We already do this for libevent; let's do it for openssl too.

For now, I'm making it always a warn, since this has caused some
problems in the past.  Later, we can see about making it less severe.

OTOH, log the Libevent and OpenSSL versions on the first line when
we're starting Tor.

We were trying to incorporate all headers in common_sha1.i, not just
the src/common ones.

This is part of bug 6778; fix on 0.2.4.1-alpha

Harmless unless we somehow generate a microdesc of more than INT_MAX
bytes.

This is just a precaution; we seem to be using var_cell_t.payload and
var_cell_t.payload_len correctly at the moment.

we can turn it into an autobool later if we have some way for it
to make a decision.

(patch possibly got lost when nickm merged #6770; or maybe nickm meant
for it to be this way. i'm not sure.)

Fix for bug 6774; bugfix on 0.2.3.17-beta.

Use router_get_advertised_{dir,or}_port() functions instead of
get_primary_{dir,or}_port().

* Add changes/ files.
* Edit the tor-fw-helper manpage.
* Fix check-spaces.
* Add prototype for get_list_of_ports_to_forward().
* Fix tor_parse_long() TCP port range.
* Improve doc. of tor_check_port_forwarding().
* Check for overflows in tor_check_port_forwarding().
* Demote successful port forwarding to LOG_INFO.

Conflicts:
	src/common/address.c
	src/or/circuitbuild.c

Conflicts:
	src/or/circuitbuild.h

Conflicts:
	src/or/transports.c

Add handle_fw_helper_output(), a function responsible for parsing the
output of tor-fw-helper. Refactor tor_check_port_forwarding() and
run_scheduled_events() accordingly too.

We now issue warnings when we get control output from tor-fw-helper,
and we log the verbose output of tor-fw-helper in LOG_INFO.

Conflicts:
	src/common/util.c

get_lines_from_handle() is a multiplatform function which drains lines
from a stream and stuffs it into a smartlist. It's useful for
line-based protocols, like the one managed proxy and the tor-fw-helper
protocols.

stdout should be clean for control messages.

Make it port-purpose-agnostic; that is it should treat each port as
equal, and not expect to get a DirPort and an ORPort.

See #4771 for rationale.

Note that this patch does not take suggested changes in #4470 into
account and keeps treating AuthDirHasIPv6Connectivity as an
AUTOBOOL. Thus, bug fixes for that are included here as well.

This is a fix on master, unreleased as of now.

This is a fix of unreleased tor. It solves ticket #6770.

In case the struct grows in the future. Shouldn't be too expensive.

Linus Nordberg authored 12 years ago and Nick Mathewson committed 12 years ago

Can it ever help? I can only see harm. What am I missing?

Linus Nordberg authored 12 years ago and Nick Mathewson committed 12 years ago

Also, make node_get_prim_orport() indicate in its return value whether
a valid OR port was copied or not.

Maybe we should make it legal to pass ap_out==NULL?

Linus Nordberg authored 12 years ago and Nick Mathewson committed 12 years ago

Also, do this only for clients, explicitly.

Also, give the flag a value every time we set consensus. We used to
touch it only when ClientPreferIPv6ORPort was set, which was wrong.

Linus Nordberg authored 12 years ago and Nick Mathewson committed 12 years ago

extend_info_from_node() used to use the primary OR port (i.e. IPv4)
unless the node had routerinfo. Now that we have IPv6 addresses in
microdescs we may want to use them.

Note that this patch changes using r->cache_info.identity_digest into
using node->identity. I count on these being well synchronised, or
things would break in other ways. Right?