[tor-bugs] #4872 [Tor]: When the valid consensus is old but still barely valid, are the descriptors referenced in it still valid?

[Tor Bug Tracker & Wiki]

#4872: When the valid consensus is old but still barely valid, are the descriptors
referenced in it still valid?
-----------------------+----------------------------------------------------
 Reporter:  Sebastian  |          Owner:                    
     Type:  defect     |         Status:  new               
 Priority:  critical   |      Milestone:  Tor: 0.2.5.x-final
Component:  Tor        |        Version:                    
 Keywords:  tor-auth   |         Parent:                    
   Points:             |   Actualpoints:                    
-----------------------+----------------------------------------------------

Comment(by nickm):

 So, microdesc validity is "for a week after it was last listed."
 Microdescs can't be at fault here.

 routerinfo_t validity is complicated; there's a lot of code that checks
 its published_on field. ROUTER_MAX_AGE is 48 hours, so might not be at
 fault, though we might have a rule to never prune a routerinfo_t in our
 current consensus if we don't have one.  There are about 10 places that
 look at the 24-hour ROUTER_MAX_AGE_TO_PUBLISH, though, and I wonder
 whether all of them are correct.

-- 
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/4872#comment:4>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online