Feedback after trying out Crowd for the first time

I have downloaded Crowd and attempted to get it working (success after a couple of hours):

1. The standalone version sort of worked. I installed it on a remote server, but the crowd.properties file had "localhost" in it so that for some actions (not all) the browser redirected to localhost instead of the actual server url which of course did not work as Crowd was not installed on localhost.

Suggestion: Check that crowd.properties is writable and update it, or at least compare it with the actual url and throw an error message during install.

2. The Crowd GUI is confusing and below the Atlassian standard:

2.1 On many pages there is no visual feedback when clicking on the "Update" buttons. Nothing happens. I find this extremely annoying and unexpected. I thought that there was a bug but actually the requested operation was in fact executed.

2.2 There should be better feedback after problems. I hope that this gets resolved when Crowd reaches 1.0.

3. I then tried to hook Crowd to Confluence. The problems were that I ended up having duplicate jar files in the Confluence WEB-INF lib directory and Confluence did not even boot anymore. Also the files crowd-core-0.4.1.jar and crowd-atlassian-user-0.4.1.jar should also go into the lib directory which is not mentioned on the "integrate with confluence" page.

Suggestion: simply make a directory "client-confluence-2.3.1" directory that has only the required files for Confluence (and another directory for JIRA etc.)

4. Another problem was an error message

2007-01-24 09:51:14,051 ERROR [TP-Processor2] [bucket.user.DefaultUserAccessor] getUser Client host is invalid: 81.169.xxx.yyy / 81
.169.xxx.yyy

after attempting to log in. I had to add my ip address to the Crowd Remote address section, the existing "localhost" entry did not work.

5. Finally, I was confused that the Confluence logins did not work after importing all Confluence users into Crowd: I had to manually reset the password.