1 Changes in version 0.0.2pre24 - 2004-03-03
3 - Fix an assertion failure in dns.c, where we were trying to dequeue
4 a pending dns resolve even if it wasn't pending
5 - Fix a spurious socks5 warning about still trying to write after the connection is finished.
6 - Hold certain marked_for_close connections open until they're finished
7 flushing, rather than losing bytes by closing them too early.
8 - Correctly report the reason for ending a stream
9 - Remove some duplicate calls to connection_mark_for_close
10 - Put switch_id and start_daemon earlier in the boot sequence, so it
11 will actually try to chdir() to options.DataDirectory
12 - Make 'make test' exit(1) if a test fails; fix some unit tests
13 - Make tor fail when you use a config option it doesn't know about,
14 rather than warn and continue.
16 - Bugfixes on the rpm spec file and tor.sh, so it's more up to date
19 Changes in version 0.0.2pre23 - 2004-02-29
21 - Print a statement when the first circ is finished, so the user
23 - If a relay cell is unrecognized at the end of the circuit,
24 send back a destroy. (So attacks to mutate cells are more
26 - New config option 'excludenodes' to avoid certain nodes for circuits.
27 - When it daemonizes, it chdir's to the DataDirectory rather than "/",
28 so you can collect coredumps there.
30 - Fix a bug in tls flushing where sometimes data got wedged and
31 didn't flush until more data got sent. Hopefully this bug was
32 a big factor in the random delays we were seeing.
33 - Make 'connected' cells include the resolved IP, so the client
34 dns cache actually gets populated.
35 - Disallow changing from ORPort=0 to ORPort>0 on hup.
36 - When we time-out on a stream and detach from the circuit, send an
37 end cell down it first.
38 - Only warn about an unknown router (in exitnodes, entrynodes,
39 excludenodes) after we've fetched a directory.
42 Changes in version 0.0.2pre22 - 2004-02-26
44 - Servers publish less revealing uname information in descriptors.
45 - More memory tracking and assertions, to crash more usefully when
47 - If the default torrc isn't there, just use some default defaults.
48 Plus provide an internal dirservers file if they don't have one.
49 - When the user tries to use Tor as an http proxy, give them an http
50 501 failure explaining that we're a socks proxy.
51 - Dump a new router.desc on hup, to help confused people who change
52 their exit policies and then wonder why router.desc doesn't reflect
54 - Clean up the generic tor.sh init script that we ship with.
56 - If the exit stream is pending on the resolve, and a destroy arrives,
57 then the stream wasn't getting removed from the pending list. I
58 think this was the one causing recent server crashes.
59 - Use a more robust poll on OSX 10.3, since their poll is flaky.
60 - When it couldn't resolve any dirservers, it was useless from then on.
61 Now it reloads the RouterFile (or default dirservers) if it has no
63 - Move the 'tor' binary back to /usr/local/bin/ -- it turns out
64 many users don't even *have* a /usr/local/sbin/.
67 Changes in version 0.0.2pre21 - 2004-02-18
69 - There's a ChangeLog file that actually reflects the changelog.
70 - There's a 'torify' wrapper script, with an accompanying
71 tor-tsocks.conf, that simplifies the process of using tsocks for
72 tor. It even has a man page.
73 - The tor binary gets installed to sbin rather than bin now.
74 - Retry streams where the connected cell hasn't arrived in 15 seconds
75 - Clean up exit policy handling -- get the default out of the torrc, so we can update it without forcing each server operator to fix
77 - Allow imaps and pop3s in default exit policy
79 - Prevent picking middleman nodes as the last node in the circuit
82 Changes in version 0.0.2pre20 - 2004-01-30
84 - We now have a deb package, and it's in debian unstable. Go to
86 - I've split the TotalBandwidth option into BandwidthRate (how many
87 bytes per second you want to allow, long-term) and
88 BandwidthBurst (how many bytes you will allow at once before the cap
89 kicks in). This better token bucket approach lets you, say, set
90 BandwidthRate to 10KB/s and BandwidthBurst to 10MB, allowing good
91 performance while not exceeding your monthly bandwidth quota.
92 - Push out a tls record's worth of data once you've got it, rather
93 than waiting until you've read everything waiting to be read. This
94 may improve performance by pipelining better. We'll see.
95 - Add an AP_CONN_STATE_CONNECTING state, to allow streams to detach
96 from failed circuits (if they haven't been connected yet) and attach
98 - Expire old streams that haven't managed to connect. Some day we'll
99 have them reattach to new circuits instead.
102 - Fix several memory leaks that were causing servers to become bloated
104 - Fix a few very rare assert triggers. A few more remain.
105 - Setuid to User _before_ complaining about running as root.
108 Changes in version 0.0.2pre19 - 2004-01-07
110 - Fix deadlock condition in dns farm. We were telling a child to die by
111 closing the parent's file descriptor to him. But newer children were
112 inheriting the open file descriptor from the parent, and since they
113 weren't closing it, the socket never closed, so the child never read
114 eof, so he never knew to exit. Similarly, dns workers were holding
115 open other sockets, leading to all sorts of chaos.
116 - New cleaner daemon() code for forking and backgrounding.
117 - If you log to a file, it now prints an entry at the top of the
118 logfile so you know it's working.
119 - The onionskin challenge length was 30 bytes longer than necessary.
120 - Started to patch up the spec so it's not quite so out of date.
123 Changes in version 0.0.2pre18 - 2004-01-02
125 - Fix endian issues with the 'integrity' field in the relay header.
126 - Fix a potential bug where connections in state
127 AP_CONN_STATE_CIRCUIT_WAIT might unexpectedly ask to write.
130 Changes in version 0.0.2pre17 - 2003-12-30
132 - Made --debuglogfile (or any second log file, actually) work.
133 - Resolved an edge case in get_unique_circ_id_by_conn where a smart
134 adversary could force us into an infinite loop.
137 - Each onionskin handshake now includes a hash of the computed key,
138 to prove the server's identity and help perfect forward secrecy.
139 - Changed cell size from 256 to 512 bytes (working toward compatibility
141 - Changed cell length to 2 bytes, and moved it to the relay header.
142 - Implemented end-to-end integrity checking for the payloads of
144 - Separated streamid from 'recognized' (otherwise circuits will get
145 messed up when we try to have streams exit from the middle). We
146 use the integrity-checking to confirm that a cell is addressed to
148 - Randomize the initial circid and streamid values, so an adversary who
149 breaks into a node can't learn how many circuits or streams have
153 Changes in version 0.0.2pre16 - 2003-12-14
155 - Fixed a bug that made HUP trigger an assert
156 - Fixed a bug where a circuit that immediately failed wasn't being
157 counted as a failed circuit in counting retries.
160 - Now we close the circuit when we get a truncated cell: otherwise we're
161 open to an anonymity attack where a bad node in the path truncates
162 the circuit and then we open streams at him.
163 - Add port ranges to exit policies
164 - Add a conservative default exit policy
165 - Warn if you're running tor as root
166 - on HUP, retry OR connections and close/rebind listeners
167 - options.EntryNodes: try these nodes first when picking the first node
168 - options.ExitNodes: if your best choices happen to include any of
169 your preferred exit nodes, you choose among just those preferred
171 - options.ExcludedNodes: nodes that are never picked in path building
174 Changes in version 0.0.2pre15 - 2003-12-03
175 o Robustness and bugfixes:
176 - Sometimes clients would cache incorrect DNS resolves, which would
177 really screw things up.
178 - An OP that goes offline would slowly leak all its sockets and stop
180 - A wide variety of bugfixes in exit node selection, exit policy
181 handling, and processing pending streams when a new circuit is
183 - Pick nodes for a path only from those the directory says are up
184 - Choose randomly from all running dirservers, not always the first one
185 - Increase allowed http header size for directory fetch.
186 - Stop writing to stderr (if we're daemonized it will be closed).
187 - Enable -g always, so cores will be more useful to me.
188 - Switch "-lcrypto -lssl" to "-lssl -lcrypto" for broken distributions.
191 - Wrote a man page. It lists commonly used options.
194 - Change default loglevel to warn.
195 - Make PidFile default to null rather than littering in your CWD.
196 - OnionRouter config option is now obsolete. Instead it just checks
198 - Moved to a single unified torrc file for both clients and servers.
201 Changes in version 0.0.2pre14 - 2003-11-29
202 o Robustness and bugfixes:
203 - Force the admin to make the DataDirectory himself
204 - to get ownership/permissions right
205 - so clients no longer make a DataDirectory and then never use it
206 - fix bug where a client who was offline for 45 minutes would never
207 pull down a directory again
208 - fix (or at least hide really well) the dns assert bug that was
209 causing server crashes
210 - warnings and improved robustness wrt clockskew for certs
211 - use the native daemon(3) to daemonize, when available
212 - exit if bind() fails
213 - exit if neither socksport nor orport is defined
214 - include our own tor_timegm (Win32 doesn't have its own)
215 - bugfix for win32 with lots of connections
216 - fix minor bias in PRNG
217 - make dirserver more robust to corrupt cached directory
220 - Wrote the design document (woo)
222 o Circuit building and exit policies:
223 - Circuits no longer try to use nodes that the directory has told them
225 - Exit policies now support bitmasks (18.0.0.0/255.0.0.0) and
226 bitcounts (18.0.0.0/8).
227 - Make AP connections standby for a circuit if no suitable circuit
228 exists, rather than failing
229 - Circuits choose exit node based on addr/port, exit policies, and
230 which AP connections are standing by
231 - Bump min pathlen from 2 to 3
232 - Relay end cells have a payload to describe why the stream ended.
233 - If the stream failed because of exit policy, try again with a new
235 - Clients have a dns cache to remember resolved addresses.
236 - Notice more quickly when we have no working circuits
239 - APPort is now called SocksPort
240 - SocksBindAddress, ORBindAddress, DirBindAddress let you configure
242 - RecommendedVersions is now a config variable rather than
243 hardcoded (for dirservers)
244 - Reloads config on HUP
245 - Usage info on -h or --help
246 - If you set User and Group config vars, it'll setu/gid to them.