#saltr (irc2p) | I2P+ 2.5.2+ >> i2pplus.github.io | skank.i2p | git.skank.i2p | purokishi.i2p | vituperative.github.io/i2pchat | natter.i2p | notbob.i2p | ramble.i2p | shreddit.i2p | cake.i2p | /msg an op for voice

dr|z3d ok, they've arrived, by a different route than I intended, but hey. svg identicons in your addressbook, available on the latest + dev build.

dr|z3d banging my head against the wall trying to get java to handle the rendering is why I was a bit testy earlier, T3s|4. Sorry.

dr|z3d Max 4 digits for in/out bandwidth in snark until we identify the root cause of the issue.

T3s|4_ sounds like a plan dr|z3d :)

zzz dr|z3d, unable to reproduce your access filter issue, see gitlab for details

dr|z3d 10-4, zzz

dr|z3d The issue was caused when performing an echo > denyfile.txt on a non-empty file while the tunnel was running.

dr|z3d Not an unreasonable thing to do, or expect, if you look at a deny file and find several thousand entries.

dr|z3d I like the idea of skipping bad lines that you suggest, either skipping them or removing them from the file. Setting an upper limit on the number of entries might also be worth considering.

zzz that would create a one byte file with a newline, not an empty file. which was it?

dr|z3d yeah, probably that.

dr|z3d apologies for the not-100% accurate report.

zzz np, so I have reproduced it, but this was completely self-inflicted ((

zzz > denyfile.txt would have been fine

dr|z3d I wasn't aware that echo > inserted a newline, my bad.

zzz echo -n > denyfile.txt would have worked too...

zzz anyway, I know enough to work on it now

dr|z3d it's a powerful tool that's probably underutilized and could do with a buff up.

uop23ip Hi, can i change my "Minimum threat points to block:" in a way to include the peers i blocked because they are slow. I got so many of them. Canon here.

zzz it's a botnet, there's thousands and thousands of them. We don't block them permanently because you'd burn a ton of memory doing it

zzz no you can't do it that way either

uop23ip wow, than better ban them i guess. I got 60+ percent of known in numbers banned, that seemed too much for me, but with your explanation it make sense to block them. Is this a "friendly" botnet (meaning they are doing their thing and don't hurt the network) and the last attack is gone? New attack?

zzz the sybil stuff is pretty much disabled because the netdb spam broke it

uop23ip is it needed for the next release? Should i let it activated or does it hurt(cpu wise or other)?

zzz there's no sybil changes so far in this cycle

zzz no such thing as a friendly botnet; ofc we don't know the details, but they're generally poorly configured and harmful, even if not doing explicit attacks

zzz working now on the one that appeared a couple weeks ago and is knocking several points off our expl. build success

uop23ip ok thanks. Wish you much success. Didn't realize that there always something going on. I thought after build success had gone up and the chinas were gone, everything is normal again and my router was just a little bit picky :)

zzz right. the 'slow' botnet still exists; not sure about the 2.5.1 botnet, it may be gone; and there's the new one I'm working on. And ofc china may come back

uop23ip Is there a danger that a massive banned bot net can fill my network view (known) to that level, that my "real" known and used peers getting that low that it would affect my anonymity?

zzz that's the kind of things we're thinking about and trying to mitigate, yes

uop23ip can i make the calculation known-banned<100-> plug out it is getting dangerous unanonymous :)

zzz ??

uop23ip i looked at my known numbers and my banned numbers and would do (known-banned). Completely wrong here?

zzz oh

zzz banned usually not included in known, so known-banned is meaningless

uop23ip ah ok good to know. But i could say that there is a ratio to the of what percentage of my network view are bad apples. in my case 2k known and 1k banned. 1/3 of my network view are bad apples, right?

uop23ip known 100 and banned 3k would be a reason to plug out the service, because 100 is to low to have anonymity. Is there some number were would say that or does have known nothing to do with anonymity??

zzz we try hard to maintain a lot more than 100

uop23ip that good that i don't have to expect that case, because i would plugout at 100/3k. that ratio doesn't feel right to me. Thanks for your time zzz and happy hunting!

T3s|4 o/ dr|z3d - your head banging on the *.svg addressbook icons was worth it - looking very snazzy on 19+. Noted on the hard coded snark down/up entry limits of four nines. Without surprise, given my recent hair pulling, new added torrents proceed as expected :D

dr|z3d thanks T3s|4, happy you're enjoying the new identicons. I was originally trying to modify the existing java identicon library, but after much headbanging and the realization that the new identicons look better, I went with those instead (javascript).

dr|z3d phase 2 is to cache all the svgs so they only need to be rendered once.

T3s|4 yeah, they really 'pop' on my 4K display :)

dr|z3d yeah, good. I suspect the old bitmap icons were looking a little blurry @ 4K (resolution, not b/w) :)

T3s|4 dr|z3d: no critique intended toward you, but WTF is up with SVG? w3.org/Graphics/SVG - Wasn't 2016 8 years ago?

dr|z3d it's a mature format.

T3s|4 dr|z3d: sure, but project maturity is not an excuse for 8 years of neglect

dr|z3d well, it's not so much a project as a defacto standard, and you don't really want to be messing with those too much once they're implemented everywhere.

dr|z3d it's a bit like png. set in stone. you want new things, you define a new file format, like webp.

T3s|4 would have been far better, even 8 years ago, to express thanks to all those who contributed to standard development and testing...then sign off

dr|z3d I think you might be right about the site, if nothing else. It's not exactly current :)

T3s|4 nah, I prefer to roll with, the svg site is current, assuming you mean by 'current'; it's currently an embarrassment to all the folks who donated many hours of their lives agreeing on the parameters of the standard. They should be explicitly recognized and thanked :)

dr|z3d :)