[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]
Re: [tor-bugs] #13192 [Tor]: Collect aggregate stats of total hidden service usage vs total exit usage in Tor network
#13192: Collect aggregate stats of total hidden service usage vs total exit usage
in Tor network
-----------------------------+---------------------------------------
Reporter: arma | Owner:
Type: enhancement | Status: needs_review
Priority: normal | Milestone: Tor: 0.2.???
Component: Tor | Version:
Resolution: | Keywords: SponsorR tor-relay tor-hs
Actual Points: | Parent ID:
Points: |
-----------------------------+---------------------------------------
Comment (by dgoulet):
Replying to [comment:51 karsten]:
> I changed the `p == 0.0` part to `p <= 0.0` because GCC thinks we
shouldn't compare floats with `==`. Not sure if we should instead do
something like `< 0.0000000000000002` (smallest possible positive double
value).
Ah right! Forgot about that! Fun with float!... :)
>
> I don't know what to do about `trunc` vs. `tor_trunc`. I figured it's
probably harmless. Want to investigate? And want to add a comment? I
think teor's idea there was to silence clang warnings which we'd get when
converting using an explicit cast and which we don't get when doing the
conversion even more explicitly.
I don't think `trunc` is any useful and I can't get any clang warnings if
it's a direct cast so maybe teor can enlight us on that.
However there is an other problem. Consider this example,
https://people.torproject.org/~dgoulet/volatile/float.c, please run it and
you'll see that even with a double well in between `[INT64_MIN,
INT64_MAX]` when cast to int64_t weird things happen. In this case here,
we always underflow even though the value is supposed to be valid which
will make `sample_laplace_distribution` return wrong values because of the
cast. (llround or/and trunc don't matter here, see example).
Yawning is playing with this as we speak to understand it...
--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/13192#comment:52>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online
_______________________________________________
tor-bugs mailing list
tor-bugs@xxxxxxxxxxxxxxxxxxxx
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-bugs