diff options
author | Christian Grothoff <christian@grothoff.org> | 2010-04-17 21:16:41 +0000 |
---|---|---|
committer | Christian Grothoff <christian@grothoff.org> | 2010-04-17 21:16:41 +0000 |
commit | 967418d0b309a8efb8d552ced8359bfe2d444e45 (patch) | |
tree | aebb8415738205197b1d45a489118fab3cb8e5ba /TODO | |
parent | 3c8c69ae5071745d3d1e810001404f752cf559c3 (diff) | |
download | gnunet-967418d0b309a8efb8d552ced8359bfe2d444e45.tar.gz gnunet-967418d0b309a8efb8d552ced8359bfe2d444e45.zip |
proper HELLO merging
Diffstat (limited to 'TODO')
-rw-r--r-- | TODO | 21 |
1 files changed, 6 insertions, 15 deletions
@@ -1,17 +1,14 @@ | |||
1 | 0.9.0pre1: | 1 | 0.9.0pre1: |
2 | * PEERINFO: | 2 | * PEERINFO: [CG] |
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 | * TOPOLOGY: | 5 | * TOPOLOGY: [CG] |
6 | - check if new HELLO learned is different from old HELLO | ||
7 | before resetting entire state! | ||
8 | - needs more testing (especially F2F topology) | ||
9 | - needs to re-try connecting after disconnect (currently, it | 6 | - needs to re-try connecting after disconnect (currently, it |
10 | initially triggers a connection request, but if that connection | 7 | initially triggers a connection request, but if that connection |
11 | fails / goes down, it does not retry in a timely fashion; | 8 | fails / goes down, it does not retry in a timely fashion; |
12 | cause seems to be the 'blacklist_after_attempt' being set to 1h, | 9 | cause seems to be the 'greylist_after_attempt' being set to 1h, |
13 | which is rather long -- and should probably be adjusted based on | 10 | which is rather long -- and should probably be adjusted based on |
14 | the number of connections / known peers) | 11 | the number of connections / known peers & use some form of back-off) |
15 | - If the topology daemon crashes, peers that were put on the | 12 | - If the topology daemon crashes, peers that were put on the |
16 | blacklist with transport will never be removed from it (until | 13 | blacklist with transport will never be removed from it (until |
17 | transport service dies); we should use the blacklist notification | 14 | transport service dies); we should use the blacklist notification |
@@ -24,9 +21,7 @@ | |||
24 | a minor issue; OTOH, we might want to be more explicit about | 21 | a minor issue; OTOH, we might want to be more explicit about |
25 | allowing/forbidding connects on pre-connect to avoid | 22 | allowing/forbidding connects on pre-connect to avoid |
26 | entering connect attempts to just be blacklisted shortly afterwards). | 23 | entering connect attempts to just be blacklisted shortly afterwards). |
27 | * DATASTORE: | 24 | - needs more testing (especially F2F topology) |
28 | - API lacks cancellation methods (needed? or is disconnect enough?) | ||
29 | - may also want to integrate request queuing here instead of gnunet-service-fs_drq.c | ||
30 | * FS: [CG] | 25 | * FS: [CG] |
31 | - support recursive download even if filename is NULL and we hence | 26 | - support recursive download even if filename is NULL and we hence |
32 | do not generate files on disk (use temp_filename) | 27 | do not generate files on disk (use temp_filename) |
@@ -180,10 +175,6 @@ | |||
180 | Optimizations: | 175 | Optimizations: |
181 | * TCP: | 176 | * TCP: |
182 | - should use hash map to look up sessions | 177 | - should use hash map to look up sessions |
183 | * PEERINFO: | ||
184 | - api creates many, many short-lived TCP connections; either some | ||
185 | clients should use the API differently or we need to change the | ||
186 | API to enable re-use of connections to the service | ||
187 | * STATISTICS: | 178 | * STATISTICS: |
188 | - should use BIO instead of mmap | 179 | - should use BIO instead of mmap |
189 | * TRANSPORT: | 180 | * TRANSPORT: |
@@ -196,7 +187,7 @@ Optimizations: | |||
196 | - should use hash map to look up Neighbours | 187 | - should use hash map to look up Neighbours |
197 | * HOSTLIST: | 188 | * HOSTLIST: |
198 | - 'server' uses 'GNUNET_PEERINFO_iterate', should probably switch to notification API | 189 | - 'server' uses 'GNUNET_PEERINFO_iterate', should probably switch to notification API |
199 | (for more instant / up-to-date hostlists at lower cost) [OPTIMIZATION] | 190 | (for more instant / up-to-date hostlists at lower cost) |
200 | * DATASTORE (?): | 191 | * DATASTORE (?): |
201 | - check for duplicates on insertion (currently, same content is frequently | 192 | - check for duplicates on insertion (currently, same content is frequently |
202 | stored again [seen with KBLOCKS and SBLOCKS]!) | 193 | stored again [seen with KBLOCKS and SBLOCKS]!) |