As you point out the security is a problem, the other issue is transactions. Alfresco's underlying API is service based and is coarse grained compaired to the JCR API so there are some challenges in mapping the two IMO. The security issues can be fixed with a patch. We're still looking at transactions.
-Russ -----Original Message----- From: Daniel Murley [mailto:[EMAIL PROTECTED] Sent: Monday, April 16, 2007 7:34 PM To: [email protected] Subject: RE: Re: JackRabbit Vs Alfresco The RMI layer for alfresco is somewhat poor, and virtually unusable if you're not using Acegi for your security. If you're not using Acegi for your security, you can expect to be able to execute 2 - 3 commands before it says you're unauthenticated (my guess would be because Acegi uses ThreadLocals, and RMI can be executed on different threads?). The Jackrabbit RMI support is significantly better, and as such after struggling with Acegi for 3-4 weeks, chucked it and had Jackrabbit doing the same thing in 2-3 days. Daniel -----Original Message----- From: Jukka Zitting [mailto:[EMAIL PROTECTED] Sent: Sunday, 15 April 2007 6:17 PM To: [email protected] Subject: Re: Re: JackRabbit Vs Alfresco Hi, On 4/15/07, Danner, Russ <[EMAIL PROTECTED]> wrote: > Another important thing to note about the implementation is that > currently the Alfresco JCR implementation has to run "in-process." Note that the same applies to by default also to Jackrabbit. You can use the jackrabbit-jcr-rmi remoting layer with any JCR implementation, so it should work fine also with Alfresco. BR, Jukka Zitting
