@eyedeekay
&kytv
&zzz
+R4SAS
+RN
+RN_
+dr|z3d
+hk
+orignal
+postman
+wodencafe
Arch
DeltaOreo
FreeRider
FreefallHeavens
Irc2PGuest19353
Irc2PGuest22478
Irc2PGuest48042
Irc2PGuest64530
Irc2PGuest77854
Nausicaa
Onn4l7h
Onn4|7h
Over
Sisyphus
Sleepy
Soni
T3s|4
T3s|4_
Teeed
aargh3
acetone_
anon4
b3t4f4c3
bak83_
boonst
cumlord
dr4wd3
eyedeekay_bnc
hagen_
khb
not_bob_afk
plap
poriori_
profetikla
r3med1tz
rapidash
shiver_1
solidx66
tr
u5657
uop23ip
w8rabbit
weko_
x74a6
orignal
zzz, if there a particulat reason that we require R for floodfills?
orignal
beside that introducers don't work well
dr|z3d
zzz: #countries path {vector-effect: non-scaling-stroke}
dr|z3d
you can thank me later :)
dr|z3d
you also probably want: paint-order:stroke;stroke-width:1.5px on those paths.
dr|z3d
combined, those will give you crisper borders and remove the russian artifacts.
dr|z3d
*crisper, but not jaggy.
dr|z3d
when I say russian artifacts, I don't mean orignal, just to qualify :)
zzz
orignal, to become a floodfill, or to use a floodfill?
orignal
to become a floodfill
orignal
why R is required?
eyedeekay
I fixed my broken snap project. Doesn't that matter much but like ~90 people will start getting updates again
dr|z3d
R is required because reliability. you don't really want a firewalled router as a floodfill. extra hop, extra latency.
dr|z3d
hooray, eyedeekay \o/
dr|z3d
not that I have any great enthusiasm for snaps, but still.
eyedeekay
I don't either but I started it years ago just to try then broke it and couldn't take it down
eyedeekay
So I had to find a way to make it useful
dr|z3d
fair enough
dr|z3d
not a fan of what ubuntu are doing with snaps. there's absolutely zero need for firefox to be a snap.
dr|z3d
it's "amusing" because mozilla are all about speed, apparently, which is why until recently you had to dig into about:config to enable userChrome.css, because it was taking 3ms to check if it was available in the filesystem, and then mozilla lobby for ubuntu to use snaps.
dr|z3d
might as well post here, zzz map remix with color heatmap and his animations: cake.i2p/file/Y0k8tDAkw2_TBAV4G1a7bzafdsMhGZc3pQrDdfr0v_YHCTo82PVk/geomap-animate.svg
eyedeekay
Yeah in general the "yet another package manager" thing seems silly, I get why they want some way of preventing applications from being able to interact with certain parts of the system and I see why package manager is a logical-seeming place to do it but I don't see why we need *another* one to accomplish this
dr|z3d
snaps are fine when there's no native build available, I get that. otherwise, apparmor should suffice.
dr|z3d
in the case of mozilla, they have native .deb builds, so snap is just a waste of space.
eyedeekay
Well now the snap is built in CI and doesn't require any weird mods to work so it's unlikely to break again and if it does I'll know
eyedeekay
I've got machines where I recklessly mix a bunch of package types because I don't do anything important with them and don't mind to wipe them
eyedeekay
that's where I use snaps
zzz
orignal, I don't think it's in the spec, do what you want, but "don't work well" seems like a good enough reason to me
zzz
orignal, be advised though, we will have countermeasures soon for the current large fU botnet, so I don't suggest making more of them
orignal
I have the code that consider Uf as non-floodfill for a long time
orignal
my point is
orignal
we have lack of floodfills now
orignal
vs. total number of routers
orignal
people have reources to run FF but don't have ipv4 address
orignal
we might want to permit such FFs with some restrictions
orignal
like interoducers must be other floodfills
zzz
we will almost never use fU floodfills now, because of all the changes to deal with attacks and botnets. but it's not "never"
zzz
but take a look at all the fU routers now, 95% of them are from a botnet that started a few weeks ago, they're useless
zzz
that's what I'm working on now
zzz
I think the botnet is knocking about 5 percentage points off our exploratory build success
orignal
because nobody published fU routers
orignal
because everybody knows that FF must be R and ipv4
orignal
I also have the requerement that publish IP must match for incoming connections
zzz
the way it works for us, must be R to become f, but if we then transition to U, we don't immediately remove f, it may happen a few hours later
zzz
so there's a few non-botnet fU routers, not too many
orignal
is Uf a valid floodfill for you?
zzz
yes
zzz
but I have to deal with this botnet so "fU" will be "not very good"
orignal
then why I explude them?
orignal
*exclude
zzz
I have no idea why you do what you do ))
orignal
I think it was the reason
orignal
seee the problem
orignal
if you don't exclude them it means we have different list of floodfills
orignal
also how about ipv6-only?
zzz
we will use ipv6-only but we will not make ourselves ff if ipv6-only
orignal
but it doesn't make sense
orignal
you publish your RI or dest to a floodfill that can't connect to others
zzz
maybe. just not something I've worried about
zzz
eyedeekay, fyi I'm going to bump to -6 after I fix this github snark issue, today or tmrw
eyedeekay
Ack, bump at will
zzz
prop. 167 updated to incorporate comments from orignal
zzz
anybody with experience on ARM servers, non-mac ARM desktops, Rasp. Pi 3/4, or Windows on ARM, please comment on the github ARM issue, we're trying to draw a line between what is "slow" and what is not