1 This file lists minor work items (also possibly called "known bugs").
2 We are not tracking them in Mantis yet since there are too many and no
3 sane end-user should care about this codebase yet anyway.
7 - container_bloomfilter: improve efficiency (see FIXME)
9 - Windows: use events instead of pipes to signal select()s [Nils]
10 - only connect() sockets that are ready (select()) [Nils]
11 [On W32, we need to select after calling socket before
16 - transport_api: support forcing disconnects through low quotas!
17 (required for working F2F support!)
18 - API: consider having core provide deadline information for each message
19 (likely important for DV plugin which wants to loop back!)
20 - implement transport API to pretty-print transport address
21 + transport_api extension (API extension!)
22 + service-transport extension (protocol extension)
23 - add calls to statistics in various places
24 - implement gnunet-transport (transport configurator / tester)
25 - UPnP-based IP detection
26 (Note: build library always, build service when libxml2/etc. are available)
27 - instantly filter addresses from *other* peers that
28 are *equal* to our own address + port (i.e., localhost:2086). We
29 no longer filter those for outgoing (helps with loopback testing
30 and keeps the code clean), but we should filter strictly *impossible*
31 incoming addresses! This is for efficiency, not correctness.
32 - We currently are happy to take any address told to us in a WELCOME
33 to our set of addresses; we should have some minimal threshold-based
34 scheme, limiting both the total number of addresses that we accept
35 this way as well as requiring multiple confirmations; also, we
36 should possibly try to confirm that the given address works for
37 us ourselves (loopback-style) before adding it to the list
39 + we may be able to simplify WELCOME messages (no need to add
40 addresses there anymore, but may help to learn them there anyway...).
41 + we probably want some kind of voting/counting for learning IP addresses
42 (maybe including IP addresses in ads proportional to how often others
43 report them? we at least need some protection against >64k HELLOs!),
44 + provide a way to give the user a list of "learned" IP addresses and
45 a way to easily "veto" addresses off the list!
46 => If MiM attacker uses vetoed address, blacklist the specific IP for
47 the presumed neighbour!
48 - not sure current way of doing ACKs works well-enough
49 with unreliable transports where the ACK maybe lost;
50 the "is_new" check would then possibly prevent future
51 ACKs to be delivered, all while we're happily
52 receiving messages from that peer! Worse, the other
53 peer won't generate another ACK since it thinks we're
54 connected just fine...
56 + How necessary is ACKing in the first place? (alternatives?)
57 + Should we transmit ACKs in response to every HELLO? (would that
58 fully address the problem?)
59 - latency measurements implemented in the transport
60 plugins makes it only work for bi-di transports
61 and results in code replication
62 - should latency be included in the ReceiveCallback and
63 NotifyConnect or passed on request?
64 - FIXME's with latency being simply set to 0 in a few places
65 - Memory leak (running valgrind --trace-children=yes on test_transport_api:
66 ==28393== 16 bytes in 1 blocks are indirectly lost in loss record 1 of 5
67 ==28393== at 0x4C2260E: malloc (vg_replace_malloc.c:207)
68 ==28393== by 0x52343E3: GNUNET_xmalloc_unchecked_ (common_allocation.c:62)
69 ==28393== by 0x5234389: GNUNET_xmalloc_ (common_allocation.c:53)
70 ==28393== by 0x524458A: GNUNET_NETWORK_socket_create_from_accept (network.c:289)
71 ==28393== by 0x524B2DA: ??? (server.c:332)
72 ==28393== by 0x524A4C7: ??? (scheduler.c:425)
73 ==28393== by 0x524A73D: GNUNET_SCHEDULER_run (scheduler.c:510)
74 ==28393== by 0x524FF8C: GNUNET_SERVICE_run (service.c:1326)
75 ==28393== by 0x405500: main (gnunet-service-transport.c:2645)
77 ==28393== 65,744 (65,728 direct, 16 indirect) bytes in 1 blocks are definitely lost in loss record 5 of 5
78 ==28393== at 0x4C2260E: malloc (vg_replace_malloc.c:207)
79 ==28393== by 0x52343E3: GNUNET_xmalloc_unchecked_ (common_allocation.c:62)
80 ==28393== by 0x5234389: GNUNET_xmalloc_ (common_allocation.c:53)
81 ==28393== by 0x524473E: GNUNET_NETWORK_socket_create_from_accept (network.c:323)
82 (rest of trace identical)
88 - auto-generate "defaults.conf" using gnunet-setup from "config.scm"
89 - integrate all options into "config.scm"
90 - change config-file writing to exclude options set to default values
92 - implement exponential back-off for service restarts
93 - better tracking of which config changes actually need to cause process restarts by ARM.
94 - have way to specify dependencies between services (to manage ARM restarts better)
95 - client-API is inefficient since it opens a TCP connection per service that is started
96 (instead of re-using connections).
98 - code currently notifies clients about "encrypted" connections being up well before
99 we get the encrypted PONG; sometimes this may be OK (for topology killing
100 unwanted connnections), but of course not in general. I suspect we want
101 to signal on PONG and have topology hook directly into transport to
102 kill plaintext connections before they have a chance to become encrypted
103 (may require minor hack in transport API)
105 - have gnunet-peerinfo print actual host addresses again
106 - add option to gnunet-peerinfo to modify trust value
108 - finish postgres implementation; simplify other SQLs using new stats
113 - active reply route caching design & implementation of service,
116 - implement advertising of hostlist URL
117 - implement learning of hostlist URLs