#dev 2018-08-22

2018-08-22 UTC
[iambismark], [matpacker], renem, [chrisaldrich], [cleverdevil], tantek__, ben_thatmustbeme, miklb_ and [eddie] joined the channel; mblaney and miklb_ left the channel
#
GWG Okay, extraction code for manifest with fallback written. Next need to hook it up to something
#
GWG Anyone have redirect_uri on their client? Want to make sure it picks them up.
#
tantek__ does anyone use any event embeds?
#
tantek__ I've seen embeds of both gcalendar and now fb events
#
tantek__ seems like something worth documenting, at least as a silo feature that may be worth replacing with indieweb functionality
#
tantek__ what is an event embed
#
Loqi It looks like we don't have a page for "event embed" yet. Would you like to create it?_ (Or just say "event embed is ____", a sentence describing the term)
#
tantek__ what is a calendar embed
#
Loqi It looks like we don't have a page for "calendar embed" yet. Would you like to create it? (Or just say "calendar embed is ____", a sentence describing the term)
#
tantek__ calendar embeds are more common so I'll go with that
#
tantek__ event embed is /calendar_embed
#
Loqi ok
#
loqi.me created /event_embed (+27) "prompted by tantek__ and redirect added by tantek__" (view diff)
#
tantek__ calendar embed is a feature of [[event]] hosting sites that allows a series of events (sometimes just one) to be embedded into another site, supported by [[silos]] like [[Google Calendar]], [[Eventbrite]], and [[Facebook]].
#
Loqi ok
#
loqi.me created /calendar_embed (+250) "prompted by tantek__ and dfn added by tantek__" (view diff)
#
tantek__ calendar embed << Example embedding of events from a Facebook page: http://www.verdiclub.net/events.html
#
Loqi ok, I added "Example embedding of events from a Facebook page: http://www.verdiclub.net/events.html" to a brand new "See Also" section of /calendar_embed https://indieweb.org/wiki/index.php?diff=51268&oldid=51267
#
loqi.me edited /calendar_embed (+106) "tantek__ added "Example embedding of events from a Facebook page: http://www.verdiclub.net/events.html" to "See Also"" (view diff)
[matpacker] joined the channel
#
miklb_ tantek__ would this be considered an event embed ? https://miklb.com/blog/2018/08/22/4184/
#
Loqi [Michael Bishop] Tonight is our first Open Hack night to pick up where we left off from National Day of Civic Hacking. We’ll be meeting every Wednesday if you can join us this week.
#
tantek__ miklb, definitely, looks like a meetup embed! would be great example to add (or maybe it's worth splitting off /event_embed as a separate thing?)
#
miklb_ not sure. Was purely a coincidence that I was creating that post when I saw your activity in the channel
#
miklb_ related, it looks just fine in Together which is exciting as I'm still using 2016 with the mf2 plugin
#
miklb_ that's coming from the json feed though
barpthewire joined the channel
#
tantek__ I'm definitely curious how readers are displaying event posts
#
tantek__ like do they have (or are they considering) RSVP buttons? [Going] [Interested] etc.
stevestreza, ichoquo0Aigh9ie, globbot and Nefertiti7 joined the channel
#
Nefertiti7 | ___ _____
cweiske joined the channel
#
Ruxton tantek__: indingenous reader on Android shows RSVP action for events, starts a new RSVP post where you select yes/no/maybe/interested - https://puu.sh/Bin2c/bdcb4f9c92.png
[kevinmarks] joined the channel
#
@pauljacobson I’m not sure what webmention badges are, but they sound interesting https://pauljacobson.me/b/AxQ (twitter.com/_/status/1032174750003605504)
#
tantek__ Ruxton: that is very cool! If you have my site in your reader, did you see such buttons for the recent indie event I posted (2nd to most recent post)
jjuran joined the channel
#
petermolnar Zegnat: you've recently had an interesting CSS experience, it seems like I'm having one as well: the 3px bottom border on my site's main menu are becoming trapezoids o.O
roger_rabbit29 joined the channel
#
roger_rabbit29 | ____
NinjaTrappeur, Davesdowntime and LooCfur joined the channel
#
Zegnat petermolnar, that usually means you also have side borders, just set to transparent or the same colour as your background.
#
Zegnat Side and top/bottom borders meet with a diagonal
#
Zegnat (I realise now that you may have figured this out in the last hour, but I was out on my bike)]
#
petermolnar I didn't
#
petermolnar thanks
#
Zegnat Looks like you added a 1px transparent border to all around on all a elements?
#
petermolnar yep, I was lazy
#
petermolnar I'll fix it :)
#
petermolnar instead of writing border-bottom, I went with border, assuming it'll be fine
#
Zegnat !tell GWG Did you find a Micropub client with redirect_uri for your testing? According to my survey there are three possible candidates: https://indieweb.org/User:Vanderven.se_martijn/IndieAuth_Client_Information
#
Loqi Ok, I'll tell them that when I see them next
calumryan, TripFandango, benwerd, KevinMarks, ben_thatmustbeme and petermolnar joined the channel
#
Zegnat For anyone following along on the IndieAuth client information thing, I just updated my survey with data about logos in various icon rels: https://indieweb.org/User:Vanderven.se_martijn/IndieAuth_Client_Information
#
Zegnat More clients added rel="apple-touch-icon" than h-x-app
#
GWG I wrote the code to pull that out too as a fallback
#
Loqi GWG: Zegnat left you a message 1 hour, 9 minutes ago: Did you find a Micropub client with redirect_uri for your testing? According to my survey there are three possible candidates: https://indieweb.org/User:Vanderven.se_martijn/IndieAuth_Client_Information
[kevinmarks] joined the channel
#
[kevinmarks] Re unsubscribes https://twitter.com/kevinmarks/status/95938217160880128?s=19
#
@kevinmarks » @Kells_Bells: The less I tweet, the more followers I acquire?!? « Marks's law: people unfollow you on a tweet but follow you on a retweet (twitter.com/_/status/95938217160880128)
#
Zegnat GWG, amazing! Could you leave a comment on https://github.com/indieweb/indieauth/issues/23 talking about that?
#
Loqi [Zegnat] #23 Client Information Discovery without relying on microformats parsers.
#
Zegnat Also, if you have time, maybe review https://indieweb.org/Web_App_Manifest#How_to_use_for_IndieAuth incase you have more details to add on how you use the web manifest? :)
barpthewire and jeremych_ joined the channel
#
GWG Zegnat: I think I should wait on the latter till I do something other than extract it
#
GWG Zegnat: Who is using meta application-name as a fallback as indicated in the manifest specification?
KevinMarks joined the channel
#
GWG I also looked for og: title and og:image as a fallback.
#
sknebel og: doesn't seem made for that
#
sknebel although I guess it's unlikely to be wrong if it is on an application page
#
GWG That was my thinking
#
GWG But only if nothing else was there
#
[kevinmarks] what do you get for kevinmarks.com/partialsilos.html ?
#
GWG It's not an application
#
Zegnat Yeah, og: was discussed as an option, the biggest downside of og: is that there wasn’t a way to identify a page as an application unless we wanted to define our own namespace :(
#
Zegnat I should maybe add og: data to the survey as well
#
Zegnat Though I don’t remember seeing it anywhere
#
Zegnat GWG: I think the only live implementation of this is [eddie]’s, and I don’t think he is doing application-name fallback. Not sure.
#
[kevinmarks] og:type defaults to website; the closet they have is product
#
Zegnat yep, and if you want to extend with different types you need your own schema for it, if I read the ogp.me docs right :(
#
Loqi *sniff*
#
GWG I just assume many web pages end up putting in og:title for other reasons, so if nothing else is there, I pull it as it is often better than the title tag
#
Zegnat Not sure what the good way is to fallback to application-name when you don’t know what language to pick. Especially if you maybe are taking other information from the manifest file, as you have no way of knowing what language manifest file you were given.
#
Zegnat There is an interesting problem with the application-name fallback as well. application-name can explicitly be multilingual per HTML spec (which is amazing!) but web app manifest has removed all multilingual aspects, recommending instead the server supplies separate manifests for separate languages.
#
sknebel something like wordpress has a configured UI langauge
#
sknebel otherwise, match browser headers?
#
sknebel (which is interesting: make the requests to find the manifest with set language headers to get the right one? is that how the manifest discovery is supposed to work?)
#
Zegnat Yes
#
sknebel https://www.w3.org/TR/appmanifest/#internationalization
#
Zegnat But its a bit double edged. Say you `accept-language: la` and get a manifest file back. You have no idea if the website really did offer a Latin version, or just gave you their default. And if they gave you their default, what language application-name fallback do you use? You don’t know their default.
#
sknebel I'm not sure why the fallback has to match the manifest?
#
Zegnat There could be text in the icon
#
Zegnat Though I guess that is a very minor inconvienience for the indieauth flow.
#
Zegnat I am more bothered in general that WAM did away with the i18n of application-name :P
#
Zegnat What is i18n?
#
Loqi Internationalization (AKA internationalisation, i18n; localization, l10n.) is the process of adapting software/content to various languages https://indieweb.org/i18n
#
sknebel it's only proposing a different mechanism for it, isn't it?
benwerd joined the channel
#
GWG I haven't gotten to language yet
[jgmac1106], icywiz5 and breitenj9 joined the channel
#
breitenj9 | /:::::::::::\
#
breitenj9 | .---------.
#
breitenj9 |
#
[kevinmarks] see https://www.scribd.com/doc/30715288/The-Open-Graph-Protocol-Design-Decisions for why ogp doesn't use title
#
[kevinmarks] https://twitter.com/kevinmarks/statuses/13080893395?tw_i=13080893395&tw_e=details&tw_p=archive
#
@kevinmarks says @daveman692 OGP has a separate clean page title, not the <title> tag, as titles have SEO noise in #www2010 (twitter.com/_/status/13080893395)
#
[kevinmarks] https://twitter.com/kevinmarks/statuses/13080932387
#
@kevinmarks says @daveman692 we wanted to use link rel="canonical", but publishers were afraid this would affect their search ranking #www2010 (twitter.com/_/status/13080932387)
[keithjgrant], barpthewire, [grantcodes] and benwerd joined the channel
#
@aminecodes Webmentions: Enabling Better Communication on the Internet "Your personal identity on the internet can end up fragmented like a Horcrux across multiple websites over which you have little if any, control." @alistapart https://alistapart.com/article/webmentions-enabling-better-communication-on-the-internet (twitter.com/_/status/1032258120330956801)
benwerd, koddsson and KevinMarks joined the channel
#
Zegnat sknebel, yeah, it is proposing a different way of doing the same i18n. I am just not sure it is a better way. How often do we see conneg go right? How many are going to craft multiple manifest files? I think there was probably a reason for the HTML5 way of offering translations within the same document.
#
Zegnat Ugh, that horcrux metaphor is wrong. Horcrux aren’t fragmented :P
#
petermolnar aren't they, though?
#
petermolnar it's a copy of the person at a certain point in time imo
barpthewire joined the channel
#
Zegnat No. A horcrux is a soul fragment.
#
Zegnat So describing all our social media profiles as horcruxes might be right
#
Zegnat But something can’t be “fragmented like a Horcrux”, because the horcrux itself isn’t fragmented ;) semantics
#
Zegnat This should probably move to chat, hahaha
#
Loqi nice
benwerd and darxun24 joined the channel
#
darxun24 | _.--"""""--._
KevinMarks, benwerd, snarfed and [grantcodes] joined the channel
#
[grantcodes] Hey aaronpk is there any way to identify the original url of a photo in the proxied image url?
#
aaronpk in aperture?
#
[grantcodes] yes sorry
#
[grantcodes] Mainly because I did have css to shrink down wordpress emoji based on the url but now the url is proxied
#
aaronpk haha oops
#
aaronpk is there no class added to those emoji?
TripFandango joined the channel
#
[grantcodes] Not the ones I am currently looking at. Maybe because it's from an rss feed
#
aaronpk xray is probably stripping the class anyway, can you check the original?
#
[grantcodes] Huh it's actually a quoted tweet embed that has come through as html
#
aaronpk i feel like i've seen that too but i can't find them now
#
[grantcodes] This is what was in my channels: https://wptavern.com/classicpress-gutenberg-not-included
#
aaronpk from what feed
#
[grantcodes] But it's the same thing, you could identify them with a css selector prior to the proxy
#
aaronpk i don't see any emoji images on that url, just actual emoji
#
aaronpk aha https://wptavern.com/feed
#
aaronpk the embedded tweet in the rss feed somehow has <img src="https://s.w.org/images/core/emoji/11/72x72/1f393.png" alt="🎓" class="wp-smiley" style="height: 1em; max-height: 1em;" />
#
[grantcodes] I'm using the json feed
#
aaronpk interesting, should I let XRay pass through the "height" and "max-height" style properties?
#
[grantcodes] Which has the same html
#
[grantcodes] I'm not sure what the best solution is but that would work.
#
aaronpk hm alternately, i could recognize an <img> tag with an alt attribute of a single emoji and replace the img tag with the emoji
#
@mdejayhawk @JLAudio Looking for replacement Grill for 10’ microsub? Help can’t find anywhere (twitter.com/_/status/1032275861599383552)
#
[grantcodes] That's an option, but you know there are people out there with alt emojis for a photo
#
[grantcodes] Even if they shouldn't 😛
#
[grantcodes] I think recognizing the url is the most bulletproof option. Dead easy in css at least `img[src^="https://s.w.org/images/core/emoji"]`
#
aaronpk ugh no i don't want to hardcode to wordpress
#
[grantcodes] In that case I'd lean towards passing some css instead of looking at the alt
#
aaronpk i already added some support for this for mastodon so i want to keep it generic
#
aaronpk let me see what it takes to pass through specific style properties
#
[grantcodes] But not a huge issue right now
KevinMarks and barpthewire joined the channel
#
aaronpk [grantcodes]: believe it or not, this issue is coming up in trying to fix this! https://github.com/microformats/microformats2-parsing/issues/21
#
Loqi [aaronpk] #21 reduce instances where photo is implied
#
[grantcodes] Reading that makes me glad I don't write a microformats parser!
#
aaronpk for real
benwerd, barpthewire, tantek__ and gRegorLove joined the channel
#
aaronpk i can't figure it out!
#
aaronpk for some reason can't get the htmlpurifier library to allow the style tag at all, much less filter out specific attributes
jackjamieson joined the channel
#
Zegnat ezyang/htmlpurifier ?
#
aaronpk yea
#
Zegnat I seem to recall checking that out and disliking it a lot, haha
#
aaronpk it's definitely weird, but it does the job super well
#
aaronpk i was able to write a filter to remove all css classes except ones that are microformats classes, which was neat
jackjamieson joined the channel
#
Zegnat This is why I didn’t like HTML Purifier: it has both ForbiddenAttributes and AllowedAttributes configurations. Hard to figure out exactly where it would be removing what
#
sknebel aaronpk: where exactly did that issue come up right now?
#
aaronpk trying to allow inline style height attributes on images
#
sknebel no, the microformats issue
#
aaronpk yea
#
aaronpk posts with a single emoji img end up getting that img turned into an implied photo
#
sknebel do you have an example?
#
Zegnat aaronpk, for style elements, do you need to enable this, maybe? http://htmlpurifier.org/live/configdoc/plain.html#Filter.ExtractStyleBlocks
#
aaronpk i was writing a simple test case based on a real example, but my html was constructed
#
aaronpk <div class="h-entry"><div class="e-content">sample emoji <img src="emoji.png"></div></div>
#
aaronpk the actual post is more like <div class="h-entry"><div class="e-content"><p>sample emoji <img src="emoji.png"></p></div></div> so doesn't trigger it
[jgmac1106] joined the channel
#
@miklb ↩️ But as it stands, with only two minor edits to the mf2 plugin & suite of Indieweb WP plugins I already had installed, I’m using default twentysixteen WordPress theme. I’m syndicating to http://micro.blog, Twitter (via http://brid.gy) and… https://miklb.com/blog/2018/08/22/4190/ (twitter.com/_/status/1032294842385223681)
[keithjgrant] joined the channel
#
[keithjgrant] When posting to a micropub endpoint, how do you distinguish between a note and a full article?
#
sknebel an article has a title (name) and content
#
[keithjgrant] that's it? that's easy enough
#
Zegnat What is PTD?
#
Loqi Post Type Discovery specifies an algorithm for determining the type of a post by what properties it has and potentially what value(s) they have, which helps avoid the need for explicit post types that are being abandoned by modern post creation UIs https://indieweb.org/PTD
KevinMarks joined the channel
#
Zegnat That’s the full answer ^^^ the micropub endpoint could run PTD on the mf2 object it recieved. This will tell it if it was an article or note (or something else entirely)
miklb joined the channel
#
[keithjgrant] awesome thanks
#
[keithjgrant] i'm guessing my mp endpoint doesn't do all that yet. But it does mean I could add support to omnibear
#
Zegnat Sounds like a plan! :D
#
Zegnat Is the omnibear website opensourced somewhere, [keithjgrant]? We should fix its webmanifest. Yours is a little wonky: https://omnibear.com/manifest.json
#
Zegnat (See also https://indieweb.org/User:Vanderven.se_martijn/IndieAuth_Client_Information - nice job on having a mask-icon!)
#
[keithjgrant] yes. at https://github.com/keithjgrant/omnibear-site
#
Loqi [keithjgrant] omnibear-site: Public website for Omnibear
#
[keithjgrant] interesting. I don't remember giving it a manifest 😆
#
Zegnat It isn’t much of a manifest ;) But I might PR an update there
#
GWG We can chase everyone to do it
#
GWG Just chasing aaronpk would get a large amount of clients
#
Zegnat aaronpk even has links to manifests just pointing at 404s :P
cloe joined the channel
#
GWG Aaronpk does a lot
#
Zegnat Should probably just open PRs for manifest files for him too... Hmm. This might be a nice little task for tomorrow morning!
snarfed joined the channel
#
tantek__ Zegnat: re: WAM and fallback, I think we should only do the fallback in the spec (meta application-name, link rel icon) and not go down the path of rando proprietary values (apple-touch-icon)
snarfed joined the channel
#
tantek__ and aaronpk noted that manifest 404 came from a copy/paste of setting up icon rels
#
tantek__ something that did favicons and gave you markup to copy/paste (also a good reason not just copy/paste a block of meta markup)
#
Zegnat I know. It just skewed the statistics on manifest usage in the first iteration of the survey. Statistics have been fixed by [eddie] though.
#
Zegnat I am just doing discovery to see what clients are publishing. Felt like the correct first step in figuring out what things may replace h-app
#
tantek__ as long as there is rel=icon, we can ignore the others
#
tantek__ what would be more interesting is if there were cases with proprietary markup *only*
#
Zegnat The survey data did surface something else that I need to bring up in mf2 parsing. I was using the rel parser from the mf2 parser to find the URLs, and rel-icon is super useless there because we aren’t keeping the `sizes` attribute
#
Zegnat Not in the survey, so far, tantek__ :) So that was a good thing to confirm too
#
tantek__ great, so it confirmed we can ignore proprietary icon markup
[grantcodes] joined the channel
#
GWG I am tracking sizes in my pending implementation as a fallback to manifest
[cleverdevil] joined the channel
#
Zegnat tantek__, do you know if there is some clever way to pick the correct language application-name fallback?
#
Zegnat meta-application-name may give a list of names for i18n: https://html.spec.whatwg.org/multipage/semantics.html#meta-application-name
#
tantek__ right, I'd expect the HTML spec to give that guidance
#
tantek__ and if not, feel free to file an issue (on the HTML spec) requesting it :)
#
Zegnat Yeah, their algorithm determines a “winning language”, ha. I guess that’ll be as good as any :)
#
Zegnat I should probably add application-name to the survey data. Though I don’t believe any one of the pages includes that metadata
#
Loqi it is probable
calumryan joined the channel
#
tantek__ tempted to introduce meta name="page-name-no-seo-nonsense"
calumryan_ and [kevinmarks] joined the channel
#
[kevinmarks] If sizes is important we should consider adding it to rel parsing - we added only attributes we saw in the wild.
#
Zegnat I am writing up the issue for that right now, [kevinmarks]
#
Zegnat https://github.com/microformats/microformats2-parsing/issues/39
#
Loqi [Zegnat] #39 Include sizes attribute when parsing rel microformats.
benwerd joined the channel
#
loqi.me edited /OkCupid (+258) "tantek__ added "2018-06-24 [https://cubicgarden.com/2018/06/24/gdpr-dating-information-update/ GDPR dating information update] <blockquote>Going to need some time to digest everything but the OKCupid data dump is full of interesting things.</blockquot..." (view diff)
eli_oat, dougbeal|mb1, rhiaro, renem, [pfefferle], KartikPrabhu and [jgmac1106] joined the channel
#
snarfed hey GWG, re micropub on localhost allowing missing token (https://github.com/snarfed/wordpress-micropub/pull/162#discussion_r212028124), i'm not sure when it was removed, but i'd like to keep it, built in, and not change to requiring users to write PHP code to handle a filter. mind adding the built in behavior back, in a new PR?
#
snarfed (asking here instead of in #-wordpress to avoid distracting them from a useful ongoing conversation :P)
KartikPrabhu joined the channel
#
tantek__ snarfed++
#
Loqi snarfed has 67 karma in this channel over the last year (121 in all channels)
[chrisaldrich] and [kevinmarks] joined the channel
#
[kevinmarks] bah, google has taken down my verify-me chrome plugin for lacking a privacy policy.
#
[kevinmarks] it does send all rel-me urls you see to indiewebify.me, whihc is not ideal
#
[kevinmarks] maybe I should do it locally
#
tantek__ what is a privacy policy?
#
Loqi A disclosure is a bit of content, typically on a home page, on an indie web site that proactively discloses some aspect about the site that the site owner wants the user to explicitly be aware of https://indieweb.org/privacy_policy
[schmarty] joined the channel
#
Zegnat [kevinmarks], you can probably just use my local version? I haven’t looked at it for a while though. Maybe I should brush it up
#
Zegnat https://github.com/Zegnat/verify-me-locally
#
Loqi [Zegnat] verify-me-locally: A spin-off of @kevinmarks’ verify-me without external dependencies.
KevinMarks_, leg, KevinMarks, KartikPrabhu and barpthewire joined the channel
#
NinjaTrappeur Hey indieweb! I'm a having a problem related to webmentions. I have been stuck for the last 2 days with this, I think I could use some of your brain power :) It's a bit too long for a chat message, here's the problem: https://alternativebit.fr/rand/webmentionioweird/
#
Loqi [NinjaTrappeur] Confusing Behaviour when Integrating with webmention.io
#
aaronpk NinjaTrappeur: did you try %20 instead of +?
#
GWG snarfed, I will restore a localhost override, but suggest WP_DEBUG be required
#
snarfed GWG: thanks! what's your concern? what problem are you hoping to avoid?
#
NinjaTrappeur aaronpk, yes, same result
#
GWG Security issues
#
GWG Shouldn't it not be used in production? As some people are on shared hosting
#
snarfed as mentioned, it only applied if the site was serving on localhost
#
snarfed but sure, feel free to add a new WP_DEBUG req't
#
aaronpk NinjaTrappeur: https://webmention.io/alternativebit.fr/webmention/NhVU5nOiUkXH2W4gP9Oj
#
aaronpk it needs to be double encoded
#
aaronpk i sent that from curl with: curl https://webmention.io/alternativebit.fr/webmention -d source="https://public.alternativebit.fr/bl%2520so.html" -d target=https://alternativebit.fr/posts/lightweight-blog/
#
aaronpk i'm pretty sure it was coming through the other end as a literal space, rather than a URL-encoded space, which is not a valid URL
#
NinjaTrappeur I'm a bit confused. Do you mean I need to first substitute spaces with %20 and encoding the resulting string as www-form?
#
aaronpk are you using a library for this or building one from scratch?
#
aaronpk think of it this way: 'https://example.com/foo bar/' is not a valid URL, but 'https://example.com/foo%20bar/' and 'https://example.com/foo+bar/' are
#
aaronpk so you need to send 'https://example.com/foo%20bar/' via a form-encoded request
#
NinjaTrappeur https://github.com/NinjaTrappeur/lobsters/commit/7eaa1543bcd6513773080c3d3bf2fa04f75d43c4#diff-322393a70f7dfdbd82984ef5aef6dcb8R64
#
aaronpk right, source = "https://public.alternativebit.fr/bl so.html" is not a valid URL
#
NinjaTrappeur alright
#
NinjaTrappeur get it
#
aaronpk try typing that in a browser, it'll auto-fix it to %20
#
NinjaTrappeur Thanks for your help \0/
#
GWG snarfed, I have to finish up a quick initial client information discovery piece for IndieAuth first.
#
snarfed GWG: sure! no worries. no deadlines or etas here, this is all fun side projects, not work
#
NinjaTrappeur fun fact: firefox is not visually fixing the URL, hence my confusion :)
#
aaronpk oh funny
#
GWG I still want to finish up the Micropub project so I can focus elsewhere for a bit. But right now, still in testing.
#
sknebel I wouldn't be surprised if the space issue in URLs depends on which URL spec you follow
#
GWG Client Information Discovery is related because I have been asking for a lot of tokens to test the flow.
#
GWG I want it to say Quill, not Quill.p3k.io
#
aaronpk sknebel: i don't think a space has ever been a valid character for a path
#
NinjaTrappeur aaronpk, you're absolutely right, the actual RFC about this is here: https://www.ietf.org/rfc/rfc2396.txt (section 2.4.1) . Interesting rabbit hole, I guess I can now go to sleep a bit less ignorant :)
#
NinjaTrappeur sknebel, ^
#
snarfed smells yak blood
#
sknebel I seem to remember the WHATWG URL spec allows a bunch of stuff like that
#
aaronpk ugh i can't read the whatwg url spec
#
aaronpk it just keeps sending me in loops whenever i try
#
sknebel snarfed: not shaved *that* deeply (to draw blood), and at least it was triggered by an actual confusion happening :D
[tantek] joined the channel
#
[tantek] Even the "+" I'm pretty sure you can only put in query param values
barpthewire, tantek__, jackjamieson, kline, [kiai] and TripFandango joined the channel
#
snarfed evidently https://prettygoodhat.com/ added his bridgy user page https://brid.gy/twitter/schussman to his reader, so aperture/watchtower have been happily refetching it every 2m for days now
#
Loqi Alan S.
#
snarfed 😂🍿
#
tantek__ pulls out his meme post re: "refetching it every 2m" http://tantek.com/2018/024/t4/websub-recommendation
#
Loqi [Tantek Çelik] #WebSub is now a @W3C Recommendation! https://www.w3.org/TR/2018/REC-websub-20180123/ Supported by @Google @WordPress etc. #implementnow Congrats @Julien51 @aaronpk! So stop polling so much, start SUBscribing to WEB pages for changes! #openweb #ind... https://i.imgflip.com/1z96ih.jpg
#
tantek__ views source onf bridgy profile pages
#
snarfed tantek__: https://brid.gy/about#api
#
snarfed relevant: https://github.com/snarfed/bridgy/issues/677#issuecomment-232683048
#
Loqi [snarfed] one low tech way to do this would be a cron job that polls [bridgy's api](https://brid.gy/about#api) for each of your user pages and looks for `p-bridgy-status="failed"` (or `error`). e.g.: ``` sh curl 'https://mf2.kylewm.com/?url=https://brid.gy/tw...
#
tantek__ snarfed, interesting, I see (visually on the page) at least two streams that could be h-feed
#
tantek__ keeps digging
#
aaronpk snarfed: if it's been fetching every 2 minutes it's because the URL has been returning new content each time and there are no caching headers
#
snarfed tantek__: there's lots of appropriate mf2, documented on that doc link, but no h-feeds. i wanted to avoid any semantics debates.
#
snarfed aaronpk: funny, yes, both of those are true
#
tantek__ snarfed, sounds like an invitation
#
tantek__ h-feed + websub could make this a better situ for such reader scenarios
#
snarfed i'd love to hear why anyone would want their bridgy user page in a reader. and what that would even mean.
#
snarfed not a use case i really care about or even understand :P
#
tantek__ but ... pull requests accepted
#
tantek__ ?
#
snarfed often yes
#
aaronpk i don't understand why you'd want that in a reader
#
snarfed yup
#
tantek__ in a "monitoring" channel maybe?
#
snarfed singpolyma's request in the issue above is maybe actually one: notifs of bridgy wm failures
#
snarfed not sure reader is the right answer...but it could be
#
tantek__ for aaronpk, reader is likely the right answer for all notifs ;)
#
aaronpk definitely
#
snarfed dreads the PRs
#
aaronpk lol
#
tantek__ knows better than to submit a Dread PR
leg and koddsson joined the channel
#
snarfed lol
#
snarfed hey i'm skeptical but if it's reasonable and good i might merge it
#
snarfed (^ basically describes the triage process for any incoming PR, anywhere)
jackjamieson joined the channel
#
tantek__ snarfed, my general philosophy is if it looks like a list of dated items, it's worth marking up as an h-feed
#
snarfed ok!
#
tantek__ which is now making me rethink resumes
#
tantek__ wow I wonder if anyone has an h-feed resume with WebSub so all the recruiters subscribing to it get notified when there's a new item added
#
snarfed former seems slightly more plausible than latter
jackjamieson joined the channel
#
tantek__ snarfed, I've met random people at conferences who told me they had copy/pasted my resume markup when they were in college 😂
#
snarfed *markup* huh
#
tantek__ yeah, because it gave them a default structure in which to put their own info
#
tantek__ oh and because apparently they thought my markup+css produced a decent result already
#
tantek__ (presentationally)
#
snarfed sorry, it was a joke, wondered if they also copied the content
#
tantek__ hmm, good reason to hide some eastereggs in the content
koddsson joined the channel
#
tantek__ checks his own markup before submitting a pull request to bridgy
#
tantek__ snarfed, FYI, one of my older pages I marked up with an h-entry stream https://tantek.com/favelets/#r2013-290
#
tantek__ and now in looking at it, looks like it could benefit from explicit h-feed + p-name
#
snarfed (minor nit, that url needs http :P)
#
tantek__ gosh darn it IDK why my browser is doing that
#
tantek__ (or maybe one of my add-ons)
[schmarty] joined the channel
#
@ndw Exploring webmentions (mostly implemented) took me back through some micro formats stuff. So angry/sad/frustrated to see namespaces “fixed” by using “-“ instead of “:” and removing any extensibility. You got to h- and u- first! You win, whatever you are! (twitter.com/_/status/1032381785047425024)
#
aaronpk 🤔
#
tantek__ Actually, we "fixed" namespaces by eliminating them. There's no URI behind the h- or u-
#
tantek__ they are merely parsing directives
#
tantek__ time to take off the namespace colored glasses
#
snarfed sknebel++ earlier for using indiemap to inform the wordpress debate with data
#
Loqi sknebel has 27 karma in this channel over the last year (81 in all channels)
#
GWG snarfed, got a minute?
#
snarfed GWG: sure
#
GWG What users are going to disable token handling?
#
snarfed anyone who uses a local wordpress to test anything
#
snarfed technically they can still try to do full auth, but it's often a hassle in local envs
#
GWG So, it should be part of that... okay.. that's different.
#
aaronpk that's interesting, i hadn't considered disabling auth while testing lol
#
GWG I may try to redesign it that way
#
aaronpk i would take the approach of making it easy to generate an access token
#
aaronpk without going through the indieauth flow
#
GWG aaronpk, neither did I
#
GWG aaronpk, on my list for IndieAuth the plugin, but Micropub without it only uses IndieAuth.com
#
aaronpk ah right
#
GWG I'd rather everyone went local
#
snarfed GWG: hopefully this doesn't rise to level of "redesign." the code before was like three lines :P
#
GWG But that's me
#
GWG snarfed, rewrite?
#
GWG Rethink?
#
snarfed add back three lines of code :P
#
snarfed maybe just semantics
koddsson joined the channel
#
GWG Using IndieAuth.com vs your own IndieAuth...what are the reasons for one over the other?
#
GWG Assuming one click install of a plugin.
gRegorLove, miklb__ and tantek__ joined the channel
#
snarfed GWG: many reasons. do they change your work at all? probably not worth enumerating them really.
tantek joined the channel
#
GWG I just wonder because I like things under my control
KartikPrabhu joined the channel
#
snarfed heh, a luxury we sadly have little of with software we ship. all we can control is the code, config, and docs, and even those aren't foolproof.
#
snarfed ...which is one reason it's easier in many ways to build services instead
#
GWG WordPress isn't built by me, but I have more control
#
GWG I am not knocking services
#
GWG I am thinking of the readme for IndieAuth
#
GWG I am trying to educate
#
snarfed .com: hosted/managed, reliable, highly supported; less control/ownership
#
gRegorLove How is browser support for SVG rel-icon?
#
snarfed plugin: fully owned/controlled; more admin work, younger, less mature
#
snarfed gRegorLove: https://caniuse.com/#feat=link-icon-svg
#
gRegorLove ah, thanks!
#
gRegorLove Zegnat: I added h-app and rel-icon to indiebookclub
#
sknebel !tell Zegnat gRegor​Love says "Zegnat: I added h-app and rel-icon to indiebookclub"
#
Loqi Ok, I'll tell them that when I see them next
#
tantek__ gRegorLove: and WAM?
#
gRegorLove something something wake me up before you go, go
#
gRegorLove (nope, no WAM yet. Not very familiar.)
#
gregorlove.com edited /h-x-app (+113) "/* Publishers */ +indiebookclub" (view diff)
#
tantek__ gRegorLove: wow. I think we know what to suggest for your next IWS karaoke
#
gRegorLove What's the "Why" for WAM? I haven't been keeping up with that conversation.
#
gRegorLove reviews https://github.com/indieweb/indieauth/issues/23
#
Loqi [Zegnat] #23 Client Information Discovery without relying on microformats parsers.
#
tantek__ that basically
#
tantek__ though then Zegnat filed an issue for mf2 parsers not picking up the "sizes" attribute in link tags, but it's like wait, if you're using rel=icon you're already not relying on a microformats parser
#
tantek__ so it doesn't make much sense to me (issue on mf2 parsing for a use-case that is specifically about *not* using an mf2 parser)
#
tantek__ gRegorLove: tl;dr - some devs would rather do another http request and parse JSON, than use mf2 parser apparently
#
tantek__ WAM is the format of that JSON
#
tantek.com edited /IndieWebCamps (+506) "/* How to organize */ a few more details/suggestions for wiki page / event" (view diff)
#
martymcgui.re uploaded /Special:Log/upload "uploaded "[[File:hwc-baltimore-2018-08-22.jpg]]" https://indieweb.org/File:hwc-baltimore-2018-08-22.jpg"
#
sknebel because some questioning the use of microformats means nobody should use a microformats parser for it, and potential issues in related areas should not be noted if they don't perfectly fit in the story?
#
tantek__ sknebel, point is, if you're using a mf2 parser, you can already use the h-ap
#
tantek__ h-app
#
tantek__ and you don't need to use the link rel=icon sizes stuff
#
tantek__ we should discourage extra work like that
#
Loqi yea!
#
sknebel so we assume that everyone will be publishing WAM and microformats, despite at least part of the reasoning for WAM just yesterday (or the day before?) was that WAM is an easier sell than microformats to e.g. the Oauth groups?
#
sknebel or rather, rel=icon I guess
#
tantek__ sknebel - different groups
#
tantek__ microformats is easier for publishers
#
tantek__ consuming a JSON side-file is *trendier* for webdevs, and apparently more palatable to OAuth groups
#
tantek__ there is no one "easier sell" to everyone
#
sknebel even in our community, the rel=icon was pretty common, on pages without h-x-app
#
sknebel if we consider taking rel=icon into account, it makes sense that a microformats parser can do that
#
sknebel (mentioning WAM was actually wrong, since discovery for that's already covered with what the parser can do. bit late here...)
#
sknebel (I personally think all reasonably widely accepted link-rel-attributes should be covered by the parsers, since it's a very easy addition to the parser and the effort for the parser users that might want a rel-parser too to add another parsing thing just to see the properties the microformats parser doesn't cover is pretty annoying in comparison)
#
tantek__ I suppose we can try to figure out an optimal order or something for discovery, e.g. look h-app, then WAM, then WAM fallbacks
#
aaronpk the fewer the better
#
tantek__ aaronpk, point being, if you (web dev) already don't want to do mf2 parsing, then by the time you get to step 3, WAM fallbacks, you're not using an mf2 parser (you started with that assumption) so there is no need for the mf2 parser to cater to you
#
tantek__ do I need to write out the proof, Geometry Class style?
#
sknebel so I as a web dev willing to use a microformats parser (since I already use it for all kinds of things) now get to find another parser just to find the size= on rel=icons if a site forces me to fall back to that, because ... ?
#
tantek__ aaronpk, also, in all your usage of h-app, did you ever get use-case / feature requests for app photos of different sizes with sizes property or something?
#
tantek__ no sknebel, you as a web dev using an mf2 parser, parse the h-app and you're done
#
tantek__ no need for additional HTTP request etc.
#
sknebel okay, so everyone is going to publish h-app
#
sknebel even in our community they don't, but they will
#
tantek__ depending on tools and consuming code iteration
#
tantek__ but yes, if you make it easier / more useful, then people will
#
tantek__ heck maybe even some sort of script that auto-converts your h-app to WAM.json
#
tantek__ so you never have to maintain that sidefile yourself
#
tantek__ like some folks do with auto-generating their RSS/Atom from their parsed mf2
#
tantek__ that's likely the easiest path for publishers
#
aaronpk i'm just saying the fewer different ways to parse things the better. less code, fewer test cases, fewer ways things can go wrong.
#
tantek__ i'm just saying history is now showing that dead-end sidefile formats are what ends up enabling proprietary silos to take-off, so keeping information publishing/consuming in evolving HTML is the longterm strategy for openness
KevinMarks joined the channel