Arvid Norberg

this is an assertion failure. It suggests that you're running a debug build (which is fine, and you probably wouldn't have discovered this bug otherwise). But the output printed to...

@master255 It's not helpful to argue that this is not an issue. It clearly is. If you want to discuss how package managers and shared libraries are wrong and should...

@glassez I agree that the ABI should transition towards being more flexible from and ABI point of view. Some examples of changes I've made in the past to support updates...

@SemiAccurate I don't think this feature trumps the pread disk I/O in priority. There is an alert for when pieces finish already, you can issue an explicit `read_piece()` when you...

my suspicion would be that there's something corrupting the heap somewhere, and this is the symptom. I would suggest running with address-sanitizer enabled

It's not so simple to add this information in an efficient way. Any peer can have been received from any tracker, and there may be very many peers and very...

> If proxy_peer_connections and proxy_tracker_connections could be enabled independently, then it would be possible to connect to udp trackers with a direct connection and use tcp peer connections with the...

the case I was getting at is that when you contact a tracker, it will look at the source IP (and port in the case of DHT) and assume that's...

the first step is to define, precisely, how it's expected to work.

I agree with your expected behavior, it would make sense. I don't think it's entirely trivial to get there though. Off the top of my head, I suspect the reason...