#dev 2017-12-14

2017-12-14 UTC
#
gRegorLove what's the websub feed Loqi subscribes to for indieweb mentions? I'd like to figure out why my mentions to gregorlove.superfeedr.com don't show up
#
gRegorLove Found https://documentation.superfeedr.com/trackers.html
#
Loqi Trackers
chrisaldrich, tantek, snarfed, [eddie] and tbbrown joined the channel
#
tbbrown gRegorLove++
#
Loqi gregorlove has 36 karma in this channel (197 overall)
[manton], tantek and renem joined the channel
#
gregorlove.com edited /person-tag (+150) "/* IndieWeb Examples */ +me" (view diff)
tantek, tbbrown, [miklb], renem and [mikirk] joined the channel
#
tantek hey aaronpk - cassis.js updated with auto_link for @-domains if you want to try it out
#
tantek do you need a new release?
#
aaronpk Nice. It'll take me a while to load that into my head. Might have to wait a while
#
tantek or do you take from latest?
#
tantek ok
#
aaronpk I don't remember actually lol
#
tantek this update was inspired / forced by your $ request
#
tantek wherein you had to separately type in-reply to tantek.com and then start your note with @tantek. whereas you should be able to start your note with @tantek.com and Quill should automatically turn that into a reply-to tantek.com
#
tantek filing issue
#
tantek https://github.com/aaronpk/Quill/issues/86
#
Loqi [tantek] #86 starting a note with @-domain should auto set reply-to that domain
[asuh] joined the channel
#
tantek aaronpk well the auto_link update to link @-domain-name is my first submission for the 2017-12 indieweb challenge
#
aaronpk 🎉
#
tantek only missed doing it yesterday because I fell asleep before I could
[eddie] and renem joined the channel
#
tantek we're discussing what to do with @-names in replies from webmentions
#
tantek Manton showed an example on aaronparecki.com which had "@aaronpk" replies which came from micro.blog
[manton] joined the channel
#
[manton] This page is a good example: https://aaronparecki.com/2017/11/02/19/
#
Loqi [Aaron Parecki] Well this was an effective demonstration of why a centralized Twitter is a problem. #indieweb https://twitter.com/twittergov/status/926267806261407744
#
[manton] The Micro.blog replies start with aaronpk and the Brid.gy ones don't. Which makes sense... But I was just kind of thinking about the inconsistency and whether Micro.blog should remove that first @-username from the content.
#
aaronpk that's a tricky one
#
aaronpk it's a recent change of twitter that they support replies without mentioning the username now
#
aaronpk I think I like that, and i've started to do it myself when i'm replying to indie posts too
#
loqi.me created /GeneratePress (+92) "prompted by gRegorLove and dfn added by gRegorLove" (view diff)
#
[manton] It wasn't clear to me on first glance if it was a side effect of that change, or if Brid.gy was being smart about detecting usernames.
#
aaronpk i'm not sure actually
#
gregorlove.com edited /GeneratePress (+3) "link dfn" (view diff)
j12t joined the channel
#
[miklb] bridgy seems to know when it’s a reply and doesn’t append the @user to your tweet reply to the content. iirc it did before the change
#
[miklb] s/to/in/
#
gRegorLove When you @-reply someone on Twitter, the API adds everyone in the canoe based on the tweet you're replying to.
#
aaronpk not anymore
#
gRegorLove Bridgy itself isn't doing anything special afaik
#
aaronpk not to the actual tweet text
#
gRegorLove Right, sorry that's what i meant
#
gRegorLove But if you're @-replying you can think it's going to one person, when it will go to everyone in the canoe.
#
gRegorLove I've stopped putting usernames at the start of my replies as a result
#
aaronpk I still need to patch silo.pub to not add the username to the text for that reason
[mikirk] and tbbrown joined the channel
#
@pmugnj Testing webmentions. Hello @khurtwilliams (twitter.com/_/status/941143519393210369)
eli_oat, [aaronpk] and [cleverdevil] joined the channel
#
@jackysee 1)有的話貼原blog link 2)不介意的話可以貼全文，加"原文貼於xxx"讓有興趣讀者跟link 自host可減少fb玩針對，貼全文也只是RSS重現，而不在意要削弱fb影響 如想回收回應可用webmention+bridgy但需要一點技術 wordpress上可用social media auto poster幫手 http://tech.azuremedia.net/2017/12/14/7545/ (twitter.com/_/status/941155392733831168)
[asuh], [jackysee], genet, tbbrown, treora and wagle joined the channel
#
xavierroy.com edited /read (+23) "/* Xavier Roy */" (view diff)
#
xavierroy.com edited /User:Xavierroy.com (+908) (view diff)
tantek, KartikPrabhu and cweiske joined the channel
#
@jackysee @siuying 用番microformat,webmention同micropub 理論上可以同其他獨立網站溝通 more on https://indieweb.org/micro.blog (twitter.com/_/status/941212138642821120)
[pfefferle], KevinMarks, tantek, KartikPrabhu and dansup joined the channel
#
petermolnar re YAML I certainly agree there should be a Strict YAML or something
#
petermolnar I mean http://www.yaml.org/spec/1.2/spec.html tells you about aliases, anchors, blocks, etc. - most of them are completely unneeded and most of yaml usage is for ini 2.0
#
petermolnar btw, isn't that TOML we're talking about? except, unfortunately, that is not a drop-in replacement
#
Zegnat What is TOML?
#
Loqi TOML is Tom's Obvious, Minimal Language, a configuration file format that should be called ini 2.0 https://indieweb.org/TOML
tantek joined the channel
#
vanderven.se martijn edited /TOML (+42) "Linkify" (view diff)
#
petermolnar although there is truth in the notion that it might just be simpler to use json
#
sknebel I'm not a fan of nested objects/lists of objects in TOML syntax
#
sknebel and yes, familarity is also important if it's for humans to touch
#
sknebel there are a few attempts at further tying down the details of JSON, one of those might be interesting, assuming there's adequate parser support (subset of JSON is nice as a concept, but if the parser isn't enforcing it because you use a normal JSON parser you loose some benefits again)
#
sknebel some of the unclear bits in JSON are quite ugly
#
sknebel see having to put 64-bit numbers in strings to make sure they get across or the recent couchDB security fail due to differing handling of duplicate keys
#
petermolnar https://github.com/crdoconnor/strictyaml/blob/master/FAQ.rst#what-features-does-strictyaml-remove
#
Zegnat Is the ECMA JSON standard also unclear? I know there were complaints about json.org as a spec, but thought that was why the ECMA doc was made?
#
Zegnat I wonder if strictyaml still allows arrays as dictionary keys. That is a YAML feature that I have never understood. Not listed as a thing they remove.
deathrow1 joined the channel
#
sknebel Zegnat: ECMA neither addresses duplicate keys nor does it specify how parsers are supposed to treat numbers in edge cases
#
Zegnat Ah, missed opportunity :(
#
Zegnat Even without ambiguities, there were enough problems abound anyway: http://seriot.ch/parsing_json.php
#
sknebel (for reference, the couchDB bug: https://justi.cz/security/2017/11/14/couchdb-rce-npm.html - the part of the system that validates permissions and the one enforcing it handled duplicate keys differently, so you could tell the second one you're an administrator without the first one stopping you)
tantek joined the channel
#
Zegnat Hmm, I wonder if I could teach XRay to parse Tumblr blogs
#
Zegnat Even on own domains with custom HTML themes, it looks like Tumblr may be injecting some SCRIPT tags at the top of the HTML. So they are recognisable.
#
Zegnat files that under interesting thoughts for later
[pfefferle], KevinMarks, jjuran, tantek and KartikPrabhu joined the channel
#
Zegnat Interesting to see that Twitter Lite has added Web Share Target API to their manifest. Are there any mobile browsers that support it? https://mobile.twitter.com/manifest.json
#
sknebel Zegnat: not officially as far as I know, doesn't mean there are no test implementations around
#
sknebel Zegnat: https://bugs.chromium.org/p/chromium/issues/detail?id=668389&desc=2 <- under development behind a flag in chrome it seems
#
sknebel or per this, in chrome dev/canary on android: https://wicg.github.io/web-share-target/demos/sharetarget.html
[kevinmarks] and tantek joined the channel
#
sebastiangreger.net edited /Akismet (+287) "alternative WP plugin compliant with strict privacy laws" (view diff)
genet, tantek, eli_oat, tbbrown, tantek_ and [miklb] joined the channel
#
[miklb] just came across this if anyone has an open project and is looking to better outline how to contribute https://github.com/nayafia/contributing-template
#
Loqi [nayafia] contributing-template: Template for writing your own contributing guide
#
tantek what is contribute
#
Loqi https://indieweb.org/Contribute
#
[miklb] feel free to add it tantek
#
tantek oh that page is about contributing to IWC
#
tantek from the 2011-2014 era
#
tantek.com edited /Contribute (+263) "update page with dfn, expansion, more evergreen links" (view diff)
genet, j12t, petermolnar and [tantek] joined the channel
#
[tantek] Wow. ICYMI https://twitter.com/patreon/status/941005364882845696
#
@Patreon Creators & patrons: We've heard you loud and clear. We’re sorry, and we’re not rolling out the fees change. https://blog.patreon.com/not-rolling-out-fees-change/ (twitter.com/_/status/941005364882845696)
snarfed, genet, tantek, eli_oat, [asteres], sgreger, [miklb], KevinMarks and tbbrown joined the channel
#
ascraeus.org edited /Patreon (+402) "/* Funding Change 2017 */" (view diff)
gRegorLove joined the channel
#
dgold i should try and do more wikistuff
#
Loqi definitely
John____ joined the channel
#
gregorlove.com edited /session_facilitating (+132) "bold dfn, note you don't have to be an expert" (view diff)
[eddie] joined the channel
#
[eddie] !tell aaronpk When compass says “before - specify a full timestamp” how is a full timestamp formatted?
#
Loqi Ok, I'll tell them that when I see them next
#
loqi.me created /Colloq (+175) "prompted by gRegorLove and dfn added by gRegorLove" (view diff)
#
gregorlove.com edited /Colloq (+42) "link dfn, Nuremberg 2018" (view diff)
genet, KartikPrabhu, tantek and eli_oat joined the channel
#
[eddie] !tell aaronpk: Nevermind. I realized its the same as the ‘localtime’ attribute returned in the geocode
#
Loqi Ok, I'll tell them that when I see them next
#
GWG snarfed, should I space out my PRs to give you a break?
#
snarfed GWG: lol no, you're fine. thanks though
#
GWG snarfed, I try not to take advantage
#
snarfed GWG: no worries! generally reviewers can throttle on their end, it's not your job to choose that for them
#
GWG snarfed, I try to jump around to evenly distribute things.
#
GWG Either way, I am excited about using Micropub. I owned a check-in yesterday and a photo today
#
snarfed GWG++
#
Loqi gwg has 18 karma in this channel (289 overall)
#
GWG snarfed, you know I would return the favor if created something you used.
#
snarfed lol yes! appreciate it
#
snarfed i use syndication links!
#
GWG snarfed, have you updated recently?
#
snarfed probably?
#
GWG snarfed, I just added a lot of icons. Khurt wanted untappd.
tantek joined the channel
#
GWG Basically, scripted adding new ones. Going to do the same to rel me
#
aaronpk [eddie]: I think many formats are accepted but iso8601 is the best
#
Loqi aaronpk: [eddie] left you a message 1 hour, 17 minutes ago: When compass says “before - specify a full timestamp” how is a full timestamp formatted?
#
Loqi aaronpk: [eddie] left you a message 35 minutes ago: Nevermind. I realized its the same as the ‘localtime’ attribute returned in the geocode
#
www.svenknebel.de edited /posts (+74) "/* Pieces of a Post */ link previews" (view diff)
#
tantek now that I'm doing a bunch of replies to multiple folks (@-names canoe in Twitter), I'm reconsidering showing the @-names as part o the reply
#
tantek rather, considering moving them to the reply-context
#
tantek does anyone do this?
#
aaronpk Most of the time when I reply to someone's indie post i don't include their @-name in the text of the reply
#
aaronpk I'm thinking about making it more explicit and consistently not including the @-name in the text for anything
#
tantek I'm thinking of just moving the display
#
tantek that is, still accepting it as user-entry as part of a reply (and setting user expectation that it is a reply to a specific person), but then just moving that part of the content to the reply-context
#
gRegorLove I show the display name in the reply-context, not the @-name. Not doing multi-replies, though.
#
gRegorLove Well, I guess Twitter is doing the multi-reply for me, but on my site I only show the reply-context for the tweet I'm replying to.
tbbrown joined the channel
#
GWG How do I decide if I've done enough for a major version bump?
#
Zegnat At every update? ;)
#
gRegorLove Using semver?
#
gRegorLove I think usually if it's breaking changes or otherwise majorly refactored, then major version
#
GWG I was trying to consider it by feature
#
gRegorLove If it's backwards compatible, I'd keep it as a minor version bump
#
gregorlove.com edited /semver (+6) "full name, link dfn" (view diff)
#
Zegnat I like semver. BC are major bumps, else not. But I see multiple projects who then start shoving feature updates on those same major bump updates, which in my mind just needlessly delays the features from launching. Because they want a major bump to feel like something meaningful and big. Which IMHO semver isn’t about.
#
gregorlove.com moved /semver to /Semantic_Versioning "canonical name"
#
GWG gRegorLove: You know a lot about the PHP-MF2 parser, correct?
#
gRegorLove A fair bit. I try. :)
#
GWG How does it handle fragments?
#
gRegorLove I'm not sure it does, but I haven't experimented with that. Have an example link?
#
gRegorLove Like "only parse the h-entry that matches this fragment?
#
gRegorLove "
#
GWG http://boffosocko.com/2017/12/11/an-update-to-read-posts-for-physical-books/#comment-40026
#
GWG This one
#
gRegorLove pin13 is up to date, so should confirm
#
gRegorLove !mf2 http://boffosocko.com/2017/12/11/an-update-to-read-posts-for-physical-books/#comment-40026
#
Kaja_ https://php.microformats.io/?url=http%3A%2F%2Fboffosocko.com%2F2017%2F12%2F11%2Fan-update-to-read-posts-for-physical-books%2F%23comment-40026
#
aaronpk I implemented that in XRay before it hands it off to the parser
#
gRegorLove Yeah, I was just wondering if the parser *should* handle fragments. Isn't that supposed to be for clients only? idk
#
GWG Bad example
#
gRegorLove But to answer your question, GWG, no, php-mf2 isn't handling the fragment. parses the entire page there.
#
GWG Chris doesn't have u-comment though
#
gRegorLove I don't think that matters
#
GWG Should it?
sgreger joined the channel
#
GWG aaronpk: Xray gives a null on that link
genet joined the channel
#
Zegnat I don't think it is part of the mf2 spec. Instead I think it is up to the consumer to find the right parsed object. Much like how you might have to sort through multiple h-card on a page to find the one with the matching u-url.
#
KartikPrabhu fragment handling could be an "extra" feature of some parser
#
Zegnat So a consumer would walk through all h-entry objects and check for a u-url that includes the fragment.
#
Zegnat Where consumer is something consuming the mf2 parser output
#
KartikPrabhu if a consumer already knows that the relevant information is in some fragment of the page, it would be more efficient to not parse the entire page
#
gRegorLove GWG, chrisaldrich: Tangent - Know where the h-as-reply is coming from in Chris' page there? That should be removed from IndieWeb plugins afaik
#
GWG gRegorLove: I know where it is coming from
#
GWG Semantic Linkbacks
#
Zegnat KartikPrabhu if the consumer knows beforehand it can do what xray does and only supply the specified part of the DOM to the mf2 parser.
#
gRegorLove I don't know how XRay is doing it, but I suspect it might be null because the vcard isn't on the same element as that fragment ID? It's a child of it.
#
Zegnat IIRC xray parses the DOM looking for the element with the specified ID and then gives that element and all its children to the mf2 parser. Thus only parsing part of the page or fetched.
#
aaronpk That's correct
#
gRegorLove GWG: Filing an issue on Semantic Linkbacks
#
GWG So, that would be my problem, not PHP-MF2
#
GWG I have to hand the right piece over
#
Zegnat The way mf2 parsing is currently implemented: yes.
#
aaronpk It seems like that'd be a good feature for libraries to implement but it's not part of the parsing spec
#
gRegorLove As it is now, I don't think fragment handling would go into php-mf2, unless the spec is updated. Otherwise it might be good for a helper lib, like Barnaby's php-mf-cleaner.
#
gRegorLove Not opposed to it being in the parsing spec, but that's a discussion for #microformats
#
GWG gRegorLove: I was just wondering for now.
#
gRegorLove Hm?
#
GWG About fragments
#
GWG Not sure what is right.
#
Zegnat I personally also think it is good that it isn't in the parsing spec. Mf2 parsing is now only concerned with turning HTML into objects. Not where the HTML is coming from, what subpart of it that needs to be read, or even if the HTML is valid.
#
gRegorLove Filed: https://github.com/pfefferle/wordpress-semantic-linkbacks/issues/142
#
Loqi [gRegorLove] #142 Recommend removing h-as-* classes
#
Zegnat Makes things a lot lighter on the mf2 parsing spec.
#
aaronpk Agreed
#
Zegnat And with that, I am off to dream land. Have a good one all!
eli_oat joined the channel
#
gRegorLove GWG: You could create an Mf2 object with the URL, run an xpath query with Mf2\query() to get the DOMElement, then call Mf2\parse() with the $DOMelement parameter to limit parsing to children of that element: https://github.com/indieweb/php-mf2/blob/master/Mf2/Parser.php#L1262
#
gRegorLove So most of the pieces are there in php-mf2, just a few lines of code "around" it to get the desired effect.
Kongaloosh and eli_oat joined the channel
#
gRegorLove Nevermind, that doesn't quite work
John____, tantek and [miklb] joined the channel