geronimo-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Alan Cabrera (JIRA)" <>
Subject [jira] Commented: (GERONIMO-420) Refactor *UserPrincipal and *GroupPrincipal
Date Tue, 02 Nov 2004 12:27:32 GMT
     [ ]
Alan Cabrera commented on GERONIMO-420:

Extraction of principals can be done by class.  It might be helpful to still have the different
principal classes.

In any case, using the the off the shelf login modules will give you many different kinds
of principals; just try logging in via Kerberos, an NT domain, or a Solaris domain. 

I guess that it's not clear to me that there is a problem here to solve.

> Refactor *UserPrincipal and *GroupPrincipal
> -------------------------------------------
>          Key: GERONIMO-420
>          URL:
>      Project: Apache Geronimo
>         Type: Improvement
>   Components: security
>     Versions: 1.0-M2
>     Reporter: Aaron Mulder

> It would be nice if the default security realms all used the same user and group principal
implementation classes.  For one thing, the code is 99% identical across all 4 existing implementation
classes.  But also, these are essentially generic, and it would be nice to provide generic
implementations of users and groups that custom login modules could take advantage of, instead
of setting the precedent that every module must have a new and different set of principal
> If we truly make the user and group principals generic, we would lose the current behavior
that equals() can distinguish between a SQLUserPrincipal "foo" and a PropertiesFileUserPrincipal
"foo", but I'm not sure that's all that important anyway -- the required class name goes into
the configuration files, so if the class name isn't correct then the principal needs to be
discarded regardless of what equals() reports.  Plus, a principal "foo" from FileRealm "bar"
is not actually the same as a principal "foo" from FileRealm "baz" even though the user principal
implementation classes and user principal names are both the same -- so the current equals
implementation isn't fully correct anyway.
> However, if we still want different classes for different realm types for any reason,
we could at least put all the code in a base class and then have empty subclasses.

This message is automatically generated by JIRA.
If you think it was sent incorrectly contact one of the administrators:
If you want more information on JIRA, or have a bug to report see:

View raw message