#dev 2017-08-15

2017-08-15 UTC
rhiaro, myfreeweb, [cleverdevil], plindner, mindB, schmarty, zoglesby, eli_oat[m], TheGillies, AlanPearce[m] and tantek joined the channel
#
tantek
is going through websub.rocks
#
tantek
as a publisher
#
tantek
"Found the Hub!"
#
tantek
"Subscription Active!"
#
Loqi
Tantek Çelik
#
tantek
"Waiting for notifications"
#
tantek
looks like I have to publish something to test the subscription / notifications
#
tantek
oh hey that worked
#
tantek
but now what
#
tantek
I don't understand what I'm supposed to do to "finish" the test - it got some notifications
#
tantek
but still says "Waiting for notifications" - for more notifications I guess?
#
tantek
presumably I don't have to do any more
#
tantek
I'm going to click Publisher to see what else to do
#
tantek
and that just asks me to Enter the URL again
#
tantek
and we're back to Found the Hub, Subscription Active, Waiting for notifications
#
tantek
so I guess I'm done?
#
aaronpk
it's more of just a tool
#
aaronpk
compared to webmention.rocks which builds the implementation report for you
#
tantek
ok I'll do that at home
#
[miklb]
what is node
#
Loqi
Node.js is a platform for running javascript applications and runs on all modern operating systems https://indieweb.org/node
eli_oat and [tantek] joined the channel
#
[miklb]
what is reacji
#
Loqi
reacji is an emoji reaction, the use of a single emoji character in response to a post, introduced as a feature by Slack[1] https://indieweb.org/reacji
#
[miklb]
so, a like :thumbsup: from FB is a like-of but any other reactions are just single emoji replies? Am I following that correctly?
#
[tantek]
Via Bridgy backfeed? Yes that sounds right.
#
[tantek]
Not sure who is displaying reacjis tho
#
[miklb]
yes, sorry I didn’t specify via bridgy
#
[miklb]
I’m aiming to add support in WordPress via a plugin.
#
[miklb]
I posted this in #wordpress but it’s probably general enough if anyone has thoughts: If I’m doing the reacji thing, and I filter all incoming replies for single emoji, I should just save it with commentmeta, with the short_name or the whole array?
#
[tantek]
Great question. Hence I was wondering who else is displaying reacjis and how
#
[tantek]
Figuring they have an answer for how to store them
tantek joined the channel
#
[miklb]
the answer to my question I believe is the whole array as will need to be able to also query skin_tone. I hadn’t scrolled far enough down to see that is also included.
#
aaronpk
Emoji are weird and honestly I'd just store the raw bytes and hope that OSs can always continue to render them
#
[miklb]
I’m looking at how to group them
#
aaronpk
You want to consider skin tone variations as the same emoji?
#
aaronpk
You could find the skin tone and other variation bytes and remove those
#
[miklb]
so my thinking was if I store the short name and skin tone, I can create arrays of all the short name and then to do a facepile, then loop through each one and a loop through an array of the skin tones for that emoji.
#
[miklb]
I wasn’t planning to to consider them all the same.
#
aaronpk
Oh like sorting?
#
[miklb]
yes. for facepiles to go along with likes and reposts
#
aaronpk
hmm I think if you sort in byte order it'll work
#
[miklb]
I should probably read up on byte order it sounds like. I was just reading the examples in your emoji detection docs.
#
[miklb]
specifically, the array output when checking the is_single_emoji
tbbrown joined the channel
#
tantek
aaronpk, it appears I'm sending a "hub.url" param in my websub POST notification instead of a "hub.topic" param
#
tantek
(inspecting my code). and websub.rocks didn't complain (either about presence of hub.url or absence of hub.topic)
#
tantek
seems to be ok per spec? that is, nothing against submitting hub.url, and nothing in the spec says how or why a publisher should set hub.topic
#
tantek
so I'm having trouble understanding how to answer the last question of the implementation report: "If your publisher uses an external hub, please describe how the publisher notifies the hub of new content:" - or rather, what's a good / the right answer? do I need to change my code to say hub.topic instead of hub.url?
#
tantek
other than that I have an implementation report ready to commit
#
tantek
for websub
Kongaloosh_, jonnybarnes, snarfed, tantek and cweiske joined the channel
#
Zegnat
sknebel your implementation for 7.2, at least for your page, seems correct? Why the doubt?
j12t and jeremycherfas joined the channel
#
vanderven.se martijn
edited /authorship (+150) "/* Questions */ Sturdy-backbone supports authorship algorithm for h-event"
(view diff)
#
jeremycherfas
Morning zegnat
#
jeremycherfas
Mind if I ask you a non IndieWeb question?
#
Zegnat
Not at all
#
jeremycherfas
Can you see any reason why the links on this page https://www.jeremycherfas.net/blog do not seem to be operating correctly. I’ve checked the source, and they look to me to be correctly formed. But they just don’t work.
#
cweiske
which links?
#
cweiske
the links in the blog posts?
#
cweiske
the blog posts are overlaid by the following h1
#
Zegnat
cweiske beat me to it
#
cweiske
actually, the next list-item already covers the previous content
#
jeremycherfas
So it is the size of the H1 div, and it is on top of the posts, but transparent?
#
cweiske
because the list-blog-padding has a float:left
#
Zegnat
jeremycherfas: you can open the inspector in your browser, and use the element picker. When you then try to hover the links you will see it highlighting a different element, because that is on top
#
jeremycherfas
OK. Thanks a bunch. It was driving me crazy, but now it should be easy enough to fix.
#
jeremycherfas
Zegnat: I thought that was just me being inaccurate in where the mouse was !
#
cweiske
if you remove that float:left from list-blog-padding it works again
j12t joined the channel
#
cweiske
there is also another problem on your site:
#
cweiske
blog:13 Mixed Content: The page at 'https://www.jeremycherfas.net/blog' was loaded over HTTPS, but requested an insecure stylesheet 'http://yui.yahooapis.com/pure/0.6.0/pure-min.css'. This request has been blocked; the content must be served over HTTPS.
#
Zegnat
and cweiske’s fix works for me too, removing the float:left :)
#
jeremycherfas
Oh yes. I see that now. Guess I better fix that too.
#
jeremycherfas
Thanks, both. It is silly how the little things are so hard to see.
#
Zegnat
Also an experience thing, jeremycherfas. When I use the element picker and it is finicky I know there is a CSS problem, I know that because of experience. If not for experience I might also have thought it was me using it wrong or pointing my mouse wrong, and then spending hours looking the wrong way.
#
sknebel
Zegnat: due to the redirect on my homepage
#
sknebel
The algorithm used the URL that was fetched, not the URL the redirect ends up at
#
Zegnat
Aah. So the question becomes if the algo should match against resolved (after redirect) URL or the URL mentioned in the author property?
j12t joined the channel
#
jeremycherfas
Fixed both those things, and while I was about it updated pure.css
#
sknebel
Zegnat: yes (or both or all steps inbetween or ...) - my site should work with 7.3 as well, that's why I was surprised it triggered 7.2 already
barpthewire joined the channel
#
sknebel
I remember with some crawling project that also caused issues (is it an authorative h-card or not?). I probably should just not redirect that page (although ability to deal with redirects also might be useful for http->https
#
Zegnat
Not a bad idea to incorporate rules for redirects
#
Zegnat
I just haven’t seen it go wrong because of it yet
#
jeremycherfas
This is disappointing. A spammer has evaded my honeypot field, probably by being a human being. There must be another way to keep bad things at bay without inconveniencing genuine people.
#
Zegnat
Not if the spammers are human beings. Unless you filter on content with an actual spam filter
#
jeremycherfas
I accept the need to be active myself if there are human beings at work. Unfortunately, the Grav plugin has no moderation facilities. Except, probably, editing the datafile directly. Which is a pain.
j12t joined the channel
#
Loqi
[Perlkonig] grav-plugin-webmention: Implements the Webmention protocol (with the Vouch extension) for the Grav CMS
#
cweiske
the command line interface also does not support deleting incoming webmentions
#
cweiske
maybe a feature request is appropriate
#
jeremycherfas
I gave up on that plugin in the end to use voxpelli’s heorku app, and there is no problem, so far, with webmentions. It is “ordinary” comments that are giving me a problem.
#
jeremycherfas
You have also reminded me that I need to think about a little script to allow people to submit the URL of a comment or reply on their own site.
#
cweiske
you actually don't need a script at all
#
cweiske
a simple POST <form> is enough
#
cweiske
if you prefill the target
#
jeremycherfas
I’ll come back to you when I get a little further … :)
tbbrown, [miklb], j12t, snarfed, tantek, dougbeal|iOS and [cleverdevil] joined the channel
#
tantek
good morning #indieweb-dev (evening to the UK / EU folks)
cweiske joined the channel
#
tantek
hey cweiske does your WebSub hub require hub.topic in the POST from a publisher, or does it accept / handle just hub.url?
#
cweiske
let me check
dougbeal|mb1 joined the channel
#
cweiske
it uses hub.url and falls back to hub.topic if hub.url is not there
#
cweiske
but hub.topic is not required
#
tantek
that sounds right. but something changed at some point PuSH -> WebSub because I cannot find "hub.url" in https://www.w3.org/TR/websub/
#
Loqi
[Julien Genestoux] WebSub
#
tantek
the WebSub spec only mentions hub.topic now, not hub.url. so I'm wondering if I have to switch to using hub.topic to be compliant / interop with future potential hub implementations
#
aaronpk
the websub/pubsubhubbub spec never specified the publisher->hub relationship https://www.w3.org/TR/websub/#publishing
#
aaronpk
so the closest thing to a spec for that was what google and superfeedr implemented
#
loqi.me
edited /uberspace (+157) "/* See Also */ new section"
(view diff)
#
www.svenknebel.de
edited /Google_Analytics (+514) "IP anonymization feature"
(view diff)
#
tantek
aaronpk, as a publisher then I feel like I don't know what is the right thing to do
#
tantek
and it feels wrong to be using hub.url when its not even in the spec
#
cweiske
just say that as a publisher, you were doing it wrong before :)
#
aaronpk
i've wanted to publish an extension that specifies the publisher->hub relationship for public hubs
#
tantek
cweiske: I'm ok to accept that, I'd just like to be able to point to an answer in case someone else asks the same question in the future
#
aaronpk
the reason for not putting it in the core spec was that some hubs are built in to the publishing software and will never need that API
#
tantek
aaronpk: I'm ok with that too. I'm just worried that no mention of hub.url means using it is non-compliant, or rather, not something you can interoperably depend on
#
aaronpk
i wonder if it'd be possible to add it to the spec as an optional feature for public hubs
#
aaronpk
or rather, external hubs (public/private distinction is not important)
#
tantek
so I'm seriously considering switching to hub.topic in my implementation, but then that breaks our spec design intention of not asking implementations to change to conform to the spec
#
tantek
so I don't know what the right answer is
#
cweiske
but it's a living spec!!
#
tantek
cweiske: if only!
#
cweiske
it was never 1.0, only 0.3 and 0.4
#
cweiske
so breaking changes are ok
#
aaronpk
i always thought the hub.url parameter didn't make sense because that same value is called hub.topic in the spec
#
tantek
aaronpk: where was the hub.url param ever defined?
#
tantek
did I just get it from copy/pasta of some code example?
#
tantek
both of those mention hub.url, but DO NOT mention hub.topic!
#
tantek
uh, that's the opposite of what the spec does
#
tantek
so now I'm *really* confused
#
aaronpk
no the spec doesn't say *anything* about how to publish
#
tantek
which is what I'm reading through now
#
tantek
"A conforming publisher MUST advertise topic and hub URLs for a given resource URL as described in Discovery."
#
aaronpk
right, nothing about pinging the hub
#
tantek
this sounds like a need for an explicit NOTE: saying Pub->Hub not defined, and or, for example, some hubs ask for "hub.url" as an informative example.
#
snarfed
so i'm currently writing a little proxy that injects a webmention endpoint into any web page, kind of like https://cors-anywhere.herokuapp.com/ etc
#
Loqi
[snarfed] bridgy-federated: A service that bridges IndieWeb and federated social networks: ActivityPub, OStatus, etc.
#
snarfed
just mentioning it in case anyone has any thoughts...
#
aaronpk
oh nice. when cloudflare announced their app store thingy i was thinking about doing that for cloudflare
#
aaronpk
so you'd check a box in the control panel to enable receiving webmentions
#
[cleverdevil]
That's a great idea.
#
Loqi
[snarfed] #420 extend blog webmentions feature to all web sites
#
sknebel
I don't think you can do that with th ecloudflare thing
#
sknebel
limited to JS and CSS injection afair
#
Loqi
[tantek] #117 informative NOTE pub->hub protocol unspecified, and examples
#
aaronpk
sknebel: i thought it let you add headers too
#
tantek
and with that I've submitted my publisher implementation report: https://github.com/w3c/websub/blob/master/implementation-reports/PUBLISHER-falcon.md
j12t, [miklb], tbbrown, j12t_, j12t__ and gRegorLove joined the channel
#
tantek.com
edited /Planning (+82) "update a few dates"
(view diff)
#
tantek
heh, /Planning edits should be white-listed to #indieweb-meta
[manton], tbbrown, j12t, snarfed, jonnybarnes, gRegorLove and j12t_ joined the channel
[miklb] joined the channel
j12t, tbbrown, j12t_ and KartikPrabhu joined the channel
#
loqi.me
created /awstats (+137) "prompted by tantek and dfn added by petermolnar"
(view diff)
#
cweiske.de
edited /website-analytics (-53) "/* Log Analyzers */"
(view diff)
#
tantek
awstats << analytics
#
Loqi
ok, I added "[[website-analytics]]" to the "See Also" section of /awstats
#
loqi.me
edited /awstats (+41) "/* See Also */ new section"
(view diff)
#
tantek
Loqi, "See also" vs. "See Also"
#
tantek.com
edited /awstats (+3) "cap See Also so Loqi can find it"
(view diff)
#
tantek.com
moved /awstats to /AWStats "canonical capitalization per site"
#
www.aaron-gustafson.com
edited /POSSE (-8) "Trying to fix"
(view diff)
[aarongustafson] and tantek joined the channel
#
tantek.com
moved /Template:aaron_gustafson to /Template:aarongustafson "use username/ircnick for templates - no spaces :)"
#
emmahodge.org
edited /Planning (+1) "/* NYC */"
(view diff)
#
tantek.com
edited /Planning (+202) "/* NYC */ update and add more dates"
(view diff)
#
tantek.com
edited /Planning (-5) "lowercase user template names, just emma :)"
(view diff)
#
tantek.com
edited /Planning (+23) "/* NYC */ end of Oct, second to end of Sept"
(view diff)
#
tantek.com
edited /Planning (+85) "/* NYC */ Venue! we have a possibility"
(view diff)
#
werd.io
edited /Planning (+238) "/* NYC */"
(view diff)
#
martymcgui.re
edited /Planning (+195) "/* NYC */ update my prefs"
(view diff)
#
ben.thatmustbe.me
edited /Planning (+64) "/* NYC */"
(view diff)
#
tantek
^^^ merge conflict!
#
martymcgui.re
edited /Planning (+195) "/* NYC */ re-add my updates"
(view diff)
j12t, KartikPrabhu, snarfed, wagle, tbbrown and GWG joined the channel