bryan Lewis <jbryanle...@gmail.com> wrote: > > When I make a trivial change to a Tapestry page in development and > refresh > > my browser to cause a hot reload, it's taking about 11 seconds which is > > getting annoying. I don't know exactly when this started (months ago), > but > > in an older workspace the page-reload time is about 0.3 second. Both are > > using Jetty and m2eclipse. I can't figure out what the key difference > is. > > > > Just now I fiddled around with Java compiler level on all my projects > > (changed all references to 1.5 to 1.6) and got the reload time down to > 3.5 > > seconds, but I'm not convinced the compiler level was the real culprit. > The > > older workspace was using uses 1.5. > > > > Does this ring any bells? What could cause the slowness? Maybe I should > > just recreate my workspace from scratch. > > > > Thanks for any help. > > > > > > -- > Howard M. Lewis Ship > > Creator of Apache Tapestry > > The source for Tapestry training, mentoring and support. Contact me to > learn how I can get you up and productive in Tapestry fast! > > (971) 678-5210 > http://howardlewisship.com > > --------------------------------------------------------------------- > To unsubscribe, e-mail: users-unsubscr...@tapestry.apache.org > For additional commands, e-mail: users-h...@tapestry.apache.org > >
