mesh
so the good news is I don't get the "unsupported encryption options" error any more
mesh
the bad news is I'm not able to connect to anything. virtually every website is unreachable
mesh
mayhaps something has gone wrong with hidden mode
dr|z3d
logs would be helpful, mesh
T3s|4_
dr|z3d: apparently not all the IRC connection issues have been resolved. While shopping, I saw 7 'irc: connection timeout (message 001 not received)' attempts with the reconnection delay at 10 minutes before it finally connection. Neither a router or a tunnel restart seem to impact this IRC problem.
T3s|4_
^running your latest -36+
dr|z3d
likely an upstream issue, T3s|4.
T3s|4_
k
dr|z3d
do you know which server you were connecting to at the time?
T3s|4_
I am still using both echelon and postman, but it finally connected to echelon
dr|z3d
when it wouldn't connect, postman or echelon?
T3s|4_
weechat never tells me that: only irc: 'connected to 127.0.0.1/6668 (127.0.0.1)' then 'irc: connection timeout (message 001 not received)' occurs multiple times
dr|z3d
ok
T3s|4_
dr|z3d: after a few more router updates and restarts, I cannot reliably reproduce 'irc: connection timeout (message 001 not received)' on both laptops, so I am just going to keep me eye on this issue for a few days
T3s|4_
*my eye
dr|z3d
10-4, T3s|4
zzz
whats up kids
y2kboy23
dr|z3d Found on the netDB page: cake.i2p/file/CIgptVeEVo_CK1Y8Nt9w4yrDonwtZJiKe1GzpGX8Q_z1BvJgVwHC/typeerror.jpg
y2kboy23
Hey zzz
dr|z3d
welcome to #saltR, zzz
dr|z3d
thanks, y2kboy23, let's have a look at that.
y2kboy23
I found something else, but forgot which page it was on.
T3s|4
zzz: nice to see about these parts :D
T3s|4
see *you about
y2kboy23
dr|z3d /tunnelpeercount: cake.i2p/file/G8o7gE5IDh_E2bGN7e8s2RCWPCWezoQ2XAXLOpzDI_bS0P6yMqco/referenceerror.jpg
dr|z3d
that second one is a typo, the first is mostly due to our flags having gone missing. will fix both, minor, thanks.
y2kboy23
Good dr
y2kboy23
Deal
zzz
yup, first checkin today in 7 months
zzz
I see we're now pushed back to thanksgiving, which feels realistic to me
zzz
thanks dr|z3d, you content these days, or itching to get a release out? or?
dr|z3d
Not exactly content, but the release schedule is not a major concern. When it's ready, it's ready.
dr|z3d
I'll be content when the things that got borked with the segmented db monster merge get unborked.
zzz
I haven't floated this w/ eyedeekay yet, but I can foresee a future where most of the codebase has stabilized, but the subdb stuff is either unproven or still borked
zzz
but the changes from 2.3.0 are finally clean enough that a simple 'DEFAULT_USE_SUBDBS = false' could be added someplace and call it a day
zzz
as a 2.3.1 or 2.4.0
zzz
just a back-pocket contingency
zzz
yeah dr|z3d you kinda sealed your fate when you swallowed the tsunami
dr|z3d
well, I did, and I didn't :)
dr|z3d
I still have a pre-tsunami branch just in case.
zzz
wouldn't be fun though
dr|z3d
This is true. None of this has been fun.
zzz
speaking of fun, I'm playing with something from my todo list in March 2005
dr|z3d
oh? do tell..
dr|z3d
fixes pushed to master, y2kboy23
zzz
Patch in development. Preliminary patch: persistent-v0.2.patch.txt (90K) 2005-03-08 against CVS 0.6.1.12-7.
dr|z3d
wassit do?
zzz
http persistent conns / pipelining
zzz
it's an 1800 line patch, from a month or less after I learned java, and mostly I pretended it was C
zzz
it's pretty shaky
zzz
I think I was trying to impress jrandom
zzz
he probably wasn't
dr|z3d
oh, pipelining.
dr|z3d
http/2 is the new pipelining :)
zzz
yeah a lot of changes in 18 years
zzz
as you said re: http/2
zzz
but persistence is expected and the default
zzz
this page has good pictures and info: developer.mozilla.org/en-US/docs/Web/HTTP/Connection_management_in_HTTP_1.x
zzz
so I'd like to do persistence
zzz
w/o pipelining it's a _lot_ easier
dr|z3d
yeah, pipelining doesn't look like it's worth the hassle.
zzz
Note: HTTP pipelining is not activated by default in modern browsers:
zzz
Q.E.D.
dr|z3d
I always remember pipelining being something you turned on manually and hoped that it didn't break websites, back when it was a thing.
dr|z3d
"... For these reasons, pipelining has been superseded by a better algorithm, multiplexing, that is used by HTTP/2."
zzz
Buggy proxies are still common and these lead to strange and erratic behaviors <--- my 2005 patch
dr|z3d
well, persistence with a sensible timeout sounds like it could be handy.
T3s|4
zzz: probably not among your top todo priorities, but do you have any plans to re-spin-up zzz.i2p to its former glory - or did you find maintaining it was too much of a time sync?
zzz
yeah. we don't have the RTT issues because streaming is basically 0-RTT. But our streaming SYN is huge compared to TCP
zzz
*sink :) no decisions made yet on that or stats.i2p. Other than if either comes back, the first would be stats
T3s|4
right = sink! thanks zzz
dr|z3d
so what state's the patch in right now, zzz? any local testing?
mesh
zzz: it would be nice to see even a read-only version of zzz.i2p come back
zzz
yeah mesh I had that thought also
mesh
there was a lot of valuable knowledge of that website. I've got a lot of bookmarks that point to nowhere hehe. next time I'll add the stuff to my notes
zzz
dr|z3d, fatally flawed and impossible to merge. Mostly useful as inspiration
dr|z3d
:)
zzz
just playing around. may turn into something, or something else may catch my eye tomorrow
zzz
45 bugs submitted in 2 weeks though doesn't leave much time for anything else. Hopefully that's about all of it, but I still haven't looked at the actual subdb machinery
zzz
the cleanups are helping make the bugs more obvious though
zzz
as predicted
dr|z3d
the question is, is it salvageable?
dr|z3d
also, when this all blows over and you want something else to look at, github.com/hyperxpro/Brotli4j
dr|z3d
that is, if you're looking at stuff that could potentially improve web access speeds.
zzz
I think the cleanups and reversions will knock several thousand lines out of the release diff, leaving something that can be reviewed and reasonably tested
zzz
I'm confident we can get to that point
dr|z3d
good, that sounds more manageable.
zzz
but the q of does (or will) the code do what it claims and provide a net security benefit? not making a prediction there
zzz
I'm hopeful, but the proposal above could be a useful temp offramp if needed
zzz
but scrap the branch and start over? absolutely not
dr|z3d
if we can be confident the protections work as intended, then it's probably been worth the pain.
zzz
I'm 110% supportive of eyedeekay and give him mad props for what he's done the last 6 months, and for keeping his chin up last 2 weeks in the face of the counter-tsunami of tickets
zzz
I see the path forward in _my_ head. Not sure how well I'm communicating it, TBD...
dr|z3d
eyedeekay has been sterling, especially considering he had the management of the project dropped in his lap :)
zzz
re: brotli, should go thru us transparently now w/o issue. If you want to teach Jetty howto brotli, that's another story
dr|z3d
brotli works fine when served from anything other than jetty, sure.
zzz
yeah I know it's partially my fault
zzz
dunno if jetty has pluggable compression facility
dr|z3d
I did a quick look re jetty, doesn't seem brotli's a priority for them, but that may have changed.
dr|z3d
They've just pushed out Jetty 12.
zzz
hmph
zzz
we're kinda at a dead end w/ 9.3/9.4, it's long dead to them but we'd lose deb/ubuntu and java 8 if we tried to catch up
zzz
not a good situation
zzz
I thought we were really really screwed anyway until debian backed out their breakage
zzz
no luck w/ ubuntu though: bugs.launchpad.net/ubuntu/+source/jetty9/+bug/2036331
dr|z3d
well, java8 is becoming less and less a thing. even android's moving on.
dr|z3d
oh, you filed the bug. :)
zzz
yeah, if the windows bundle has edged out the regular install, and ofc android doesn't use jetty, we could keep java 8 for our code to keep android happy but do the other builds with something higher
dr|z3d
it's probably about time. isn't android effectively its own branch anyways?
zzz
no because it uses core, router, i2ptunnel, etc. from i2p.i2p
zzz
if android has any JRE 9/10/11 features it's news to me. They never did all of 8 last I looked
zzz
we're still requiring java 7 compatibility in those subsystems, although we could probably move up to 8 without losing too much android userbase at this point
zzz
right now we're at lollipop (2014) minimum, java 8 would be oreo (2017)
dr|z3d
may have misread, thought I saw somewhere that android was getting java12.
zzz
the issue with relying on anything in android is, how much of your userbase do you want to abandon in order to require it
zzz
and I don't think google updates their version pie charts anymore
dr|z3d
abandoning lollipop is a no brainer. anything running that isn't likely to be able to run i2p too well.
zzz
sure, if we had any good reason to I'm sure we would
zzz
I just replaced the battery on my old phone today, it was exploding a little, maybe I'm part of the problem...
zzz
but it is on android 10 so I guess not
dr|z3d
in other news, you've got a couple of easy optimizations with html5.
dr|z3d
no need to for the type attribute for <script> or <style>
dr|z3d
and although they still work, align= and <center> are technically deprecated.
eyedeekay
The only reasons I've accepted to bump android dependencies were google-imposed, open to changing that but like you said, gotta have a good reason
zzz
yup
eyedeekay
but also, 100% of my pre android 8 hardware has now failed in some critical way, screen, battery, something, there may be billions of android 5 devices that were sold not getting updates, but how many of them still power on?
zzz
re: "html optimizations", not a good use of our current time/brainpower right now imho
dr|z3d
sure, just throwing it out there. file it away under "maybe later".
zzz
I've never had an android die like that. Other than battery so shot it has to stay plugged in
zzz
the one I changed the battery on today is 5 years old
zzz
they functionally die because the storage gets eaten up, you have to delete most of the apps just to have enough room for the others to update, so they become functionally useless, but not dead. That's my experience
eyedeekay
Well if the battery is so shot it has to be plugged in does it cease to be a mobile? I suppose I did mostly buy a bunch of burners with cheap screens. And I have 4 cats.
zzz
a plugin-only tablet isn't terrible but the old ones are so damn slow
zzz
just plugged in my motorola slider-keyboard model from like 2007, let's see...
dr|z3d
if you want something affordable and reasonably performant, with detatchable pogo pin keyboard, you could do a lot worse than this: xiaomitoday.com/2021/12/26/alldocube-x-game-review
dr|z3d
sure, it's not rocking the absolutely latest chips, but it's pretty capable for the ~200$ pricepoint, screen's good, ram's adequate, etc.
dr|z3d
thanks for confirming, major.