Jeff
80a7b788bc
Merge pull request #1089 from majestrate/dht-introset-lookup-relayed-try-closer-2020-02-07
...
try closer router if relayed
5 years ago
Jeff Becker
00260555bd
logic simplification
5 years ago
Jeff Becker
e4a16dfdd5
fix comment
5 years ago
Jeff Becker
b211450cc8
always recurse to second closest if we don't have it
5 years ago
Jeff Becker
5b87a9419e
default recursion depth to 2 not 12
5 years ago
Jeff Becker
19d91a440f
move declaration of peer
5 years ago
Jeff Becker
3bea7327cb
handle end case by telling requester that it's not there
5 years ago
Jeff Becker
7e1a6236be
if relayed try closer router if we don't have an intro
5 years ago
Jeff Becker
a9331a1431
fix unit test
5 years ago
Stephen Shelton
2279ebeb40
Add IWP session stats to JSON API
5 years ago
Jeff Becker
cd27b33849
dont flood dht with requests, if we know of a router don't look it up
5 years ago
Jeff Becker
c60a83f7f4
update lokinet monitor to work with iwp
5 years ago
Jeff
0271f178d2
Merge pull request #1078 from jagerman/travis-bionic-and-arm
...
New travis arches + required endian fixes, libsodium build replacement
5 years ago
Jason Rhinelander
a1813b6164
Remove s390x from allowed_failures
...
It now builds and passes all tests.
5 years ago
Jason Rhinelander
0497fb9ccb
Another big-endian fix
5 years ago
Jason Rhinelander
3be7eb789b
Endian defines fix for macOS
...
It seems `__BYTE_ORDER`/`__LITTLE_ENDIAN`/`__BIG_ENDIAN` aren't defined
on macOS, so `if __BYTE_ORDER == __BIG_ENDIAN` was true which made macOS
take the big endian path *twice* (which cancelled out the big endian
conversion).
This makes util/endian.hpp define __LITTLE_ENDIAN__ or __BIG_ENDIAN__
everywhere, and errors if it can't be set.
5 years ago
Jason Rhinelander
9283aac4b6
Big endian fixes
5 years ago
Jason Rhinelander
04fa117399
Use -mcpu=native on ppc64le
5 years ago
Jason Rhinelander
8735073cc8
Disable router docker image
5 years ago
Jason Rhinelander
c12367d463
DOWNLOAD_SODIUM in docker travis builds
5 years ago
Jason Rhinelander
66ebe3995e
Remove -Wshadow; it is far too aggressive under gcc
5 years ago
Jason Rhinelander
f8201fcf40
Turn on -DDOWNLOAD_SODIUM for gcc trunk build
5 years ago
Jason Rhinelander
1c17d50550
Disable stupid native windows builds
...
The builds just sit at a `C:\blah\blah>` prompt until travis kills them.
5 years ago
Jason Rhinelander
e7902d2e45
Move ppc64le/s390x to allow_failures and simplify it
5 years ago
Jason Rhinelander
7ab6220a41
Older cmake workaround
5 years ago
Jeff
e44449b2d5
Merge pull request #1084 from notlesh/reduce-client-log-output-2020-02-06
...
s/LogInfo/LogDebug/ @ explore message
5 years ago
Jason Rhinelander
74cc73d46b
DOWNLOAD_SODIUM=FORCE to always download libsodium
...
This bypasses the check for a compatible local libsodium.
5 years ago
Jason Rhinelander
5a4700d868
Use ExternalProject to build downloaded libsodium
5 years ago
Jeff Becker
8c698a1cd1
* rename i to foundIntro
...
* check for zero timestamp
* use > not >=
5 years ago
Jeff Becker
41210a8ae1
check for current introset being zero (initial state)
5 years ago
Stephen Shelton
37e4f26be1
s/LogInfo/LogDebug @ 'signed/verify encrypted introset'
5 years ago
Stephen Shelton
2cade16c0e
s/LogInfo/LogDebug/ @ 'updating SN list'
5 years ago
Stephen Shelton
bc3184a126
s/LogInfo/LogDebug/ @ explore message
5 years ago
Jason Rhinelander
5f642e4f01
Zero-initialize test array
5 years ago
Jeff
4bc5b52762
Merge pull request #1082 from majestrate/log-build-after-send-2020-02-06
...
only mark path build success if delivered to first hop
5 years ago
Jason Rhinelander
228a3573f7
Remove travis_wait
...
We don't have anything that takes >10 minutes without producing output,
and travis_wait is really annoying by hiding output until the job
completely finishes.
5 years ago
Jason Rhinelander
0c69cad451
Add arm64/s390x/ppc64le builds, reduce apt deps
...
The apt deps is installing a pile of stuff that isn't needed (and isn't
available on non-amd64); reduce the set and use a YAML reference to use
it only in the jobs doing a native build (none of the docker builds need
to install anything since everything happens inside docker).
5 years ago
Jason Rhinelander
18705ba9fe
Switch primary builds from xenial to bionic
5 years ago
Jeff Becker
499e346da6
notify delivery of path builds
5 years ago
Jeff
8230b1152e
Merge pull request #1072 from majestrate/2020-01-25-router-version
...
add router version
5 years ago
Jeff
702ab29ce4
Merge pull request #1080 from majestrate/fix-kdf-2020-02-03
...
don't derive x25519 key from ed25519 key
5 years ago
Jeff Becker
1318846a08
test/service/test_llarp_service_identity.cpp:
...
* remove failing irrelevent test
* add expect_call in EnsureKeys to match expected bevahior
5 years ago
Jeff Becker
e6174efa21
don't derive x25519 key from ed25519 key
5 years ago
Jeff
c122b460e1
Merge pull request #1075 from jagerman/dht-blinding
...
Dht blinding
5 years ago
Jeff Becker
80f8363386
don't drop messages with pathid zero which belong to snode to snode dht and path build requests.
5 years ago
Jason Rhinelander
73239f4c39
Use glob to find sodium source files
...
Manually changing the list every time we want to change the libsodium
version is painful.
I tested this glob against libsodium 1.0.17 and verified it produces the
same set of files (aside from the generated version.h), and that in
1.0.18 it produces the same set modulo upstream files removed/added.
5 years ago
Jason Rhinelander
b72a4296b7
Fix libsodium 1.0.18 URL
5 years ago
Jeff
c8417a7270
Merge pull request #1076 from majestrate/plug-memory-leak-2020-02-01
...
use std::shared_ptr not a bare pointer to prevent memleak
5 years ago
Jason Rhinelander
2b93ccb393
Bump libsodium required version to 1.0.18
...
We use crypto_core_ed25519_scalar_mul which is only exported by
libsodium >= 1.0.18.
5 years ago
Jason Rhinelander
8cd95c01b3
Abstract signing hash data location
...
This makes the interface a little cleaner by not requiring the class
user to know how we laid out the data in the object.
5 years ago