1 PHASE #1: (Goal: settle key design questions)
4 * testcase fails the first time when run with fresh /tmp
5 1) CONNECT fails because no validated HELLO is available
6 (actual validations are scheduled a few cycles later)
7 2) client waits for SEND_OK, which never comes (due to
8 CONNECT request failing; TCP signals error, but
9 "transmit_send_continuation" in gnunet-service-transport.c
10 does not (and currently has no way to) signal this to
11 the test-process => test-process eventually times out with
13 3) in the meantime, gnunet-service-transport validates the
14 HELLOs => testcase passes next time...
17 * improve disk API [Nils] (Nils, is this done? -Christian)
18 * Windows: use events instead of pipes to signal select()s [Nils]
19 * only connect() sockets that are ready (select()) [Nils]
20 [On W32, we need to select after calling socket before
24 For these functions, it would be nice if we had testcases ("make check")
25 that would cause them to be executed and check that they are working:
26 * gnunet-service-peerinfo:
27 - change_host_trust / flush_trust
29 - discard_hosts_helper / cron_clean_data_hosts
30 * gnunet-service-transport:
31 - try_unvalidated_addresses
32 - lookup_address_callback
33 - lookup_hello_callback
34 - plugin_env_lookup_address
35 - notify_clients_disconnect
36 - list_validated_addresses
38 - disconnect_neighbour
40 * plugin_transport_tcp.c:
42 - tcp_plugin_address_pretty_printer / append_port
43 - tcp_plugin_set_receive_quota
46 - GNUNET_TRANSPORT_set_qutoa / send_set_quota
50 - remove_from_any_list / remove_neighbour
51 - GNUNET_TRANSPORT_notify_transmit_ready_cancel
54 - solicit_traffic / copy_and_free
55 - GNUNET_CORE_peer_configure / produce_configure_message
56 * gnunet-service-core:
59 - handle_client_request_configure
62 - handle_transport_notify_disconnect
63 * hostlist (everything)
64 * topology (everything)
68 PHASE #2: (Goal: recover basic file-sharing functionality)
71 * implement sqlite-based sqstore/datastore service
72 + implement service (datastore + talks to plugin)
73 + implement library (talks to service)
74 + implement sqlite plugin (talks to DB)
75 + fix testcases (make them use CPS, complete their inner workings...)
76 * implement sqlite-based dstore services
78 + implement library (talks to service)
79 + implement service (talks to plugin)
80 + implement sqlite plugin (talks to DB)
88 PHASE #3: (Goal: ready for pre-release) [completion-goal: end of 2009]
98 * UDP backend (need LIBRARY to support (de)fragmentation!)
101 + Testcases for set_quota, timeouts, disconnects, transmit_ready_cancel
102 + Better coverage of gnunet-service-transport (hello validation)
103 + direct test of plugins compliance to plugin API
106 * sqlite, mysql & postgres backend
112 - crypto_hash: use libgcrypt (supports SHA-512 since 2003)
113 - container_bloomfilter: improve efficiency (see FIXME)
115 - inefficient memmove
117 - transport_api: support forcing disconnects through low quotas!
118 (required for working F2F support!)
119 - API: consider having core provide deadline information for each message
120 (likely important for DV plugin which wants to loop back!)
121 - implement transport API to pretty-print transport address
122 + transport_api extension (API extension!)
123 + service-transport extension (protocol extension)
124 - add calls to statistics in various places
125 - implement gnunet-transport (transport configurator / tester)
126 - UPnP-based IP detection
127 (Note: build library always, build service when libxml2/etc. are available)
128 - instantly filter addresses from *other* peers that
129 are *equal* to our own address + port (i.e., localhost:2086). We
130 no longer filter those for outgoing (helps with loopback testing
131 and keeps the code clean), but we should filter strictly *impossible*
132 incoming addresses! This is for efficiency, not correctness.
133 - We currently are happy to take any address told to us in a WELCOME
134 to our set of addresses; we should have some minimal threshold-based
135 scheme, limiting both the total number of addresses that we accept
136 this way as well as requiring multiple confirmations; also, we
137 should possibly try to confirm that the given address works for
138 us ourselves (loopback-style) before adding it to the list
140 + we may be able to simplify WELCOME messages (no need to add
141 addresses there anymore, but may help to learn them there anyway...).
142 + we probably want some kind of voting/counting for learning IP addresses
143 (maybe including IP addresses in ads proportional to how often others
144 report them? we at least need some protection against >64k HELLOs!),
145 + provide a way to give the user a list of "learned" IP addresses and
146 a way to easily "veto" addresses off the list!
147 => If MiM attacker uses vetoed address, blacklist the specific IP for
148 the presumed neighbour!
149 - not sure current way of doing ACKs works well-enough
150 with unreliable transports where the ACK maybe lost;
151 the "is_new" check would then possibly prevent future
152 ACKs to be delivered, all while we're happily
153 receiving messages from that peer! Worse, the other
154 peer won't generate another ACK since it thinks we're
155 connected just fine...
157 + How necessary is ACKing in the first place? (alternatives?)
158 + Should we transmit ACKs in response to every HELLO? (would that
159 fully address the problem?)
160 - latency measurements implemented in the transport
161 plugins makes it only work for bi-di transports
162 and results in code replication
163 - should latency be included in the ReceiveCallback and
164 NotifyConnect or passed on request?
166 - auto-generate "defaults.conf" using gnunet-setup from "config.scm"
167 - integrate all options into "config.scm"
168 - change config-file writing to exclude options set to default values
170 - implement exponential back-off for service restarts
171 - better tracking of which config changes actually need to cause process restarts by ARM.
172 - have way to specify dependencies between services (to manage ARM restarts better)
173 - client-API is inefficient since it opens a TCP connection per service that is started
174 (instead of re-using connections).
176 - code currently notifies clients about "encrypted" connections being up well before
177 we get the encrypted PONG; sometimes this may be OK (for topology killing
178 unwanted connnections), but of course not in general. I suspect we want
179 to signal on PONG and have topology hook directly into transport to
180 kill plaintext connections before they have a chance to become encrypted
181 (may require minor hack in transport API)
183 - have gnunet-peerinfo print actual host addresses again
184 - add option to gnunet-peerinfo to modify trust value
186 - finish postgres implementation; simplify other SQLs using new stats
191 - active reply route caching design & implementation of service,
194 - implement advertising of hostlist URL
195 - implement learning of hostlist URLs
201 PHASE #4: [completion-goal: mid 2010]
209 * Features eliminated from util:
210 - threading (goal: good riddance!)
211 - complex logging features [ectx-passing, target-kinds] (goal: good riddance!)
212 - complex configuration features [defaults, notifications] (goal: good riddance!)
213 - network traffic monitors (goal: eliminate)
214 - IPC semaphores (goal: d-bus? / eliminate?)
216 - DNS lookup (goal: have async service; issue: still need synchronous resolution in places, current code may not be portable)
217 => code shrunk from 61 files to 34, 22k LOC to 15k LOC, 470k to 330k (with symbols)
218 * New features in util:
220 - service and program boot-strap code
221 * Major changes in util:
222 - more expressive server (replaces selector)
224 - how to integrate scheduler with GTK event loop!