1 PHASE #1: (Goal: settle key design questions)
5 * improve disk API [Nils] (Nils, is this done? -Christian)
6 * Windows: use events instead of pipes to signal select()s [Nils]
7 * only connect() sockets that are ready (select()) [Nils]
8 [On W32, we need to select after calling socket before
12 For these functions, it would be nice if we had testcases ("make check")
13 that would cause them to be executed and check that they are working:
14 * gnunet-service-peerinfo:
15 - change_host_trust / flush_trust
17 - discard_hosts_helper / cron_clean_data_hosts
18 * gnunet-service-transport:
19 - try_unvalidated_addresses
20 - lookup_address_callback
21 - lookup_hello_callback
22 - plugin_env_lookup_address
23 - notify_clients_disconnect
24 - list_validated_addresses
26 - disconnect_neighbour
28 * plugin_transport_tcp.c:
30 - tcp_plugin_address_pretty_printer / append_port
31 - tcp_plugin_set_receive_quota
34 - GNUNET_TRANSPORT_set_qutoa / send_set_quota
38 - remove_from_any_list / remove_neighbour
39 - GNUNET_TRANSPORT_notify_transmit_ready_cancel
42 - solicit_traffic / copy_and_free
43 - GNUNET_CORE_peer_configure / produce_configure_message
44 * gnunet-service-core:
47 - handle_client_request_configure
50 - handle_transport_notify_disconnect
51 * hostlist (everything)
52 * topology (everything)
56 PHASE #2: (Goal: recover basic file-sharing functionality)
59 * implement sqlite-based sqstore/datastore service
60 + implement service (datastore + talks to plugin)
61 + implement library (talks to service)
62 + implement sqlite plugin (talks to DB)
63 + fix testcases (make them use CPS, complete their inner workings...)
64 * implement sqlite-based dstore services
66 + implement library (talks to service)
67 + implement service (talks to plugin)
68 + implement sqlite plugin (talks to DB)
76 PHASE #3: (Goal: ready for pre-release) [completion-goal: end of 2009]
86 * UDP backend (need LIBRARY to support (de)fragmentation!)
89 + Testcases for set_quota, timeouts, disconnects, transmit_ready_cancel
90 + Better coverage of gnunet-service-transport (HELLO validation)
91 + direct test of plugins compliance to plugin API
94 * sqlite, mysql & postgres backend
100 - crypto_hash: use libgcrypt (supports SHA-512 since 2003)
101 - container_bloomfilter: improve efficiency (see FIXME)
103 - inefficient memmove
105 - transport_api: support forcing disconnects through low quotas!
106 (required for working F2F support!)
107 - API: consider having core provide deadline information for each message
108 (likely important for DV plugin which wants to loop back!)
109 - implement transport API to pretty-print transport address
110 + transport_api extension (API extension!)
111 + service-transport extension (protocol extension)
112 - add calls to statistics in various places
113 - implement gnunet-transport (transport configurator / tester)
114 - UPnP-based IP detection
115 (Note: build library always, build service when libxml2/etc. are available)
116 - instantly filter addresses from *other* peers that
117 are *equal* to our own address + port (i.e., localhost:2086). We
118 no longer filter those for outgoing (helps with loopback testing
119 and keeps the code clean), but we should filter strictly *impossible*
120 incoming addresses! This is for efficiency, not correctness.
121 - We currently are happy to take any address told to us in a WELCOME
122 to our set of addresses; we should have some minimal threshold-based
123 scheme, limiting both the total number of addresses that we accept
124 this way as well as requiring multiple confirmations; also, we
125 should possibly try to confirm that the given address works for
126 us ourselves (loopback-style) before adding it to the list
128 + we may be able to simplify WELCOME messages (no need to add
129 addresses there anymore, but may help to learn them there anyway...).
130 + we probably want some kind of voting/counting for learning IP addresses
131 (maybe including IP addresses in ads proportional to how often others
132 report them? we at least need some protection against >64k HELLOs!),
133 + provide a way to give the user a list of "learned" IP addresses and
134 a way to easily "veto" addresses off the list!
135 => If MiM attacker uses vetoed address, blacklist the specific IP for
136 the presumed neighbour!
137 - not sure current way of doing ACKs works well-enough
138 with unreliable transports where the ACK maybe lost;
139 the "is_new" check would then possibly prevent future
140 ACKs to be delivered, all while we're happily
141 receiving messages from that peer! Worse, the other
142 peer won't generate another ACK since it thinks we're
143 connected just fine...
145 + How necessary is ACKing in the first place? (alternatives?)
146 + Should we transmit ACKs in response to every HELLO? (would that
147 fully address the problem?)
148 - latency measurements implemented in the transport
149 plugins makes it only work for bi-di transports
150 and results in code replication
151 - should latency be included in the ReceiveCallback and
152 NotifyConnect or passed on request?
153 - FIXME's with latency being simply set to 0 in a few places
154 - Memory leak (running valgrind --trace-children=yes on test_transport_api:
155 ==28393== 16 bytes in 1 blocks are indirectly lost in loss record 1 of 5
156 ==28393== at 0x4C2260E: malloc (vg_replace_malloc.c:207)
157 ==28393== by 0x52343E3: GNUNET_xmalloc_unchecked_ (common_allocation.c:62)
158 ==28393== by 0x5234389: GNUNET_xmalloc_ (common_allocation.c:53)
159 ==28393== by 0x524458A: GNUNET_NETWORK_socket_create_from_accept (network.c:289)
160 ==28393== by 0x524B2DA: ??? (server.c:332)
161 ==28393== by 0x524A4C7: ??? (scheduler.c:425)
162 ==28393== by 0x524A73D: GNUNET_SCHEDULER_run (scheduler.c:510)
163 ==28393== by 0x524FF8C: GNUNET_SERVICE_run (service.c:1326)
164 ==28393== by 0x405500: main (gnunet-service-transport.c:2645)
166 ==28393== 65,744 (65,728 direct, 16 indirect) bytes in 1 blocks are definitely lost in loss record 5 of 5
167 ==28393== at 0x4C2260E: malloc (vg_replace_malloc.c:207)
168 ==28393== by 0x52343E3: GNUNET_xmalloc_unchecked_ (common_allocation.c:62)
169 ==28393== by 0x5234389: GNUNET_xmalloc_ (common_allocation.c:53)
170 ==28393== by 0x524473E: GNUNET_NETWORK_socket_create_from_accept (network.c:323)
171 (rest of trace identical)
172 => SVN 8727 may fix this; should test it again!
175 - auto-generate "defaults.conf" using gnunet-setup from "config.scm"
176 - integrate all options into "config.scm"
177 - change config-file writing to exclude options set to default values
179 - implement exponential back-off for service restarts
180 - better tracking of which config changes actually need to cause process restarts by ARM.
181 - have way to specify dependencies between services (to manage ARM restarts better)
182 - client-API is inefficient since it opens a TCP connection per service that is started
183 (instead of re-using connections).
185 - code currently notifies clients about "encrypted" connections being up well before
186 we get the encrypted PONG; sometimes this may be OK (for topology killing
187 unwanted connnections), but of course not in general. I suspect we want
188 to signal on PONG and have topology hook directly into transport to
189 kill plaintext connections before they have a chance to become encrypted
190 (may require minor hack in transport API)
192 - have gnunet-peerinfo print actual host addresses again
193 - add option to gnunet-peerinfo to modify trust value
195 - finish postgres implementation; simplify other SQLs using new stats
200 - active reply route caching design & implementation of service,
203 - implement advertising of hostlist URL
204 - implement learning of hostlist URLs
210 PHASE #4: [completion-goal: mid 2010]
218 * Features eliminated from util:
219 - threading (goal: good riddance!)
220 - complex logging features [ectx-passing, target-kinds] (goal: good riddance!)
221 - complex configuration features [defaults, notifications] (goal: good riddance!)
222 - network traffic monitors (goal: eliminate)
223 - IPC semaphores (goal: d-bus? / eliminate?)
225 - DNS lookup (goal: have async service; issue: still need synchronous resolution in places, current code may not be portable)
226 => code shrunk from 61 files to 34, 22k LOC to 15k LOC, 470k to 330k (with symbols)
227 * New features in util:
229 - service and program boot-strap code
230 * Major changes in util:
231 - more expressive server (replaces selector)
233 - how to integrate scheduler with GTK event loop!