Speaking of service-now...
cantor.2 at osu.edu
Wed Sep 14 18:03:07 BST 2011
On 9/14/11 12:55 PM, "Christopher Bongaarts" <cab at umn.edu> wrote:
>Sounds like what we saw when we were getting things up and running in
>test, when something would go wrong we got the "hall of mirrors" effect
>of infinitely nested frames.
That comes up too, though I haven't really finished working on error
handling (and it's rather difficult from their docs to figure out what
they mean with the variuous settings). But this is more of a routine
occurrence, rather than an error result.
>I don't know if anyone ever tested timeouts, so it's likely we have the
>same problem there. If it's SP/app timeouts at issue, we might have
>them turned off.
The app, yes. Unless your sessions just last forever, you would presumably
encounter it. You won't see it unless you try something from within the UI
after the session dies, though, since going to the main page will work.
I do believe that Safari uses some non-trivial definition of "third
party". I run FF that way, and it's pretty much a full block, but I think
if Safari were that way, more people would be aware of the problems with
third party cookies. I think they must have some exceptions.
In fact, I had a dead session handy in Safari, and sure enough, the IdP
worked. So Safari doesn't always block them. That probably explains why
people don't have the violent objection I do to front channel logout UIs.
But, despite it "working", Service-Now actually ends up creating a whole
frameset of itself inside the inner frame, like the hall of mirrors thing
you mentioned. So that's still not acceptable behavior, but that's a
straight app problem, and I'll take it up with them.
More information about the users