1 PHASE #1: (Goal: settle key design questions)
4 * hostlist server (MHD-side)
5 * hostlist client (CURL-side); monitoring of number of active connections (to establish need for bootstrapping)
6 * hostlist server URL advertising & learning via P2P
9 * Selecting peers from peerinfo for connects; blacklisting
10 * Managing connections, F2F configuration obedience, rejecting prohibited connections
11 * Forwarding of known HELLOs to neighbours (advertising)
12 [ Inbound HELLOs are processed by transport, right?
13 But what about inbound encrypted HELLOs? ]
16 * Make sure ARM supports daemons (non-service background processes, see hostlist/topology)
19 * improve disk API [Nils]
20 * Windows: use events instead of pipes to signal select()s [Nils]
23 For these functions, it would be nice if we had testcases ("make check")
24 that would cause them to be executed and check that they are working:
25 * gnunet-service-peerinfo:
26 - change_host_trust / flush_trust
28 - discard_hosts_helper / cron_clean_data_hosts
29 * gnunet-service-transport:
30 - try_unvalidated_addresses
31 - lookup_address_callback
32 - lookup_hello_callback
33 - plugin_env_lookup_address
34 - notify_clients_disconnect
35 - list_validated_addresses
37 - disconnect_neighbour
39 * plugin_transport_tcp.c:
41 - tcp_plugin_address_pretty_printer / append_port
42 - tcp_plugin_set_receive_quota
45 - GNUNET_TRANSPORT_set_qutoa / send_set_quota
49 - remove_from_any_list / remove_neighbour
50 - GNUNET_TRANSPORT_notify_transmit_ready_cancel
53 - solicit_traffic / copy_and_free
54 - GNUNET_CORE_peer_configure / produce_configure_message
55 * gnunet-service-core:
58 - handle_client_request_configure
61 - handle_transport_notify_disconnect
64 PHASE #2: (Goal: recover basic file-sharing functionality)
67 * implement sqlite-based sqstore/datastore service
68 + implement library (talks to service)
69 + implement service (datastore + talks to plugin)
70 + implement sqlite plugin (talks to DB)
71 * implement sqlite-based dstore services
72 + implement library (talks to service)
73 + implement service (talks to plugin)
74 + implement sqlite plugin (talks to DB)
77 * implement hostlist service (need to bootstrap!)
82 * have ONE backend working
85 PHASE #3: (Goal: ready for pre-release) [completion-goal: end of 2009]
95 * UDP backend (need LIBRARY to support (de)fragmentation!)
98 + Testcases for set_quota, timeouts, disconnects, transmit_ready_cancel
99 + Better coverage of gnunet-service-transport (hello validation)
100 + direct test of plugins compliance to plugin API
103 * sqlite, mysql & postgres backend
109 - inefficient memmove
111 - transport_api: support forcing disconnects through low quotas!
112 - API: consider having core provide deadline information for each message
113 (likely important for DV plugin which wants to loop back!)
114 - implement transport API to pretty-print transport address
115 + transport_api extension (API extension!)
116 + service-transport extension (protocol extension)
117 - add calls to statistics in various places
118 - implement gnunet-transport (transport configurator / tester)
119 - UPnP-based IP detection
120 (Note: build library always, build service when libxml2/etc. are available)
121 - instantly filter addresses from *other* peers that
122 are *equal* to our own address + port (i.e., localhost:2086). We
123 no longer filter those for outgoing (helps with loopback testing
124 and keeps the code clean), but we should filter strictly *impossible*
125 incoming addresses! This is for efficiency, not correctness.
126 - We currently are happy to take any address told to us in a WELCOME
127 to our set of addresses; we should have some minimal threshold-based
128 scheme, limiting both the total number of addresses that we accept
129 this way as well as requiring multiple confirmations; also, we
130 should possibly try to confirm that the given address works for
131 us ourselves (loopback-style) before adding it to the list
133 - not sure current way of doing ACKs works well-enough
134 with unreliable transports where the ACK maybe lost;
135 the "is_new" check would then possibly prevent future
136 ACKs to be delivered, all while we're happily
137 receiving messages from that peer! Worse, the other
138 peer won't generate another ACK since it thinks we're
139 connected just fine...
141 + How necessary is ACKing in the first place? (alternatives?)
142 + Should we transmit ACKs in response to every HELLO? (would that
143 fully address the problem?)
144 - latency measurements implemented in the transport
145 plugins makes it only work for bi-di transports
146 and results in code replication
147 - should latency be included in the ReceiveCallback and
148 NotifyConnect or passed on request?
150 - auto-generate "defaults.conf" using gnunet-setup from "config.scm"
151 - integrate all options into "config.scm"
152 - change config-file writing to exclude options set to default values
154 - implement exponential back-off for service restarts
155 - better tracking of which config changes actually need to cause process restarts by ARM.
156 - have way to specify dependencies between services (to manage ARM restarts better)
158 - have gnunet-peerinfo print actual host addresses again
159 - add option to gnunet-peerinfo to modify trust value
161 - finish postgres implementation; simplify other SQLs using new stats
166 - active reply route caching design & implementation of service,
171 PHASE #4: [completion-goal: mid 2010]
179 * Features eliminated from util:
180 - threading (goal: good riddance!)
181 - complex logging features [ectx-passing, target-kinds] (goal: good riddance!)
182 - complex configuration features [defaults, notifications] (goal: good riddance!)
183 - network traffic monitors (goal: eliminate)
184 - IPC semaphores (goal: d-bus? / eliminate?)
186 - DNS lookup (goal: have async service; issue: still need synchronous resolution in places, current code may not be portable)
187 => code shrunk from 61 files to 34, 22k LOC to 15k LOC, 470k to 330k (with symbols)
188 * New features in util:
190 - service and program boot-strap code
191 * Major changes in util:
192 - more expressive server (replaces selector)
194 - how to integrate scheduler with GTK event loop!
199 * UTIL: 75%, 4914 out of 6463