zzz
eyedeekay, what's the best practice for embedding links in the blog markdown?
zzz
is it `text link`_ and is it one _ or two? or should the link be in the <% trans ... %> section like we do for some other docs?
zzz
or a link down to [0] reference section like we do for specs?
zzz
what works best for translators?
zzz
(thinking about finally starting the long-promised SSU2 post)
eyedeekay
zzz I don't know what works best for translators but the way that I have been doing it is with a `text link` in blog posts with one _ unless it's a situation where there are a bunch of links, when the reference section is easier to manage and makes more sense
zzz
well do we want the translators to see the link?
zzz
also, it's easier on the translators if the newlines are at natural breaks in the text, like at sentences or links
zzz
when there's newlines in the middle of `...` stuff it just makes it more likely they will accidentally break things
zzz
do you know if it's one _ or two or what's the difference? I see some of each
eyedeekay
Usually it's a single _, __ is for "indirect" hyperlink targets in RestructuredText
eyedeekay
Blinded message
eyedeekay
Re: should translators see links if it's stuff on our own site, i.e. the SSU2 spec itself, I think they should probably see that partly for context and partly because they might need to translate/want to revise the material being referenced too?
zzz
in my experience, translators mostly just break complex formatting by accident
zzz
skimmed that reference and I'm even more confused
zzz
anyway, I don't know what you're using to format blog posts, but think about putting the newlines between sentences, not randomly, to make it a little easier on the translators
zzz
eyedeekay, look at this one from zab:
zzz
8/3/Apple-Silicon-Easy-Install.rst:`{% trans %}I2P Mac OS Easy Install bundles{% endtrans %}`__
zzz
8/3/Apple-Silicon-Easy-Install.rst:__ geti2p.net/en/download/mac
zzz
that's completely different from the way you do it, e.g.:
zzz
{% trans -%}
zzz
To learn more about StormyCloud visit their `web site
zzz
<https://www.stormycloud.org/>`_.
zzz
{%- endtrans %}
zzz
but I guess zab's way only works if you're linkifying the whole text
zzz
and here's the third way that we do in the html pages:
zzz
<p>{% trans anonbib="https://gitweb.torproject.org/anonbib.git" -%}
zzz
The source code for this page was adapted from
zzz
<a href="{{ anonbib }}">Free Haven's anonymity bibliography</a>.
zzz
{%- endtrans %}</p>
zzz
the benefit of that way is we can change the link without breaking the translation
zzz
and it prevents the translators from changing the link
zzz
or seeing it at all
eyedeekay
Then if there's a way to do it without breaking the links, we should keep them out of the trans tags in the restructured text posts too, but I haven't been able to keep the actual links outside the trans tags in rst yet without weird formatting stuff, except for zab's way which as you pointed out linkifies the whole preceding section
zzz
yeah I assume that last way works in rst, because it's the translation code that's doing the substitution, the rst code won't see it
zzz
I think?
zzz
oh, it would have to be like this I think?
zzz
{% trans foo="https://..." =%}
zzz
Blah blah `link text <{{ foo }}>`_ more text.
zzz
<%- endtrans %>
eyedeekay
Yeah I think that should work
zzz
would that work?
zzz
ok I think that's going to be the winner for partial link text, if it works, and zab's is the winner for full link text
zzz
I'll give it a try when I write the draft and we'll see
zzz
I'll also update the README
robin
Is there a maximum of one SAM PRIMARY SESSION per HELLO?
mesh
Anybody else having trouble reaching exit.stormycloud.i2p?
robin
I just tried a .com address and it timed out
mesh
robin: can you reach exit.stormycloud.i2p?
mesh
the service is literally unreachable to me. It's been unreachable for a good chunk of the last 24 hours.
robin
There is something THERE because it returns a 500 error
mesh
I don't get a 500. I get the I2P "Website Unreachable"
robin
I was trying the top address, not the /outproxy one
mesh
Anyways I guess it's not just me having problems with the Stormycloud outproxy. That's a little comforting.
robin
The /outproxy.html one does timeout
mesh
It's too bad. Up until today stormycloud has been pretty great. Fast and totally uncensored. Anyways we need many more outproxies.
robin
It is working now
robin
My test is to go to whatismyip.com and if it says "Texas", it is working
mesh
it is working now
zzz
eyedeekay, can you update your javadoc site or even better put it on a cron job? idk.i2p/javadoc-i2p/overview-summary.html
eyedeekay
Yeah I'll get it scheduled
zzz
thanks ))
StormyCloud
mesh: Checking on it now
StormyCloud
mesh: I just checked and outproxy is working for me
StormyCloud
cake.i2p/file/1nr5PVCdmY_pkjhP0f68UYrKyyoeAvO46dAhgZLo0_NpRKFbp8yL/Screenshot%202022-09-12%20125638.png
eyedeekay
zzz that's one of those bespoke scripts that's going to have to stick around after I get everything else moved onto a static-site generator and automated properly
eyedeekay
Updated it and checked in the script though
eyedeekay
apeace I updated the CONTRIBUTING.md document today, any suggestions? github.com/go-i2p/go-i2p/blob/master/CONTRIBUTING.md
zzz
yeah since it's liked throughout the specs on our site it should be kept reasonably up to date
robin
eyedeekay speaking of specs, my reading of the SAM3 spec implies that there can only be one PRIMARY session per HELLO, since they share the same TCP socket. . Is that correct?
eyedeekay
IIRC that is the case, although it's been a while since I tried it in practice
robin
Ok.
robin
The spec does not come right out and SAY this, but reading between the lines it makes sense.
eyedeekay
Let me see if I can write a quick test for it and make sure
eyedeekay
Just fiddling with telnet right now it looks that's how it works, I'll get more details this evening though
zzz
eyedeekay, still showing 1.6.1-7
eyedeekay
github side updated, eepsite side isn't yet, I'll trigger it manually
dr|z3d
got a deadlock for you, zzz.. logs to follow.
eyedeekay
the eepsite side is automated, but the github side isn't
eyedeekay
and the eepsite side follows the github
eyedeekay
There it goes
dr|z3d
this deadlock might have already been addressed, don't remember, logs from 2nd Sept. cake.i2p/view/1MAf5nxioL_NHDizQJjIz7eo239MQlPUFLRhhSXC9_4xoUFTrf80/deadlock
zzz
dr|z3d, what version was it running at that time?
dr|z3d
looks like 4.
zzz
we've fixed 2 deadlocks recently, on 8/6 and 8/25, the latter was probably -1 ?
zzz
4 threads, will take a minute to untangle
dr|z3d
ok, so that's maybe new then. didn't notice it earlier, sorry.
zzz
np, thx
mesh
h
mesh
/win 3
zzz
dr|z3d, I think it's a new one, will look more closely tomorrow
dr|z3d
ok zzz
dr|z3d
the deadlock detector.. any more thoughts on making that auto-restart the router, on a parallel note?
dr|z3d
I was also thinking it might tie in with the router watchdog.. if there are x watchdoig barking events within a given period, also auto-restart.