wiki:org/roadmaps/Tor/025/TicketTriage025

Temporary page for 0.2.5 ticket triage

Possible status values: "must do in 0.2.5", "should try to do in 0.2.5", "could do in 0.2.5; tag as deferrable", "move to unspecified", and "move to unspecified, tag as lorax". IMO.

Tickets tagged 0.2.5-final

  • #936: Race in picking connections for create cell
    • nm: Could do diagnostic in 0.2.5. Should add a situation to notice when we have two live channels to one node.
    • as: Should do; diagnostics at minimum
  • #1354: Configuring Tor with --with*dir gives wrong directories
    • nm: Should do in 0.2.5. Related to #6311.
    • as: Should do, along with #6311 referred to in comments; it's 2013 and we should have pkg-config damn it
  • #1523: [Error] We couldn't read a descriptor that is supposedly mmaped in our cache
    • nm: could do. seems uncommon these days.
    • as: Could do - is still an issue with the recent changes to the microdescriptor code for that other bug?
  • #1776: Allow regular relays to be used as bridges
    • nm: must do; it's a deliverable.
    • as: Must do; wasn't this something we were trying to do for one of the sponsors?
  • #2045: Make tor_check_port_forwarding handle incomplete lines
    • nm: could do. should do if we're promoting tor-fw-helper.
    • as: Could do
  • #2068: Add a guard-status getinfo
    • nm: could do.
    • as: Could do
  • #2106: Controller can't unset httpsproxy if it doesn't resolve
    • nm: could do.
    • as: Should do - this sounds more like a defect than an enhancement
  • #2129: Allow Tor client functionality even when we hibernate
    • nm: could do.
    • as: could do
  • #2149: new 'extra dormant' mode for people who never use their tor
    • nm: Should do. Very important as we get shipped to people who aren't using us.
    • as: Should do; looks like we've had patches hanging around forever for this that we should deal with
  • #2191: Send all error-level log messages to controllers
    • nm: should do; deferrable.
    • as: should do
  • #2297: fetching certs for legacy keys?
    • nm: should do if this still happens.
    • as: should do; how does this interact with the #5595 changes?
  • #2454: We should check our IP immediately when cbt notes the network is live again
    • nm: should do.
    • as: should do; patch has been awaiting review a while
  • #2536: Disable outgoing token bucket and reduce token bucket refill interval
    • nm: Needs investigation, research, refactoring. Should try to advance that on this timeframe. Perhaps we need an active research component.
    • as: agree with nick's assessment
  • #2579: Hidden service directory permissions prevent Vidalia from displaying the hidden service hostname
    • nm: defer.
    • as: could do
  • #2665: Create a dirauth DoS response procedure
    • nm: I think we need a new component for this.
    • as: could do
  • #2667: Exits should block reentry into the tor network
    • nm: Major, not critical. Should do, at least to the point of writing a proposal and seeing whether it matters.
    • as: could do, but sounds hairy
  • #2681: brainstorm ways to let Tor clients use yesterday's consensus more safely
    • nm: defer.
    • as: defer
  • #2715: Is rephist-calculated uptime the right metric for HSDir assignment?
    • nm: should-do or must-do.
    • as: should do
  • #2743: safelogging should cover hidden service name and intro-points too
    • nm: must-do
    • as: must doi
  • #2767: Another possible directory-handling bug
    • nm: lorax
    • as: could do
  • #3029: We should save received documents before parsing them
    • nm: lorax
    • as: defer
  • #3037: Internal checks to detect client streams/circuits whose sock request vanished
    • nm: should-do, deferrable.
    • as: should do
  • #3080: Bug: The application request to "jabber.org":5222 has launched 10 circuits without finding one it likes.
    • nm: could-do, needs insight.
    • as: could do but the bug report gives little indication of cause
  • #3511: Automatically chosen published ports should be stable
    • nm: should-do
    • as: should do
  • #3521: Allow controllers to retrieve HS descriptors from Tor
    • nm: could-do, deferrable, semi-lorax.
    • as: could do
  • #3566: Should controller events respect SafeLogging 1 torrc option?
    • nm: unspecified. leaning towards wontfix.
    • as: could do
  • #3569: Refactor socks parsing
    • nm: could-do as part of testing.
    • as: could do
  • #3725: Implement the wildcard "*" protocol in {Client,Server}TransportPlugin lines
    • nm: unsure. ask the pt people if they need this?
    • as: could do
  • #3948: "fooport auto" should be able to get a hint
    • nm: should-do. Related to #3511.
    • as: should do
  • #3972: Implement proposal 179: TLS certificate and handshake normalization
    • nm: must-do: merge the already finished parts.
    • as: should do

31:

  • #4019: Tor warns about public SocksPort addresses twice on startup
    • nm: easy, and we have code. Not a must-do, but an easy fix.
  • #4085: We don't count accounting bytes when time moves backward?
    • nm: should fix; deferrable.
  • #4241: MAX_REND_FAILURES should not be 30
    • nm: Must fix.
  • #4243: rend_consider_services_upload waits up to 4 hours to publish the first HS descriptor by default
    • nm: Should fix as part of hidden service stuff.
  • #4244: Tor changes default value of DirReqStatistics, then wants to SAVECONF the new default
    • nm: Should fix.
  • #4296: Trivial issues in tor_spawn_background()/tor_check_port_forwarding()
    • nm: Has code; should revise and merge.
  • #4347: Geoip parsing fails on user accounts under windows
    • nm: should fix.
  • #4378: Tor guesses IP address when Address is 127.0.0.1
    • nm: should fix
  • #4390: The rotation of the TLS context can act as a fingerprint for bridges
    • nm: should probably fix, though maybe deferrable given pluggable xport
  • #4436: Bridges should be able to disable v1 and v2 link handshakes
    • nm: should probably fix, though maybe deferrable given pluggable xport
  • #4520: Update 'tor-win32-mingw-creation.txt' for Windows Tor build
    • nm: must fix or remove.
  • #4549: Implement user-defined certificate strings through torrc (part of the proposal 179 efforts)
    • nm: deferrable in favor of #4550.
  • #4550: Allow bridge operators to specify their own link certificates (part of proposal 179)
    • nm: should merge patch.
  • #4580: Some Tor clients go nuts requesting the consensus if there is no recent enough consensus
    • nm: investigate more, though we don't have much insight here.
  • #4581: Dir auths should defend themselves from too many begindir requests per address
    • nm: deferrable.
  • #4587: Bugs in tor_tls_got_client_hello()
    • nm: Can we just disable support for the v2 handshake?
  • #4594: tor_tls_state_changed_callback(): detects of ClientHello is too late
    • nm: Can we just disable support for the v2 handshake?
  • #4600: Spec doesn't mention password quotes
    • nm: Should fix.
  • #4636: Add GETINFO interface_address
    • nm: Deferrable, especially unless somebody wants it now.
  • #4647: Tor needs to parse its command line exactly once
    • nm: Should do; there's initial code.
  • #4664: ./autogen.sh puts stuff on stderr
    • nm: Closed as fixed-in-0.2.4.
  • #4670: More bugs on renegotiation limiting code.
    • nm: Should fix, deferrable.
  • #4677: master won't build without V2_HANDSHAKE_SERVER
    • nm: Should fix. (Can we disable the v2 handshake?)
  • #4692: If only a working static OpenSSL is available, ./configure fails
    • nm: Must fix as part of build work.
  • #4712: Review and update any existing patches for proposal 182
    • nm: Should test, investigate, etc. Status pending outcome.
  • #4773: Implement Extended OR port (part of proposal 180)
    • nm: must do for deliverable
  • #4806: Detect and warn when running IPv6-using client without IPv6 address privacy
    • nm: must do.
  • #4817: Control port authentication failures don't differentiate failure types
    • nm: could do; deferrable.
  • #4826: Write proposal for improved consensus voting schedules
    • nm: should do; deferrable.

61:

  • #4847: Bridges binding only to an IPv6 address doesn't work
    • nm: must do
  • #4872: When the valid consensus is old but still barely valid, are the descriptors referenced in it still valid?
    • nm: should investigate; could defer
  • #4885: Make all paths absolute
    • nm: could do; might help with sandbox; deferrable.
  • #4900: Use a more randomized hash function for our hash tables
    • nm: must do.
  • #5018: don't start ClientTransportPlugin proxies until we have a bridge that wants them
    • nm: should do; deferrable
  • #5040: Make public bridges add obfsproxy stats to their extra-info descriptors
    • nm: must do for deliverable
  • #5048: cbtmintimeout should have a lower maximum
    • nm: could do; deferrable
  • #5081: autoreload ServerTransportPlugin process?
    • nm: could do: deferrable I think
  • #5129: Avoid fcntl(O_NONBLOCK) calls on Linux
    • nm: should do; we have code.
  • #5166: 198.18.0.0/15 is reserved and in use by home routers
    • nm: should do something here; should at least have a solution
  • #5190: Collect Rob's patch for throttling flows at guards
    • nm: we should ask rob whether he advocates this
  • #5286: Demote “We stalled too much while trying to write 512 bytes to [moria1].”? (on a client)
    • nm: should do.
  • #5287: Turn LogMessageDomains on by default?
    • nm: do or close as wontfix.
  • #5298: Relay does not pick the right IP addr of local node
    • nm: Should fix.
  • #5390: The SSL3_CK_SCSV ciphersuite makes us look unlike firefox
    • nm: good to do; deferrable now that we focus on PT for anticensorship.
  • #5408: Implement Extended ORPort and TransportControlPort
    • nm: must do for deliverable
  • #5456: Defend against path bias and tagging attacks
    • nm: should do, code quality permitting.
  • #5462: Clients should alert the user if many guards are unreachable
    • nm: should do.
  • #5506: Do we just keep downloading a consensus if our clock is wrong?
    • nm: Should investigate even more.
  • #5526: Expose current Accounting calculation
    • nm: should do; deferrable.
  • #5528: router->address is redundant with router->addr
    • nm: should do; have code to review.
  • #5532: Merge 4561 leftover code that adds wrappers for address-access functions
    • nm: could do; also could abandon.
  • #5601: Tor complains that it can't find a ClientTransportPlugin line, when managed proxy dies.
    • nm: could do.
  • #5609: Bridges should announce transport listening addresses to the control port
    • nm: should do.
  • #5707: Use end to end stream timing data to further prune circuits
    • nm: More investigation needed. Not sure if this is a good idea.
  • #5708: Don't make too many circuits once we're separating streams by domain
    • nm: Must do. Prereq for implementing IsoStreams
  • #5756: Seccomp system call whitelisting on Linux
    • nm: Already in progress.
  • #5788: Add support for relays without an IPv4 address
    • nm: Defer.
  • #5824: Stop reporting relay-only statistics as a bridge
    • nm: Could do; deferrable.
  • #5825: Bridges without geoip file report empty statistics
    • nm: could do.

91:

  • #5902: The whole "choose a doomed exit at random" business is busted
    • nm: could-defer. I'd suggest deferring until we do some big circuitbuild refactor,which we know we need.
  • #5903: Restore ExcludeEntryNodes feature
    • nm: could do; deferrable.
  • #5940: Figure out own IPv6 address
    • nm: should do. (I'd expect that this will interoperate will with #4806.)
  • #5966: Can we speed up bootstrapping by starting circuit building earlier?
    • nm: deferrable; doable.
  • #5968: Improve onion key and TLS management
    • nm: needs-proposal. Can do if the proposal and code land early. I don't grok mike's reasoning well enough here.
  • #5978: Dead code, possible races, in NT service code
    • nm: could-do; needs windows knowledge. deferrable. Is anyone using this?
  • #6027: Directory authorities on IPv6
    • nm: must-do; sponsor item.
  • #6031: Distinguish when a Tor HS is "not found" vs "not reachable" (exists / does not exists)
    • nm: lorax; needs-proposal.
  • #6055: Re-enable TLS 1.1 and TLS 1.2 once they are fixed
    • nm: must do.
  • #6153: Make circ_times static, and abstract most of its accessors.
    • nm: should do; deferrable.
  • #6268: Implement user-defined SSL Ciphers and TLS version
    • nm: must do; related to sponsor item.
  • #6311: Migrate TOR_SEARCH_LIBRARY to use pkg-config
    • nm: should do.
  • #6313: Many of Tor's complex functions should be refactored
    • nm: should do, teim permitting.
  • #6418: Drop support for older versions of the hidden service protocol
    • nm: should do.
  • #6419: is it really a protocolwarn when connection_or_client_learned_peer_id() finds a different keyid?
    • nm: should do.
  • #6456: Merge parse_client_transport_line() and parse_server_transport_line()
    • nm: should do.
  • #6515: Treat timeout of all name lookup tests as also meaning that DNS is broken
    • nm: should do; deferrable.
  • #6525: Add ExcludeDirectDirectoryFetchNodes option
    • nm: should do; could roll into #5903
  • #6546: Replace check.tp.o with internal mapaddress + JSON/XML object
    • nm: not sure if we still want this. I suspect not. Ask mike.
  • #6572: “Bug: Circuit somehow completed a hop while the network was not live.”
    • nm: should investigate.
  • #6609: Proposal to add tor-connect utility to tor-core distribution
    • nm: should consider; there's code.
  • #6622: Tor link against static zlib broken by -pie switch
    • nm: should do something; not sure what to do.
  • #6623: --enable-static-tor cannot succeed
    • nm: should do
  • #6761: PDS_NO_EXISTING_SERVERDESC_FETCH is somewhat archaic
    • nm: should do
  • #6769: Relays (and bridges) don't use microdescriptors
    • nm: could do; deferrable (I think, unless there's subtlety)
  • #6773: DirServer lines should take more than one "orport="
    • nm: must do; easy to do given #8351
  • #6777: add config option to not rate limit authority dir conns
    • nm: could do. (Should do if authorities still want this)
  • #6783: should not serve old v2 statuses
    • nm: must do
  • #6800: An attacker can flood network with new relays to make us stop using bwauth weights
    • nm: must do

121:

  • #6836: Chop functionality out of routerlist.c
    • nm: must do.
  • #6837: More fine-grained modular decomposition
    • nm: must do
  • #6852: bridges (especially unpublished ones) should include usage info in their heartbeats
    • nm: should do; deferrable.
  • #6877: Finally replace all char[] buffers with uint8_t[] buffers
    • nm: should do; deferrable.
  • #6878: Make outbound DNS requests honor OutboundBindAddressIPv6
    • nm: should do; deferrable.
  • #6936: why link to librt and libdl?
    • nm: could do; deferrable.
  • #6938: Bridge DirPort warning not logged to Tor log file
    • nm: should do; deferrable.
  • #6939: Missing IPv6 ORPort reachability check
    • nm: should do.
  • #7003: Wipe relay key material from memory on common crash conditions
    • nm: should do; deferrable.
  • #7009: Handle unstable relays better
    • nm: should do; there are easy fixes here that would improve fetch size.
  • #7059: TorControl: Unrecognized (node) key errors
    • nm: should do; deferrable.
  • #7077: Tor exits with 'descriptor at <address> begins with unexpected string ""'
    • nm: must-do; should be fixed now?
  • #7145: Evaluate, possibly revise, and then implement ideas for TLS certificate normalization
    • nm: must-do; sponsor item.
  • #7148: Even better parameter voting protocol
    • nm: should-do
  • #7164: Bug: microdesc_free() called, but md was still referenced 1 node(s); held_by_nodes == 1
    • nm: must-do
  • #7172: Come up with configuration options and instructions for running private Tor networks in testing and simulation environments
    • nm: must do
  • #7193: Tor's sybil protection doesn't consider IPv6
    • nm: should do
  • #7282: Create consensus-info directory commands
    • nm: must do
  • #7317: torspec: The listening addrport of pluggable transports should be announced on the control port
    • nm: should do
  • #7356: Make channel state test macros
    • nm: should do
  • #7359: Design/implement method for collecting/reporting statistics
    • nm: could do. I think this isn't an nm/as item though.
  • #7457: Add client-side log indicator that an obfsbridge works
    • nm: should do
  • #7471: circuit_unlink_all_from_channel() is brain-damaged
    • nm: must do
  • #7472: Audit all calls to connection_mark_for_close()
    • nm: should do.
  • #7478: routerset membership tests should consider nodes' addr6 fields
    • nm: should do
  • #7481: There should be some way to indicate [2001:...].foo.exit
    • nm: should do.
  • #7482: Discard nonsense in address.c about v4-mapped addresses
    • nm: must do
  • #7484: We allow */bits as an address-and-mask pattern
    • nm: should do
  • #7488: Make sure transparent proxies work with IPv6
    • nm: should do

151:

  • #7509: Publish and use circuit success rates in extrainfo descriptors
    • nm: could do
  • #7553: [simple patch] Expose ISO_STREAM via isolation flag config option
    • nm: not clear. needs safety features. We should at least resolve this.
  • #7555: MapAddress from FQDN to .onion fails because resolve requests for hidden services are not allowed.
    • nm: should do.
  • #7572: Make relay crypto run on multiple CPU cores
    • nm: should do
  • #7583: Distribute Tor Browser version and other information through consensus and internal mapaddress + JSON/XML object
    • nm: must do. (The consensus part, at least)
  • #7590: [PATCH] New option LocalOutboundBindAddress
    • nm: should do, since there's code.
  • #7646: fix/enhance getinfo ns/id/* commands
    • nm: should do
  • #7660: Refactor circuit cannibalization into one place
    • nm: should do
  • #7678: Check if a stream could send a cell but the circuit hasn't asked it to
    • nm: should do
  • #7704: Tor NT service: control request exceptions
    • nm: should do something here, but needs windows clue
  • #7707: Impose a minimum write size for TLS writes
    • nm: should do, if the measured TLS overhead is ever high. Otherwise defer.
  • #7727: Simplify some costly Tor functions (by profile)
    • nm: should do.
  • #7729: Reading pending TLS bytes can take us over at_most
    • nm: should do
  • #7733: Two channels required for bootstrap
    • nm: should do. (Evaluate; is this still happening?)
  • #7734: Something tried to close an or_connection_t without going through channels at src/or/connection.c:3138
    • nm: must do
  • #7743: Avoid needless wasted space in cells
    • nm: should do, if the measured cell overhead is ever high. Otherwise defer.
  • #7751: torspec: Announce name of active pluggable transport through the Extended ORPort
    • nm: must do.
  • #7764: Reimplement circuit_get_by_rend_token_and_purpose with an associative array
    • nm: must do.
  • #7789: Also display configuration file (torrc) location when running as Windows service
    • nm: should do
  • #7803: Clients shouldn't send timestamps in INTRODUCE1 cells
    • nm: must do, high priority, 0.2.4 backport!
  • #7899: We forget to set tls_error sometimes
    • nm: must do
  • #7952: Control port method to get the exit policy
    • nm: should do
  • #7956: Tor uses Roaming (remote) AppData, not Local
    • nm: MUST do.
  • #7961: Publish transports that bind on IPv6 addresses
    • nm: must do
  • #7962: Which address should a multi-ORPort Tor put in its "transport" extra-info line?
    • nm: should do
  • #7971: review address lists in tor_addr_is_internal_()
    • nm: must do
  • #7986: Lengthen the consensus validity interval
    • nm: should do (needs-proposal, needs-investigation)
  • #8001: obfsproxy makes tor warn when one bridge is down
    • nm: should do, deferrable.

181:

  • #8018: path-spec does not discuss guard rotation
    • nm: must-do; spec
  • #8045: torspec: Make public bridges add obfsproxy stats to their extra-info descriptors
    • nm: should do
  • #8093: Jan 29 23:32:20.066 [Warning] Bug/attack: unexpected sendme cell from client. Closing circ.
    • nm: needs proposal to get this right; should do; could defer.
  • #8106: Make .onion addresses harder to harvest by directory servers
    • nm: should do, pending omnibus hidden service improvement ideas being done in a reasonable timeframe
  • #8111: Refactor our checking of whether we should be reading/writing on a connection to use a set of reason-flags
    • nm: should do: let's avoid bugs here.
  • #8140: Review and update rend-spec.txt
    • nm: should do; but better still would be to get the protocol right.
  • #8159: Consolidate thresholds for path build bias and use bias
    • nm: could do; I don't really understand this.
  • #8160: Create separate pending counters during circuit construction
    • nm: could do; I don't much understand this.
  • #8163: It is no longer deterministic which Sybils we omit
    • nm: should do
  • #8170: get independent from host clock time / insecure NTP
    • nm: should do, needs proposal,
  • #8184: Verify path bias code plays well with optimistic data
    • nm: should do; not really 0.2.5-relevant.
  • #8185: circuit_package_relay_cell(): Bug: outgoing relay cell has n_chan==NULL. Dropping.
    • nm: must do
  • #8188: Introduce MaxCircuitDirtiness per listener
    • nm: should do.
  • #8195: tor and capabilities
    • nm: should do; there is code
  • #8197: Do something about policies_parse_exit_policy()'s arguments
    • nm: could do, could defer.
  • #8214: "getinfo address" should work more consistently soon after startup
    • nm: could do
  • #8239: Hidden services should try harder to reuse their old intro points
    • nm: should do.
  • #8243: Getting the HSDir flag should require more effort
    • nm: should do.
  • #8244: The HSDirs for a hidden service should not be predictable indefinitely into the future
    • nm: should/must do.
  • #8247: Some day soon, a 50KB bandwidth test will be too low for the Fast flag
    • nm: should do
  • #8291: configuration file and folder /etc/torrc + /etc/tor.d/
    • nm: could do.
  • #8297: Do not start reading connection if any blocking reason still present
    • nm: should do
  • #8298: Do not start write to connection if blocked on bw
    • nm : should do
  • #8323: Missing 'GETINFO md/all'
    • nm: could do
  • #8351: Refactor our controller-command/torrc-option processing logic into a data-driven function
    • nm: should do (code in progress)
  • #8368: Add tor.service (for systemd) to upstream package
    • nm: should do (we have code)
  • #8369: Option to limit information Tor's control port discloses
    • nm: could-do, needs proposal
  • #8380: DNS A records returned when asking for AAAA
    • nm: must-do, IPv6
  • #8387: Unbuilt one-hop circuits sometimes hang around forever
    • nm: must-do
  • #8402: Tor should help its transport proxy use a proxy, if needed.
    • nm: should do

210:

  • #8405: Provide a control port command to query the circuit used for SOCKS u+p
    • nm: should-do, there's code
  • #8415: Use event_set_mem_functions
    • nm: should do, it's easy.
  • #8453: Alter flag-weight balancing equations
    • nm: could do, needs-proposal
  • #8533: Improve instructions for how to run a testing Tor network
    • nm: should do, orthogonal to release cycle
  • #8546: Make a copy-able connection-config type to limit copy burden of isolation flags, etc
    • nm: should do, refactoring.
  • #8564: Newnym - HS is unavailable
    • nm: unsure. Is this reasonable to do?
  • #8568: use TOR_SIMPLEQ for cell_queue_t
    • (implemented)
  • #8593: Circuit build timeout not respected?
    • nm: should do
  • #8625: Do not call networkstatus_reset_download_failures() hourly
    • nm: must do; authority load.
  • #8710: Sybil selection should prefer measured over advertised bw
    • nm: should do
  • #8712: Authorities should not vote against Fast just because they vote against Running
    • nm: should do, needs-proposal, easy
  • #8727: ServerTransportListenAddr validation could be better
    • nm: could do
  • #8742: Byte history leaks information about local usage/hidden services
    • nm: should do. Or maybe just recomment that an HS not run on the same Tor instance as a node?
  • #8746: Tor tries to kill nonexistent proxy PID on second SIGINT
    • nm: should do; correctness
  • #8749: Return information about the leaking application
    • nm: could do, not sure if it would help
  • #8766: Tor never recovers when started with skewed clock
    • nm: should do, and other timing issues too
  • #8771: GEO IP file directory is wrong
    • nm: must do
  • #8782: Don't give up so easily on your guards if the consensus calls them Running
    • nm: must do
  • #8786: Add extra-info line that tracks the number of consensus downloads of each pluggable transports
    • nm: should do
  • #8787: Check return values for more unix functions
    • nm: should do
  • #8788: We can crash on a bad resolv.conf file
    • nm: should do
  • #8793: Resolve clang scan-build issues
    • nm: should do
  • #8795: Make #8822 survivable
    • nm: could do or must do, depending on whether #8822 reports persist.
  • #8809: Describe key extraction correctly in rend-spec.txt
    • nm: must do
  • #8846: Useless log message when using IPv6 address on SocksPort without IPv6
    • nm: should do
  • #8897: Faster curve25519 implementation for ntor
    • nm: should do
  • #8902: Rumors that hidden services have trouble scaling to 100 concurrent connections
    • nm: could do
  • #8908: Tor systemd socket activation support
    • nm: should do; there is code. See also #8368
  • #8913: update v3 dir spec to say that v2 is obsolete
    • nm: must do
  • #8948: Write a "code review guidelines" page
    • nm: must do
  • #8950: Comments on the formula calculating the number of Introduction Points
    • nm: should do

241:

  • #8954: Need a better convention for testing option variable names
    • nm: should do
  • #8960: replaycache_add_test_and_elapsed uses int for the length of a cache entry
    • nm: done.
  • #8961: src/or/replaycache.c hashes entries with SHA-1
    • nm: must do
  • #8964: doc/HACKING is out of date
    • nm: must do
  • #8966: contrib/ directory should be cleaned up
    • nm: must do
  • #8976: rend_service_introduce() doesn't notice if the rendezvous point is on 127.0.0.1
    • nm: must do
  • #9001: Slow Guard Discovery of Hidden Services and Clients
    • nm: must do
  • #9004: tt_str_op() core-dumps if third argument is NULL
    • nm: should do
  • #9024: add supplementary groups when changing uid
    • nm: could do
  • #9045: Don't let the call frequency of update_networkstatus_downloads() depend on TorTestingNetwork
    • nm: should do.
  • #9062: Authorities should describe their bwauth version in their votes
    • nm: should do
  • #9066: use a macro like CHECK_PRINTF for tor_sscanf
    • nm: must do
  • #9067: Choice of address and match of fascist_firewall_allows_address* need to consider ipv6
    • nm: should do
  • #9068: Unify reachableaddresses and IPv6 settings
    • nm: should do
  • #9088: Lower the voting interval in a test network
    • nm: should do
  • #9093: We should have better, fairer OOM handling
    • nm: must do
  • #9097: Hidden service v0 and v1 descriptor code should go away
    • nm: must do
  • #9099: Comment on rend_service_descriptor_t lies shamelessly
    • nm: must do
  • #9105: make heartbeat messages consistent
    • nm: should do
  • #9107: Drop unused circuit_dump_by_chan function
    • nm: could do
  • #9108: Make circuitbuild.c: global_circuitlist a proper doubly-linked list
    • nm: must do; there's code.
  • #9113: severity too high for message: Something tried to close
    • nm: should do
  • #9119: Reachability detection not relieable
    • nm: should do
  • #9125: Allow reporting different obfs ports to bridge db
    • nm: must do
  • #9126: microdesc_cache_rebuild(): Bug
    • nm: closed as duplicate
  • #9162: get_configured_bridge_by_addr_port_digest() does not work well if digest == NULL
    • nm: must do
  • #9163: Remove PT SOCKS argument length limit when SOCKS4 is used
    • nm: should do
  • #9176: Make the predicted circuits timeout after a configurable amount of time
    • nm: should do/could do?
  • #9188: Our weighting algorithms probably behave strangely when guards are disabled
    • nm: should do
  • #9206: 'Guard' flags only assigned to first nodes started in a private Tor network
    • nm: should do (must do?)
  • #9207: Outdated configure advice concerning libssl
    • nm: should do

271:

  • #9213: configuration files used in wrong order
    • nm: Must do
  • #9222: Torrc's bandwidth entries are more flexible than documented
    • nm: Should do, easy
  • #9229: While bootstrapping, Tor clients stall for 60s when obfsproxy bridges are used.
    • nm: should do
  • #9241: Abstract and decouple path selection from circuit construction
    • nm: should do
  • #9262: Refactor cell scheduling to consider all connections at once
    • nm: must do
  • #9267: Usability tweaks for code coverage
    • nm: could do
  • #9273: Brainstorm tradeoffs from moving to 2 (or even 1) guards
    • nm: must do, security-related
  • #9286: ordb1 uses milliseconds in its descriptor, spec says it can't
    • nm: must do, stability-related
  • #9290: Use something other than "known relay" to decide on rate in connection_or_update_token_buckets_helper()
    • nm: should do
  • #9297: Error while compiling the latest git (seccomp2-related)
    • nm: must do
  • #9299: Dump stack traces on assertion, crash, or general trouble
    • nm: should do (code written!)
  • #9320: Assign bandwidth to new relays faster
    • nm: should-do
  • #9321: Load balance right when we have higher guard rotation periods
    • nm: must-do so we can get the security benefit of keeping guards longer
  • #9333 Illegal nickname "PPriv@last-listed" in family line
    • nm: must-do (conformance)
  • #9363 test_pt_configure_proxy() tests should be improved.
    • nm: must do (testing)
  • #9368 Turn static throttling on in the live network
    • nm: must-do (performance)
  • #9386 Use CREATE_FAST less, or not at all
    • nm: must do
  • #9390 Warn if you're being a public relay but have too-low file descriptor limit
    • nm: should do
  • #9393 Broken weblink in Tor "running as a server, this probably won't work" message
    • nm: must do
Last modified 9 months ago Last modified on Aug 6, 2013 7:51:49 PM