]> git.ipfire.org Git - thirdparty/tor.git/log
thirdparty/tor.git
15 years agoInitial conversion to use node_t throughout our codebase.
Nick Mathewson [Wed, 29 Sep 2010 19:00:41 +0000 (15:00 -0400)] 
Initial conversion to use node_t throughout our codebase.

A node_t is an abstraction over routerstatus_t, routerinfo_t, and
microdesc_t.  It should try to present a consistent interface to all
of them.  There should be a node_t for a server whenever there is
  * A routerinfo_t for it in the routerlist
  * A routerstatus_t in the current_consensus.
(note that a microdesc_t alone isn't enough to make a node_t exist,
since microdescriptors aren't usable on their own.)

There are three ways to get a node_t right now: looking it up by ID,
looking it up by nickname, and iterating over the whole list of
microdescriptors.

All (or nearly all) functions that are supposed to return "a router"
-- especially those used in building connections and circuits --
should return a node_t, not a routerinfo_t or a routerstatus_t.

A node_t should hold all the *mutable* flags about a node.  This
patch moves the is_foo flags from routerinfo_t into node_t.  The
flags in routerstatus_t remain, but they get set from the consensus
and should not change.

Some other highlights of this patch are:

  * Looking up routerinfo and routerstatus by nickname is now
    unified and based on the "look up a node by nickname" function.
    This tries to look only at the values from current consensus,
    and not get confused by the routerinfo_t->is_named flag, which
    could get set for other weird reasons.  This changes the
    behavior of how authorities (when acting as clients) deal with
    nodes that have been listed by nickname.

  * I tried not to artificially increase the size of the diff here
    by moving functions around.  As a result, some functions that
    now operate on nodes are now in the wrong file -- they should
    get moved to nodelist.c once this refactoring settles down.
    This moving should happen as part of a patch that moves
    functions AND NOTHING ELSE.

  * Some old code is now left around inside #if 0/1 blocks, and
    should get removed once I've verified that I don't want it
    sitting around to see how we used to do things.

There are still some unimplemented functions: these are flagged
with "UNIMPLEMENTED_NODELIST()."  I'll work on filling in the
implementation here, piece by piece.

I wish this patch could have been smaller, but there did not seem to
be any piece of it that was independent from the rest.  Moving flags
forces many functions that once returned routerinfo_t * to return
node_t *, which forces their friends to change, and so on.

15 years agoTry to make most routerstatus_t interfaces const
Nick Mathewson [Wed, 29 Sep 2010 05:35:08 +0000 (01:35 -0400)] 
Try to make most routerstatus_t interfaces const

15 years agoTry to make most routerinfo_t interfaces const
Nick Mathewson [Wed, 29 Sep 2010 04:38:32 +0000 (00:38 -0400)] 
Try to make most routerinfo_t interfaces const

15 years agoImplement a basic node and nodelist type
Nick Mathewson [Tue, 28 Sep 2010 17:29:31 +0000 (13:29 -0400)] 
Implement a basic node and nodelist type

The node_t type is meant to serve two key functions:

  1) Abstracting difference between routerinfo_t and microdesc_t
     so that clients can use microdesc_t instead of routerinfo_t.

  2) Being a central place to hold mutable state about nodes
     formerly held in routerstatus_t and routerinfo_t.

This patch implements a nodelist type that holds a node for every
router that we would consider using.

15 years agoInitialize fd values in tor_port_check_forwarding to -1
Nick Mathewson [Fri, 1 Oct 2010 22:14:17 +0000 (18:14 -0400)] 
Initialize fd values in tor_port_check_forwarding to -1

15 years agoMerge branch 'maint-0.2.2'
Roger Dingledine [Fri, 1 Oct 2010 18:12:11 +0000 (14:12 -0400)] 
Merge branch 'maint-0.2.2'

15 years agofix comment
Roger Dingledine [Fri, 1 Oct 2010 18:11:08 +0000 (14:11 -0400)] 
fix comment

15 years agolog when we guess our ip address, not just when we fail
Roger Dingledine [Fri, 1 Oct 2010 17:32:38 +0000 (13:32 -0400)] 
log when we guess our ip address, not just when we fail

15 years agoMerge branch 'maint-0.2.2'
Roger Dingledine [Fri, 1 Oct 2010 09:00:02 +0000 (05:00 -0400)] 
Merge branch 'maint-0.2.2'

Conflicts:

configure.in
contrib/tor-mingw.nsi.in
src/win32/orconfig.h

15 years agobump to 0.2.2.17-alpha-dev
Roger Dingledine [Fri, 1 Oct 2010 08:59:11 +0000 (04:59 -0400)] 
bump to 0.2.2.17-alpha-dev

15 years agoretroactively declare the cbt fixes to be major
Roger Dingledine [Fri, 1 Oct 2010 08:53:23 +0000 (04:53 -0400)] 
retroactively declare the cbt fixes to be major

also put in release blurbs

15 years agoThe build stuff for tor-fw-helper means we need automake 1.6 or later
Nick Mathewson [Fri, 1 Oct 2010 04:01:08 +0000 (00:01 -0400)] 
The build stuff for tor-fw-helper means we need automake 1.6 or later

15 years agoMerge remote branch 'sebastian/fw-helper-manpage'
Nick Mathewson [Fri, 1 Oct 2010 02:24:18 +0000 (22:24 -0400)] 
Merge remote branch 'sebastian/fw-helper-manpage'

15 years agorenumber, clean whitespace
Roger Dingledine [Thu, 30 Sep 2010 23:24:04 +0000 (19:24 -0400)] 
renumber, clean whitespace

15 years agogive steven's everybody-a-bridge proposal a number
Roger Dingledine [Thu, 30 Sep 2010 23:22:52 +0000 (19:22 -0400)] 
give steven's everybody-a-bridge proposal a number

15 years agoAdd comments from nickm and arma, from IRC
Steven Murdoch [Tue, 30 Mar 2010 15:57:49 +0000 (16:57 +0100)] 
Add comments from nickm and arma, from IRC

15 years agoAdd algorithm and rationale for performance measurement
Steven Murdoch [Mon, 29 Mar 2010 19:49:30 +0000 (20:49 +0100)] 
Add algorithm and rationale for performance measurement

15 years agoIn the migration plan, mention how to prevent overloading the bridge authority
Steven Murdoch [Wed, 17 Mar 2010 11:56:59 +0000 (11:56 +0000)] 
In the migration plan, mention how to prevent overloading the bridge authority

15 years agoIntegrate more feedback from IRC
Steven Murdoch [Mon, 15 Mar 2010 17:50:50 +0000 (17:50 +0000)] 
Integrate more feedback from IRC

- For now we are only talking about moving clients to be bridges

- Some questions on how we should inform users

15 years agoAdd some open questions, and mention Roger's idea about asking for consent via email
Steven Murdoch [Mon, 15 Mar 2010 01:07:12 +0000 (01:07 +0000)] 
Add some open questions, and mention Roger's idea about asking for consent via email

15 years agoNote that we only can't handle automatic demotion. Users can always change their...
Steven Murdoch [Mon, 15 Mar 2010 00:23:12 +0000 (00:23 +0000)] 
Note that we only can't handle automatic demotion. Users can always change their state manually.

15 years agoChange "server" to "relay", so as to match existing terminology
Steven Murdoch [Mon, 15 Mar 2010 00:14:41 +0000 (00:14 +0000)] 
Change "server" to "relay", so as to match existing terminology

15 years agoStart idea xxx-automatic-node-promotion
Steven Murdoch [Sun, 14 Mar 2010 23:06:48 +0000 (23:06 +0000)] 
Start idea xxx-automatic-node-promotion

- Initial draft of overview and motivation
- Start of design

15 years agoMerge branch 'maint-0.2.2'
Roger Dingledine [Fri, 1 Oct 2010 02:02:54 +0000 (22:02 -0400)] 
Merge branch 'maint-0.2.2'

15 years agoimprove rpm init script by borrowing from Debian tor-0.2.2.17-alpha
Erinn Clark [Fri, 1 Oct 2010 00:38:28 +0000 (17:38 -0700)] 
improve rpm init script by borrowing from Debian

15 years agoUpdate rpm spec file so that it will build without manual intervention on all rpm...
Erinn Clark [Sat, 21 Aug 2010 07:12:26 +0000 (00:12 -0700)] 
Update rpm spec file so that it will build without manual intervention on all rpm-based distributions

15 years agoAUTHORS doesn't exist, so stop trying to cp it. Add 'sudo' to a mv that needs higher...
Erinn Clark [Sat, 21 Aug 2010 07:07:04 +0000 (00:07 -0700)] 
AUTHORS doesn't exist, so stop trying to cp it. Add 'sudo' to a mv that needs higher permissions.

15 years agoMerge branch 'maint-0.2.2'
Roger Dingledine [Thu, 30 Sep 2010 21:55:01 +0000 (17:55 -0400)] 
Merge branch 'maint-0.2.2'

Conflicts:

configure.in
contrib/tor-mingw.nsi.in
src/win32/orconfig.h

15 years agobump to 0.2.2.17-alpha
Roger Dingledine [Thu, 30 Sep 2010 21:49:11 +0000 (17:49 -0400)] 
bump to 0.2.2.17-alpha

15 years agowrite up a changelog file
Roger Dingledine [Thu, 30 Sep 2010 21:32:17 +0000 (17:32 -0400)] 
write up a changelog file

15 years agoAdd more tor-fw-helper files to gitignore
Sebastian Hahn [Thu, 30 Sep 2010 21:10:37 +0000 (23:10 +0200)] 
Add more tor-fw-helper files to gitignore

15 years agoBuild the tor-fw-helper manpage
Sebastian Hahn [Thu, 30 Sep 2010 20:44:11 +0000 (22:44 +0200)] 
Build the tor-fw-helper manpage

So far we just had the asciidoc manpage, but didn't build it.

15 years agoAdd 'compile' to .gitignore
Nick Mathewson [Thu, 30 Sep 2010 20:41:12 +0000 (16:41 -0400)] 
Add 'compile' to .gitignore

15 years agoMerge branch 'tor-fw-squashed2'
Nick Mathewson [Thu, 30 Sep 2010 20:22:39 +0000 (16:22 -0400)] 
Merge branch 'tor-fw-squashed2'

Conflicts:
src/common/util.c

15 years agoFix space and formatting issues
Nick Mathewson [Thu, 30 Sep 2010 19:21:37 +0000 (15:21 -0400)] 
Fix space and formatting issues

15 years agoMerge remote branch 'origin/maint-0.2.2'
Nick Mathewson [Thu, 30 Sep 2010 19:30:14 +0000 (15:30 -0400)] 
Merge remote branch 'origin/maint-0.2.2'

15 years ago#if-out the fw-helper code in util.c when building on windows
Nick Mathewson [Thu, 30 Sep 2010 16:58:48 +0000 (12:58 -0400)] 
#if-out the fw-helper code in util.c when building on windows

15 years agoTurn select() usage in tor-fw-helper into a function.
Nick Mathewson [Thu, 30 Sep 2010 15:49:41 +0000 (11:49 -0400)] 
Turn select() usage in tor-fw-helper into a function.

15 years agoStart tor-fw-helper in the background, and log whatever it outputs
Steven Murdoch [Wed, 16 Jun 2010 18:47:06 +0000 (19:47 +0100)] 
Start tor-fw-helper in the background, and log whatever it outputs

15 years agoChanges to tor-fw-helper, some based on Nick's review
Jacob Appelbaum [Fri, 24 Sep 2010 19:52:07 +0000 (12:52 -0700)] 
Changes to tor-fw-helper, some based on Nick's review

 * MINIUPNPC rather than the generic UPNP
 * Nick suggested a better abstraction model for tor-fw-helper
 * Fix autoconf to build with either natpmp or miniupnpc
 * Add AM_PROG_CC_C_O to fix automake complaint
 * update spec to address nickm's concern
 * refactor nat-pmp to match upnp state
 * we prefer tor_snprintf to snprintf
 * link properlty for tor_snprintf
 * rename test_commandline_options to log_commandline_options
 * cast this uint as an int
 * detect possible FD_SETSIZE errors
 * make note about future enhancements for natpmp
 * add upnp enhancement note
 * ChangeLog entry
 * doxygen and check-spaces cleanup
 * create tor-fw-helper.1.txt

15 years agoFirst implementation of tor-fw-helper.
Jacob Appelbaum [Sat, 17 Apr 2010 00:45:12 +0000 (17:45 -0700)] 
First implementation of tor-fw-helper.

tor-fw-helper is a command-line tool to wrap and abstract various
firewall port-forwarding tools.

This commit matches the state of Jacob's tor-fw-helper branch as of
23 September 2010.

  (commit msg by Nick)

15 years agoWhitespace fixes on recent merges to master.
Nick Mathewson [Thu, 30 Sep 2010 05:36:36 +0000 (01:36 -0400)] 
Whitespace fixes on recent merges to master.

15 years agoAdd a changes file for the bug1912 fix
Sebastian Hahn [Thu, 30 Sep 2010 04:16:41 +0000 (06:16 +0200)] 
Add a changes file for the bug1912 fix

15 years agoNote an XXX about potential overflow
Sebastian Hahn [Thu, 30 Sep 2010 04:11:51 +0000 (06:11 +0200)] 
Note an XXX about potential overflow

15 years agoUse an upper and lower bound for bridge weights
Sebastian Hahn [Thu, 30 Sep 2010 04:10:56 +0000 (06:10 +0200)] 
Use an upper and lower bound for bridge weights

When picking bridges (or other nodes without a consensus entry (and
thus no bandwidth weights)) we shouldn't just trust the node's
descriptor. So far we believed anything between 0 and 10MB/s, where 0
would mean that a node doesn't get any use from use unless it is our
only one, and 10MB/s would be a quite siginficant weight. To make this
situation better, we now believe weights in the range from 20kB/s to
100kB/s. This should allow new bridges to get use more quickly, and
means that it will be harder for bridges to see almost all our traffic.

15 years agoFix check-spaces
Sebastian Hahn [Thu, 30 Sep 2010 04:10:25 +0000 (06:10 +0200)] 
Fix check-spaces

15 years agoMerge branch 'maint-0.2.2'
Roger Dingledine [Thu, 30 Sep 2010 04:10:14 +0000 (00:10 -0400)] 
Merge branch 'maint-0.2.2'

15 years agoMerge commit 'mikeperry/bug1772' into maint-0.2.2
Roger Dingledine [Thu, 30 Sep 2010 04:00:06 +0000 (00:00 -0400)] 
Merge commit 'mikeperry/bug1772' into maint-0.2.2

15 years agoNominaly lower the minimum timeout value to 1500.
Mike Perry [Thu, 30 Sep 2010 03:53:50 +0000 (20:53 -0700)] 
Nominaly lower the minimum timeout value to 1500.

This won't change any behavior, since it will still be rounded back
up to 2seconds, but should reduce the chances of some extra warns.

15 years agoMerge branch 'bug1772' into maint-0.2.2
Roger Dingledine [Thu, 30 Sep 2010 03:52:18 +0000 (23:52 -0400)] 
Merge branch 'bug1772' into maint-0.2.2

15 years agofix two casts
Roger Dingledine [Thu, 30 Sep 2010 03:51:25 +0000 (23:51 -0400)] 
fix two casts

15 years agoComment network liveness and change detection behavior.
Mike Perry [Thu, 30 Sep 2010 02:35:40 +0000 (19:35 -0700)] 
Comment network liveness and change detection behavior.

15 years agono measurement circs if not enough build times
Roger Dingledine [Wed, 29 Sep 2010 22:05:10 +0000 (18:05 -0400)] 
no measurement circs if not enough build times

In the first 100 circuits, our timeout_ms and close_ms
are the same. So we shouldn't transition circuits to purpose
CIRCUIT_PURPOSE_C_MEASURE_TIMEOUT, since they will just timeout again
next time we check.

15 years agorefactor and recomment; no actual changes
Roger Dingledine [Wed, 29 Sep 2010 22:01:22 +0000 (18:01 -0400)] 
refactor and recomment; no actual changes

15 years agoMerge branch 'maint-0.2.2'
Roger Dingledine [Wed, 29 Sep 2010 21:21:43 +0000 (17:21 -0400)] 
Merge branch 'maint-0.2.2'

15 years agoMerge commit 'mikeperry/bug1739' into maint-0.2.2
Roger Dingledine [Wed, 29 Sep 2010 21:17:59 +0000 (17:17 -0400)] 
Merge commit 'mikeperry/bug1739' into maint-0.2.2

15 years agoMerge commit 'mikeperry/bug1740' into maint-0.2.2
Roger Dingledine [Wed, 29 Sep 2010 21:05:38 +0000 (17:05 -0400)] 
Merge commit 'mikeperry/bug1740' into maint-0.2.2

15 years agoAdd changes file.
Mike Perry [Wed, 29 Sep 2010 16:15:08 +0000 (09:15 -0700)] 
Add changes file.

15 years agoCap the circuit build timeout to the max time we've seen.
Mike Perry [Wed, 29 Sep 2010 17:06:31 +0000 (10:06 -0700)] 
Cap the circuit build timeout to the max time we've seen.

Also, cap the measurement timeout to 2X the max we've seen.

15 years agoDo away with the complexity of the network liveness detection.
Mike Perry [Wed, 29 Sep 2010 15:55:11 +0000 (08:55 -0700)] 
Do away with the complexity of the network liveness detection.

We really should ignore any timeouts that have *no* network activity for their
entire measured lifetime, now that we have the 95th percentile measurement
changes. Usually this is up to a minute, even on fast connections.

15 years agoFix state checks on liveness handling.
Mike Perry [Wed, 29 Sep 2010 15:47:35 +0000 (08:47 -0700)] 
Fix state checks on liveness handling.

If we really want all this complexity for these stages here, we need to handle
it better for people with large timeouts. It should probably go away, though.

15 years agoFix non-live condition checks.
Mike Perry [Wed, 29 Sep 2010 18:49:31 +0000 (11:49 -0700)] 
Fix non-live condition checks.

Rechecking the timeout condition was foolish, because it is checked on the
same codepath. It was also wrong, because we didn't round.

Also, the liveness check itself should be <, and not <=, because we only have
1 second resolution.

15 years agoAdd changes file.
Mike Perry [Wed, 29 Sep 2010 09:37:53 +0000 (02:37 -0700)] 
Add changes file.

15 years agoSend control port events for timeouts.
Mike Perry [Sat, 24 Jul 2010 16:19:20 +0000 (18:19 +0200)] 
Send control port events for timeouts.

We now differentiate between timeouts and cutoffs by the REASON string and
the PURPOSE string.

15 years agoAdd changes file.
Mike Perry [Tue, 27 Jul 2010 16:30:52 +0000 (09:30 -0700)] 
Add changes file.

15 years agoOnly count timeout data for 3 hop circuits.
Mike Perry [Wed, 29 Sep 2010 18:41:27 +0000 (11:41 -0700)] 
Only count timeout data for 3 hop circuits.

Use 4/3 of this timeout value for 4 hop circuits, and use half of it for
canabalized circuits.

15 years agomake c612ddee compile with old libevents
Roger Dingledine [Wed, 29 Sep 2010 06:50:46 +0000 (02:50 -0400)] 
make c612ddee compile with old libevents

15 years agoMerge remote branch 'origin/maint-0.2.2'
Nick Mathewson [Wed, 29 Sep 2010 04:38:44 +0000 (00:38 -0400)] 
Merge remote branch 'origin/maint-0.2.2'

15 years agoeven more comment
Roger Dingledine [Wed, 29 Sep 2010 03:50:56 +0000 (23:50 -0400)] 
even more comment

15 years agoRemove the has_old_dnsworkers flag.
Nick Mathewson [Wed, 29 Sep 2010 03:37:45 +0000 (23:37 -0400)] 
Remove the has_old_dnsworkers flag.

15 years agoimprove code comments, based on comments from nick
Roger Dingledine [Wed, 29 Sep 2010 03:27:00 +0000 (23:27 -0400)] 
improve code comments, based on comments from nick

15 years agoa changelog entry for the entrynodes retry mess
Roger Dingledine [Wed, 29 Sep 2010 02:49:36 +0000 (22:49 -0400)] 
a changelog entry for the entrynodes retry mess

15 years agohandle ugly edge case in retrying entrynodes
Roger Dingledine [Wed, 29 Sep 2010 02:32:38 +0000 (22:32 -0400)] 
handle ugly edge case in retrying entrynodes

Specifically, a circ attempt that we'd launched while the network was
down could timeout after we've marked our entrynodes up, marking them
back down again. The fix is to annotate as bad the OR conns that were
around before we did the retry, so if a circuit that's attached to them
times out we don't do anything about it.

15 years agoActually notice when our last entrynode goes down
Roger Dingledine [Wed, 29 Sep 2010 01:59:31 +0000 (21:59 -0400)] 
Actually notice when our last entrynode goes down

Otherwise we'd never set have_minimum_dir_info to false, so the
"optimistic retry" would never trigger.

15 years agooptimistically retry EntryNodes on socks request
Roger Dingledine [Fri, 24 Sep 2010 02:41:01 +0000 (22:41 -0400)] 
optimistically retry EntryNodes on socks request

We used to mark all our known bridges up when they're all down and we
get a new socks request. Now do that when we've set EntryNodes too.

15 years agoremove a redundant assert
Roger Dingledine [Fri, 24 Sep 2010 02:10:30 +0000 (22:10 -0400)] 
remove a redundant assert

15 years agorefactor; no actual changes
Roger Dingledine [Fri, 24 Sep 2010 02:10:13 +0000 (22:10 -0400)] 
refactor; no actual changes

15 years agoMerge branch 'maint-0.2.2'
Roger Dingledine [Tue, 28 Sep 2010 22:42:04 +0000 (18:42 -0400)] 
Merge branch 'maint-0.2.2'

15 years agoMerge branch 'maint-0.2.1' into maint-0.2.2
Roger Dingledine [Tue, 28 Sep 2010 22:37:55 +0000 (18:37 -0400)] 
Merge branch 'maint-0.2.1' into maint-0.2.2

15 years agoactually retry bridges when your network goes away
Roger Dingledine [Tue, 28 Sep 2010 22:36:15 +0000 (18:36 -0400)] 
actually retry bridges when your network goes away

15 years agoMerge branch 'iocp_option'
Nick Mathewson [Tue, 28 Sep 2010 18:42:52 +0000 (14:42 -0400)] 
Merge branch 'iocp_option'

15 years agoMerge remote branch 'origin/maint-0.2.2'
Nick Mathewson [Tue, 28 Sep 2010 18:42:50 +0000 (14:42 -0400)] 
Merge remote branch 'origin/maint-0.2.2'

15 years agoAutodetect the number of CPUs when possible if NumCPUs==0
Nick Mathewson [Tue, 28 Sep 2010 18:36:28 +0000 (14:36 -0400)] 
Autodetect the number of CPUs when possible if NumCPUs==0

This is needed for IOCP, since telling the IOCP backend about all
your CPUs is a good idea.  It'll also come in handy with asn's
multithreaded crypto stuff, and for people who run servers without
reading the manual.

15 years agoAdd a new option to enable/disable IOCP support
Nick Mathewson [Tue, 28 Sep 2010 18:01:45 +0000 (14:01 -0400)] 
Add a new option to enable/disable IOCP support

15 years agoRemove the completely outdated Win32Build directory
Nick Mathewson [Tue, 28 Sep 2010 17:35:06 +0000 (13:35 -0400)] 
Remove the completely outdated Win32Build directory

If you want to try to build Tor with a recent MSVC, you are better
off starting from scratch than trying to use the project files that
we used to build 2006 versions of Tor using 2006 versions of MSVC.

15 years agoMerge branch 'microdesc_dl_v2'
Nick Mathewson [Tue, 28 Sep 2010 04:46:01 +0000 (00:46 -0400)] 
Merge branch 'microdesc_dl_v2'

15 years agoMerge branch 'maint-0.2.2'
Roger Dingledine [Mon, 27 Sep 2010 22:33:12 +0000 (18:33 -0400)] 
Merge branch 'maint-0.2.2'

15 years agoa dir-spec entry for refuseunknownexits
Roger Dingledine [Mon, 27 Sep 2010 22:32:09 +0000 (18:32 -0400)] 
a dir-spec entry for refuseunknownexits

plus quiet a log line

15 years agoClean long-unlisted microdescs from the cache
Nick Mathewson [Tue, 21 Sep 2010 20:16:56 +0000 (16:16 -0400)] 
Clean long-unlisted microdescs from the cache

15 years agoDownload microdescriptors if you're a cache
Nick Mathewson [Tue, 11 May 2010 21:20:33 +0000 (17:20 -0400)] 
Download microdescriptors if you're a cache

This commit adds some functions to see what microdescriptors we're missing,
and adds fetch-microdesc/store-microdesc logic to the directory code.

15 years agoWhen updating routerdesc downloads, look specifically at the ns consensus
Nick Mathewson [Mon, 20 Sep 2010 18:18:17 +0000 (14:18 -0400)] 
When updating routerdesc downloads, look specifically at the ns consensus

This will be needed when "get_live_consensus" potentially returns a
microdesc consensus.

15 years agoActually parse the m lines from a microdesc consensus
Nick Mathewson [Thu, 16 Sep 2010 16:55:48 +0000 (12:55 -0400)] 
Actually parse the m lines from a microdesc consensus

15 years agoCode to download, parse, and store microdesc consensuses
Nick Mathewson [Fri, 17 Sep 2010 02:12:03 +0000 (22:12 -0400)] 
Code to download, parse, and store microdesc consensuses

15 years agoNew function to grab a directory connection fetching a resource
Nick Mathewson [Mon, 20 Sep 2010 15:24:01 +0000 (11:24 -0400)] 
New function to grab a directory connection fetching a resource

We need this to tell if a given consensus flavor is being downloaded
or not.

15 years agoMerge remote branch 'origin/maint-0.2.2'
Nick Mathewson [Mon, 27 Sep 2010 21:52:20 +0000 (17:52 -0400)] 
Merge remote branch 'origin/maint-0.2.2'

Conflicts:
src/or/config.c

15 years agothe actual changelog entry this time
Roger Dingledine [Mon, 27 Sep 2010 21:44:51 +0000 (17:44 -0400)] 
the actual changelog entry this time

15 years agochangelog entry for bug1751
Roger Dingledine [Mon, 27 Sep 2010 21:44:00 +0000 (17:44 -0400)] 
changelog entry for bug1751

15 years agoMerge remote branch 'sebastian/bug1964' into maint-0.2.2
Nick Mathewson [Mon, 27 Sep 2010 21:26:32 +0000 (17:26 -0400)] 
Merge remote branch 'sebastian/bug1964' into maint-0.2.2

15 years agoFix a bridge segfault
Sebastian Hahn [Mon, 27 Sep 2010 21:15:43 +0000 (23:15 +0200)] 
Fix a bridge segfault

When we enabled support to change statistic options without restarting
Tor we forgot to initialize geoip_countries. Fix that.

15 years agoMerge branch 'refuseunknown' into maint-0.2.2
Nick Mathewson [Mon, 27 Sep 2010 21:15:57 +0000 (17:15 -0400)] 
Merge branch 'refuseunknown' into maint-0.2.2