#dev 2019-01-01

2019-01-01 UTC
#
@ChrisAldrich
↩️ This sounds a lot like the Microsub spec which abstracts and separates the parsing and displaying of content. There are already several separate server and reader implementations if you're interested in tinkering. https://indieweb.org/Microsub https://boffosocko.com/2018/12/31/55739705/
(twitter.com/_/status/1079891569291849728)
eli_oat1, [kevinmarks], snarfed and [tantek] joined the channel
#
Loqi
[tantek]: [schmarty] left you a message 5 hours, 58 minutes ago: looks like your 🕸 links are out of date. possibly you added them very early before I had the URL normalization and hashing down pat? either way: your site is not currently in the 🕸 database. you can log in again to get fresh links.
#
@t
#microformats2 parsing spec updated: http://microformats.org/wiki/index.php?title=microformats2-parsing&diff=66969&oldid=66967 * always parse alt with img src for accessibility info by default https://github.com/microformats/microformats2-parsing/issues/2 Implemented in current version of: * mf2py - behind flag, in use by http://Brid.gy This is a ... http://tantek.com/t4yG5
(twitter.com/_/status/1079984280539942914)
[nick], gRegorLove, leg and barpthewire joined the channel
sebsel, kisik21, jjuran and [kevinmarks] joined the channel
#
Loqi
[Kevin Marks] I updated mention.tech to output webmention.io compatible JSON for mentions eg https://mention.tech/listmentions?target=https%3A%2F%2Ftumblelog.xyz&json=on and modified fluffy's script to show them on https://tumblelog.xyz/ at https://tumblelog.xyz/...
#
kisik21
does mf2util output jf2? and why does jf2 not support multiple authors? (a completely sensible setting for a collaborative document)
#
kisik21
I wasted SO MUCH TIME on supporting multiple authors in pyindieblog templates and now it is kinda sad for me to see that nobody else supports this stuff
#
[kevinmarks]
Hm. Jf2 doesn't inherently not support multiple authors, but it will make a children thing if there are multiple
#
kisik21
children?
#
[kevinmarks]
That's the tradeoff for things not being lists
#
kisik21
if I sort by published as string, am I guaranteed to get the correct ordering of posts?
#
Zegnat
Will it do that even with the strict restriction that, “If present, [author] MUST be an object and it MUST contain at least a name, url, or photo property.”, [kevinmarks]? https://www.w3.org/TR/jf2/#jf2feed_required_fields
#
Zegnat
I seem to recall discussing this with sknebel
#
[kevinmarks]
Hm. That may have got tighter since I wrote my jf2 converter
#
sknebel
[kevinmarks]: what do multiple elements in a document have to do with multiple values for a property?
#
sknebel
and ugh, annoying that jf2 explicitly forbids that :/
#
[kevinmarks]
Because my converter is recursive
#
Zegnat
Outside of Microsub (and I guess XRay?), does anything use jf2?
#
sknebel
wm.io has a mode for it
#
Zegnat
Ah, I didn’t know that, cool
#
Zegnat
wonders what it will return for a webmention with multiple authors
#
sknebel
I'd guess it uses xray, so whatever that does
#
sknebel
which probably is "the first it finds"
#
[kevinmarks]
Do you have a multiple author post?
#
Zegnat
I barely have any normal posts
#
Zegnat
Hoping to change that this year
#
kisik21
Zegnat: I can try to make a test post on my staging instance and send a webmention for it. Just so you know, it uses my deadname though... >< i tried to showcase it to my web design professor in college
#
kisik21
(he was impressed)
#
Zegnat
I don’t use wm.io to receive webmentions. Not sure if there is something akin to a test endpoint
#
[kevinmarks]
If you use mention.tech to send it you can compare what it thinks to webmention.io
#
kisik21
Zegnat: well, I use webmention.io
#
GWG
Good morning all
metbril, jjuran and wagle joined the channel; metbril left the channel
#
aaronpk
oh wow PHP 5.6 is officially EOL now
#
aaronpk
as is 7.0
#
Zegnat
Yep! <3
#
Zegnat
Moved to 7.3 locally about a week ago, going to look at upgrading my server
#
aaronpk
guess I should do the same
[tantek] joined the channel
#
Zegnat
Once I got past the whole 5.6 to 7.0 move, I’ve found the relatively quick iterations from 7.0 through 7.1 and 7.2 to 7.3 pretty easy
#
GWG
I should upgrade too, but my site is still 5.3/5.4 compatible
#
[tantek]
feels like yak-shaving
#
GWG
Probably
#
GWG
I like to keep my yaks presentable
#
aaronpk
doing these small upgrades more often is a good way to avoid getting into the situation we're in with the microformats wiki
#
Zegnat
And if you are in charge of your own server updates, keeps you from ending up running EOL’d libraries that aren’t getting security patches anymore
#
GWG
I am considering what improvements writing 5.3/5.4 compatible code might get me
#
GWG
I have been sticking to 5.2 functionality
#
[tantek]
Zegnat, good reason to minimize # of external libraries you run
#
Zegnat
For sure. But I can’t get around the webserver passing requests over (Fast-)CGI to PHP. As long as my number one dependency is PHP itself, I like to keep it up to date ;)
[kevinmarks] joined the channel
#
sknebel
yeah, PHP is something where I wouldn't want to run something that's not being patched anymore
#
sknebel
+ the 7.x line is quite a bit faster by now, so better performance for little effort
weborganics joined the channel
#
Zegnat
The 5.6 -> 7.0 step can be some effort. But after that I have found it pretty painless.
#
Zegnat
Also happy I haven’t ran into any regex issues with the 7.3 upgrade (so far)
#
[tantek]
I like to keep my PHP use minimal (per CASSIS) so I shouldn't have any version # based dependencies
#
[tantek]
not really into all the fancy features
#
GWG
I want the null coalescing operator
#
GWG
Zegnat, what features were an effort from 5.6 to 7?
eli_oat1 joined the channel
#
Zegnat
They changed chained variables ($a->$b[]) interpertation. I mostly stopped writing code that did that, because it isn’t clear when you read back, but it broke in some places with old code
#
Zegnat
And those were a little tricky to spot. Not hard to fix, just manual finding (or waiting for stuff to break in use).
#
Zegnat
If I had unit tests, I would have seen it fail immediately after running the tests after upgrading. But alas, I’ve only recently started doing testing for real
#
Zegnat
The migration code on php.net should tell you all you need to know though
#
GWG
I was reading that
#
[kevinmarks]
Json style array syntax is somewhere in there
[sebsel] joined the channel
#
Zegnat
Not sure what you mean by that, [kevinmarks]
sebsel joined the channel
#
[kevinmarks]
I mean being able to do $v = {'a':2} instead of $v = array( a => 2)
#
sknebel
Zegnat: curious, did you look for tools to help with the variable order thing? that looks like something tooling could detect or even fix...
#
[tantek]
kevinmarks that's neat - sounds like it may expand the CASSIS subset!
#
Zegnat
That is not a supported syntax anywhere in PHP, AFAIK
#
Zegnat
PHP has $v = ['a'=>2]; which has been there since 5.4
#
Zegnat
sknebel, I didn’t look for tools, no. Nowadays I almost always run some sort of syntax checker so I’d think they might catch it for me. Does sound possible
#
[kevinmarks]
Ah, maybe I was doing it with lists
#
[kevinmarks]
So $v =[1,2,[3, 4]] works, but still need => rather than : for dicts?
#
[kevinmarks]
I am writing php, python and js so having common syntax is nice.
#
[tantek]
kevinmarks, would appreciate your shared snippets when you find such common syntax 🙂
#
Zegnat
[kevinmarks], that is correct
#
Zegnat
The colon rather than fat arrow was proposed way back, but PHP chose to go with the arrows
#
Zegnat
I don’t think they ever brought up switching again after that, but not sure. I think it was discussed somewhere around the time of the short array syntax (without array()) was discussed, so PHP 5.4
#
[kevinmarks]
I suppose I could call JSON to init arrays
#
aaronpk
wtf why is it so hard to get a systemd script for nginx working
#
sknebel
aaronpk: ?
#
[tantek]
what is systemd
#
Loqi
It looks like we don't have a page for "systemd" yet. Would you like to create it? (Or just say "systemd is ____", a sentence describing the term)
#
sknebel
(first question: what's wrong with the one the distro provided?)
#
aaronpk
i'm compiling nginx from source
#
aaronpk
so there isn't one provided
#
sknebel
why that? (although you probably can still use any ready-made one)
jjuran joined the channel
#
aaronpk
the one from the docs doesn't work, the systemctl command hangs even though nginx is started
#
aaronpk
I need to include some custom modules
#
sknebel
share the unit file somewhere and I can take a look
#
aaronpk
i've tried both with and without the -g 'daemon on; master_process on;' part
#
[tantek]
what is systemctl
#
Loqi
It looks like we don't have a page for "systemctl" yet. Would you like to create it? (Or just say "systemctl is ____", a sentence describing the term)
#
sknebel
does it start if you run it manually?
#
sknebel
and create the pidfile in that location?
#
aaronpk
oh let me check that
#
aaronpk
aha it does not
#
sknebel
then set it in config or look up what default location you compiled it with and change the path
#
sknebel
(curious, what extra modules does the indieweb server need?)
#
aaronpk
sknebel++ that was it!
#
Loqi
sknebel has 34 karma in this channel over the last year (88 in all channels)
#
aaronpk
the nginx-push-stream module which is how all the realtime stuff works like for chat.indieweb.org
#
Zegnat
Isn’t the auth thing LASSO depends on also non-default?
#
Zegnat
Hmm, thought that needed a separate flag during building to enable. But I might be misremembering
eli_oat1 joined the channel
#
aaronpk
it's built into nginx but it does need a compile-time flag to enable it
#
aaronpk
not sure how to do that with a distro install either
eli_oat joined the channel
#
sknebel
(side note, lasso helpfully decided to rebrand to "vouch", which isn't confusing at all in our context :D)
#
aaronpk
oh yeah lol
#
[tantek]
what is lasso
#
Loqi
Lasso is a Single-Sign On OAuth client that works with the Nginx auth_request module and supports IndieAuth https://indieweb.org/Lasso
#
sknebel
aaronpk: did you add full IndieAuth support to it or should that read IndieLogin?
#
aaronpk
yeah I guess that's using indielogin's API which also happens to look a lot like IndieAuth/OAuth
[schmarty] and eli_oat joined the channel
#
[tantek]
If you made a commitment in https://indieweb.org/2019-01-01-commitments, and haven't quite finished it, you still have today to do so! (y.t. included 😂) And if you finished yours, mark it done! [jgmac1106] [eddie] [gRegorLove] [sknebel] Jacky [schmarty]
#
[schmarty]
haha oops i should document my finished commitment.
#
[tantek]
yes yes! it helps inspire the rest of us to finish 😄
#
[schmarty]
i have been avoiding writing about it because it feels like a pyrrhic victory. 😂
#
[schmarty]
avoiding writing got my some procrastination energy to work on 🕸, so that's nice, lol
[kevinmarks] joined the channel
#
[tantek]
what is indiewebring
#
Loqi
Indiewebring is part of the great rebirth of of webrings powered by #IndieWeb technology that made 2018 the year of the webring and not the year of the microsub reader https://indieweb.org/indiewebring
#
KartikPrabhu
wow! that is some hype^ :)
#
[schmarty]
[jgmac1106] is all about the webrings lol
#
[tantek]
(shade thrown)
#
[tantek]
[schmarty] now that you've finished your commitment, any plans to start a 100 days project today? https://indieweb.org/100DaysOfIndieWeb
#
[schmarty]
[tantek] no plans for a 100 days challenge at the moment. i am too in flux! it is nice to feel un-blocked, though.
#
[tantek]
hey when did we lose the emoji-pairs and have to go to emoji-triples in indiewebring?
#
sknebel
oh right, I didn't finish the second half of it
eli_oat joined the channel
#
[schmarty]
tantek: i moved from single-emoji to emoji-pairs after i implemented URL normalization (e.g. it was initially possible to log in as both http://tantek.com and http://tantek.com/). from emoji-pairs to emoji-triples when i removed flags from the emoji hash set.
#
Loqi
Tantek Çelik
jjuran joined the channel
#
[schmarty]
currently thinking about yet another pass of emoji removal 😒. not sure what schema i'll choose after that. maybe i'll lock one emoji in place and go back to assigning pairs.
#
[tantek]
ah that makes sense
#
[tantek]
hmm so far I've clicked "right arrow" then back 5 times and not found a site with the indieweb ring nav links 😞
#
[tantek]
the first 3 had zero webring links. the 4th had a microcast or blog webring link pair, but no "indieweb webring" links. the 5th was a dead site
#
[schmarty]
dang, so many ring breakers.
#
[tantek]
ah number 6 finally
#
[schmarty]
i am refactoring stuff just now to make it easier to re-crawl for ring breakers and disable them.
#
sknebel
[schmarty]: did you ever remove ringbreakers?
#
[tantek]
would also be good if people could report them
#
[schmarty]
haha, i am totally gonna make a report page that will check up on them. :}
#
sknebel
(cause I know I logged in when the link first appeared just to see what was going on and was kind of surprised it immediately stored me)
#
[schmarty]
sknebel: my IWC hack day projects often assume extreme good faith 😂
#
sknebel
sure, to be expected, but cases like that might explain why there's relatively many breakers, if the DB was never cleaned
jonnybarnes and eli_oat joined the channel
#
jacky
ooh I totally hit my committments!
#
jacky
lemme write about them
swentel joined the channel
#
[schmarty]
watches 🕸 garden itself for the first time
jjuran joined the channel
#
[schmarty]
82 sites active out of 171!
eli_oat, wagle, [Khurt], eli_oat1, [kevinmarks], [davidmead], tbbrown, [chrisaldrich], [tantek], [benatwork], psaa and maxandersen1 joined the channel