https://github.com/ring-clojure/ring/blob/1.9.0/ring-core/src/ring/middleware/session.clj#L61 hm, the logic is there. But hard to say what could be changing the :session-cookie-attrs, or put the :recreate key in the response

Could something be setting the session age, for example, all the time?

I don't think core ring has anything that sets those, so it should be in other parts

You could start by printing the relevant parts of the response right before the session middleware checks it, so you know which path it goes down to

Then move your debug middleware up layer by layer, so you catch where it happens

(or if the stack is deep, binary search 😉

thanks Hukka!

Are you using reitit?

yep

we also have a lib-noir server (from years ago) that we're still using and its cookie value is stable from page to page, so i lost half a day trying to figure out why, until i recognised that the session was still working as i wanted it to

you may be hitting this bug? https://github.com/metosin/reitit/issues/205

What I had to do was to have a global store, and not put the wrap-session in the route middleware.

oh interesting, going to give that a close read

also i should read Issue logs 😅

thanks for this, i don't have this specific issue; i implemented my own interceptor instead of using their middleware

np, glad you sorted it out 🙂