Neil Bartlett wrote > [...] > > It does seem to suggest that resolution time is a factor. I wonder if > your test harness is capable of testing on Equinox and Knopflerfish as > well? The comparison could be enlightening. In particular I believe > that Equinox is able to persist resolution state and therefore might > start more quickly on the second and subsequent launches. > [...] > --------------------------------------------------------------------- > To unsubscribe, e-mail:
> [email protected] > For additional commands, e-mail: > [email protected] I also think that resolving and creating wires between bundles consumes time. It depends on how many thirdparty-bundles are required. I need about 20 libraries. Most of the thirdparty-bundles do not implement a BundleActivator or register a Service but just export some packages (e.g. log4j, slf4j, commons.io,...). So, maybe we can save time if we can somehow store the *wires* in the bundle-cache after the fist startup (Serialization?), so that they are already / immediately available at a restart. Or start the bundles fast without resolving and resolve them later. I'm just thinking out loud. -- View this message in context: http://apache-felix.18485.x6.nabble.com/How-to-improve-the-start-time-of-Apache-Felix-tp5004833p5005073.html Sent from the Apache Felix - Users mailing list archive at Nabble.com. --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
