1 $Id: TODO 16258 2008-07-30 13:04:38Z nickm $
4 SPEC - Spec not finalized
23 =======================================================================
25 Things Roger would be excited to see:
28 - Finish buffer stuff in libevent; start using it in Tor.
29 - Tors start believing the contents of NETINFO cells.
30 . Work with Steven and Roger to decide which parts of Paul's project
32 - respond to Steven's red-team TLS testing (a.k.a, look at a packet
36 - Fit Vidalia in 640x480 again.
37 - When user changes the language in Vidalia, have it change right then.
38 - Vidalia should display/edit PlaintextPorts events/config.
39 . Vidalia's GUI should let you specify an http proxy that it launches
40 for you. Maybe in the general config window next to which Tor it
42 - Vidalia should avoid stomping on your custom exit policy lines
43 just because you click on 'save' for a totally different config thing.
44 - How much space do we save in TBB by stripping symbols from Vidalia
45 first? Good idea or crazy idea?
48 - gmail auto responder so you send us an email and we send you a Tor
49 binary. Probably needs a proposal first.
50 - document how it works / what its interface is
51 - set it up so people can 'get tor' in many languages, and so the
52 docs we send back are in many languages.
53 - weather.torproject.org should go live.
54 - Keep advocating new Tor servers and working with orgs like Mozilla
56 - solve the huge green onion on check.tp.o
57 - Start converting critical wiki pages into real Tor wml pages. E.g.,
58 https://wiki.torproject.org/noreply/TheOnionRouter/VerifyingSignatures
59 - Find out what happened to the buildbot and get it back up:
60 http://tor-buildbot.freehaven.net:8010/
61 - Learn about locking memory pages that have sensitive content. Get
64 - Vidalia html help files
65 - should we i18nize polipo's error messages too?
66 - Some of our translated wml files are very old -- so old that they
67 are harmful to leave in place. We need some sort of way to notice
68 this and disable them.
71 - Figure out (or give up on) how to run Tor Browser and ordinary
73 - Enumerate and analyze traces left when running from USB
74 - Write a list of research items Tor would like to see done, for the
75 volunteer page. Pick a few you'd like to work on yourself.
76 - Move proposal 131 or equivalent forward.
77 - Keep bugging us about exploits on the .exit notation.
78 - If relays have 100KB/s but set relaybandwidthrate to 10KB/s, do your
79 interference attacks still work?
80 - Mike's question #3 on https://www.torproject.org/volunteer#Research
81 - Worthwhile shipping TBB with some local html help files that come
83 - Decide whether TBB should use Torbutton's "lock" feature.
84 http://archives.seul.org/or/cvs/Jun-2008/msg00186.html
87 - Which bundles include Torbutton? Change the docs/tor-doc-foo pages
88 so they admit that Torbutton is in them too. Change the download
90 - The OS X bundle screenshots are from forever ago -- they don't
91 include Torbutton, they still say it's tor.eff.org, etc.
92 - Should we still be telling you how to use Safari on OS X for Tor,
93 given all the holes that Torbutton-dev solves on Firefox?
96 - Figure out how to make Vidalia and Tor play nicely on Debian, make
97 the necessary modifications, and make some Vidalia debs that pass
100 - Get oftc to switch to Tor dns bulk exitlist. Or tell us why it's
102 - Move proposal 134 forward.
103 - putting port predictions in state file
104 - if tor hasn't been used in a while it stops fetching consensus
105 documents. Retain that state over restarts.
108 - Finish tor-doc-bridge.wml
109 . Fix FAQ entry on setting up private Tor network
110 - Did we actually apply Steven's dkimproxy patch?
111 - Brainstorm about safe but effective ways for vidalia to
112 auto-update its user's bridges via Tor in the background.
115 - Roger wants to get an email every time there's a blog change,
116 e.g. a comment. That way spam doesn't go undetected for weeks.
117 - Or, maybe just disable linking from blog comments entirely?
118 - Get blog.torproject.org a favico
120 =======================================================================
122 Bugs/issues for Tor 0.2.0.x:
123 . we should have an off-by-default way for relays to dump geoip data to
124 a file in their data directory, for measurement purposes.
125 o Basic implementation
126 N - Include probability-of-selection
127 R d let bridges set relaybandwidthrate as low as 5kb
128 R - bridge communities
131 - man page entries for Alternate*Authority config options
133 Documentation for Tor 0.2.0.x:
135 . 111: Prioritize local traffic over relayed.
136 R - Merge into tor-spec.txt.
137 - 113: mark as closed close.
138 o document the "3/4 and 7/8" business in the clients fetching consensus
140 R - then document the bridge user download timeline.
141 - HOWTO for DNSPort. See tup's wiki page.
142 . Document transport and natdport in a good HOWTO.
143 - Quietly document NT Service options: revise (or create) FAQ entry
145 =======================================================================
148 R d bug: if we launch using bridges, and then stop using bridges, we
149 still have our bridges in our entryguards section, and may use them.
150 R d add an event to report geoip summaries to vidalia for bridge relays,
151 so vidalia can say "recent activity (1-8 users) from sa".
152 R - investigate: it looks like if the bridge authority is unreachable,
153 we're not falling back on querying bridges directly?
154 R - if "no running bridges known", an application request should make
155 us retry all our bridges.
156 R d Setting DirPort when acting as bridge will give false Warnings
160 o 110: avoid infinite-length circuits
162 - Internal code support for ipv6:
163 o Clone ipv6 functions (inet_ntop, inet_pton) where they don't exist.
164 o Many address variables need to become tor_addr_t
165 o addr in connection_t
166 o n_addr in extend_info_t
167 - Teach resolving code how to handle ipv6.
168 . Teach exit policies about ipv6 (consider ipv4/ipv6 interaction!)
169 o Use IPv6 in connect/connected/failed-exitpolicy cells
170 o accept ipv6 from socks
171 o Generate END_REASON_EXITPOLICY cells right
172 . ... and parse them right
173 . Generate new BEGIN cell types and parse them right
174 - Detect availability of ipv6
175 - Advertise availability of ipv6.
176 - Geoip support, if only to add a zone called "ipv6"
178 - 118: Listen on and advertise multiple ports:
179 - Tor should be able to have a pool of outgoing IP addresses that it is
180 able to rotate through. (maybe. Possible overlap with proposal 118.)
181 - config option to publish what ports you listen on, beyond
182 ORPort/DirPort. It should support ranges and bit prefixes (?) too.
183 - Need to figure out the right format for routerinfo_t on this.
184 K . 121: Hidden service authentication:
185 missing: delayed descriptor publication for 'stealth' mode.
186 R d 128: families of private bridges
187 - 134: handle authority fragmentation.
188 o 135: simplify configuration of private tor networks. Th
189 - 140: Provide diffs betweeen consensuses
190 K - 143: Improvements of Distributed Hidden Service Descriptor Storage:
191 only easy parts for 0.2.1.x, defer complex ones to 0.2.2.x.
192 - 147: Eliminate the need for v2 directories in generating v3 directories
193 - 148: Stream end reasons from the client side should be uniform.
194 K - 155: Four Improvements of Hidden Service Performance
197 - 145: Separate "suitable from a guard" from "suitable as a new guard"
198 - 146: Adding new flag to reflect long-term stability
199 - 149: Using data from NETINFO cells
201 - Proposals to write:
202 - Fix voting to handle bug 608 case when multiple servers get
204 R d Do we want to maintain our own set of entryguards that we use as
205 next hop after the bridge?
206 d Possibly: revise link protocol to allow big circuit IDs,
207 variable-length cells, proposal-110 stuff, and versioned CREATES?
208 N . Draft proposal for GeoIP aggregation (see external constraints *)
209 . Figure out how to make good use of the fallback consensus file. Right
210 now many of the addresses in the fallback consensus will be stale,
211 so it will take dozens of minutes to bootstrap from it. This is a
212 bad first Tor experience. But if we check the fallback consensus
213 file *after* we fail to connect to any authorities, then it may
214 still be valuable as a blocking-resistance step.
215 o Write the proposal.
216 - Patch our tor.spec rpm package so it knows where to put the fallback
218 d Something for bug 469, to limit connections per IP.
219 . Put bandwidth weights in the networkstatus? So clients get weight
220 their choices even before they have the descriptors; and so
221 authorities can put in more accurate numbers in the future.
222 d Fetch an updated geoip file from the directory authorities.
224 - Tiny designs to write:
225 . Better estimate of clock skew; has anonymity implications. Clients
226 should estimate their skew as median of skew from servers over last
227 N seconds, but for servers this is not so easy, since a server does
228 not choose who it connects to.
229 - Do TLS connection rotation more often than "once a week" in the
231 (One reason not to do it more often is because the old TLS conn
232 probably has a circuit on it, and we don't really want to build up
233 dozens of TCP connections to all the other extra-stable relays.)
234 - If a relay publishes a new descriptor with a significantly lower
235 uptime or with a new IP address, then we should consider its current
236 "running" interval to have ended even if it hadn't yet failed its
237 third reachability test. the interval ended when the new descriptor
238 appeared, and a new interval began then too.
241 - Optimize cell pool allocation.
242 d Support (or just always use) jemalloc (if it helps)
244 - Look into pulling serverdescs off buffers as they arrive.
246 - Use if-modified-since to download consensuses
247 - Handle multi-core cpus better
248 - Split circuit AES across cores?
249 - Split TLS across cores? This will be harder.
250 - Use information from NETINFO cells
251 - Don't extend a circuit over a noncanonical connection with
253 - Learn our outgoing IP address from netinfo cells?
254 - Learn skew from netinfo cells?
256 - Better unit test coverage
257 - Refactor unit tests into multiple files
258 - Verify that write limits to linked connections work.
259 - Use more mid-level and high-level libevent APIs
264 - Get IOCP patch into libevent *
266 - Security improvements
267 - make is-consensus-fresh-enough check way tighter.
268 - If we haven't tried downloading a consensus for ages since we're tired,
269 try getting a new one before we use old descriptors for a circuit.
272 - Feature removals and deprecations:
273 - Get rid of the v1 directory stuff (making, serving, and caching)
274 - First verify that the caches won't flip out?
275 - If they will, just stop the caches from caching for now
276 - perhaps replace it with a "this is a tor server" stock webpage.
277 - The v2dir flag isn't used for anything anymore, right? If so, dump it.
278 - Even clients run rep_hist_load_mtbf_data(). Does this waste memory?
280 - Unless we start using ftime functions, dump them.
281 - can we deprecate 'getinfo network-status'?
282 - can we deprecate the FastFirstHopPK config option?
283 - Can we deprecate controllers that don't use both features?
284 - Dump most uint32_t addr functions.
286 Nice to have for 0.2.1.x:
288 - steven's plan for replacing check.torproject.org with a built-in
289 answer by tor itself.
292 P - Make documentation realize that location of system configuration file
293 will depend on location of system defaults, and isn't always /etc/torrc.
295 - Small controller features
296 - A status event for when tor decides to stop fetching directory info
297 if the client hasn't clicked recently: then make the onion change too.
298 - Add a status event when new consensus arrives
301 P - Figure out why dll's compiled in mingw don't work right in WinXP.
302 P - create a "make win32-bundle" for vidalia-privoxy-tor-torbutton bundle
305 - Refactor the HTTP logic so the functions aren't so large.
306 - Refactor buf_read and buf_write to have sensible ways to return
307 error codes after partial writes
308 o Router_choose_random_node() has a big pile of args. make it "flags".
309 - Streamline how we pick entry nodes: Make choose_random_entry() have
310 less magic and less control logic.
311 - Don't call time(NULL) so much; instead have a static time_t field
312 that gets updated only a handful of times per second.
313 - Move all status info out of routerinfo into local_routerstatus. Make
314 "who can change what" in local_routerstatus explicit. Make
315 local_routerstatus (or equivalent) subsume all places to go for "what
317 - deprecate router_digest_is_trusted_dir() in favor of
318 router_get_trusteddirserver_by_digest()
320 - Make Tor able to chroot itself
321 o allow it to load an entire config file from control interface
323 - log rotation (and FD passing) via control interface
324 - chroot yourself, including inhibit trying to read config file
325 and reopen logs, unless they are under datadir.
328 - Base relative control socket paths (and other stuff in torrc) on datadir.
329 - Tor logs the libevent version on startup, for debugging purposes.
330 This is great. But it does this before configuring the logs, so
331 it only goes to stdout and is then lost.
332 o Make TrackHostExits expire TrackHostExitsExpire seconds after their
333 *last* use, not their *first* use.
334 - enforce a lower limit on MaxCircuitDirtiness and CircuitBuildTimeout.
335 - Make 'safelogging' extend to info-level logs too.
336 - don't do dns hijacking tests if we're reject *:* exit policy?
337 (deferred until 0.1.1.x is less common)
338 - More consistent error checking in router_parse_entry_from_string().
339 I can say "banana" as my bandwidthcapacity, and it won't even squeak.
341 - Interface for letting SOAT modify flags that authorities assign.
342 (How to keep the authority from clobbering them afterwards?