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