When a CometActor is in use in a web application, the long-running requests to that Actor delays Jetty's restart when it detects a change in class files. This leads to logs that look like this:

[INFO] restarting org.mortbay.jetty.plugin.Jetty6PluginWebAppContext@6bb93c
{/,/home/debio/workspace/liftapp/src/main/webapp}
[delay ranging from a few seconds to over a minute]
INFO - Service request (GET) /comet_request/99063002597/n9frn2ckf67v took
120002 Milliseconds
[INFO] Webapp source directory =
/home/debio/workspace/liftapp/src/main/webapp
[INFO] Reload Mechanic: automatic
[INFO] Classes = /home/debio/workspace/liftapp/target/classes
[INFO] Context path = /
[INFO] Tmp directory = determined at runtime
[INFO] Web defaults = org/mortbay/jetty/webapp/webdefault.xml
[INFO] Web overrides = none
[INFO] web.xml file =
/home/debio/workspace/liftapp/src/main/webapp/WEB-INF/web.xml
[INFO] Webapp directory = /home/debio/workspace/liftapp/src/main/webapp
2010-03-20 01:00:50.737:INFO::No Transaction manager found - if your webapp
requires one, please configure one.
INFO - The CometActor lifttest.comet.Prob@1cf33a6 Received Shutdown
[INFO] Restart completed at Sat Mar 20 01:00:51 CDT 2010

Thie behavior isn't necessarily incorrect, but it does tend to hinder development as a restart that should take less than 2 seconds is extended to around a minute.
I can provide a basic code example if necessary, but this bug occurs any time a Jetty restart is attempted while a Comet request is pending.