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