aboutsummaryrefslogtreecommitdiff
path: root/TODO
diff options
context:
space:
mode:
authorChristian Grothoff <christian@grothoff.org>2010-04-17 18:22:03 +0000
committerChristian Grothoff <christian@grothoff.org>2010-04-17 18:22:03 +0000
commit6032eb391ed917368d05d697aa2e441e0908d8bb (patch)
tree588c26883d38c2af51b4de6b79910e29fb63f42d /TODO
parent1152a709e8783e8c0c879b7f407d26ea141a1e4e (diff)
downloadgnunet-6032eb391ed917368d05d697aa2e441e0908d8bb.tar.gz
gnunet-6032eb391ed917368d05d697aa2e441e0908d8bb.zip
prioritize
Diffstat (limited to 'TODO')
-rw-r--r--TODO27
1 files changed, 14 insertions, 13 deletions
diff --git a/TODO b/TODO
index f10447169..406eb4f64 100644
--- a/TODO
+++ b/TODO
@@ -2,10 +2,18 @@
2* PEERINFO: 2* PEERINFO:
3 - trust: need *fast* way to check/update trust in peers 3 - trust: need *fast* way to check/update trust in peers
4 (async peerinfo would not be right; certainly not with the current API) 4 (async peerinfo would not be right; certainly not with the current API)
5* UTIL:
6 - only connect() sockets that are ready (select()) [Nils]
7 [On W32, we need to select after calling socket before doing connect etc.]
8* TOPOLOGY: 5* TOPOLOGY:
6 - move code to use hash table instead of linked list
7 - check if new HELLO learned is different from old HELLO
8 before resetting entire state!
9 - the code uses the term 'blacklist' for both peers that are forbidden
10 to connect (i.e. F2F mode) as well as peers that we currently
11 won't try to actively connect to ourselves (since we just tried);
12 This is confusing. We need two distinct terms (greylist?).
13 - instead of periodically discarding blacklisted entries,
14 simply add task that is triggered at the right time (earlier free,
15 more balanced load)
16
9 - needs more testing (especially F2F topology) 17 - needs more testing (especially F2F topology)
10 - needs to re-try connecting after disconnect (currently, it 18 - needs to re-try connecting after disconnect (currently, it
11 initially triggers a connection request, but if that connection 19 initially triggers a connection request, but if that connection
@@ -25,16 +33,6 @@
25 a minor issue; OTOH, we might want to be more explicit about 33 a minor issue; OTOH, we might want to be more explicit about
26 allowing/forbidding connects on pre-connect to avoid 34 allowing/forbidding connects on pre-connect to avoid
27 entering connect attempts to just be blacklisted shortly afterwards). 35 entering connect attempts to just be blacklisted shortly afterwards).
28 - the code uses the term 'blacklist' for both peers that are forbidden
29 to connect (i.e. F2F mode) as well as peers that we currently
30 won't try to actively connect to ourselves (since we just tried);
31 This is confusing. We need two distinct terms (greylist?).
32 - move code to use hash table instead of linked list
33 - instead of periodically discarding blacklisted entries,
34 simply add task that is triggered at the right time (earlier free,
35 more balanced load)
36 - check if new HELLO learned is different from old HELLO
37 before resetting entire state!
38* DATASTORE: 36* DATASTORE:
39 - API lacks cancellation methods (needed? or is disconnect enough?) 37 - API lacks cancellation methods (needed? or is disconnect enough?)
40 - may also want to integrate request queuing here instead of gnunet-service-fs_drq.c 38 - may also want to integrate request queuing here instead of gnunet-service-fs_drq.c
@@ -66,6 +64,9 @@
66 - implement DV transport plugin 64 - implement DV transport plugin
67 - implement testcases 65 - implement testcases
68 - implement performance tests 66 - implement performance tests
67* UTIL:
68 - only connect() sockets that are ready (select()) [Nils]
69 [On W32, we need to select after calling socket before doing connect etc.]
69* GNUNET-GTK: 70* GNUNET-GTK:
70 - use g_main_context_set_poll_func to integrate GTK with GNUnet Scheduler!? (YUCK!) 71 - use g_main_context_set_poll_func to integrate GTK with GNUnet Scheduler!? (YUCK!)
71 - OR: add scheduler API to enable integration with GTK main loop instead of doing our own select 72 - OR: add scheduler API to enable integration with GTK main loop instead of doing our own select