code review planning
Rod Widdowson
rdw at steadingsoftware.com
Mon Jul 1 05:59:23 EDT 2013
Presenting further feedback from the work following on from this review:
I have made many of the changes suggested, and annotated the wiki
appropriately.
One change is not trivial and I am still considering the options. I made a
note in the JIRA case
(https://issues.shibboleth.net/jira/browse/IDP-287#comment-15495) about
that.
I also took an action to look at the three potential implementation of
AttributeRule
1) Flat, with explicit permit and deny rules as members
2) Flat with a single rule member and a boolean to flag permit or deny.
3) As a base class with two implementations ("PermitAttributeRule" and
"DenyAttributeRule").
The current mainline reflects (1) and over on GitHub
(https://github.com/rodwiddowson/java-identity-provider) there is a recent
duplicate of SVN with a couple of branches showing the two implementations.
I haven't got the tests clean, but
There's not a huge difference in the implementations or the config parsing.
I rather think that I prefer (3), Commit
e3ab82a539a3ed9c10d92d969cff8f612858f2fa has the parsing change for that and
5433f71f712a87a332b4d5918ed72c652bdfbb4b (much of) the implement changes.
Based on this, I think that I am convinced of the need (and ease involved)
in separation the implementations of the Policy Applicability predicate from
the Matcher implementation. I want to play a little bit with the
implementation (and critically the parsing) of this before I commit to it,
but I'll keep the list posted.
Thanks
Rod
> -----Original Message-----
> From: dev-bounces at shibboleth.net [mailto:dev-bounces at shibboleth.net]
> On Behalf Of Rod Widdowson
> Sent: 22 June 2013 11:29
> To: 'Shib Dev'
> Subject: RE: code review planning
>
> Thanks to all of you who say through that review yesterday.
>
> I was not scribing, but I took some notes and I have transferred them up
to
> the wiki at https://wiki.shibboleth.net/confluence/x/94O3
>
> Please feel free to amend or add to them. Further discussion should
> probably take place in this mail thread or the next meeting. I'll report
on
> my action items in this mail thread.
>
> Rod
>
> --
> To unsubscribe from this list send an email to dev-
> unsubscribe at shibboleth.net
More information about the dev
mailing list