Dependencies / activation conditions

[Rod Widdowson]

It will be pure accident.  The dependency code is substantially unchanged from v2 (kinda) and the activation conditions were added later.

As you observe that code is delicate, but I cannot see a reason why moving them around would matter , but I’ll try to remember to look the code over when I get back to my desk

One thing to note is that dependency analysis will recursively call the resolvers and that might impact (but I cannot imagine how)

Sent from my iPhone

> On 20 Nov 2017, at 16:37, Cantor, Scott <cantor.2 at osu.edu> wrote:
> 
> Rod, do you recall if we intentionally for some reason had the code resolving dependencies before actually testing the activation condition for an attribute definition?
> 
> It was a bit non-intuitive but I know there are some complications in that code.
> 
> -- Scott
> 
> -- 
> To unsubscribe from this list send an email to dev-unsubscribe at shibboleth.net