Bug 40876: Handle temporary usage restrictions

The primary-guard-to-use list should not grow when we exclude a relay temporarily for a specific circuit. This caused tor#40876 (closed).

(I am not sure if this is the right way to do a markdown MR, or if I botched any syntax).

requested review from @nickm

assigned to @mikeperry

mentioned in merge request tor!778 (merged)

mentioned in issue tor#40876 (closed)

This is probably right? but I don't understand what the difference is between temporary and permanent restrictions. That is, how can you tell what is "temporary" and what is "permanent"? (I get that you have to treat them differently, but how do you decide which restrictions need to be treated in which way?) Maybe there are better words we can use.

In other words, I'm hoping we can figure out how to write this down in a way that will still make sense in 5 years.

added 1 commit

• 4e1a5e94 - Bug 40876: Handle guard usage restrictions

Compare with previous version

added 1 commit

• 0eb22ded - Bug 40876: Handle guard usage restrictions

Compare with previous version

Ok @nickm, I tried to revamp the discussed bullet point to describe restrictions in the latest force-push, as we discussed. I did not describe the restrictions in detail, other than to mention the notion of "per-circuit". My thinking is you can add the description of restrictions later, above this, when we get to the bottom of this microdesc restriction (and if it even still matters).

This bullet point should match the diff in tor!778 (a073a1bb), which implements what I (tried to) describe.

resolved all threads

mentioned in commit bf1bf17e

merged

Thanks! I'll do the followups we talked about and pass them to you for review.

mentioned in merge request !184 (merged)

mentioned in issue arti#1091