Mark Qvist
e543d5c27f
Implemented basic channel windowing
2023-05-10 19:15:20 +02:00
Mark Qvist
01c59ab0c6
Cleanup
2023-05-10 18:44:05 +02:00
Mark Qvist
a4c64abed4
Initial framework for channel windowing
2023-05-10 18:43:17 +02:00
Mark Qvist
7df11a6f67
Fixed missing isolation of packet delivery callback
2023-05-10 18:40:46 +02:00
Mark Qvist
1bd6020163
Cleanup
2023-05-10 18:40:18 +02:00
Mark Qvist
b3ac3131b5
Updated version
2023-05-09 23:07:47 +02:00
Mark Qvist
f522cb1db1
Added per-packet compression to buffer
2023-05-09 22:13:57 +02:00
Mark Qvist
d96a4853fe
Fixed version display
2023-05-09 22:13:23 +02:00
Mark Qvist
52a0447fea
Fixed resent packets not getting repacked
2023-05-09 22:12:49 +02:00
Mark Qvist
e82e6d56f1
Added ability to trust external signing keys to rnodeconf
2023-05-09 15:31:02 +02:00
Mark Qvist
dcf33e125b
Cleanup
2023-05-05 10:43:27 +02:00
Mark Qvist
01600b96a4
Fix import paths
2023-05-05 10:37:22 +02:00
Mark Qvist
64bdc4c18c
Fix import paths
2023-05-05 10:25:15 +02:00
Mark Qvist
5e5d89cc92
Removed dependency on netifaces.
2023-05-04 23:19:43 +02:00
Mark Qvist
a3bee4baa9
Removed netifaces dependency from AutoInterface
2023-05-04 17:55:58 +02:00
Mark Qvist
fab83ec399
Restructured library
2023-05-04 17:55:38 +02:00
Mark Qvist
b740e36985
Added ifaddr module
2023-05-04 17:46:56 +02:00
Mark Qvist
53b325d34d
Added support for T3 v1.0 to rnodeconf
2023-05-03 15:56:19 +02:00
Mark Qvist
d31cf6e297
Added ability to configure RNode display intensity
2023-05-03 14:26:47 +02:00
Mark Qvist
e386a5d08b
Use native Python unzip for updates
2023-05-03 12:57:38 +02:00
Mark Qvist
d467ed9ece
Merge branch 'master' of github.com:markqvist/Reticulum
2023-05-03 12:27:10 +02:00
Mark Qvist
892a467d74
Update version
2023-05-03 12:26:48 +02:00
markqvist
4366e71f34
Merge pull request #272 from VioletEternity/windows
...
Improve Windows compatibility for rnodeconf
2023-05-03 12:26:36 +02:00
Mark Qvist
7e9998b4fd
Use included platform detection method
2023-05-03 12:21:57 +02:00
markqvist
79abe93139
Merge pull request #278 from VioletEternity/windows-so_reuseaddr
...
Use SO_EXCLUSIVEADDRUSE instead of SO_REUSEADDR on Windows
2023-05-03 12:18:49 +02:00
Mark Qvist
d69d4b3920
Fixed firmware extraction for unverifiable devices. Fixes #266 .
2023-05-02 18:10:04 +02:00
Mark Qvist
3300541181
Fixed invalid error code in conditional. Fixes #284 .
2023-05-02 17:45:30 +02:00
Mark Qvist
3848059f19
Only use ifname for link-local discovery scopes. Fixes #283 .
2023-05-02 17:39:06 +02:00
Mark Qvist
30021d89cb
Fixed header bits in get_packed_flags(). Fixes #275 .
2023-05-02 17:33:38 +02:00
Mark Qvist
29019724bd
Added verbosity argument to Reticulum instantiation. Fixes #238 .
2023-05-02 16:42:04 +02:00
Maya
ba7838c04e
Use SO_EXCLUSIVEADDRUSE instead of SO_REUSEADDR on Windows.
...
On Linux, SO_REUSEADDR is used so that a socket in TIME-WAIT state can
be rebound after a listening process is restarted. It does not allow two
processes to listen on the exact same (addr, port) combination. However,
on Windows, it does, and SO_EXCLUSIVEADDRUSE is required to reproduce
the Linux behavior.
Reticulum relies on an error being returned by bind() that reuses
the same (addr, port) combination as another process to detect whether
there is a shared instance already running. Setting SO_EXCLUSIVEADDRUSE
makes this detection process work on Windows as well.
2023-04-19 03:03:15 +01:00
Maya
af16c68e47
Make esptool.py invocation compatible with Windows.
2023-04-13 18:17:14 +01:00
Maya
bda5717051
Use standard Python zipfile module to decompress firmware
2023-04-13 18:10:21 +01:00
Mark Qvist
fac4973329
Fixed potential race condition in announce queue handling for AutoInterface
2023-03-09 18:32:14 +01:00
Mark Qvist
159c7a9a52
Fixed rnstatus JSON output error
2023-03-08 14:10:33 +01:00
Mark Qvist
226004da94
Ignore lo0 in all cases. Fixes #237 .
2023-03-07 16:43:10 +01:00
Aaron Heise
9c9f0a20f9
Handle sequence overflow when checking incoming message
2023-03-04 23:54:07 -06:00
Aaron Heise
6d9d410a70
Address multiple issues with Buffer and Channel
...
- StreamDataMessage now packed by struct rather than umsgpack for a more predictable size
- Added protected variable on LocalInterface to allow tests to simulate a low bandwidth connection
- Retry timer now has exponential backoff and a more sane starting value
- Link proves packet _before_ sending contents to Channel; this should help prevent spurious retries especially on half-duplex links
- Prevent Transport packet filter from filtering out duplicate packets for Channel; handle duplicates in Channel to ensure the packet is reproven (in case the original proof packet was lost)
- Fix up other tests broken by these changes
2023-03-04 23:37:58 -06:00
Mark Qvist
d8f3ad8d3f
Temporarily disabled extra-level log statement
2023-03-04 19:30:47 +01:00
Mark Qvist
a1b75b9746
Increased per-hop timeout
2023-03-04 19:30:23 +01:00
Mark Qvist
80f3bfaece
Adjusted StreamDataMessage overhead calculation
2023-03-04 19:06:47 +01:00
Mark Qvist
37b2d8a6ec
Fixed Link MDU output in phyparams()
2023-03-04 18:37:28 +01:00
Mark Qvist
777fea9cea
Differentiate exception between link establishment callback, and internal RTT packet handling
2023-03-04 18:32:36 +01:00
Mark Qvist
bbfdd37935
Added check for link state before sending
2023-03-04 18:31:07 +01:00
Mark Qvist
28e6302b3d
Updated versions
2023-03-04 17:56:30 +01:00
Aaron Heise
aac2b9f987
Buffer: send and receive binary data over Channel
...
(also some minor fixes in channel)
2023-03-02 17:17:18 -06:00
markqvist
a6551fc019
Merge pull request #246 from gdt/fix-transmit-hash
...
AutoInterface: Drop embedded scope identifier on fe80::
2023-03-02 11:34:00 +01:00
markqvist
a06ae40797
Merge pull request #236 from faragher/master
...
Additional error messages for offline flashing.
2023-03-02 11:31:31 +01:00
markqvist
1db08438df
Merge pull request #248 from Erethon/hkdf-remove-dead-code
...
hkdf: Remove duplicate check if the salt is None
2023-03-02 11:29:18 +01:00
Dionysis Grigoropoulos
ddb7a92c15
hkdf: Remove duplicate check if the salt is None
...
The second if isn't needed since we initialize the salt with zeroes
earlier. If instead we meant to pass an empty bytes class to the HMAC
implementation, the end result would be the same, since it's gonna get
padded with zeroes in the HMAC code.
2023-03-01 16:22:51 +02:00
