]> git.ipfire.org Git - thirdparty/tor.git/log
thirdparty/tor.git
14 years agofix a bridge edge case similar to 2511
Roger Dingledine [Wed, 1 Jun 2011 00:43:55 +0000 (20:43 -0400)] 
fix a bridge edge case similar to 2511

If you had configured a bridge but then switched to a different bridge
via the controller, you would still be willing to use the old one.

14 years agostop asserting at boot
Roger Dingledine [Tue, 31 May 2011 03:50:37 +0000 (23:50 -0400)] 
stop asserting at boot

The patch for 3228 made us try to run init_keys() before we had loaded
our state file, resulting in an assert inside init_keys. We had moved
it too early in the function.

Now it's later in the function, but still above the accounting calls.

14 years agoMerge branch 'bug3216_v2' into maint-0.2.2
Nick Mathewson [Mon, 30 May 2011 20:12:53 +0000 (16:12 -0400)] 
Merge branch 'bug3216_v2' into maint-0.2.2

14 years agoMerge branch 'bug3289' into maint-0.2.2
Nick Mathewson [Mon, 30 May 2011 19:50:35 +0000 (15:50 -0400)] 
Merge branch 'bug3289' into maint-0.2.2

14 years agoWarn when two hs use the same directory
Sebastian Hahn [Sat, 28 May 2011 22:54:59 +0000 (00:54 +0200)] 
Warn when two hs use the same directory

This simple implementation has a few issues, but it should do for
0.2.2.x. We will want to revisit this later and make it smarter.

14 years agominor cleanups while reviewing 3216
Roger Dingledine [Mon, 30 May 2011 19:31:06 +0000 (15:31 -0400)] 
minor cleanups while reviewing 3216

14 years agowhitespace fixes
Nick Mathewson [Mon, 30 May 2011 19:21:06 +0000 (15:21 -0400)] 
whitespace fixes

14 years agoMerge branch 'bug3045' into maint-0.2.2
Nick Mathewson [Mon, 30 May 2011 19:18:59 +0000 (15:18 -0400)] 
Merge branch 'bug3045' into maint-0.2.2

Conflicts:
src/or/circuitbuild.c

14 years agoImprove comments and defensive programming for 3045
Nick Mathewson [Mon, 30 May 2011 19:15:10 +0000 (15:15 -0400)] 
Improve comments and defensive programming for 3045

The comment fixes are trivial.  The defensive programming trick is to
tolerate receiving NULL inputs on the describe functions. That should
never actually happen, but it seems like the likeliest mistake for us
to make in the future.

14 years agoMerge remote-tracking branch 'public/bug3270' into maint-0.2.2
Nick Mathewson [Mon, 30 May 2011 18:49:49 +0000 (14:49 -0400)] 
Merge remote-tracking branch 'public/bug3270' into maint-0.2.2

14 years agoMerge branch 'bug3228_squashed' into maint-0.2.2
Nick Mathewson [Mon, 30 May 2011 16:42:08 +0000 (12:42 -0400)] 
Merge branch 'bug3228_squashed' into maint-0.2.2

14 years agoReinit keys at the start of options_act().
Nick Mathewson [Mon, 23 May 2011 20:38:35 +0000 (16:38 -0400)] 
Reinit keys at the start of options_act().

Previously we did this nearer to the end (in the old_options &&
transition_affects_workers() block).  But other stuff cares about
keys being consistent with options... particularly anything which
tries to access a key, which can die in assert_identity_keys_ok().

Fixes bug 3228; bugfix on 0.2.2.18-alpha.

14 years agoMerge branch 'bug1297a' into maint-0.2.2
Nick Mathewson [Mon, 30 May 2011 16:25:07 +0000 (12:25 -0400)] 
Merge branch 'bug1297a' into maint-0.2.2

14 years agoUse the normal four-hop CBT for client intro circuits
Robert Ransom [Sun, 29 May 2011 15:21:43 +0000 (08:21 -0700)] 
Use the normal four-hop CBT for client intro circuits

Fixes another part of bug 1297.

14 years agoSet timestamp_dirty on HS circuits as circuit_expire_building requires
Robert Ransom [Sun, 29 May 2011 15:03:41 +0000 (08:03 -0700)] 
Set timestamp_dirty on HS circuits as circuit_expire_building requires

Fixes part of #1297; bugfix on 48e0228f1e031a709c1deb149c7dfd187c3609cf,
when circuit_expire_building was changed to assume that timestamp_dirty
was set when a circuit changed purpose to _C_REND_READY.  (It wasn't.)

14 years agoanswer an XXX nickm asked in aa950e6c4
Roger Dingledine [Sun, 29 May 2011 22:52:37 +0000 (18:52 -0400)] 
answer an XXX nickm asked in aa950e6c4

14 years agoMerge remote-tracking branch 'origin/maint-0.2.1' into maint-0.2.2
Nick Mathewson [Sat, 28 May 2011 06:13:14 +0000 (02:13 -0400)] 
Merge remote-tracking branch 'origin/maint-0.2.1' into maint-0.2.2

14 years agoFix typo in changes/bug2574. Thanks, rransom
Nick Mathewson [Sat, 28 May 2011 06:13:04 +0000 (02:13 -0400)] 
Fix typo in changes/bug2574. Thanks, rransom

14 years agoMerge remote-tracking branch 'origin/maint-0.2.1' into maint-0.2.2
Nick Mathewson [Sat, 28 May 2011 06:08:39 +0000 (02:08 -0400)] 
Merge remote-tracking branch 'origin/maint-0.2.1' into maint-0.2.2

14 years agoMerge branch 'bug2574' into maint-0.2.1
Nick Mathewson [Sat, 28 May 2011 06:07:48 +0000 (02:07 -0400)] 
Merge branch 'bug2574' into maint-0.2.1

14 years agoDon't try to build descriptors when router_get_advertised_or_port()==0
Nick Mathewson [Tue, 24 May 2011 15:55:08 +0000 (11:55 -0400)] 
Don't try to build descriptors when router_get_advertised_or_port()==0

The previous attempt was incomplete: it told us not to publish a
descriptor, but didn't stop us from generating one.  Now we treat an
absent OR port the same as not knowing our address.  (This means
that when we _do_ get an OR port, we need to mark the descriptor
dirty.)

More attempt to fix bug3216.

14 years agoDon't build descriptors if ORPort auto is set and we have no OR listener
Nick Mathewson [Wed, 18 May 2011 01:42:35 +0000 (21:42 -0400)] 
Don't build descriptors if ORPort auto is set and we have no OR listener

This situation can happen easily if you set 'ORPort auto' and
'AccountingMax'.  Doing so means that when you have no ORPort, you
won't be able to set an ORPort in a descriptor, so instead you would
just generate lots of invalid descriptors, freaking out all the time.

Possible fix for 3216; fix on 0.2.2.26-beta.

14 years agoWork correctly if your nameserver is ::1
Nick Mathewson [Mon, 23 May 2011 21:42:38 +0000 (17:42 -0400)] 
Work correctly if your nameserver is ::1

We had all the code in place to handle this right... except that we
were unconditionally opening a PF_INET socket instead of looking at
sa_family.  Ow.

Fixes bug 2574; not a bugfix on any particular version, since this
never worked before.

14 years agoFix GCC 4.6's new -Wunused-but-set-variable warnings.
Nick Mathewson [Mon, 23 May 2011 21:04:38 +0000 (17:04 -0400)] 
Fix GCC 4.6's new -Wunused-but-set-variable warnings.

Most instances were dead code; for those, I removed the assignments.
Some were pieces of info we don't currently plan to use, but which
we might in the future.  For those, I added an explicit cast-to-void
to indicate that we know that the thing's unused.  Finally, one was
a case where we were testing the wrong variable in a unit test.
That one I fixed.

This resolves bug 3208.

14 years agoRemove the -F option from tor-resolve.
Nick Mathewson [Mon, 23 May 2011 20:59:41 +0000 (16:59 -0400)] 
Remove the -F option from tor-resolve.

It used to mean "Force": it would tell tor-resolve to ask tor to
resolve an address even if it ended with .onion.  But when
AutomapHostsOnResolve was added, automatically refusing to resolve
.onion hosts stopped making sense.  So in 0.2.1.16-rc (commit
298dc95dfd8), we made tor-resolve happy to resolve anything.

The -F option stayed in, though, even though it didn't do anything.
Oddly, it never got documented.

Found while fixing GCC 4.6 "set, unused variable" warnings.

14 years agoUnbreak the build on libevent 1.x systems
Robert Ransom [Mon, 23 May 2011 05:54:02 +0000 (22:54 -0700)] 
Unbreak the build on libevent 1.x systems

14 years agoMerge branch 'feature3049-v2' into maint-0.2.2
Nick Mathewson [Mon, 23 May 2011 05:19:04 +0000 (01:19 -0400)] 
Merge branch 'feature3049-v2' into maint-0.2.2

Conflicts:
src/common/Makefile.am

14 years agoThe first argument for a libevent callback should be evutil_socket_t
Nick Mathewson [Mon, 23 May 2011 05:12:00 +0000 (01:12 -0400)] 
The first argument for a libevent callback should be evutil_socket_t

14 years agoAppease make check-spaces wrt procmon.h
Nick Mathewson [Mon, 23 May 2011 05:10:49 +0000 (01:10 -0400)] 
Appease make check-spaces wrt procmon.h

14 years agoUse a 64-bit type to hold sockets on win64.
Nick Mathewson [Mon, 23 May 2011 04:17:48 +0000 (00:17 -0400)] 
Use a 64-bit type to hold sockets on win64.

On win64, sockets are of type UINT_PTR; on win32 they're u_int;
elsewhere they're int.  The correct windows way to check a socket for
being set is to compare it with INVALID_SOCKET; elsewhere you see if
it is negative.

On Libevent 2, all callbacks take sockets as evutil_socket_t; we've
been passing them int.

This patch should fix compilation and correctness when built for
64-bit windows.  Fixes bug 3270.

14 years agonaked constants are ugly
Roger Dingledine [Sat, 21 May 2011 22:55:23 +0000 (18:55 -0400)] 
naked constants are ugly

14 years agofinish a comment nickm started in 8ebceeb3
Roger Dingledine [Sat, 21 May 2011 22:34:55 +0000 (18:34 -0400)] 
finish a comment nickm started in 8ebceeb3

14 years agoremove some (confusing) dead code
Roger Dingledine [Sat, 21 May 2011 20:12:37 +0000 (16:12 -0400)] 
remove some (confusing) dead code

14 years agoMerge branch 'bug1810' into maint-0.2.2
Roger Dingledine [Sat, 21 May 2011 20:09:55 +0000 (16:09 -0400)] 
Merge branch 'bug1810' into maint-0.2.2

14 years agoDon't recreate descriptor on sighup
Sebastian Hahn [Sat, 21 May 2011 17:10:11 +0000 (19:10 +0200)] 
Don't recreate descriptor on sighup

We used to regenerate our descriptor whenever we'd get a sighup. This
was caused by a bug in options_transition_affects_workers() that would
return true even if the options were exactly the same. Down the call
path we'd call init_keys(), which made us make a new descriptor which
the authorities would reject, and the node would subsequently fall out
of the consensus.

This patch fixes only the first part of this bug:
options_transition_affects_workers() behaves correctly now. The second
part still wants a fix.

14 years agowhat's up with this trailing whitespace
Roger Dingledine [Sat, 21 May 2011 03:30:37 +0000 (23:30 -0400)] 
what's up with this trailing whitespace

14 years agoSplit out owning-controller-loss shutdown code into a function
Robert Ransom [Fri, 20 May 2011 15:21:11 +0000 (08:21 -0700)] 
Split out owning-controller-loss shutdown code into a function

14 years agoAdd changes file for #3049
Robert Ransom [Thu, 19 May 2011 23:45:09 +0000 (16:45 -0700)] 
Add changes file for #3049

14 years agoImplement TAKEOWNERSHIP command
Robert Ransom [Thu, 19 May 2011 23:34:40 +0000 (16:34 -0700)] 
Implement TAKEOWNERSHIP command

14 years agoSplit control connection cleanup out of connection_free
Robert Ransom [Thu, 19 May 2011 23:27:51 +0000 (16:27 -0700)] 
Split control connection cleanup out of connection_free

14 years agoFix comment typo
Robert Ransom [Wed, 18 May 2011 11:35:20 +0000 (04:35 -0700)] 
Fix comment typo

14 years agoAdd an XXX
Robert Ransom [Wed, 18 May 2011 11:33:48 +0000 (04:33 -0700)] 
Add an XXX

14 years agoFix trailing asterisk in the output of "GETINFO info/names"
Robert Ransom [Wed, 18 May 2011 11:13:21 +0000 (04:13 -0700)] 
Fix trailing asterisk in the output of "GETINFO info/names"

14 years agoExit immediately if we can't monitor our owning controller process
Robert Ransom [Mon, 16 May 2011 17:25:59 +0000 (10:25 -0700)] 
Exit immediately if we can't monitor our owning controller process

tor_process_monitor_new can't currently return NULL, but if it ever can,
we want that to be an explicitly fatal error, without relying on the fact
that monitor_owning_controller_process's chain of caller will exit if it
fails.

14 years agoFix some comments
Robert Ransom [Sun, 15 May 2011 15:23:29 +0000 (08:23 -0700)] 
Fix some comments

14 years agoImplement __OwningControllerProcess option
Robert Ransom [Sun, 15 May 2011 15:23:04 +0000 (08:23 -0700)] 
Implement __OwningControllerProcess option

Implements part of feature 3049.

14 years agolog the reason for publishing a new relay descriptor
Roger Dingledine [Fri, 20 May 2011 03:36:20 +0000 (23:36 -0400)] 
log the reason for publishing a new relay descriptor

now we have a better chance of hunting down the root cause of bug 1810.

14 years agoFix a compile warning on OSX 10.6
Sebastian Hahn [Fri, 20 May 2011 03:52:36 +0000 (05:52 +0200)] 
Fix a compile warning on OSX 10.6

14 years agodiscard circuits when we change our bridge configuration
Roger Dingledine [Wed, 18 May 2011 00:51:47 +0000 (20:51 -0400)] 
discard circuits when we change our bridge configuration

otherwise we might reuse circuits from the previous configuration,
which could be bad depending on the user's situation

14 years agorefetch bridge descriptors in a timely fashion
Roger Dingledine [Mon, 16 May 2011 16:58:44 +0000 (12:58 -0400)] 
refetch bridge descriptors in a timely fashion

When we configure a new bridge via the controller, don't wait up to ten
seconds before trying to fetch its descriptor. This wasn't so bad when
you listed your bridges in torrc, but it's dreadful if you configure
your bridges via vidalia.

14 years agoMerge branch 'bug2752' into maint-0.2.2
Nick Mathewson [Tue, 17 May 2011 23:51:53 +0000 (19:51 -0400)] 
Merge branch 'bug2752' into maint-0.2.2

14 years agoAdd credit to bug2752; correct the issue number in the changes file
Nick Mathewson [Tue, 17 May 2011 23:50:52 +0000 (19:50 -0400)] 
Add credit to bug2752; correct the issue number in the changes file

14 years agoHandle NULL argument to get_configured_bridge_by_addr_port_digest
Nick Mathewson [Tue, 17 May 2011 23:46:47 +0000 (19:46 -0400)] 
Handle NULL argument to get_configured_bridge_by_addr_port_digest

Fixes bug 2313; bugfix on 0.2.2.26-beta.

14 years agoOops; that function got renamed.
Nick Mathewson [Tue, 17 May 2011 23:45:05 +0000 (19:45 -0400)] 
Oops; that function got renamed.

14 years agoCheck onion keys in microdescriptors, too
Robert Ransom [Thu, 28 Apr 2011 22:16:54 +0000 (15:16 -0700)] 
Check onion keys in microdescriptors, too

14 years agoRemove from maint-0.2.2 changes files merged in release-0.2.2 and master
Nick Mathewson [Mon, 16 May 2011 20:43:04 +0000 (16:43 -0400)] 
Remove from maint-0.2.2 changes files merged in release-0.2.2 and master

14 years agoMerge remote-tracking branch 'origin/maint-0.2.1' into maint-0.2.2
Nick Mathewson [Mon, 16 May 2011 20:40:57 +0000 (16:40 -0400)] 
Merge remote-tracking branch 'origin/maint-0.2.1' into maint-0.2.2

Conflicts:
changes/bug2190
changes/forget-rend-descs-on-newnym

14 years agorm files for changes merged in all of release021, release022, master
Nick Mathewson [Mon, 16 May 2011 20:17:35 +0000 (16:17 -0400)] 
rm files for changes merged in all of release021, release022, master

14 years agoFix bug2752 : 48-char HTTPProxyAuthenticator limitation
Michael Yakubovich [Mon, 16 May 2011 20:09:35 +0000 (16:09 -0400)] 
Fix bug2752 : 48-char HTTPProxyAuthenticator limitation

Bumped the char maximum to 512 for HTTPProxyAuthenticator &
HTTPSProxyAuthenticator. Now stripping all '\n' after base64
encoding in alloc_http_authenticator.

14 years agoMerge remote-tracking branch 'origin/maint-0.2.1' into maint-0.2.2
Nick Mathewson [Mon, 16 May 2011 18:49:55 +0000 (14:49 -0400)] 
Merge remote-tracking branch 'origin/maint-0.2.1' into maint-0.2.2

Fixed trivial conflict due to headers moving into their own .h files
from or.h.

Conflicts:
src/or/or.h

14 years agosquash! Add crypto_pk_check_key_public_exponent function
Nick Mathewson [Mon, 16 May 2011 18:44:23 +0000 (14:44 -0400)] 
squash! Add crypto_pk_check_key_public_exponent function

Rename crypto_pk_check_key_public_exponent to crypto_pk_public_exponent_ok:
it's nice to name predicates s.t. you can tell how to interpret true
and false.

14 years agoRequire that certain public keys have public exponent 65537
Robert Ransom [Thu, 28 Apr 2011 21:35:03 +0000 (14:35 -0700)] 
Require that certain public keys have public exponent 65537

14 years agoAdd crypto_pk_check_key_public_exponent function
Robert Ransom [Thu, 28 Apr 2011 19:13:03 +0000 (12:13 -0700)] 
Add crypto_pk_check_key_public_exponent function

14 years agoCheck fetched rendezvous descriptors' service IDs
Robert Ransom [Wed, 27 Apr 2011 20:37:08 +0000 (13:37 -0700)] 
Check fetched rendezvous descriptors' service IDs

14 years agoMerge remote-tracking branch 'public/bug2850' into maint-0.2.2
Nick Mathewson [Mon, 16 May 2011 15:10:17 +0000 (11:10 -0400)] 
Merge remote-tracking branch 'public/bug2850' into maint-0.2.2

Fixed a trivial conflict where this and the ControlSocketGroupWritable
code both added different functions to the same part of connection.c.

Conflicts:
src/or/connection.c

14 years agoIncrease the length of the buffer in smartlist_string_num_isin().
Nick Mathewson [Mon, 16 May 2011 02:13:53 +0000 (22:13 -0400)] 
Increase the length of the buffer in smartlist_string_num_isin().

This was harmless, since we only used this for checking for lists of
port values, but it's the principle of the thing.

Fixes 3175; bugfix on 0.1.0.1-rc

14 years agochanges file for bug3045
Nick Mathewson [Mon, 16 May 2011 02:02:24 +0000 (22:02 -0400)] 
changes file for bug3045

14 years agoLog descriptions of nodes, not just nicknames.
Nick Mathewson [Mon, 16 May 2011 01:58:46 +0000 (21:58 -0400)] 
Log descriptions of nodes, not just nicknames.

This patch introduces a few new functions in router.c to produce a
more helpful description of a node than its nickame, and then tweaks
nearly all log messages taking a nickname as an argument to call these
functions instead.

There are a few cases where I left the old log messages alone: in
these cases, the nickname was that of an authority (whose nicknames
are useful and unique), or the message already included an identity
and/or an address.  I might have missed a couple more too.

This is a fix for bug 3045.

14 years agoMerge branch 'bug2972' into maint-0.2.2
Nick Mathewson [Mon, 16 May 2011 00:20:36 +0000 (20:20 -0400)] 
Merge branch 'bug2972' into maint-0.2.2

14 years agoMake check_private_dir check for group ownership as appropriate
Nick Mathewson [Sun, 15 May 2011 17:11:48 +0000 (13:11 -0400)] 
Make check_private_dir check for group ownership as appropriate

14 years agoFix up some comment issues spotted by rransom
Nick Mathewson [Sun, 15 May 2011 17:03:15 +0000 (13:03 -0400)] 
Fix up some comment issues spotted by rransom

14 years agoCheck permissions on the directory holding a control socket
Nick Mathewson [Fri, 13 May 2011 19:46:53 +0000 (15:46 -0400)] 
Check permissions on the directory holding a control socket

14 years agoAdd a new flag to check_private_dir to make it _not_ change permissions
Nick Mathewson [Fri, 13 May 2011 19:40:03 +0000 (15:40 -0400)] 
Add a new flag to check_private_dir to make it _not_ change permissions

We'll need this for checking permissions on the directories that hold
control sockets: if somebody says "ControlSocket ~/foo", it would be
pretty rude to do a chmod 700 on their homedir.

14 years agoAdd a function to pull off the final component of a path
Nick Mathewson [Fri, 13 May 2011 19:15:41 +0000 (15:15 -0400)] 
Add a function to pull off the final component of a path

14 years agoMake check_private_dir accept g+rx dirs if told to do so.
Nick Mathewson [Fri, 13 May 2011 18:55:40 +0000 (14:55 -0400)] 
Make check_private_dir accept g+rx dirs if told to do so.

14 years agoClean up the 2972 implementation a little
Sebastian Hahn [Sat, 23 Apr 2011 00:57:53 +0000 (02:57 +0200)] 
Clean up the 2972 implementation a little

14 years agoAdd UnixSocketsGroupWritable config flag
Jérémy Bobbio [Sat, 23 Apr 2011 00:35:02 +0000 (02:35 +0200)] 
Add UnixSocketsGroupWritable config flag

When running a system-wide instance of Tor on Unix-like systems, having
a ControlSocket is a quite handy mechanism to access Tor control
channel.  But it would be easier if access to the Unix domain socket can
be granted by making control users members of the group running the Tor
process.

This change introduces a UnixSocketsGroupWritable option, which will
create Unix domain sockets (and thus ControlSocket) 'g+rw'. This allows
ControlSocket to offer same access control measures than
ControlPort+CookieAuthFileGroupReadable.

See <http://bugs.debian.org/552556> for more details.

14 years agoPreserve bridge download status across SETCONF, HUP
Nick Mathewson [Sun, 15 May 2011 16:46:26 +0000 (12:46 -0400)] 
Preserve bridge download status across SETCONF, HUP

This code changes it so that we don't remove bridges immediately when
we start re-parsing our configuration.  Instead, we mark them all, and
remove all the marked ones after re-parsing our bridge lines.  As we
add a bridge, we see if it's already in the list.  If so, we just
unmark it.

This new behavior will lose the property we used to have that bridges
were in bridge_list in the same order in which they appeared in the
torrc.  I took a quick look through the code, and I'm pretty sure we
didn't actually depend on that anywhere.

This is for bug 3019; it's a fix on 0.2.0.3-alpha.

14 years agoFixup whitespace issues from 3122 commit
Nick Mathewson [Mon, 16 May 2011 00:12:01 +0000 (20:12 -0400)] 
Fixup whitespace issues from 3122 commit

14 years agoMerge remote-tracking branch 'origin/maint-0.2.1' into maint-0.2.2
Nick Mathewson [Mon, 16 May 2011 00:09:10 +0000 (20:09 -0400)] 
Merge remote-tracking branch 'origin/maint-0.2.1' into maint-0.2.2

Conflicts:
src/or/networkstatus.c

14 years agoFixup whitespace issues from 3122 commit
Nick Mathewson [Mon, 16 May 2011 00:06:36 +0000 (20:06 -0400)] 
Fixup whitespace issues from 3122 commit

14 years agoMerge remote-tracking branch 'origin/maint-0.2.1' into maint-0.2.2
Nick Mathewson [Sun, 15 May 2011 23:56:27 +0000 (19:56 -0400)] 
Merge remote-tracking branch 'origin/maint-0.2.1' into maint-0.2.2

14 years agoAnother doc tweak on tor_memcmp: <b>b</b>, not <b>.
Nick Mathewson [Sun, 15 May 2011 23:56:05 +0000 (19:56 -0400)] 
Another doc tweak on tor_memcmp: <b>b</b>, not <b>.

14 years agominor tweaks to 4b19730c8234d
Roger Dingledine [Sun, 15 May 2011 23:20:42 +0000 (19:20 -0400)] 
minor tweaks to 4b19730c8234d

14 years agoMerge branch 'bug3026' into maint-0.2.2
Nick Mathewson [Sun, 15 May 2011 16:18:23 +0000 (12:18 -0400)] 
Merge branch 'bug3026' into maint-0.2.2

14 years agoReplace a nasty add-malloc-snprintf with a nice clean asprintf
Nick Mathewson [Sun, 15 May 2011 15:41:49 +0000 (11:41 -0400)] 
Replace a nasty add-malloc-snprintf with a nice clean asprintf

14 years agoMerge branch 'bug1345' into maint-0.2.2
Nick Mathewson [Sun, 15 May 2011 15:40:14 +0000 (11:40 -0400)] 
Merge branch 'bug1345' into maint-0.2.2

14 years agoBetter doc for consider_recording_trackexithost
Nick Mathewson [Sun, 15 May 2011 15:37:33 +0000 (11:37 -0400)] 
Better doc for consider_recording_trackexithost

14 years agoMerge branch 'bug2732-simpler' into maint-0.2.2
Nick Mathewson [Sun, 15 May 2011 15:17:54 +0000 (11:17 -0400)] 
Merge branch 'bug2732-simpler' into maint-0.2.2

14 years agoRip out more of hid_serv_acting_as_directory
Nick Mathewson [Sun, 15 May 2011 15:14:28 +0000 (11:14 -0400)] 
Rip out more of hid_serv_acting_as_directory

rransom notes correctly that now that we aren't checking our HSDir
flag, we have no actual reason to check whether we are listed in the
consensus at all when determining if we should act like a hidden
service directory.

14 years agoHandle transitions in Automap*, VirtualAddrNetwork correctly
Nick Mathewson [Fri, 13 May 2011 20:59:31 +0000 (16:59 -0400)] 
Handle transitions in Automap*, VirtualAddrNetwork correctly

Previously, if they changed in torrc during a SIGHUP, all was well,
since we would just clear all transient entries from the addrmap
thanks to bug 1345.  But if you changed them from the controller, Tor
would leave old mappings in place.

The VirtualAddrNetwork bug has been here since 0.1.1.19-rc; the
AutomapHosts* bug has been here since 0.2.0.1-alpha.

14 years agoWhen TrackExitHosts changes, remove all no-longer-valid mappings
Nick Mathewson [Fri, 13 May 2011 20:22:58 +0000 (16:22 -0400)] 
When TrackExitHosts changes, remove all no-longer-valid mappings

This bug couldn't happen when TrackExitHosts changed in torrc, since
the SIGHUP to reload the torrc would clear out all the transient
addressmap entries before.  But if you used SETCONF to change
TrackExitHosts, old entries would be left alone: that's a bug, and so
this is a bugfix on Tor 0.1.0.1-rc.

14 years agoRaise the TrackHostExits membership code into its own function
Nick Mathewson [Fri, 13 May 2011 20:22:10 +0000 (16:22 -0400)] 
Raise the TrackHostExits membership code into its own function

14 years agoDon't clear out transient addressmap entries on HUP
Nick Mathewson [Fri, 13 May 2011 20:20:01 +0000 (16:20 -0400)] 
Don't clear out transient addressmap entries on HUP

If you really want to purge the client DNS cache, the TrackHostExits
mappings, and the virtual address mappings, you should be using NEWNYM
instead.

Fixes bug 1345; bugfix on Tor 0.1.0.1-rc.

Note that this needs more work: now that we aren't nuking the
transient addressmap entries on HUP, we need to make sure that
configuration changes to VirtualAddressMap and TrackHostExits actually
have a reasonable effect.

14 years agoNew smartlist function to see if two lists of strings are equal.
Nick Mathewson [Fri, 13 May 2011 20:18:53 +0000 (16:18 -0400)] 
New smartlist function to see if two lists of strings are equal.

We'll use this to detect changes in CSV options.

14 years agoMerge branch 'feature3076_squashed' into maint-0.2.2
Nick Mathewson [Fri, 13 May 2011 14:43:41 +0000 (10:43 -0400)] 
Merge branch 'feature3076_squashed' into maint-0.2.2

14 years agoAdd a ControlPortFileGroupWritable option
Nick Mathewson [Thu, 12 May 2011 23:17:48 +0000 (19:17 -0400)] 
Add a ControlPortFileGroupWritable option

14 years agoWrite automatically-chosen control ports to a file.
Nick Mathewson [Mon, 9 May 2011 16:13:37 +0000 (12:13 -0400)] 
Write automatically-chosen control ports to a file.

14 years agonew GETINFO command to return list of listeners of a given type
Nick Mathewson [Mon, 2 May 2011 20:44:37 +0000 (16:44 -0400)] 
new GETINFO command to return list of listeners of a given type

14 years agoAdd a new "tor_sockaddr_to_str()" function
Nick Mathewson [Mon, 2 May 2011 20:43:11 +0000 (16:43 -0400)] 
Add a new "tor_sockaddr_to_str()" function

It does what it says on the tin.  It turns out I'll want this in a couple
of places.