1 PHASE #1: (Goal: settle key design questions)
4 * improve disk API [Nils] (Nils, is this done? -Christian)
5 * Windows: use events instead of pipes to signal select()s [Nils]
6 * only connect() sockets that are ready (select()) [Nils]
7 [On W32, we need to select after calling socket before
11 For these functions, it would be nice if we had testcases ("make check")
12 that would cause them to be executed and check that they are working:
13 * gnunet-service-peerinfo:
14 - change_host_trust / flush_trust
16 - discard_hosts_helper / cron_clean_data_hosts
17 * gnunet-service-transport:
18 - try_unvalidated_addresses
19 - lookup_address_callback
20 - lookup_hello_callback
21 - plugin_env_lookup_address
22 - notify_clients_disconnect
23 - list_validated_addresses
25 - disconnect_neighbour
27 * plugin_transport_tcp.c:
29 - tcp_plugin_address_pretty_printer / append_port
30 - tcp_plugin_set_receive_quota
33 - GNUNET_TRANSPORT_set_qutoa / send_set_quota
37 - remove_from_any_list / remove_neighbour
38 - GNUNET_TRANSPORT_notify_transmit_ready_cancel
41 - solicit_traffic / copy_and_free
42 - GNUNET_CORE_peer_configure / produce_configure_message
43 * gnunet-service-core:
46 - handle_client_request_configure
49 - handle_transport_notify_disconnect
50 * hostlist (everything)
51 * topology (everything)
55 PHASE #2: (Goal: recover basic file-sharing functionality)
58 * implement sqlite-based sqstore/datastore service
59 + implement service (datastore + talks to plugin)
60 + implement library (talks to service)
61 + implement sqlite plugin (talks to DB)
62 + fix testcases (make them use CPS, complete their inner workings...)
63 * implement sqlite-based dstore services
65 + implement library (talks to service)
66 + implement service (talks to plugin)
67 + implement sqlite plugin (talks to DB)
75 PHASE #3: (Goal: ready for pre-release) [completion-goal: end of 2009]
85 * UDP backend (need LIBRARY to support (de)fragmentation!)
88 + Testcases for set_quota, timeouts, disconnects, transmit_ready_cancel
89 + Better coverage of gnunet-service-transport (HELLO validation)
90 + direct test of plugins compliance to plugin API
93 * sqlite, mysql & postgres backend
99 - crypto_hash: use libgcrypt (supports SHA-512 since 2003)
100 - container_bloomfilter: improve efficiency (see FIXME)
102 - inefficient memmove
104 - transport_api: support forcing disconnects through low quotas!
105 (required for working F2F support!)
106 - API: consider having core provide deadline information for each message
107 (likely important for DV plugin which wants to loop back!)
108 - implement transport API to pretty-print transport address
109 + transport_api extension (API extension!)
110 + service-transport extension (protocol extension)
111 - add calls to statistics in various places
112 - implement gnunet-transport (transport configurator / tester)
113 - UPnP-based IP detection
114 (Note: build library always, build service when libxml2/etc. are available)
115 - instantly filter addresses from *other* peers that
116 are *equal* to our own address + port (i.e., localhost:2086). We
117 no longer filter those for outgoing (helps with loopback testing
118 and keeps the code clean), but we should filter strictly *impossible*
119 incoming addresses! This is for efficiency, not correctness.
120 - We currently are happy to take any address told to us in a WELCOME
121 to our set of addresses; we should have some minimal threshold-based
122 scheme, limiting both the total number of addresses that we accept
123 this way as well as requiring multiple confirmations; also, we
124 should possibly try to confirm that the given address works for
125 us ourselves (loopback-style) before adding it to the list
127 + we may be able to simplify WELCOME messages (no need to add
128 addresses there anymore, but may help to learn them there anyway...).
129 + we probably want some kind of voting/counting for learning IP addresses
130 (maybe including IP addresses in ads proportional to how often others
131 report them? we at least need some protection against >64k HELLOs!),
132 + provide a way to give the user a list of "learned" IP addresses and
133 a way to easily "veto" addresses off the list!
134 => If MiM attacker uses vetoed address, blacklist the specific IP for
135 the presumed neighbour!
136 - not sure current way of doing ACKs works well-enough
137 with unreliable transports where the ACK maybe lost;
138 the "is_new" check would then possibly prevent future
139 ACKs to be delivered, all while we're happily
140 receiving messages from that peer! Worse, the other
141 peer won't generate another ACK since it thinks we're
142 connected just fine...
144 + How necessary is ACKing in the first place? (alternatives?)
145 + Should we transmit ACKs in response to every HELLO? (would that
146 fully address the problem?)
147 - latency measurements implemented in the transport
148 plugins makes it only work for bi-di transports
149 and results in code replication
150 - should latency be included in the ReceiveCallback and
151 NotifyConnect or passed on request?
152 - FIXME's with latency being simply set to 0 in a few places
154 - auto-generate "defaults.conf" using gnunet-setup from "config.scm"
155 - integrate all options into "config.scm"
156 - change config-file writing to exclude options set to default values
158 - implement exponential back-off for service restarts
159 - better tracking of which config changes actually need to cause process restarts by ARM.
160 - have way to specify dependencies between services (to manage ARM restarts better)
161 - client-API is inefficient since it opens a TCP connection per service that is started
162 (instead of re-using connections).
164 - code currently notifies clients about "encrypted" connections being up well before
165 we get the encrypted PONG; sometimes this may be OK (for topology killing
166 unwanted connnections), but of course not in general. I suspect we want
167 to signal on PONG and have topology hook directly into transport to
168 kill plaintext connections before they have a chance to become encrypted
169 (may require minor hack in transport API)
171 - have gnunet-peerinfo print actual host addresses again
172 - add option to gnunet-peerinfo to modify trust value
174 - finish postgres implementation; simplify other SQLs using new stats
179 - active reply route caching design & implementation of service,
182 - implement advertising of hostlist URL
183 - implement learning of hostlist URLs
189 PHASE #4: [completion-goal: mid 2010]
197 * Features eliminated from util:
198 - threading (goal: good riddance!)
199 - complex logging features [ectx-passing, target-kinds] (goal: good riddance!)
200 - complex configuration features [defaults, notifications] (goal: good riddance!)
201 - network traffic monitors (goal: eliminate)
202 - IPC semaphores (goal: d-bus? / eliminate?)
204 - DNS lookup (goal: have async service; issue: still need synchronous resolution in places, current code may not be portable)
205 => code shrunk from 61 files to 34, 22k LOC to 15k LOC, 470k to 330k (with symbols)
206 * New features in util:
208 - service and program boot-strap code
209 * Major changes in util:
210 - more expressive server (replaces selector)
212 - how to integrate scheduler with GTK event loop!