shibboleth-dev - Re: More defined custom extensions mechanism
Subject: Shibboleth Developers
List archive
- From: Chad La Joie <>
- To: Shibboleth Developers <>
- Subject: Re: More defined custom extensions mechanism
- Date: Wed, 06 Jul 2005 12:42:47 -0400
- Organization: UIS - Project Sentinel
I think I understand what you're saying Tom. Let me offer a suggestion. Anyone running an IdP has to configure the IdP, the build doesn't do that. Instead, why don't you treat the configuration of your name mapper the same way that all the other components like this (name mappers, artifact mappers, protocol handlers, etc) are configured in the IdP? Just include some number of elements within the NameMapping element in the config file the use them in your mapper. So you might have a name mapper deceleration like this:
<NameMapping id="some_string" class="gridshib.name.mapper.class">
<GridMapsDirectory>/home/gridshib/gridmaps/</GridMapsDirectory>
<GridMapFile>/etc/gridshib/gridmap-default</GridMapFile>
</NameMapping>
You can include any number of sub-elements in the NameMapping element, they're just passed to the Mapper when it's initialized.
This way your name mapper behaves just like the standard ones that come with Shib and you can put your files or directory of files wherever you want.
Tom Scavo wrote:
I guess the best thing for me to do at this point is to fall back on
first principles and try to describe (again) the GridShib requirements
with respect to installation, configuration, maintenance, and so
forth. If the requirements can't be met, that's fine, but at least it
won't be because they're not well defined. I'll try to keep this as
simple as I can.
GridShib has its own metadata file, which contains role descriptors
for every grid service provider the IdP trusts. Initially, we will
ship GridShib with a metadata stub containing a single role
descriptor, that of a test grid service provider. Over time, other
role descriptors will be added to this metadata file. In fact, we
expect this metadata file to be actively maintained by the IdP
sysadmin.
The GridShib plugin, an implementation of NameIdentifierMapping, loads
name mappings from files called gridmap files. Gridmap files are
ordinary text files, loaded and parsed at startup. There is (or will
be) a refresh capability so that gridmap files are auto-reloaded
on-the-fly as needed.
Like the metadata file, the gridmap files are actively maintained by
the IdP sysadmin. Entries may be added, modified, or deleted from
existing gridmap files. New gridmap files may be created (by
definition gridmap filenames start with the characters "gridmap")
while old gridmap files may be deleted, all of which will be handled
dynamically by the plugin.
Okay, I really should stop here because now you can mostly infer the
requirements by reading between the lines. I will emphasize, however,
the obvious requirement that the definitive versions of the metadata
and gridmap files must be totally separate from Shibboleth. They can
not live in the Shibboleth installation directory, nor can they reside
in a tomcat webapp directory (both of which are volatile). Instead,
the metadata and gridmap files must be stored in a home directory
known only to GridShib (otherwise we run the risk of some process
clobbering them).
I appreciate what you're trying to do with the custom/ directory, and
I'm sure many extension developers will find it useful, but unless I'm
missing something it doesn't help our use case very much since the
definitive versions of the metadata and gridmap files are not in the
custom/ directory but rather in the GridShib home directory. Unless
we can work out a system where each extension has its own home
directory, I'm afraid we won't be able to make much use of the build
process you propose.
I apologize for not making our requirements more clear before (this
has been and continues to be an ongoing process). I realize you're
trying to get Shib 1.3 out of the door, so if this isn't doable in the
short timeframe you have before you, or if you feel ours is not a
generally applicable use case, that's fine, we'll figure something
else out.
Tom
--
Chad La Joie 315Q St. Mary's Hall
Project Sentinel 202.687.0124
- Re: More defined custom extensions mechanism, (continued)
- Re: More defined custom extensions mechanism, Tom Scavo, 07/05/2005
- Re: More defined custom extensions mechanism, Chad La Joie, 07/05/2005
- Re: More defined custom extensions mechanism, Walter Hoehn, 07/05/2005
- Re: More defined custom extensions mechanism, Tom Scavo, 07/06/2005
- RE: More defined custom extensions mechanism, Scott Cantor, 07/06/2005
- Re: More defined custom extensions mechanism, Tom Scavo, 07/06/2005
- RE: More defined custom extensions mechanism, Scott Cantor, 07/06/2005
- Re: More defined custom extensions mechanism, Tom Scavo, 07/06/2005
- RE: More defined custom extensions mechanism, Scott Cantor, 07/06/2005
- Re: More defined custom extensions mechanism, Tom Scavo, 07/06/2005
- Re: More defined custom extensions mechanism, Tom Scavo, 07/06/2005
- RE: More defined custom extensions mechanism, Scott Cantor, 07/06/2005
- Re: More defined custom extensions mechanism, Tom Scavo, 07/06/2005
- Re: More defined custom extensions mechanism, Walter Hoehn, 07/05/2005
- Re: More defined custom extensions mechanism, Chad La Joie, 07/06/2005
- Re: More defined custom extensions mechanism, Tom Scavo, 07/06/2005
- Re: More defined custom extensions mechanism, Walter Hoehn, 07/06/2005
- Re: More defined custom extensions mechanism, Tom Scavo, 07/06/2005
- Re: More defined custom extensions mechanism, Walter Hoehn, 07/06/2005
- Re: More defined custom extensions mechanism, Tom Scavo, 07/06/2005
- Re: More defined custom extensions mechanism, Tom Scavo, 07/06/2005
- Re: More defined custom extensions mechanism, Chad La Joie, 07/05/2005
- Re: More defined custom extensions mechanism, Tom Scavo, 07/05/2005
- Re: More defined custom extensions mechanism, Ian Young, 07/06/2005
- Re: More defined custom extensions mechanism, Chad La Joie, 07/06/2005
- RE: More defined custom extensions mechanism, Scott Cantor, 07/06/2005
- Re: More defined custom extensions mechanism, Ian Young, 07/08/2005
- Re: More defined custom extensions mechanism, Tom Scavo, 07/08/2005
Archive powered by MHonArc 2.6.16.