Permalink
bug3199
Commits on Sep 11, 2013
-
-
Changed how periodic events are handled.
No longer returning zero to keep interval. Functions are expected to always return their intended interval (or a time they want to be polled at). Now if a precondition fails then it will be polled every second, e.g. servermode is enabled then all server events should be run in the next second.
-
-
-
-
Commits on Sep 10, 2013
Commits on Sep 9, 2013
Commits on Sep 8, 2013
Commits on Sep 5, 2013
-
-
Roger Dingledine committed
Sep 5, 2013 -
collect and log statistics about onionskins received/processed
we skip onionskins that came from non-relays, so we're less likely to run into privacy troubles. starts to implement ticket 9658.
Roger Dingledine committedSep 5, 2013 -
Revert e443bef and solve it a different way
Now we explicitly check for overflow. This approach seemed smarter than a cascade of "change int to unsigned int and hope nothing breaks right before the release". Nick, feel free to fix in a better way, maybe in master.
Roger Dingledine committedSep 5, 2013 -
don't let recently_chosen_ntors overflow
with commit c6f1668 we let it grow arbitrarily large. it can still overflow, but the damage is very small now.
Roger Dingledine committedSep 5, 2013 -
nickm wants us to prioritize tap in a currently-rare edge case
Roger Dingledine committedSep 5, 2013 -
add a changes entry for ticket 9574
Roger Dingledine committedSep 5, 2013 -
Be more general in calculating expected onion queue processing time
Now we consider the TAP cells we'll process while draining the NTor queue, and vice versa.
Roger Dingledine committedSep 5, 2013 -
let the NumNTorsPerTAP consensus param override our queue choice
Roger Dingledine committedSep 5, 2013 -
do a lopsided round-robin between the onion queues
that way tap won't starve entirely, but we'll still handle ntor requests quicker.
Roger Dingledine committedSep 5, 2013