0e49051f17
make create_random_files() take a span instead of pointer, length parameters
2018-06-11 01:35:31 +02:00
b26467c04e
deprecate network-threads setting
2018-06-10 19:34:07 +02:00
e8f971a54a
scrape_reply_alert should be high priority, since it's triggered by the client
2018-06-10 19:28:45 +02:00
6647b0988d
add support for multi-home NAT-PMP and PCP
2018-06-10 16:55:42 +02:00
1b62aff9d1
minor improvements to test_checking
2018-06-10 16:25:10 +02:00
247672820c
report transport version in NAT-PMP send/receive logs
2018-06-09 16:40:46 -07:00
e7fe7f1b89
fix boost::get<> invocation for variant
2018-06-10 00:56:55 +02:00
f039cffeb3
relax packet length check prior to checking response code
...
Don't assume that an unsupported version response will contain anything past
the response code.
2018-06-09 08:55:49 -07:00
a357e490ed
minor code refactor, more use of auto
2018-06-08 21:56:06 +02:00
2cf1018fa0
enum_if: add default gateway column to interface table
2018-06-07 18:18:31 -07:00
fbaa1afd62
return a vector of mapping ids from add_port_mapping
2018-06-07 18:18:31 -07:00
e523566622
add support for the Port Control Protocol (PCP)
2018-06-07 18:18:31 -07:00
f6720e31ed
don't re-map all listen sockets when changing listen_interfaces
...
New sockets will be mapped regardless of the remap parameter. Requesting
full re-mapping just creates unecessary network traffic.
2018-06-07 18:16:08 -07:00
78be52112f
create a natpmp instance for each listen socket
2018-06-07 18:16:08 -07:00
496cf46b25
enum_routes fixes on Linux
2018-06-07 18:09:51 -07:00
4a578b1450
fix scope_id in enum_routes on Windows
2018-06-07 18:09:51 -07:00
72cf4fd9ea
Remove UNC prefixes from device names on Windows
2018-06-07 18:09:51 -07:00
603e206f36
fix netmask of routes on Windows
...
The build_netmask function was incorrectly dividing the size of the bytes_type
by 8. Also use b.size() instead of sizeof for consistency.
I'm not 100% certain what the difference between DestinationPrefix.PrefixLength
and SitePrefixLength is. It looks like SitePrefixLength is something specific to
site-local addresses. Emperically I found DestinationPrefix.PrefixLength
contains reasonable values while SitePrefixLength is always zero.
2018-06-07 18:09:51 -07:00
e88339378c
merged RC_1_1 into master
2018-06-07 00:30:23 +02:00
049d867c12
fixed race condition in random number generator
2018-06-06 09:54:49 +02:00
2408200475
fix race condition in stat_cache
2018-06-05 19:07:33 +02:00
794ccf4f79
remove unused disk stats counters
2018-06-05 19:03:04 +02:00
95cfc16bca
fix integer overflow in alert_manager
2018-06-02 15:16:08 +02:00
267ca40b28
merged RC_1_1 into master
2018-06-02 13:15:13 +02:00
c55bc7dd42
improve error handling of failing to change file priority
2018-05-31 00:47:17 +02:00
8a0ac58658
some lint oriented refactor, more use of const
2018-05-30 23:52:28 +02:00
c56f6abf72
deliver notification of alerts being dropped via alerts_dropped_alert
2018-05-30 11:26:02 +02:00
f9b43f3511
some cast cleanup, const modifier and lint refactor
2018-05-30 00:06:52 +02:00
c2ea38fdfe
Add set_alert_notify to python bindings ( #3066 )
2018-05-29 22:30:59 +02:00
6c4869f044
moved instantiate_connection to aux namespace
2018-05-29 09:25:25 +02:00
396c5dd3af
remove use of deprecated function readdir_r
...
Ports 140b8ace
2018-05-28 16:06:41 +02:00
3b0ce8d2e1
merged RC_1_1 into master
2018-05-28 15:56:22 +02:00
4cef1814b0
Add 1.1.7 release in the ChangeLog
2018-05-28 09:44:26 +02:00
e15077fc3f
Fix pkg-config branch in the example CMake find module
...
1. The imported target properties may not contain -D switches in the
INTERFACE_COMPILE_DEFINITIONS property, because CMake will always
prepend them to each item. Therefore we have to remove those switches
and also pass other options (e.g. -fexceptions) in INTERFACE_COMPILE_OPTIONS
property, otherwise CMake creates "mutants" '-D-fexceptions' alike.
2. FindOpenSSL.cmake variable for include directory is
OPENSSL_INCLUDE_DIR, not OPENSSL_INCLUDE_DIRS.
2018-05-27 23:01:22 +02:00
b1d7dd89c1
Reflect Boost version and C++11/14 interconnection in CMake build
...
C++ standard and Boost requirements are connected:
1. With C++11 onward, we require Boost system component, with
C++03 we need chrono and random components too
2. When building against boost 1.66 and newer, C++11 is required.
Closes #3011 , partially closes #2966 .
2018-05-27 22:59:13 +02:00
35d7672432
raise priority of fastresum_rejected_alert
2018-05-27 22:33:24 +02:00
287f299314
cmake: pass target C++ standard to pkgconfig generator
2018-05-27 09:55:13 +02:00
d6cecee8c0
some code cleanup and minor refactor
2018-05-24 23:46:37 +02:00
6ec6c1b340
since moving an existing file into the part file isn't supported, just don't use the partfile for existing files when their priority is set to 0
2018-05-24 23:45:28 +02:00
534276e7a5
refactor to use the utility function for bind_socket_to_device() instead of using the bind_to_device socket option directly, to make it best effort. include operation enum in udp_error alert.
2018-05-23 12:11:04 +02:00
ff300ed224
some cleanup and refactor in torrent_info code ( #3054 )
2018-05-23 10:39:55 +02:00
77191a3820
fix hidden attribute in file class
2018-05-23 10:39:10 +02:00
ed8fffc1a6
set the hidden attribute when creating the part file
2018-05-23 10:39:10 +02:00
55a0a6e0df
fix production asserts
2018-05-22 12:12:33 +02:00
92a7188163
only start a new accept request on new listen sockets
...
Existing sockets already have an outstanding accept.
Also fix mapping new listen sockets when re-mapping is not requested.
2018-05-22 11:06:33 +02:00
1a66a8b285
merged RC_1_1 into master
2018-05-20 12:08:46 +02:00
c986d93913
fix properly ignoring failure to get local endpoint of a socket while printing an error
2018-05-20 10:47:46 +02:00
a137bbfdd4
simplify some using statements to pull in the address types
2018-05-20 08:46:14 +02:00
54f00ce09d
don't change state to downloading if the torrent is finished
...
Move the transition to downloading after the torrent has been checked
for being finished. This prevents a spurious state_changed_alert when
adding a torrent which is already completed.
Also move the posting of the torrent_checked_alert to maintain the
historical ordering of that alert with respect to the transition out of
checking_resume_data.
2018-05-20 08:45:24 +02:00
5f74b6e353
Bump `file_error_alert` priority
...
This alert's priority must be higher than `block_downloading_alert`'s
to allow custom handling of file errors.
With progress notifications enabled, network thread might fully load
alerts queue, so file errors from disk thread won't have a chance to
be delivered.
2018-05-17 00:55:07 +02:00
Arvid Norberg