11 - trust: need *fast* way to check/update trust in peers
12 (async peerinfo would not be right; certainly not with the
15 - scheduler should change OS process priority based on task priority;
16 should make better use of task priorities in general
17 - only connect() sockets that are ready (select()) [Nils]
18 [On W32, we need to select after calling socket before
21 - design & implement new setup tool
23 - good to have for transport/DV evaluation!
25 - write DV API (need to move declarations from dv_api.c to gnunet_dv_service.h!)
26 - implement DV service
27 - implement DV library (looks done)
28 - implement DV transport plugin
30 - implement performance tests
32 - does not seem to work with timeouts (especially if service is not running)
34 - needs more testing (especially F2F topology)
35 - needs to re-try connecting after disconnect (currently, it
36 initially triggers a connection request, but if that connection
37 fails / goes down, it does not retry in a timely fashion;
38 cause seems to be the 'blacklist_after_attempt' being set to 1h,
39 which is rather long -- and should probably be adjusted based on
40 the number of connections / known peers)
41 - If the topology daemon crashes, peers that were put on the
42 blacklist with transport will never be removed from it (until
43 transport service dies); we should use the blacklist notification
44 API to learn about the exact set of blacklisted peers at all times
45 (FIXME: the transport_api implementation of blacklisting
46 also does not work nicely for this since it won't let us know about
47 disconnect-reconnect events and the implicit whitelisting
48 that might happen here; that's not so bad since we will
49 re-blacklist on pre-connect attempts anyway, so this is
50 a minor issue; OTOH, we might want to be more explicit about
51 allowing/forbidding connects on pre-connect to avoid
52 entering connect attempts to just be blacklisted shortly afterwards).
53 - the code uses the term 'blacklist' for both peers that are forbidden
54 to connect (i.e. F2F mode) as well as peers that we currently
55 won't try to actively connect to ourselves (since we just tried);
56 This is confusing. We need two distinct terms (greylist?).
57 - move code to use hash table instead of linked list
58 - instead of periodically discarding blacklisted entries,
59 simply add task that is triggered at the right time (earlier free,
61 - check if new HELLO learned is different from old HELLO
62 before resetting entire state!
64 - support recursive download even if filename is NULL and we hence
65 do not generate files on disk (use temp_filename)
66 - bound parallelism (# fs downloads)
67 - distinguish in performance tracking and event signalling between
68 downloads that are actually running and those that are merely in the queue
69 - persistence support (publish, unindex, search, download)
70 - active migration support (in fs or in datastore or new daemon?)
71 - gnunet-service-fs (hot-path routing, load-based routing, nitpicks)
72 - [gnunet-service-fs.c:208]: member 'LocalGetContext::results_bf_size' is never used
73 - [gnunet-service-fs.c:501]: member 'PendingRequest::used_pids_size' is never used
74 - [gnunet-service-fs.c:654]: member 'ConnectedPeer::last_client_replies' is never used
75 - [gnunet-service-fs.c:669]: member 'ConnectedPeer::avg_delay' is never used
76 - [gnunet-service-fs.c:675]: member 'ConnectedPeer::avg_priority' is never used
77 - [gnunet-service-fs.c:688]: member 'ConnectedPeer::pending_requests' is never used
78 - [gnunet-service-fs.c:694]: member 'ConnectedPeer::last_p2p_replies_woff' is never used
79 - [gnunet-service-fs.c:700]: member 'ConnectedPeer::last_client_replies_woff' is never used
81 + active reply route caching design & implementation of service; gap extension!
83 - API lacks cancellation methods (needed? or is disconnect enough?);
84 may also want to integrate request queuing here instead of gnunet-service-fs_drq.c
86 - use g_main_context_set_poll_func to integrate GTK with GNUnet Scheduler!? (YUCK!)
87 - OR: add scheduler API to enable integration with GTK main loop instead of doing our own select
88 - use g_main_context_pending, g_main_context_query / g_main_context_check / g_main_context_dispatch
89 and NEVER g_main_loop_run (can this be done? might be the clean way to do this! But how
90 to integrate this with "tak_main"? Docu says:
91 "It's OK to use the GLib main loop directly instead of gtk_main(), though it involves
92 slightly more typing. See GMainLoop in the GLib documentation."
93 => so maybe it "just works"?
97 - good to have for DV/DHT evaluation!
99 - implement DHT service
100 - implement DHT library
101 - implement testcases
102 - implement performance tests
106 - need to get rid of synchronous API for service starts (cause all kinds of problems)
107 [=> eliminate for need to tell ARM about service starts most of the time!] [Safey]
108 - better tracking of which config changes actually need to cause process restarts by ARM.
109 - listen for requests to discover dependencies between services (and avoid
110 having to explicitly program start requests)
111 - better crash management (attach debugging support, capture and analyze
112 debug output, detect random vs. deterministic crashes)
115 - datastore reservation (publishing)
116 - search: availability probes
117 - location URIs (publish, search, download)
118 - non-anonymous FS service (needs DHT)
119 + DHT integration for search
120 + CS-DHT-functions (DHT-put of LOC)
121 + P2P-functions (DHT-get)
122 - collection API & tests
123 + gnunet-pseudonym (collection support)
124 - implement FS performance tests
131 * Determine RC bugs and fix those!
133 - modify configuration to allow controlling connections for non-local starts
134 - testbed creation with topology (needs working F2F topology)
136 - implement testcases for distributed testing
137 - test basic peer re-configure
138 - test topology creation
139 - test churn generation
140 - consider changing API for peer-group termination to
141 call continuation when done
143 - finalize API design
146 - integration with transport service
147 * MYSQL database backends: [CG]
153 - expand bibliography
154 - convert documentation pages to books
155 - update books (especially for developers)
156 - create good Drupal theme for GNUnet
157 - make a NICE download page and figure out how to
158 enable developers to publish TGZs nicely
159 - port "contact" page
160 - add content type for "todo" items?
161 * POSTGRES database backends: [CG]
164 * Determine RC bugs and fix those!
168 - SMTP transport backend
169 - HTTPS transport backend
170 + improved HTTPS support in MHD
172 - Implement method of learning our external addresses from
173 other peers; need some kind of threshold-based
174 scheme, limiting both the total number of addresses that we accept
175 this way as well as requiring multiple confirmations; also, we
176 should possibly try to confirm that the given address works for
177 us ourselves (loopback-style) before adding it to the list
178 + we may be able to simplify WELCOME messages (no need to add
179 addresses there anymore, but may help to learn them there anyway...).
180 + we probably want some kind of voting/counting for learning IP addresses
181 (maybe including IP addresses in ads proportional to how often others
182 report them? we at least need some protection against >64k HELLOs!),
183 + provide a way to give the user a list of "learned" IP addresses and
184 a way to easily "veto" addresses off the list!
185 => If MiM attacker uses vetoed address, blacklist the specific IP for
186 the presumed neighbour!
187 - implement gnunet-transport (transport configurator / tester)
188 - UPnP-based IP detection
189 (Note: build library always, build service when libxml2/etc. are available)
191 - Remove KBlocks in gnunet-unindex (see discussion with Kenneth Almquist on gnunet-devs in 9/2009)
192 - Allow checking of presence of search results and/or content via command-line tools
193 (add options to gnunet-search / gnunet-download to limit search to local peer)
195 - expire 'ancient' HELLOs (those without valid addresses AND that
196 we have not 'used' (for their public keys) in a while; need a way
197 to track actual 'use')
198 - make sue we also trigger notifications whenever HELLOs expire
205 - should use hash map to look up sessions
207 - api creates many, many short-lived TCP connections; either some
208 clients should use the API differently or we need to change the
209 API to enable re-use of connections to the service
211 - should use BIO instead of mmap
213 - need to periodically probe latency/transport cost changes & possibly switch transport
214 - instantly filter addresses from *other* peers that
215 are *equal* to our own address + port (i.e., localhost:2086). We
216 no longer filter those for outgoing (helps with loopback testing
217 and keeps the code clean), but we should filter strictly *impossible*
218 incoming addresses! This is for efficiency, not correctness.
219 - should use hash map to look up Neighbours
221 - 'server' uses 'GNUNET_PEERINFO_iterate', should probably switch to notification API
222 (for more instant / up-to-date hostlists at lower cost) [OPTIMIZATION]
224 - check for duplicates on insertion (currently, same content is frequently
225 stored again [seen with KBLOCKS and SBLOCKS]!)
229 - repeatedly resolve hostname and look up interfaces to determine our own IP
230 - [./transport/plugin_transport_tcp.c:391]: (style) struct or union member 'Plugin::address_update_task' is never used (related to issue above)
232 - [./transport/gnunet-service-transport.c:173]: (style) struct or union member 'TransportPlugin::rebuild' is never used (related to TCP not refreshing external addresses?)
234 - add stats (# bytes available, # bytes used, # PUTs, # GETs, # GETs satisfied)
236 - support inline data in directories for recursive file downloads (fs_download)