Re: Article on removing codebase property
Dan Creswell <[email protected]>
| Newsgroups | gmane.comp.java.sun.jini |
|---|---|
| Message-ID | <[email protected]> |
Hi Calum, I think my preferred option is to bring the discussion over here as we're talking a Jini specific issue that we perhaps want to address in a project of our own or perhaps as part of River at some point. IMHO, the TSS audience won't have much interest in the debate itself but might be interested in anything that results from it such as a new API or project or whatever. Thoughts, Dan. Calum Shaw-Mackay wrote: > Perhaps we might do well putting some of this information on the > serverside thread, and see what the author thinks about > PreferredClassLoader over his approach? > > --Calum > > On 20/02/07, Gregg Wonderly <[email protected]> wrote: >> >> >> Calum Shaw-Mackay wrote: >> > Any thoughts? >> > >> > http://www.theserverside.com/news/thread.tss?thread_id=44288 >> >> The main issue, is making the client utilize an unchanging codebase so >> that you >> can update the content without having to worry about changing the >> codebase spec >> on the server deployment and/or codesource security specs used by the >> client. >> >> I utilize the class dependency tools to build a composite jar for my >> services >> -dl.jars. I do include jsk-dl.jar explicitly in the codebase. And, >> there are a >> couple of larger jars that I sometimes include in Class-Path: meta >> data too. >> >> There are multiple ways that exist to day to minimize changes in your >> build/deployment as your service evolves. >> >> As Dan said, we've essentially "solved" that problem. What we need to >> really >> work on is some better handling of failover and such. I have an HTTP >> server in >> the startnow project which I haven't deployed, but which I started >> work on, and >> have done some testing with. It is, in fact a service. It provides >> facilities >> for a service to ask "who can serve jar XYZ.jar" when that service is >> building >> its codebase. >> >> It might be more interesting for the client to ask this question so that >> failover is very close to the event that needs to mediate the >> failure. But, I >> thought it would be less pain and security configuration for the >> client if the >> server made that decision as the codebase annotation is queried for at >> marshalling time. >> >> Gregg Wonderly >> > > -------------------------------------------------------------------------- > Getting Started: http://www.jini.org/wiki/Category:Getting_Started > Community Web Site: http://jini.org > jini-users Archive: http://archives.java.sun.com/archives/jini-users.html > Unsubscribing: email "signoff JINI-USERS" to [email protected] > -------------------------------------------------------------------------- Getting Started: http://www.jini.org/wiki/Category:Getting_Started Community Web Site: http://jini.org jini-users Archive: http://archives.java.sun.com/archives/jini-users.html Unsubscribing: email "signoff JINI-USERS" to [email protected]