@RN
@StormyCloud
@T3s|4
@T3s|4_
@eyedeekay
@not_bob_afk
@orignal
@postman
@zzz
%Liorar
+FreefallHeavens_
+Xeha
+acetone
+bak83_
+hk
+mareki2p
+profetikla
+r00tobo_BNC
+snex
+uop23ip
+weko
Arch
Danny
DeltaOreo
Irc2PGuest12249
Irc2PGuest25220
Irc2PGuest88061
Irc2PGuest90883
Irc2PGuest91305
Leopold_
Meow
Nausicaa
Onn4l7h
Onn4|7h
aisle1
anu
ardu
boonst
deadface
dickless
dr|z3d_
duck
evasiveStillness
foobar_
mareki2pb
maylay
poriori_
qend-irc2p
radakayot_
shiver_
simprelay
solidx66
tr
u5657
usr001
woodwose
mareki2p
Also, Java does not have unsigned integer types.
dr|z3d
ok, so we now have a tooltip for the sidebar hosts count that presents links to all the new hosts of the day.
dr|z3d
dev build shortly...
mareki2p
Java is weird language...
dr|z3d
you should ping zzz so he knows to take a look
dr|z3d
except I just did it for you :)
mareki2p
Here you have. Working cha cha function in Java. As specified in RFC 7539, sections 2.1, 2.1.1, 2.2, 2.2.1, 2.3, 2.3.1, 2.3.2. Test vector works fine, but I'm not using bytes or any other unsigned variables. cake.i2p/view/2qWZmRLczv_R8bNC1KXoeAR2uYued5scIhiwiV4Cl_nSrbRE2eQa/Main.java
dr|z3d
ok, a few iterations later, we have something a little less clunky re sidebar new hosts count: cake.i2p/file/BBf7Vw3uYL_jALb932j0qi5z6jHbzk409s1qOjuV5_IYIUhZz5JU/newhosts.webp
orignal
<ESR> orignal: ты с ява разрабами связь держись? А то они для андроида глючное говно выложили. Оно не работает. Я им уже писал, но чего то они не того
orignal
people complain about Java-I2P for andorid
orignal
it just doesn't work
orignal
<ESR> 14, 13, 12 трубы и версии андроида.
orignal
versions of android
orignal
eyedeekay I think andorid is yours
eyedeekay
Yeah it is, I test on Android 10, 14, and GrapheneOS so it should work but I'll look into the situation
eyedeekay
Thanks for letting me know orignal
dr|z3d
orignal: as an interim measure, termux with normal i2p/i2p+ is also an option.
zzz
results of changing the lookup timeout
zzz
fail rate went from ~30 to ~40%
zzz
the change was from fixed 3000 ms to min(3000, max(500, 3*avgResponseTime))
zzz
I think 40% is probably ok but I'm also open to increasing the min to maybe 750, dunno
orignal
I use 1600
orignal
eyedeekay it's no me it's guy
zzz
what did your change to 1600 do to your fail rate?
orignal
const int MAX_LEASESET_REQUEST_TIMEOUT = 12000; // in milliseconds
orignal
sorry
orignal
const int LEASESET_REQUEST_TIMEOUT = 1600; // in milliseconds
orignal
lookup is just faster
dr|z3d
can we do something with the y values I wonder? rather than 0.4 etc, perhaps percent.
orignal
also onon has an idea
zzz
orignal, but what did it do to the per-peer fail rate, or the peers-per-lookup count?
orignal
if we have lookup timeout
orignal
try that FF again and another one at the same time
orignal
it's for LS through tunnels
orignal
I didn't not measure
orignal
my goal was to wait less
zzz
you have any data at all? peers per lookup? average response time? average time-to-success?
orignal
no
orignal
what I could say no failed lookup after 7 attempts
zzz
you will if the dest isn't up
orignal
ofc I when I know for sure it's up