]> git.ipfire.org Git - thirdparty/tor.git/log
thirdparty/tor.git
16 years agoSplit crypto tests into a separate module.
Nick Mathewson [Tue, 22 Sep 2009 17:06:47 +0000 (13:06 -0400)] 
Split crypto tests into a separate module.

16 years agoRemove a warning on ./autogen.sh
Sebastian Hahn [Tue, 22 Sep 2009 16:50:09 +0000 (18:50 +0200)] 
Remove a warning on ./autogen.sh

16 years agoMove testing code into new src/test directory.
Nick Mathewson [Tue, 22 Sep 2009 16:31:15 +0000 (12:31 -0400)] 
Move testing code into new src/test directory.

16 years agoRefactor unit tests to use the tinytest framework.
Nick Mathewson [Mon, 21 Sep 2009 18:23:13 +0000 (14:23 -0400)] 
Refactor unit tests to use the tinytest framework.

"Tinytest" is a minimalist C unit testing framework I wrote for
Libevent.  It supports some generally useful features, like being able
to run separate unit tests in their own processes.

I tried to do the refactoring to change test.c as little as possible.
Thus, we mostly don't call the tinytest macros directly.  Instead, the
test.h header is now a wrapper on tinytest.h to make our existing
test_foo() macros work.

The next step(s) here will be:
  - To break test.c into separate files, each with its own test group.
  - To look into which things we can test
  - To refactor the more fiddly tests to use the tinytest macros
    directly and/or run forked.
  - To see about writing unit tests for things we couldn't previously
    test without forking.

16 years agoBump version to 0.2.2.3-alpha-dev
Nick Mathewson [Wed, 23 Sep 2009 04:24:37 +0000 (00:24 -0400)] 
Bump version to 0.2.2.3-alpha-dev

16 years agofinishing touches on the changelog tor-0.2.2.3-alpha
Roger Dingledine [Wed, 23 Sep 2009 04:13:57 +0000 (00:13 -0400)] 
finishing touches on the changelog

16 years agobump to 0.2.2.3-alpha, plus add a changelog for bug 1103
Roger Dingledine [Wed, 23 Sep 2009 02:15:56 +0000 (22:15 -0400)] 
bump to 0.2.2.3-alpha, plus add a changelog for bug 1103

16 years agoBe more robust to bad circwindow values
Roger Dingledine [Wed, 23 Sep 2009 02:09:33 +0000 (22:09 -0400)] 
Be more robust to bad circwindow values

If the networkstatus consensus tells us that we should use a
negative circuit package window, ignore it. Otherwise we'll
believe it and then trigger an assert.

Also, change the interface for networkstatus_get_param() so we
don't have to lookup the consensus beforehand.

16 years agoMerge commit 'mikeperry/circuittimeout-1103'
Nick Mathewson [Tue, 22 Sep 2009 03:35:49 +0000 (23:35 -0400)] 
Merge commit 'mikeperry/circuittimeout-1103'

16 years agoFix Bug 1103.
Mike Perry [Tue, 22 Sep 2009 03:01:20 +0000 (20:01 -0700)] 
Fix Bug 1103.

Don't pass in a quantile that is too high during pretimeout
calcualtion.

16 years agoMerge commit 'debian-tor-0.2.2.2-alpha-1'
Nick Mathewson [Mon, 21 Sep 2009 20:06:01 +0000 (16:06 -0400)] 
Merge commit 'debian-tor-0.2.2.2-alpha-1'

16 years agoBump version to 0.2.2.2-alpha-dev
Nick Mathewson [Mon, 21 Sep 2009 19:51:08 +0000 (15:51 -0400)] 
Bump version to 0.2.2.2-alpha-dev

16 years agoBuild debian's micro-revision into the binary
Peter Palfrader [Mon, 21 Sep 2009 12:51:26 +0000 (14:51 +0200)] 
Build debian's micro-revision into the binary

If we have a debian/micro-revision.i, replace the one in src/or
with our copy so that this will be the revision that ends up in
the binary.  This is an informational only version string, but
it'd be kinda nice if it was (more) accurate nonetheless.

Of course this won't help if people manually patch around but
it's still preferable to claiming we are exactly upstream's source.

If we are building directly out of a git tree, update
debian/micro-revision.i in the clean target.

16 years agoRemove common_sha1.i and or_sha1.i in clean
Peter Palfrader [Mon, 21 Sep 2009 11:39:40 +0000 (13:39 +0200)] 
Remove common_sha1.i and or_sha1.i in clean

The files src/common/common_sha1.i src/or/or_sha1.i get changed during
the build - they contain the checksums of the individual files that end
up in the binary.  Of couse changes only end up in the debian diff.gz
after building a second time in the same directory.  So, remove those
files in clean to get both a cleaner diff.gz and idempotent builds.

16 years agoMerge branch 'debian-merge' into debian
Peter Palfrader [Mon, 21 Sep 2009 11:16:44 +0000 (13:16 +0200)] 
Merge branch 'debian-merge' into debian

* debian-merge: (95 commits)
  New upstream version: 0.2.2.2-alpha
  downgrade a log severity, since this event has been known
  Update to the "September 4 2009" ip-to-country file.
  bump to 0.2.2.2-alpha
  Revert "Teach connection_ap_can_use_exit about Exclude*Nodes"
  fix grammar / add changelog for the torify commit
  Fix compile on Snow Leopard
  Fix build warnings on OSX 10.5.8
  Change the condition on the nonlive timeout counting.
  Add a couple of time helper functions.
  Fix typos and comments, plus two bugs
  Implement and document new network liveness algorithm.
  Fix some precision-related asserts in unit tests.
  replace contrib/auto-naming with a readme saying where it went
  clarify our rules for assigning the Named flag
  disable the end of circuitbuildtimeout units tests
  draw in a lot of 0.2.1.20 changelog items into 0.2.2.2-alpha
  Fix compile on freebsd
  Let our config abbreviations rewrite more than once
  a mish-mash of stuff in my sandbox
  ...

16 years agoNew upstream version: 0.2.2.2-alpha
Peter Palfrader [Mon, 21 Sep 2009 11:16:05 +0000 (13:16 +0200)] 
New upstream version: 0.2.2.2-alpha

16 years agoMerge commit 'tor-0.2.2.2-alpha' into debian-merge
Peter Palfrader [Mon, 21 Sep 2009 11:14:39 +0000 (13:14 +0200)] 
Merge commit 'tor-0.2.2.2-alpha' into debian-merge

* commit 'tor-0.2.2.2-alpha': (94 commits)
  downgrade a log severity, since this event has been known
  Update to the "September 4 2009" ip-to-country file.
  bump to 0.2.2.2-alpha
  Revert "Teach connection_ap_can_use_exit about Exclude*Nodes"
  fix grammar / add changelog for the torify commit
  Fix compile on Snow Leopard
  Fix build warnings on OSX 10.5.8
  Change the condition on the nonlive timeout counting.
  Add a couple of time helper functions.
  Fix typos and comments, plus two bugs
  Implement and document new network liveness algorithm.
  Fix some precision-related asserts in unit tests.
  replace contrib/auto-naming with a readme saying where it went
  clarify our rules for assigning the Named flag
  disable the end of circuitbuildtimeout units tests
  draw in a lot of 0.2.1.20 changelog items into 0.2.2.2-alpha
  Fix compile on freebsd
  Let our config abbreviations rewrite more than once
  a mish-mash of stuff in my sandbox
  give proposal 151 a changelog and other touchups
  ...

16 years agodowngrade a log severity, since this event has been known tor-0.2.2.2-alpha
Roger Dingledine [Mon, 21 Sep 2009 07:32:28 +0000 (03:32 -0400)] 
downgrade a log severity, since this event has been known
to happen and there's nothing the user can do about it

16 years agoUpdate to the "September 4 2009" ip-to-country file.
Roger Dingledine [Mon, 21 Sep 2009 05:44:59 +0000 (01:44 -0400)] 
Update to the "September 4 2009" ip-to-country file.

16 years agobump to 0.2.2.2-alpha
Roger Dingledine [Mon, 21 Sep 2009 05:31:39 +0000 (01:31 -0400)] 
bump to 0.2.2.2-alpha

16 years agoMerge branch 'maint-0.2.1'
Roger Dingledine [Mon, 21 Sep 2009 03:53:03 +0000 (23:53 -0400)] 
Merge branch 'maint-0.2.1'

16 years agoRevert "Teach connection_ap_can_use_exit about Exclude*Nodes"
Roger Dingledine [Mon, 21 Sep 2009 03:50:48 +0000 (23:50 -0400)] 
Revert "Teach connection_ap_can_use_exit about Exclude*Nodes"

This reverts commit dc3229313b6d2aaff437c6fc7fa55ead4409e93d.

We're going to do this more thoroughly in 0.2.2.x, and not in
maint-0.2.1.

16 years agofix grammar / add changelog for the torify commit
Roger Dingledine [Mon, 21 Sep 2009 03:24:46 +0000 (23:24 -0400)] 
fix grammar / add changelog for the torify commit

16 years agoFix compile on Snow Leopard
Sebastian Hahn [Mon, 21 Sep 2009 03:12:37 +0000 (05:12 +0200)] 
Fix compile on Snow Leopard

16 years agoMerge commit 'sebastian/weasel_torify'
Roger Dingledine [Mon, 21 Sep 2009 03:11:46 +0000 (23:11 -0400)] 
Merge commit 'sebastian/weasel_torify'

16 years agoFix build warnings on OSX 10.5.8
Nick Mathewson [Mon, 21 Sep 2009 03:01:43 +0000 (23:01 -0400)] 
Fix build warnings on OSX 10.5.8

16 years agoChange the condition on the nonlive timeout counting.
Mike Perry [Mon, 21 Sep 2009 01:20:10 +0000 (18:20 -0700)] 
Change the condition on the nonlive timeout counting.

Try to clarify things in the comment too.

16 years agoAdd a couple of time helper functions.
Mike Perry [Mon, 21 Sep 2009 01:03:39 +0000 (18:03 -0700)] 
Add a couple of time helper functions.

Also add rounding support to tv_mdiff().

16 years agoFix typos and comments, plus two bugs
Roger Dingledine [Sun, 20 Sep 2009 23:50:44 +0000 (19:50 -0400)] 
Fix typos and comments, plus two bugs

A) We were considering a circuit had timed out in the special cases
where we close rendezvous circuits because the final rendezvous
circuit couldn't be built in time.
B) We were looking at the wrong timestamp_created when considering
a timeout.

16 years agoImplement and document new network liveness algorithm.
Mike Perry [Fri, 18 Sep 2009 09:01:39 +0000 (02:01 -0700)] 
Implement and document new network liveness algorithm.

Based on irc discussion with arma.

16 years agoFix some precision-related asserts in unit tests.
Mike Perry [Fri, 18 Sep 2009 08:48:07 +0000 (01:48 -0700)] 
Fix some precision-related asserts in unit tests.

Mostly by storing the timeout as milliseconds and not seconds
internally.

16 years agoreplace contrib/auto-naming with a readme saying where it went
Roger Dingledine [Sat, 19 Sep 2009 10:22:50 +0000 (06:22 -0400)] 
replace contrib/auto-naming with a readme saying where it went

16 years agoclarify our rules for assigning the Named flag
Roger Dingledine [Fri, 18 Sep 2009 07:05:52 +0000 (03:05 -0400)] 
clarify our rules for assigning the Named flag

16 years agodisable the end of circuitbuildtimeout units tests
Roger Dingledine [Fri, 18 Sep 2009 06:00:20 +0000 (02:00 -0400)] 
disable the end of circuitbuildtimeout units tests

until mike figures out some other way to make them pass reliably

16 years agodraw in a lot of 0.2.1.20 changelog items into 0.2.2.2-alpha
Roger Dingledine [Fri, 18 Sep 2009 02:44:11 +0000 (22:44 -0400)] 
draw in a lot of 0.2.1.20 changelog items into 0.2.2.2-alpha

16 years agoFix compile on freebsd
Sebastian Hahn [Fri, 18 Sep 2009 00:43:45 +0000 (02:43 +0200)] 
Fix compile on freebsd

16 years agoLet our config abbreviations rewrite more than once
Roger Dingledine [Fri, 18 Sep 2009 00:32:42 +0000 (20:32 -0400)] 
Let our config abbreviations rewrite more than once

16 years agoa mish-mash of stuff in my sandbox
Roger Dingledine [Thu, 17 Sep 2009 05:58:39 +0000 (01:58 -0400)] 
a mish-mash of stuff in my sandbox

16 years agogive proposal 151 a changelog and other touchups
Roger Dingledine [Thu, 17 Sep 2009 05:42:33 +0000 (01:42 -0400)] 
give proposal 151 a changelog and other touchups

16 years agoChoose early circuit idle timeout better.
Roger Dingledine [Thu, 17 Sep 2009 05:41:47 +0000 (01:41 -0400)] 
Choose early circuit idle timeout better.

Don't discard all circuits every MaxCircuitDirtiness, because the
user might legitimately have set that to a very lower number.

Also don't use up all of our idle circuits with testing circuits,
since that defeats the point of preemptive circuits.

16 years agoMerge commit 'origin/maint-0.2.1'
Nick Mathewson [Thu, 17 Sep 2009 04:42:41 +0000 (00:42 -0400)] 
Merge commit 'origin/maint-0.2.1'

16 years agoWork around a memory leak in openssl 0.9.8g (and maybe others)
Nick Mathewson [Thu, 17 Sep 2009 04:01:20 +0000 (00:01 -0400)] 
Work around a memory leak in openssl 0.9.8g (and maybe others)

16 years agoMerge commit 'origin/maint-0.2.1'
Nick Mathewson [Thu, 17 Sep 2009 03:37:27 +0000 (23:37 -0400)] 
Merge commit 'origin/maint-0.2.1'

[Do not take conflicting change to rendclient.c]

16 years agoMerge commit 'karsten/fix-1073' into maint-0.2.1
Nick Mathewson [Thu, 17 Sep 2009 03:36:01 +0000 (23:36 -0400)] 
Merge commit 'karsten/fix-1073' into maint-0.2.1

16 years agoMerge commit 'mikeperry/circuitbuildtimeout-final'
Roger Dingledine [Thu, 17 Sep 2009 01:43:31 +0000 (21:43 -0400)] 
Merge commit 'mikeperry/circuitbuildtimeout-final'

16 years agoClarify use of magic number 0.98 with #define.
Mike Perry [Thu, 17 Sep 2009 01:41:22 +0000 (18:41 -0700)] 
Clarify use of magic number 0.98 with #define.

16 years agoMerge branch 'maint-0.2.1'
Roger Dingledine [Thu, 17 Sep 2009 01:28:49 +0000 (21:28 -0400)] 
Merge branch 'maint-0.2.1'

16 years agoMerge commit 'karsten/no-time-maint-0.2.1'
Roger Dingledine [Thu, 17 Sep 2009 00:45:01 +0000 (20:45 -0400)] 
Merge commit 'karsten/no-time-maint-0.2.1'

16 years agoMerge commit 'sebastian/specconformance'
Roger Dingledine [Thu, 17 Sep 2009 00:37:43 +0000 (20:37 -0400)] 
Merge commit 'sebastian/specconformance'

16 years agoMerge commit 'sebastian/manpage'
Roger Dingledine [Thu, 17 Sep 2009 00:29:37 +0000 (20:29 -0400)] 
Merge commit 'sebastian/manpage'

16 years agoFix unit tests and compile issues on Snow Leopard
Sebastian Hahn [Wed, 16 Sep 2009 22:20:25 +0000 (00:20 +0200)] 
Fix unit tests and compile issues on Snow Leopard

16 years agoChange liveness value to be a function of the timeout.
Mike Perry [Thu, 17 Sep 2009 00:14:01 +0000 (17:14 -0700)] 
Change liveness value to be a function of the timeout.

And also the number of recent circuits used to decide
when the network changes.

16 years agoAdd log message so we have accurate build time values.
Mike Perry [Wed, 16 Sep 2009 11:55:43 +0000 (04:55 -0700)] 
Add log message so we have accurate build time values.

16 years agoAddress nickm's issues from his review #1.
Mike Perry [Wed, 16 Sep 2009 01:00:48 +0000 (18:00 -0700)] 
Address nickm's issues from his review #1.

16 years agoUpdate proposal to match implementation.
Mike Perry [Thu, 17 Sep 2009 00:03:54 +0000 (17:03 -0700)] 
Update proposal to match implementation.

16 years agoMove circuitbuildtimeout config check.
Mike Perry [Mon, 14 Sep 2009 11:03:57 +0000 (04:03 -0700)] 
Move circuitbuildtimeout config check.

We want it to be under our control so it doesn't mess
up initialization. This is likely the cause for
the bug the previous assert-adding commit (09a75ad) was
trying to address.

16 years agoTime for some debugging by asserts.
Mike Perry [Fri, 11 Sep 2009 05:12:46 +0000 (22:12 -0700)] 
Time for some debugging by asserts.

Got a negative timeout value on startup. Need to narrow it down.

16 years agoFix bugs relating to not counting timeouts as circuit builds.
Mike Perry [Wed, 9 Sep 2009 07:01:57 +0000 (00:01 -0700)] 
Fix bugs relating to not counting timeouts as circuit builds.

Also use bin midpoints for time values.

16 years agoDocument functions and constants.
Mike Perry [Tue, 8 Sep 2009 08:31:29 +0000 (01:31 -0700)] 
Document functions and constants.

16 years agoRemove an assert.
Mike Perry [Mon, 7 Sep 2009 04:05:17 +0000 (21:05 -0700)] 
Remove an assert.

It seems to fire because of precision issues. Added
more debug info to the warn to try to figure out for sure.

16 years agoFix issues found by arma in review.
Mike Perry [Mon, 7 Sep 2009 03:43:02 +0000 (20:43 -0700)] 
Fix issues found by arma in review.

16 years agospace/indent cleanups, plus point out three bugs
Roger Dingledine [Mon, 7 Sep 2009 03:14:13 +0000 (23:14 -0400)] 
space/indent cleanups, plus point out three bugs

16 years agoWoops. Fix a couple memory leaks.
Mike Perry [Fri, 4 Sep 2009 20:42:58 +0000 (13:42 -0700)] 
Woops. Fix a couple memory leaks.

Also change the max timeout quantile to 0.98, so we can
avoid huge synthetic timeout values.

16 years agoRemove trailing spaces. As if bytes were free...
Karsten Loesing [Thu, 3 Sep 2009 12:44:01 +0000 (14:44 +0200)] 
Remove trailing spaces. As if bytes were free...

Also correct some typos.

16 years agoResolve mode ties in favor of the higher (slower) mode.
Mike Perry [Wed, 2 Sep 2009 22:29:34 +0000 (15:29 -0700)] 
Resolve mode ties in favor of the higher (slower) mode.

16 years agoMore detail for some log msgs.
Mike Perry [Wed, 2 Sep 2009 04:12:47 +0000 (21:12 -0700)] 
More detail for some log msgs.

16 years agoUse our variable directly for timeout.
Mike Perry [Wed, 2 Sep 2009 03:27:43 +0000 (20:27 -0700)] 
Use our variable directly for timeout.

Using CircuitBuildTimeout is prone to issues with SIGHUP, etc.
Also, shuffle the circuit build times array after loading it
in so that newer measurements don't replace chunks of
similarly timed measurements.

16 years agoUpdate proposal to bring it more in-line with implementation.
Mike Perry [Wed, 2 Sep 2009 03:13:52 +0000 (20:13 -0700)] 
Update proposal to bring it more in-line with implementation.

16 years agoFix a couple of assert bugs.
Mike Perry [Tue, 1 Sep 2009 22:40:54 +0000 (15:40 -0700)] 
Fix a couple of assert bugs.

16 years agoFix timeout edge case when we get enough samples.
Mike Perry [Tue, 1 Sep 2009 15:07:26 +0000 (08:07 -0700)] 
Fix timeout edge case when we get enough samples.

Also switch Xm calculation to mode, not min.

16 years agoFix the math.h log() conflict.
Mike Perry [Tue, 1 Sep 2009 06:09:54 +0000 (23:09 -0700)] 
Fix the math.h log() conflict.

It was compiling, but causing segfaults.

Also, adjust when the timer starts for new test circs
and save state every 25 circuits.

16 years agoRecover from changing network connections.
Mike Perry [Tue, 1 Sep 2009 01:10:27 +0000 (18:10 -0700)] 
Recover from changing network connections.

Also add code to keep creating circuits every minute until we
hit our minimum threshhold.

16 years agoFactor out the pretimeout handling code.
Mike Perry [Fri, 28 Aug 2009 09:05:02 +0000 (02:05 -0700)] 
Factor out the pretimeout handling code.

We need to also call it if we're going to calculate alpha
after a normal circuit build.

16 years agoWrite unit tests and fix issues they uncovered.
Mike Perry [Fri, 28 Aug 2009 06:28:20 +0000 (23:28 -0700)] 
Write unit tests and fix issues they uncovered.

16 years agoImplement the pareto fitting and timeout calculating bits.
Mike Perry [Thu, 27 Aug 2009 08:46:06 +0000 (01:46 -0700)] 
Implement the pareto fitting and timeout calculating bits.

16 years agoClean up Fallon's partially complete GSoC project.
Mike Perry [Wed, 26 Aug 2009 00:13:12 +0000 (17:13 -0700)] 
Clean up Fallon's partially complete GSoC project.

The code actually isn't that bad. It's a shame she didn't finish.
Using it as the base for this feature.

16 years agoMerge branch 'tmp'
Roger Dingledine [Wed, 16 Sep 2009 21:18:35 +0000 (17:18 -0400)] 
Merge branch 'tmp'

16 years agoTeach connection_ap_can_use_exit about Exclude*Nodes
Sebastian Hahn [Wed, 16 Sep 2009 00:23:04 +0000 (02:23 +0200)] 
Teach connection_ap_can_use_exit about Exclude*Nodes

To further attempt to fix bug 1090, make sure connection_ap_can_use_exit
always returns 0 when the chosen exit router is excluded. This should fix
bug1090.

16 years agoMerge commit 'origin/maint-0.2.1'
Nick Mathewson [Tue, 15 Sep 2009 23:37:26 +0000 (19:37 -0400)] 
Merge commit 'origin/maint-0.2.1'

Resolved conflicts in:
src/or/circuitbuild.c

16 years agoMerge commit 'sebastian/memleak' into maint-0.2.1
Nick Mathewson [Tue, 15 Sep 2009 23:35:16 +0000 (19:35 -0400)] 
Merge commit 'sebastian/memleak' into maint-0.2.1

16 years agomake some bug 1090 warnings go away
Sebastian Hahn [Fri, 11 Sep 2009 13:40:09 +0000 (15:40 +0200)] 
make some bug 1090 warnings go away

When we excluded some Exits, we were sometimes warning the user that we
were going to use the node regardless. Many of those warnings were in
fact bogus, because the relay in question was not used to connect to
the outside world.

Based on patch by Rotor, thanks!

16 years agoMerge commit 'origin/maint-0.2.1'
Nick Mathewson [Tue, 15 Sep 2009 16:53:44 +0000 (12:53 -0400)] 
Merge commit 'origin/maint-0.2.1'

16 years agoit is cached-descriptors now, not cached-routers
Sebastian Hahn [Wed, 2 Sep 2009 02:50:08 +0000 (04:50 +0200)] 
it is cached-descriptors now, not cached-routers

16 years agoFix compile on Snow Leopard
Sebastian Hahn [Tue, 15 Sep 2009 08:33:58 +0000 (10:33 +0200)] 
Fix compile on Snow Leopard

16 years agosome cleanups:
Sebastian Hahn [Thu, 3 Sep 2009 01:01:23 +0000 (03:01 +0200)] 
some cleanups:

documentation fix for get_uint64
remove extra "." from a log line
fix a long line

16 years agorevert the month in the man page, so we don't drive weasel mad
Roger Dingledine [Tue, 15 Sep 2009 10:52:23 +0000 (06:52 -0400)] 
revert the month in the man page, so we don't drive weasel mad

16 years agoForward port patches/06_add_compile_time_defaults.dpatch
Peter Palfrader [Thu, 3 Sep 2009 13:10:47 +0000 (15:10 +0200)] 
Forward port patches/06_add_compile_time_defaults.dpatch

16 years agoForward port patches/03_tor_manpage_in_section_8.dpatch
Peter Palfrader [Thu, 3 Sep 2009 13:07:08 +0000 (15:07 +0200)] 
Forward port patches/03_tor_manpage_in_section_8.dpatch

16 years agoNew upstream version
Peter Palfrader [Thu, 3 Sep 2009 12:53:51 +0000 (14:53 +0200)] 
New upstream version

16 years agoRead "circwindow=x" from the consensus and use it
Roger Dingledine [Tue, 15 Sep 2009 10:33:33 +0000 (06:33 -0400)] 
Read "circwindow=x" from the consensus and use it

Tor now reads the "circwindow" parameter out of the consensus,
and uses that value for its circuit package window rather than the
default of 1000 cells. Begins the implementation of proposal 168.

16 years agoConsensusParams config option lists key=value params
Roger Dingledine [Tue, 15 Sep 2009 08:40:08 +0000 (04:40 -0400)] 
ConsensusParams config option lists key=value params

finishes the authority-operator interface side of proposal 167.

16 years agoParameter access function, with unit tests.
Nick Mathewson [Tue, 15 Sep 2009 03:39:08 +0000 (23:39 -0400)] 
Parameter access function, with unit tests.

16 years agoMark proposal 167 as implemented.
Nick Mathewson [Tue, 15 Sep 2009 03:24:39 +0000 (23:24 -0400)] 
Mark proposal 167 as implemented.

16 years agoImplement proposal 167: Authorities vote on network parameters.
Nick Mathewson [Tue, 15 Sep 2009 02:15:57 +0000 (22:15 -0400)] 
Implement proposal 167: Authorities vote on network parameters.

This code adds a new field to vote on: "params".  It consists of a list of
sorted key=int pairs.  The output is computed as the median of all the
integers for any key on which anybody voted.

Improved with input from Roger.

16 years agoAdd a median_int32 and find_nth_int32
Nick Mathewson [Tue, 15 Sep 2009 03:14:03 +0000 (23:14 -0400)] 
Add a median_int32 and find_nth_int32

16 years agoFix a memory leak when parsing a ns
Sebastian Hahn [Sun, 13 Sep 2009 19:47:55 +0000 (21:47 +0200)] 
Fix a memory leak when parsing a ns

Adding the same vote to a networkstatus consensus leads to a memory leak
on the client side. Fix that by only using the first vote from any given
voter, and ignoring the others.

Problem found by Rotor, who also helped writing the patch. Thanks!

16 years agoFix a spec conformance issue when parsing a ns vote
Sebastian Hahn [Sun, 13 Sep 2009 19:47:55 +0000 (21:47 +0200)] 
Fix a spec conformance issue when parsing a ns vote

A vote may only contain exactly one signature. Make sure we reject
votes that violate this.

Problem found by Rotor, who also helped writing the patch. Thanks!

16 years agoFix torify: Do not complain if we cannot find all socksifiers
Peter Palfrader [Wed, 9 Sep 2009 19:33:32 +0000 (21:33 +0200)] 
Fix torify: Do not complain if we cannot find all socksifiers

No longer complain if we cannot find both torify and tsocks.  As long as
we have one we are happy.

Do not rely on which, it's not POSIX.

Catch error if torsocks fails and print an error message.

16 years agominor fixes in some comments
Roger Dingledine [Mon, 7 Sep 2009 00:09:08 +0000 (20:09 -0400)] 
minor fixes in some comments

16 years agoMerge branch 'debian-merge' into debian
Peter Palfrader [Thu, 3 Sep 2009 13:19:37 +0000 (15:19 +0200)] 
Merge branch 'debian-merge' into debian

* debian-merge: (190 commits)
  Forward port patches/06_add_compile_time_defaults.dpatch
  Forward port patches/03_tor_manpage_in_section_8.dpatch
  New upstream version
  HiddenServiceVersion must be set to 2 currently.
  put karsten's changelog in the right place
  When Tor fails to parse a descriptor of any kind, dump it to disk.
  update fetch-all with dir auth
  changelog and spec changes for the .exit fix
  bump to 0.2.2.1-alpha
  typos in dir-spec
  mark off a done proposal
  clean up the changelog for 0.2.2.1-alpha
  A changelog entry and a bit more documentation for socks-client
  Clean up a couple of style issues in the socks-client branch.
  new proposals: params in consensus, and lower circwindow
  Add some fixes after discussion with Nick.
  Refactor geoip_get_dirreq_history() some more.
  Fix a memory leak in summarizing directory request timing.
  Add the first 8 bytes of the git commit digest to our versions.
  autoconf 2.59 appears not to support AC_PROG_SED
  ...