Opened 13 years ago

Closed 13 years ago

Last modified 13 years ago

#1123 closed Enhancement (fixed)

unreachable peers should be flagged better

Reported by: charles Owned by: charles
Priority: Normal Milestone: 1.30
Component: libtransmission Version: 1.22
Severity: Normal Keywords:
Cc:

Description

Currently we mark a peer as unreachable if our outbound peer-io constructor fails. That's a bad yardstick: since the constructor is nonblocking, it's unlikely to fail due to a socket connection error. It's more likely to fail if we've maxed out our peer connections, and that unreachable doesn't follow from that condition.

It would be better to do this in the "handshake done" callback when we test for errors. For example, if during the handshake we got a connection timeout or the peer never sent us even one byte of data, we could flag it as unreachable.

Attachments (1)

unreachable.patch (3.6 KB) - added by charles 13 years ago.
patch against r6413.

Download all attachments as: .zip

Change History (4)

Changed 13 years ago by charles

patch against r6413.

comment:1 Changed 13 years ago by charles

  • Resolution set to fixed
  • Status changed from new to closed

committed in [6415]

comment:2 Changed 13 years ago by charles

  • Milestone changed from None Set to 1.30

comment:3 Changed 13 years ago by charles

  • Summary changed from unreachable peers need to be flagged better to unreachable peers should be flagged better
Note: See TracTickets for help on using tickets.