This way we can use the linux eventfd extension where available.
Using EVFILT_USER on the BSDs will be a teeny bit trickier, and will
require libevent hacking.

Also, re-enable the #if'd out condition-variable code.

Work queues are going to make us hack on all of this stuff a bit more
closely, so it might not be a terrible idea to make it easier to hack.

(Thanks to toralf for noting that the build was broken, and to
Sebastian for telling me that my fix still wasn't right.)

Fixes bug 14170.

Tomasz Torcz authored 10 years ago and Nick Mathewson committed 10 years ago

  There were following problems:
  - configure.ac wrongly checked for defined HAVE_SYSTEMD; this
    wasn't working, so the watchdog code was not compiled in.
    Replace library search with explicit version check
  - sd_notify() watchdog call was unsetting NOTIFY_SOCKET from env;
    this means only first "watchdog ping" was delivered, each
    subsequent one did not have socket to be sent to and systemd
    was killing service
  - after those fixes, enable Watchdog in systemd unit with one
    minute intervals

When I applied patch fcc78e5f, I somehow broke
stack trace symbols on Linux.  I'll leave it to others to figure out
why that happens.  This should be better.  Really.

Fixes bug 14162; bug not in any released version of Tor.

In systemd 209, they deprecated -lsystemd-daemon in favor of
-lsystemd.  So we'd better actually look at the pkg-config output,
or we'll get warnings on newer distributions.

For some as-yet-unknown-to-me reason, setting CFLAGS so early makes
it so -O2 -g doesn't get added to it later.  So, adding it myself
later.  Perhaps a better fix here can be found.

Fixes 14072; bugfix on 0.2.6.2-alpha.  Based on a patch from h.venev

Michael Scherer authored 10 years ago and Nick Mathewson committed 10 years ago

It work by notifying systemd on a regular basis. If
there is no notification, the daemon is restarted.
This requires a version newer than the 209 version
of systemd, as it is not supported before.

Michael Scherer authored 11 years ago and Nick Mathewson committed 10 years ago

This permit for now to signal readiness in a cleaner way
to systemd.

Francisco Blas Izquierdo Riera (klondike) authored 10 years ago and Nick Mathewson committed 10 years ago

The original call to getsockopt to know the original address on transparently
proxyed sockets using REDIRECT in iptables failed with IPv6 addresses because
it assumed all sockets used IPv4.

This patch fixes this by using the appropriate options and adding the headers
containing the needed definitions for these.

This patch is released under the same license as the original file as
long as the author iscredited.

Signed-off-by: Francisco Blas Izquierdo Riera (klondike) <klondike@gentoo.org>

Because in 95 years, we or our successors will surely care about
enforcing the BSD license terms on this code.  Right?

By now, support in the network is widespread and it's time to require
more modern crypto on all Tor instances, whether they're clients or
servers. By doing this early in 0.2.6, we can be sure that at some point
all clients will have reasonable support.

Right now this is only needed for test_util_format_time_interval, so
define it as a static function.  We can move it into compat later if
we need to.

Also, make sure we will compile correctly on systems where they
finally rip it out.

Fixes issue #13325.  Caused by this openbsd commit:

   ​http://marc.info/?l=openbsd-cvs&m=140768179627976&w=2

Reported by Fredzupy.

Uses libscrypt when found; otherwise, we don't have scrypt and we
only support openpgp rfc2440 s2k hashing, or pbkdf2.

Includes documentation and unit tests; coverage around 95%. Remaining
uncovered code is sanity-checks that shouldn't be reachable fwict.

Apparently some compilers want extra switches.

c99 lets us do neat stuff like:

    {
      int j, k;
      foo(&j, &k);
      int z = j + k;
    }

and also
    struct point { int x; int y; };
    struct point pt = { .x=5, .y=5 };

This commit makes the configure scripts check to make sure your
compiler implements them.  It also disables our longstanding warning
about midblock declarations.

Closes ticket 13233.

torrc.minimal is now the one that should change as infrequently as
possible.  To schedule an change to go into it eventually, make your
change to torrc.minimal.in-sample.

torrc.sample is now the volatile one: we can change it to our hearts'
content.

Closes ticket #11144

We added some AS_VAR_IF-based checks to detect whether we have
managed to compile (but not link) with stack-protector.  On autoconf
before 2.63, we don't have AS_VAR_IF, so we just have to let the
user get a compile error rather than a helpful "find libssp" error.

Fixes bug 12693; bugfix on 0.2.5.2-alpha (commit 21ac2928)

Making the text file work out-of-tree didn't work on windows.

Maybe we can get it working that way later.

(This was too hard. Silly autoconf.)

Anthony G. Basile authored 10 years ago and Nick Mathewson committed 10 years ago

When building with bufferevents enabled, configure.ac throws an
error if "$ac_cv_header_event2_bufferevent_ssl_h" is not set to
"yes".  However, nowhere was AC_CHECK_HEADERS(event2/bufferevent_ssl.h)
done.  This commit adds the check.

Long ago we supported systems where there was no support for
threads, or where the threading library was broken. We shouldn't
have do that any more: on every OS that matters, threads exist, and
the OS supports running threads across multiple CPUs.

This resolves tickets 9495 and 12439.  It's a prerequisite to making
our workqueue code work better, since sensible workqueue
implementations don't split across multiple processes.

In the unit tests I want to loop with a delay, but I want less than
a 1 second delay.  This, sadly, requires compatibility code.