54e483647c
scale down the DHT simulation even more
2015-08-13 23:07:32 -04:00
225b82d9a0
move test_trackers_extensions over to a simulation and some minor metadata extension and tracker re-announce fixes
2015-08-13 23:07:04 -04:00
2c89fa40e2
fix up i2p merge
2015-08-13 22:22:20 -04:00
68bc4f2cf4
Merge pull request #12 from mlt/wip/i2p
...
Few i2p fixes
2015-08-13 21:55:27 -04:00
f73d1a95dc
Merge pull request #83 from aldenml/utp-stream-log
...
Added utp-stats bjam option to enable utp stream log (utp.log file).
2015-08-13 20:53:19 -04:00
d60ddf6786
Minor change: #ifdef -> #if
2015-08-13 18:43:58 -04:00
9e29eb9978
Introduced TORRENT_UTP_LOG_ENABLE macro to guarantee that TORRENT_UTP_LOG and TORRENT_VERBOSE_UTP_LOG are always defined (no need to change a lot of macro if).
2015-08-13 18:43:58 -04:00
7d11198911
Fixed definition of utp log macros.
2015-08-13 18:43:58 -04:00
12fb68e8e3
Refactor of utp stream logging functions.
...
Renamed utp-stats option to utp-log.
2015-08-13 18:43:58 -04:00
02afc8b640
Added utp-stats bjam option to enable utp stream log (utp.log file).
2015-08-13 18:43:58 -04:00
24b20a59cd
remove some randomness from the unit tests
2015-08-13 02:06:30 -04:00
411be1c017
forward port Changelog update
2015-08-12 23:43:38 -04:00
213958f98d
uTP timeout fix. introduced when fixing the recent vulnerability
2015-08-12 23:05:33 -04:00
d2819cd1b9
fix uninitialized variable in test_pe_crypto simulation
2015-08-12 20:18:47 -04:00
243b31b56b
trouble shoot segfault on travis
2015-08-12 18:54:57 -04:00
b703a8fb89
Merge pull request #82 from AlexandruCostache/master
...
Added missing dependencies to fix examples build
2015-08-12 08:07:07 -04:00
34aa1a0ab7
Added missing dependencies to fix examples build
...
Main examples like simple_client failed to build as they
were not linked against boost chrono and boost random libraries
2015-08-12 11:36:29 +03:00
ca793121e0
Merge pull request #81 from aldenml/no-full-namespace-udp-endpoint
...
Removed unnecessary full namespace reference to udp::endpoint in new dht_direct_request feature.
2015-08-12 01:15:53 -04:00
b303a36ac2
Removed unnecessary full namespace reference to udp::endpoint in new dht_direct_request feature.
...
Add default value to userdata argument in dht_direct_request.
Minor spelling fix.
2015-08-12 00:49:09 -04:00
1a25448b64
Merge pull request #80 from aldenml/port-mapping-log-notification
...
Created a new port_mapping_log_notification alert category
2015-08-12 00:25:39 -04:00
2457b16d09
Documented that portmap_log_alert is only posted if port_mapping_log_notification is enabled.
2015-08-12 00:10:32 -04:00
1805c96f44
Created a new port_mapping_log_notification alert category to separate portmap_log_alert.
2015-08-12 00:07:08 -04:00
36cf20b19f
Merge pull request #69 from ssiloti/direct-dht-req
...
Direct dht reqests
2015-08-11 23:32:25 -04:00
1b4b1c4403
fix error handling for DHT direct requests
2015-08-10 20:33:05 -07:00
135f4761e1
add more comments about DHT direct requests
2015-08-10 20:33:04 -07:00
8e00172af7
return a bdecode_node rather than taking one by reference
2015-08-10 20:33:04 -07:00
1ff48aef60
add comment
2015-08-10 20:33:04 -07:00
43dc784cf2
add direct_request.hpp to Makefile.am
2015-08-10 20:33:04 -07:00
3564562dab
pass raw buffer in dht_direct_response_alert
2015-08-10 20:33:04 -07:00
182c30f2d6
fix typo in class name
2015-08-10 20:33:04 -07:00
86fd8c3048
enable arbitrary direct dht requests and extension dht requests
...
If you're wondering why I used an alert for handling responses but not requests,
it's because the former was easy to do that way but the latter would have
involved some major refactoring or kludging. The DHT node is written with the
(IMO reasonable) assumption that all responses will be generated immediately, so
there was no way to easily accommodate getting a response asynchronously via
an alert.
2015-08-10 20:33:04 -07:00
a29e64965c
add some basic operators to peer_connection_handle
2015-08-10 20:33:04 -07:00
0dd4ff4f3c
Merge pull request #78 from ssiloti/test_dht-fix
...
Test dht fix
2015-08-10 23:25:16 -04:00
2d9cf0f8b3
use test_name rather than __LINE__ for hidden test structs
...
When building tests from mutliple files together, the line number the test is
declared on is not guaranteed to be unique. If two tests happen to be on the
same line, one of them will not be run. You will then go crazy trying to figure
out why seemingly unrelated changes make the problem go away.
2015-08-10 20:05:16 -07:00
29ba738876
fix_dht mutable get test
2015-08-10 20:05:04 -07:00
a81ca2391b
fix some warnings
2015-08-10 20:03:24 -04:00
12136563c4
use boost-build --hash instead of abbreviate paths on travis and appveyor, to never make the filenames overflow the filesystem limit
2015-08-10 10:21:44 -04:00
874bfcc89e
transitioned test_super_seeding to a simulation (and remove the live test for test_swarm, which already has been transitioned to a simulation)
2015-08-10 10:18:36 -04:00
f2455c8689
made test_metadata_extension a bit more reliable
2015-08-10 10:02:01 -04:00
a75699eff1
move test_metadata_extension to a simulation instead of a live test
2015-08-10 09:17:41 -04:00
a3dbbd6a37
transition the test_pe_crypto system tests to be simulation
2015-08-10 06:57:51 -04:00
e557e63d05
scale down the dht simulation, to try to make it pass travis and appveyor time limits
2015-08-09 22:07:27 -04:00
663dc5030d
fix typo in appveyor.yml
2015-08-09 21:50:06 -04:00
c99ad7b962
fix python binding Jamfile to detect BOOST_ROOT environment variable too
2015-08-09 21:23:11 -04:00
e38f461197
disable another flapping test on appveyor/windows
2015-08-09 18:23:01 -04:00
b9b17ca3bf
pull the submodules on appveyor
2015-08-09 18:10:39 -04:00
0103e23084
run simulations on appveyor too
2015-08-09 17:58:57 -04:00
d977f2e68f
attempt to make appveyor happy
2015-08-09 17:54:10 -04:00
4fa0aa62da
simulate IP routes, interfaces and page size when building in simulation mode
2015-08-09 15:01:01 -04:00
6068418def
fix and re-enable simulations
2015-08-09 00:56:37 -04:00
arvidn