r15599@catbus: nickm | 2007-10-09 16:44:36 -0400

 Fix the bug that was making moria1 set valid-after wrong in its votes: we were looking at the preferred timing when we should have been looking at the consensus timing.


svn:r11818

ChangeLog

@@ -40,6 +40,8 @@ Changes in version 0.2.0.8-alpha - 2007-??-??
       gets rid of annoying "400 OK" log messages, which may have been masking
       some deeper issue.  Bugfix on 0.2.0.7-alpha.
     - When we get a valid consensus, recompute the voting schedule.
+    - Base the valid-after time of a vote on the consensus voting schedule,
+      not on our preferred schedule.
 
   o Minor bugfixes (performance):
     - Use a slightly simpler string hashing algorithm (copying Python's

src/or/dirserv.c

@@ -2017,8 +2017,18 @@ generate_networkstatus_vote_obj(crypto_pk_env_t *private_key,
   v3_out->is_vote = 1;
   dirvote_get_preferred_voting_intervals(&timing);
   v3_out->published = now;
-  v3_out->valid_after =
-    dirvote_get_start_of_next_interval(now, timing.vote_interval);
+  {
+    networkstatus_vote_t *current_consensus =
+      networkstatus_get_live_consensus(now);
+    time_t consensus_interval;
+    if (current_consensus)
+      consensus_interval = current_consensus->fresh_until -
+        current_consensus->valid_after;
+    else
+      consensus_interval = timing.vote_interval;
+    v3_out->valid_after =
+      dirvote_get_start_of_next_interval(now, consensus_interval);
+  }
   v3_out->fresh_until = v3_out->valid_after + timing.vote_interval;
   v3_out->valid_until = v3_out->valid_after +
     (timing.vote_interval * timing.n_intervals_valid);