#dev 2018-04-24

2018-04-24 UTC
#
aaronpk right that's why there's a media endpoint
#
skippy right, but not everything supports that, yet. :(
#
skippy and the spec specifically says that multi-part uploads are okay.
#
aaronpk yea that's for photo posts
#
aaronpk where the photo isn't in the html content
#
skippy okay, I guess; but there's no real indication in the client UIs that I've seen that they distringuish between "photo posts" and "other kinds of posts" ... so it's up to the server to deterine "if h-entry AND a photo is present, this is a photo post"?
#
aaronpk yea
#
aaronpk what is post type discovery?
#
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/post-type-discovery
snarfed joined the channel
#
skippy ok. back to the drawing board for me, then. Thanks.
eli_oat, iasai and tantek joined the channel
#
tantek I'm liking this theme of indieweb parody songs
#
www.boffosocko.com edited /lost_infrastructure (+60) "Timeshifting/Reading: microsub projects; micro.blog" (view diff)
[miklb], AngeloGladding, iasai, renem, eli_oat, Mandrake, [mrkrndvs], [chrisaldrich], chrisaldrich, cweiske, [pfefferle], tantek, [unoabraham], swentel and [kevinmarks] joined the channel
#
[kevinmarks] http://calacanis.com/2018/04/20/openbookchallenge/
#
swentel interesting
#
swentel should we make an indieweb case and submit as a team ?
#
Zegnat I am never sure such drives actually work for IndieWeb. “traction stats”? What’s that? Number of Webmentions sent? 6% payback to the incubator? From what? I’d guess things like Bridgy are running at a nett loss right now, and they would probably want that to be turned around?
#
Zegnat But that may just be me being a cynical grouch :P
#
microblog.ducamp.me edited /Facebook-fr () "(-868) [fr] : relecture en cours - typos" (view diff)
#
swentel hehe, it's not bad to think that way, I'd actually do that too :)
#
swentel but it might be a good start to figure out of we can make a case :)
#
Zegnat The closest you can get to an IndieWeb “social network service” right now is probably hosting a multi-user Known instance. That still lets you communicate outside of the network, while also applying policies for the content you are allowing within the network.
#
[kevinmarks] Not 6% payback, 6% equity in the company formed.
#
Zegnat Ah, misunderstood that sentence, thanks for correcting [kevinmarks].
#
[kevinmarks] There is still a presumption that a company is the right organisational form.
#
Zegnat Still not sure what that 6% is going to mean. I guess that might be property rights of code written? But as long as you want to feed back to the IndieWeb with open-source projects, that doesn’t actually have a lot of value.
#
Zegnat Of course you could go the micro.blog route and just build a closed-source service, but that is never what I think of when talking about “an indieweb case”
[unoabraham] and jeremycherfas joined the channel
#
[kevinmarks] Known did an incubator like this (that Ben now runs) - the code was still open source, but the company was partly owned by the incubator, the rest by staff and other investors.
#
[kevinmarks] The thing that makes me wary is that Jason's explicit goal is a billion users, which is missing the point
#
Zegnat Maybe. On the other hand, a billion IndieWeb users would be sweet.
#
swentel hehe
#
Zegnat The thing is, a billion IndieWeb users doesn’t mean a billion users for 1 service, because that isn’t what we want to encourage.
#
Zegnat Alright, for all you Micropub enthusiasts in here, I think I finished up describing the microformats JSON syntax: http://microformats.org/wiki/microformats2-json !
#
Zegnat !tell skippy I think I have now covered the entire MF2 JSON syntax on http://microformats.org/wiki/microformats2-json - please have a read and let me know if it still leaves you with questions :)
#
Loqi Ok, I'll tell them that when I see them next
#
Zegnat Things you might learn reading that page: a key/value pair in a JSON object is called a “member”
#
swentel oh, something I've been wondering
#
[kevinmarks] [aaronpk] does that clarify the rels +rel-urls for you? (as you asked about it the other day)
#
swentel oh, ok, so micropub json request is using microformats 2 json , interesting
#
swentel I've always wondered why it was so different compared to url encoded request
#
@m4d_z ↩️ @edasfr Ça dépend de tes pré-requis je dirai (l'i18n, webmentions, images responsives…). Tu as besoin de quoi, de base ? (twitter.com/_/status/988716451057033216)
#
Zegnat Yes, swentel. When Micropub is talking about the JSON syntax being “the parsed Microformats 2 JSON format”, it means the format I describe there.
#
swentel good to know
#
swentel I was kind of annoyed while testing quill to have drupal support to see a difference between some requests
#
Zegnat But there has never really been a page describing it. It is just the default serialisation for mf2 parsers
#
Zegnat So people developing Micropub stuff, that aren’t already versed with mf2 and its parsers, don’t have a place to go
#
swentel and I was like, what a random spec, this is annoying :) But I know where it comes from now (and I need to start using aaron's php libraries for parsing incoming requests though)
#
swentel s/though/too
#
swentel (it's all very new too me, so I still need to read a lot heh)
#
swentel (but it's extremely exciting)
#
swentel s/too/to
schmarty1, [pfefferle], jjuran, loicm and [xavierroy] joined the channel
#
Zegnat swentel, lets see if we can make it even easier by adding a reference to the mf2 json page from the Micropub spec :) https://github.com/w3c/Micropub/issues/106
#
Loqi [Zegnat] #106 Add a link to a description of the microformats2 JSON format.
#
swentel +1
[jgmac1106] joined the channel
#
@edasfr ↩️ @m4d_z Import depuis WordPress. Les webmentions seraient un plus. Logiciel bidouillable. Templates sympa par défaut. Édition simple (twitter.com/_/status/988727158318256128)
Mandrake joined the channel
#
[jgmac1106] Incubators annoy me. I have three kids, coach a bazillion teams, run cub scouts... I can't attend your business camp no matter who I can dine with at your networking meals
#
[jgmac1106] And as a call back to [kevinmarks] I think a co-op model where users of network owns the network is more sustainable for Jason's goals
#
loqi.me created /Glossary (+157) "prompted by petermolnar and dfn added by [unoabraham]" (view diff)
#
[jgmac1106] Though I also think we can make the web the social network of the users. Not sure a centralized home will ever be as user centric as owning your domain
iasai joined the channel
#
jeremycherfas.net edited /Microsub (+17) "Change wording of definition to focus on what it does rather than that it is a spec." (view diff)
[xavierroy], [kevinmarks] and tantek joined the channel
#
jeremycherfas.net edited /Microsub (+263) "New defn and introductory remarks" (view diff)
klez and tomasparks joined the channel
#
Zegnat What is Microsub?
#
Loqi Microsub is a technology for servers https://indieweb.org/microsub
#
Zegnat Ha
#
Zegnat jeremycherfas, did you want to also have the next sentence show in the Loqi definition? ;)
[snarfed] joined the channel
#
[snarfed] re bridgy finances, i know it was just an example, but cost and other numbers are public: https://snarfed.org/2018-01-02_bridgy-stats-update-3aa
#
[snarfed] no revenue so yes it operates at a loss :P hosting is $55-60/mo right now
#
swentel hmm
#
swentel page not found :)
#
[snarfed] oops, https://snarfed.org/2018-01-02_bridgy-stats-update-3
#
[snarfed] related: https://brid.gy/about#cost
#
[snarfed] i have no plans to ever try to make it a business, or charge users money
#
[snarfed] (or advertise)
#
[snarfed] er sorry, misquoted the current numbers, actually $45/mo right now
#
jeremycherfas Yes, zegnat, I did. Can loqi handle only 1 sentence defns?
#
Zegnat No, but if you want more you need to wrap it in a .p-summary
#
Zegnat <span class="p-summary">…dfn…</span>
#
jeremycherfas Will do.
#
Zegnat To stop big paragraphs from being spammed Loqi will, by default, only take the first sentence :)
#
Zegnat There is probably a page with a good example, but don’t have one to link right this second
#
jeremycherfas Your instructions worked fine. But Loqi misses the final full stop.
[kaushal_modi] joined the channel
#
Zegnat It might just always skip the period ...
#
Zegnat I like your plumbing comparison to HTTP for web browsers jeremycherfas... Wondering if I could illustrate it some way like this: https://wiki.zegnat.net/media/plumbing.html
#
jeremycherfas Oh that's excellent. Zegnat++
#
jeremycherfas Zegnat ++
#
jeremycherfas Oh come on Loqi, rise and shine.
#
Zegnat Hahaha
#
Loqi hahahaha
#
jeremycherfas Zegnat++
#
Loqi zegnat has 55 karma in this channel (199 overall)
#
jeremycherfas Zegnat ++
#
jeremycherfas Third time lucky. I remebered something about spaces, and something about being at the start of a statement.
#
Zegnat 5 second delay because Loqi was laughing at you first? Wow. Loqi is coming into a rebelious phase? :P
#
jeremycherfas You could even add FTP to that.
#
Zegnat FTP could be another publishing example, for sure
#
Zegnat It is not totally fair to compare them to HTTP, because Micropub is actually a protocol that goes over HTTP, but for illustrative purposes I think this did fine
#
Zegnat MetaWeblog is over XML-RPC which is over HTTP. But who has the time to dig all that up? ;)
snarfed joined the channel
#
jeremycherfas Not at all necessary. Get into which transfer protocol the information protocol uses and we may never see you again.
DenSchub_, [miklb], ancarda, iasai_, leg, wagle, alex_mayorga and cweiske joined the channel
#
aaronpk well that's a funny definition
#
aaronpk microsub is not just for servers
#
aaronpk let's try this again
#
vanderven.se martijn edited /dfn (+117) "This probably still needs a good wiki editting example…" (view diff)
kaushalmodi joined the channel
#
kaushalmodi kevinmarks: I unintentionally spammed your http://www.kevinmarks.com/fragmentions.html post because I changed the Permalink of my https://scripter.co/fragmentions/ post.
#
Loqi [Kevin Marks] Fragmentions - linking to any text 2014-04-17 05:25-07:00
#
kaushalmodi Is there a good way to prevent that when changing URL (not that I do that frequently).
#
kaushalmodi KartikPrabhu: Interestingly that Permalink change did not spam another Webmention to you: https://kartikprabhu.com/notes/fragmentioner
#
Loqi [Kartik Prabhu] Javascript UI to get the fragmention (http://indiewebcamp.com/fragmention) link of selected text. https://github.com/kartikprabhu/fragmentioner #indieweb
#
Zegnat Have the old URL redirect to the new URL, and send an update webmention for them, is probably the answer to that, kaushalmodi
#
Zegnat So when the webmention receiver goes to fetch the URL, it sees the redirect, and can drop the old one
#
kaushalmodi > Have the old URL redirect to the new URL
#
kaushalmodi > Have the old URL redirect to the new URL
#
kaushalmodi Yes I am doing that
[pfefferle] joined the channel
#
kaushalmodi > send an update webmention for them
#
kaushalmodi what is that?
#
Loqi That! (or "that ^" or "that ^^^") is a rarely seen reply often emphasizing agreement with a This post, but sometimes[1] merely emphasizing agreement with a previous reply https://indieweb.org/that
#
kaushalmodi (I wish I can shift+return to insert newlines here -- using IRC over web)
[kevinmarks] joined the channel
#
[kevinmarks] send the webmention again
#
kaushalmodi yes, that's what I did
#
aaronpk there are no newlines in IRC
#
kaushalmodi so now you have the wm's from both URLs (old and new)
#
aaronpk but yes an "update webmention" just means sending the webmention again
#
aaronpk if your old URL redirects to the new URL, then sending a webmention for the old URL should cause the receiver to update the URL
cweiske_ joined the channel
#
aaronpk the key is to send the webmention with the old URL as the source, otherwise the receiver wouldn't know that it changed
#
kaushalmodi Ah, I understand now.. I am not technically redirecting it.. I just created an alias in Hugo
#
aaronpk does that mean it shows the post at both URLs?
#
[kevinmarks] yes, both return 200
#
petermolnar that is one of my pains with static sites
#
kaushalmodi ok, I take that back.. hugo is causing the generation of the redirection meta
#
petermolnar and the reason why I have a PHP file built with 410 and 302 redirects
#
kaushalmodi <meta http-equiv="refresh" content="0; url=https://scripter.co/fragmentions/" />
#
petermolnar that's not a redirect-redirect
#
petermolnar as in it's not http 301 or 302
#
Zegnat Ah, yes, that refresh doesn’t say anything about the resource being replaced by the other one
#
kaushalmodi hmm, time to learn then
#
aaronpk hm there's also a meta tag for setting the http code
#
petermolnar ???
#
Zegnat I guess you could, maybe, as a receiver, check for rel-canonical to overwrite as well? Not sure if anyone is doing that?
#
aaronpk what is 410?
#
Loqi 410 is the HTTP status code for GONE, meaning whatever URL was requested is not there any more, and used in webmention CRUD to indicate that a source has been deleted https://indieweb.org/410
#
petermolnar aaronpk: that shouldn' work
#
aaronpk i thought there were some examples of saying http 410 in a meta tag
#
petermolnar the http code is sent before the content
#
kaushalmodi I can do that in the _redirects file for Netlify (where I host the site)
#
petermolnar (it doesn't mean it can't work, but it shouldn't)
#
Zegnat “that shouldn't work” - sure petermolnar, but this is HTML you are talking about ;)
#
petermolnar hence the correction
#
Zegnat What is meta http-equiv status?
#
Loqi meta http-equiv status is an HTML extension specification that adds the "Status" value to the existing meta element's http-equiv attribute as a method for representing the HTTP Status code to be returned by a web server serving that HTML document, and for consuming applications like webmention receivers to parse and interpret as such for use-cases such as representing a 410 Gone deleted resource on static hosting that often disallow setting HTTP headers directly https://indieweb.org/meta_http-equiv_status
#
Zegnat There you go aaronpk
#
Zegnat wow that dfn
#
aaronpk that's the one
#
kaushalmodi As I don't rename URLs often, not a problem, but I wished there were a programmable way..
#
petermolnar bloeh
#
Zegnat `<meta http-equiv="Status" content="410 Gone">`
#
petermolnar that's one ugly standards hack
#
aaronpk kaushalmodi: depending on the web server you're using you could do it at that level as well
#
aaronpk for example putting redirects in an .htaccess file for apache
#
petermolnar or build a very simple php file, like I do, if there's php on the server
#
aaronpk yeah totally
#
aaronpk i actually have my redirects in a text file and a php script reads them
#
kaushalmodi aaronpk: Yes, that's what I'll do https://www.netlify.com/docs/redirects/
#
petermolnar aaronpk: I build the whole PHP, the potential redirect values are in an array
#
kaushalmodi petermolnar: Netlify serves only static files, so PHP is not an answer (also I don't know PHP)
#
vanderven.se martijn edited /meta_http-equiv_status (+31) "Shorten the definition, this is probably enough." (view diff)
#
aaronpk oh cool that looks reasonable
#
petermolnar https://github.com/petermolnar/nasg/blob/master/templates/MagicPHP.html
#
aaronpk petermolnar: lol that sounds like my timezone lookup script hack. it's a giant PHP array with a bunch of lat/lng->timezone mappings
#
petermolnar well, I do try a bit more hacking there, as in checking for a few extra things, but the idea sounds similar
#
petermolnar it also handles 404
#
Zegnat ... That is an HTML file that will go through your template engine and then gets run as PHP? I don’t know what to say of that, petermolnar ;)
#
aaronpk kaushalmodi: yeah looks like using their _redirects folder is the best bet
#
kaushalmodi Thanks. So that solves 1/2 the problem.
#
kaushalmodi Now is there a way to override the "source" of the webmention?
#
aaronpk once you have a proper redirect set up for the old URLs, then you should send webmentions with the old URLs as the source. if kevin's site doesn't look like it's fixed at that point then that's a problem on his side
#
kaushalmodi Right now wm's are auto-sent with the URL's of the containing post as the source
#
[kevinmarks] well, on voxpelli's side
#
aaronpk [kevinmarks]: your side, you just happen to be using voxpelli's service :)
#
[kevinmarks] you can manually send a webmention with mention.tech or telegram
#
kaushalmodi correct, but I am sending it in batch each time
#
kaushalmodi cweiske is kind enough to lend me his wm sending server: https://github.com/cweiske/stapibas
#
Loqi [cweiske] stapibas: Standalone pingback server. Mirror of http://git.cweiske.de/stapibas.git/
#
kaushalmodi So I just send a POST to that endpoint, and stapibas crawls through my site Atom feed to send all the wm's
#
aaronparecki.com edited /Microsub (+261) "rephrase dfn again, microsub is not just for servers" (view diff)
#
Zegnat kaushalmodi, have you tested if it is able to do it for the old URL still, after making it a proper redirect?
#
aaronpk kaushalmodi: yeah you're going to need to do something special for this then. maybe make something that looks through the redirects folder and sends them for those URLs
#
kaushalmodi Zegnat: I haven't tested that yet.. that will just make the redirection more correctly
#
kaushalmodi As it stands, wm's won't be sent at all from the old url, only from the new url.
#
Zegnat The idea behind the update mention (that is: resending the webmention) is that you do it when a post changes, no matter what about the post changes. So when you change the URL from example.com/a to example.com/b you resend all the webmentions you previously sent, completely unchanged (still with example.com/a as the source).
#
Zegnat Now if you do further updates to the post, because the post has previously changed URL, it is safe to use example.com/b as the source. Receivers should have already gotten the message that a is now b.
#
Zegnat ... if that makes sense
#
kaushalmodi Yes, it makes sense. Just wondering why it didn't affect Kartik's post.. may be he manually gates the WM's?
#
kaushalmodi I'll go ahead and fix the redirect, and then let's see that happens
#
cweiske stapibas has no special redirect support
#
kaushalmodi So about the redirect, should 301 be good?
#
sknebel yeah
#
petermolnar Zegnat: do you have a problem with my jinja2 python template, named .html, that becomes a php and gets executed as PHP? :P
#
Zegnat No comment. Hahahaha
#
Loqi Zegnat: lol
#
petermolnar oh, btw they are associative arrays, eg url => true - it turns out looking up a key is ~3-4 multitude faster, than in_array
#
petermolnar magnitude
tantek joined the channel
#
kaushalmodi OK, 301 set on https://scripter.co/replies/fragmentions/
#
Zegnat Yes, I have used keys as the place for my array items in PHP before too. I don’t get why they can’t offer an actually optimised List or Set internal to use for that sort of thing
#
kaushalmodi kevinmarks: That duplicate wm is still there.. looks like you will need to delete the older one manually.
snarfed, [jgmac1106], tantek, dougbeal|iOS, maingo, [unoabraham] and swentel joined the channel
#
skippy micropub.rocks test 300 (Create an h-entry with a photo (multipart)) seems to be POSTing the multi-part data in a different format than that used by Indigenous for Android.
#
Loqi skippy: Zegnat left you a message 7 hours, 2 minutes ago: I think I have now covered the entire MF2 JSON syntax on http://microformats.org/wiki/microformats2-json - please have a read and let me know if it still leaves you with questions :)
#
aaronpk orly
#
skippy https://gist.github.com/skpy/2f1ae4296e22199c9d6f0ed46750afef
#
dgold if I wanted to do some sort of onboarding with nanopub, particularly creating user tokens in twitter &c, how do I get around the "Don't put the App Consumer token in the App" thing?
kaushalmodi joined the channel
#
kaushalmodi dgold: Wouldn't the suggestion in that issue work?
#
kaushalmodi i.e. have the user set the tokens in env vars, and then have PHP read those env vars
#
dgold kaushalmodi: i'm looking at going down the generations issues
#
skippy dgold: provide a sample config file for nanopub, and require users to fill in their own config file.
#
skippy dgold: eventually you could build the config from a web form that prompts the user (see how WordPress hndles this); but to start just require them to copy config.php.sample to config.php and plug in their own values.
#
aaronpk skippy: the two different output formats in that gist are throwing me off, but is the difference that there seems to be an extra array level in the indigenous one?
#
dgold skippy: that's, um, that's what I do now...
#
skippy aaronpk: yes. indigenous has an array of array values.
#
skippy dgold: apologies. i must have misunderstood your question.
#
aaronpk skippy: looks like if you set the name "file[0]" then PHP expands that into what you're seeing in the first example
#
aaronpk er, photo[0] in that example
#
dgold ah, there's no API for the `create a twitter app` stage of the process, never mind
#
skippy i tried to format that gist better. take another loo.
#
aaronpk https://gist.github.com/aaronpk/931ff687510b7d43d0d289268b6e0e78
#
dgold i'll get me coat
#
skippy aaronpk: eeeenteresting. that's helpful, PHP! Thanks! :troll:
#
aaronpk that seems like a weird decision for how PHP presents that, but oh well
#
aaronpk i would have expected $_FILES['photo'] to be a bunch of the normal looking file uploads
#
aaronpk but in any case, you should be able to handle both because if the app uploads two photos you'll need to handle that case
#
swentel yeah, it's to prepare for multiple uploads
#
skippy i'm trying to handle this with one photo right now. it works from Indigenous, but now fails the micropub.rocks test.
#
Zegnat Ah, the PHP $_FILES array. I am always happy that PSR-7 put some time into standardising that one.
#
swentel yeah, never really thought about the format either to be honest
#
swentel drupal uses symfony internally to handle files
#
swentel it's been ages since I've ever had to mess with that
#
swentel (luckily)
#
Zegnat I try to offload all my request handling into PSR-7 compatible libs these days. For the smallest projects I just grab https://github.com/Nyholm/psr7 which means I need no full Zend/Symfony/w.e.
#
Loqi [Nyholm] psr7: A super PSR-7 implementation
#
skippy that would be nice to use, if it had any documentation
#
swentel ooooh
#
swentel woow
#
swentel that is very nice
tantek joined the channel
#
Zegnat skippy, right, the documentation is basically “read the PSR-7 spec”. Which is not at all helpfull to non-veterans.
#
skippy not helpful at all.
#
aaronpk i can't keep up with all the PSR stuff
#
aaronpk everyone seems to defer elsewhere and it takes forever to actually find out what things are anymore
#
Zegnat I only know the ones I regularly use and/or have contributed small comments to
#
aaronpk i filed an issue on one project's docs asking for some better sample code or more concrete info, and they closed it without change saying that i should go read the PSR docs
#
Zegnat Yes, example code is basically non-existent. They debate (for a loooong time) on what a class interface should look like, and then they put that online and take their hands off of it.
#
Zegnat One of the reasons that PSR7 implementation I just linked exists: minimal files, only implementing the methods described in the PSR, easy to read.
#
aaronpk and then people write packages that follow it but then assume you know how to use it without any docs
#
skippy but nyholm has no description of how to invoke any of the PSR7 classes...?
#
skippy so i'm no better off than i was before.
#
skippy Zegnat: do you have existing code that demonstrates how to use Nyholm?
#
Zegnat Uuhm. Let me check.
#
skippy also, Zegnat, thanks for the microformats2-json wiki page. very nice work. and helpful.
[chrisaldrich] joined the channel
#
Zegnat Not ready on GitHub, skippy. I have https://github.com/Zegnat/php-website-starter/ which combines a lot of PSR stuff into a basic PHP website skeleton, but that is less than helpful :P
[matthilt] joined the channel
#
aaronpk the diactoros stuff drives me insane too. it takes me so long to figure out how to use it every time i try
#
aaronpk the docs are basically nonexistent
#
aaronpk and it's like come on just show me how to return an http redirect
#
Zegnat Hahaha
#
Loqi nice
#
Zegnat That is actually pretty easy with Diactoros, as it has special factories for creating redirects.
#
Zegnat Don’t need to build the entire PSR-7 response yourself
#
aaronpk i dont think ive seen that documented ::P
#
tantek what is Diactoros
#
Loqi It looks like we don't have a page for "Diactoros" yet. Would you like to create it? (Or just say "Diactoros is ____", a sentence describing the term)
#
aaronpk i don't even know how to describe that lol
#
aaronpk some of the new php stuff really annoys me, but i like that things are improving overall
#
Zegnat https://gist.github.com/Zegnat/baf88c68b517076a1baf6d67ce3a5364
#
aaronpk wtf "moveTo"?
#
Zegnat That gist is some (untested) code that you could do to save uploads
#
aaronpk oh that's not the redirect
#
Zegnat It is called that because PHP has always called that
#
Zegnat Hahaha, no, it isn’t
#
aaronpk *whew*
#
Zegnat https://zendframework.github.io/zend-diactoros/custom-responses/#redirects - is redirects in Diactoros, aaronpk
#
aaronpk gosh
#
aaronpk header('Location: /redirect') is just so much easier
#
Zegnat Yes, but the idea is to have your controller generate the full HTTP response, which is then send to the visitor at once.
romangeeko joined the channel
#
Zegnat So your controller would just `return new RedirectResponse('/user/login', 301)`, and then your dispatchers handles that.
#
aaronpk i know, and it's testable that way too, which is good
#
aaronpk can $uri be a string? that isn't clear from that doc page either
#
Zegnat Also gets rid of that nasty “headers are already sent” thing PHP likes to do so much
#
Zegnat Their “Typical usage” shows it as a string, right?
#
aaronpk so it does. i don't know how i missed that
#
Zegnat “The only required argument is a URI, which may be provided as either a string or Psr\Http\Message\UriInterface instance.” - is pretty clear in the doc text too :P
#
Zegnat Though I do not like these doc pages that cramp everything in one page. Why isn’t it a separate page for Redirects. Give the text some air to breath...
#
sknebel dgold: what do you use the Twitter API for? syndication? reply contexts?
#
Zegnat is not going to define Diactoros for IndieWeb wiki, because there would be no end to documenting symfony and zend libraries...
#
dgold sknebel: its for POSSE
#
dgold currently POSSEing notes, likes & retweets
#
dgold it _should_ be POSSEing replies, but that appears not to be working properly for some reason
#
dgold it sends them, it just doesn't associate them with the original tweet
#
sknebel dgold: then you could offer bridgy as an option, if people also get a webmention solution with it (afaik you can send them to the bridgy login through some documented URL). might be more hassle than it's worth though, given the webmentions and static site effort
leg joined the channel
#
dgold sknebel: that's all true - but the point is to try to do as much as possible without reliance on external services
#
sknebel makes sense. just wanted to have mentioned it in case you didn't think about it
#
Zegnat (Apart from all the PHP woes, I am happy you found the mf2 JSON page helpful, skippy!)
#
dgold ok - so, my twitter logic works for replies when replying to _my_ own tweets, but doesn't work when replying to someone elses
#
dgold i'm stumped
#
dgold it uses tantek's tw_url_to_status_id($url)
#
tantek perks up
#
dgold hey :)
#
dgold My nanopub is set up to do POSSE replies to twitter; it does this by using your tw_url&c to get the status
#
dgold this works perfectly when replying to _my_ tweets, it fails when replying to others
#
sknebel dgold: do you have an @mention in your replies?
#
aaronpk what's failing about it? you don't need to @-mention people in the text of the reply anymore
#
sknebel oh? I wasn't sure and checked, and twitter API docs still claim you have to...
#
dgold sknebel: that would explain it.
#
aaronpk weird. i definitely don't anymore
#
snarfed you may need to add `auto_populate_reply_metadata=true` to the API call. details in https://github.com/snarfed/bridgy/issues/703
#
Loqi [singpolyma] #703 Duplicate @reply prepended to tweet
#
aaronpk aha yes
#
dgold Sokath! His eyes uncovered!
#
aaronpk https://github.com/kylewm/silo.pub/pull/50
#
Loqi [aaronpk] #50 set auto_populate_reply_metadata=true
#
aaronpk that's where i fixed silo.pub to do that
#
dgold i'll goto j7mbo's docs and see if that's covered
KartikPrabhu and tantek joined the channel
#
dgold anyone have a spare tweet I can use?
#
KartikPrabhu dgold: https://twitter.com/kartik_prabhu/status/979175639268888577 ;)
#
@kartik_prabhu B: We need a better social network. A: Do you like ads? B: No! A: Can I sell your data? B: No! A: Can I have your data anyway? B: No! A: Do you want to host it yourself? B: No! A: Do you want to pay for this better social network? B: No! A: OK, bye! #indieweb (twitter.com/_/status/979175639268888577)
[tantek] joined the channel
#
dgold Temba! His arms open!
#
dgold https://twitter.com/ROUZealot/status/988843109931995136
#
@ROUZealot @kartik_prabhu If you see this, it worked! Sorry, I'll delete it in a second.… https://steelyglint.ascraeus.org/micro/1524593542 (twitter.com/_/status/988843109931995136)
#
KartikPrabhu dgold: where am I supposed to see this?
#
KartikPrabhu dgold: got a like and a reply on Twitter
#
dgold as a response to yours, I haven't written a display engine for steelyglint yet.
#
dgold the linkbacks are for me for debugging purposes
#
KartikPrabhu dgold: bridgy found the like
#
KartikPrabhu but not the reply
#
dgold hmmm
#
dgold that's a future problems issue tho'
#
snarfed dgold deleted the reply :P
#
dgold yes, I did, as promised
#
snarfed hence bridgy didn't find it :P
#
KartikPrabhu aah yes, it is also gone from my twitter UI
#
dgold thanks for that bit of info re the API call
#
dgold snarfed++
#
Loqi snarfed has 58 karma in this channel (379 overall)
#
KartikPrabhu seems like bridgy was ahead of the Twitter web UI
#
dgold aaronpk++
#
Loqi aaronpk has 139 karma in this channel (1624 overall)
#
dgold I would throw one at j7mbo, but I don't think he plays
snarfed, swentel, [cleverdevil], KartikPrabhu, [miklb], [pfefferle], tantek, loicm and kaushalmodi joined the channel
#
kaushalmodi So a generic Webmention question.. I am using webmention.io as my wm endpoint. If I understand correctly, it does not store the wm content but just the source/dest info... correct?
#
aaronpk It parses the page and stores what it finds
#
kaushalmodi And it is caching all those webmentions source/dest info ever since I started using it. So if I switch the endpoint, do I lose all old webmentions?
#
aaronpk I don't understand the question
jackjamieson joined the channel
#
aaronpk It receives webmentions for you and gives you an api to retrieve them
#
kaushalmodi I use webmention.io as the endpoint. Tomorrow if I move the endpoint to a different server, do I lose all the old mentions?
#
aaronpk if you stop using it it'll still have all the old stuff it received
#
kaushalmodi So if I move to a different server, I will then need to make API calls to both old and new endpoints? Sorry, I am trying to create a mental picture
[jgmac1106] joined the channel
#
aaronpk Or download everything from webmention.io so you can stop making api requests since it would never get new data after you switch
#
kaushalmodi I get it that technically I don't need to keep on making API calls to the old endpoint.. I can just do it once and download all the data as JSON.
#
kaushalmodi Yes
#
kaushalmodi But then I need to stitch together the mentions from the cache and fresh mentions from the new end-point..
[kimberlyhirsh] joined the channel
#
aaronpk Yes, that's always going to be part of changing any part of the backend of a site
#
kaushalmodi Thank you.. while making this mental picture for a blog post, I am preparing for questions that could come up from others, like "so is this moving your comments from Disqus to webmention.io?"
[kevinmarks] joined the channel
#
gregorlove.com edited /Flickr (+489) "/* History */ bought by smugmug citation" (view diff)
#
klezlab.it edited /Pelican (+117) "/* IndieWeb Examples */" (view diff)
#
klezlab.it edited /Pelican (-117) "/* IndieWeb Examples */" (view diff)
#
klez Sorry, I messed up, don't mind me :/
leg and tantek joined the channel
#
klezlab.it edited /Pelican (+117) "Undo revision 47209 by [[Special:Contributions/Klezlab.it|Klezlab.it]] ([[User talk:Klezlab.it|talk]])" (view diff)
#
dgold kaushalmodi: you should take a look at schmarty's Morris: https://github.com/martymcguire/morris
#
Loqi [martymcguire] morris: PHP webhook for caching webmention.io webmentions for static sites.
#
kaushalmodi dgold: Thanks! I'll save that for reference till I get started with my own server and learning PHP :)
#
kaushalmodi I believe an old school cron + API based fetch from webmention.io will also work.
#
kaushalmodi Note that I anyways do the API fetch -> JSON -> Using Hugo Data feature to convert those mentions to static HTML at the moment
#
dgold coolio
#
dgold schmarty's implementation is really rather nice, tho
#
kaushalmodi webmention.io -> triggers web hook -> Zappier -> POST to Netlify -> hugo fetches JSON from webmention.io -> Site rebuilds with new mentions
#
kaushalmodi (things I end up doing while trying to avoid JS and PHP :P)
#
kaushalmodi *Zapier
#
snarfed zapier++
#
Loqi zapier has 1 karma
KartikPrabhu joined the channel
#
kaushalmodi Draft of a diagram for a blog post on the above explains webmention publishing flow on Hugo: https://svgshare.com/i/6Qn.svg
[cleverdevil] joined the channel
#
aaronparecki.com edited /GoDaddy (+552) "add list of guest posts" (view diff)
#
KartikPrabhu kaushalmodi: by "user" oyu mean someone sending a webmention to the "website" ?
#
aaronparecki.com edited /Posts_about_the_IndieWeb (+346) "/* 2018 */" (view diff)
#
kaushalmodi KartikPrabhu: yes
#
GWG Does anyone put their reply context below their reply?
#
aaronpk micro.blog does sorta
#
aaronpk tho it's confusing and he's considering changing it
#
KartikPrabhu yeah it sounds confusing given current patterns
#
GWG aaronpk, someone commented on this on their Slack
#
aaronpk yea
#
snarfed GWG: i do sometimes. eg https://snarfed.org/2018-04-15_a-short-post-peter-stuifzand-2
#
kaushalmodi KartikPrabhu: OK, so I think that I should have "Sends a Webmention" arrow between User and Website
#
Loqi [Ryan Barrett] yup! bridgy publish converts HTML to plain text formatting in general, and to markdown specifically for github, before posting.
#
kaushalmodi Then what is that "thing" called that goes between Website and webmention.io?
#
GWG Have to think if it is worth adding an option to display it below
#
KartikPrabhu kaushalmodi: I don't use webmention.io so I really don't know :P
#
kaushalmodi well in general..
#
kaushalmodi any site would have a rel=webmention, right
#
KartikPrabhu not unless the rel value is put in the website
#
kaushalmodi when someone sends a webmention to a target URL, that endpoint looks up the rel=webmention in that target?
#
KartikPrabhu hmmm the actual webmention POST is sent to the webmention endpoint
#
kaushalmodi Basically I would like to properly diagram the initial part.. the interaction between User, Website and WM Endpoint.. before WM Endpoint creates that POST
#
KartikPrabhu WM endpoints don't create a POST request
#
KartikPrabhu the webmention sender creates and sends a POST request to the wm endpoint of the website
#
kaushalmodi then I am a bit confused to join those dots with the fact that anyone can send me a WM via https://telegraph.p3k.io/
#
kaushalmodi So where does telegraph link with my endpoint i.e. webmention.io?
#
KartikPrabhu telegraph looks at your website finds the rel=webmention to get your WM endpoint and then sends a POST request to that endppoint
#
kaushalmodi About "WM endpoints don't create a POST request", webmention.io specifically does that
#
kaushalmodi ah OK, thanks
#
kaushalmodi So if someone doesn't use Telegraph, they would need to find out what endpoint I am using manually?
#
KartikPrabhu or through their own sending software
#
kaushalmodi understood
#
kaushalmodi .. or through those WM sending forms which we usually have on the target posts
#
kaushalmodi and those forms will be "hard-coded" to that site's endpoint
#
kaushalmodi hmm
#
KartikPrabhu yes
#
kaushalmodi KartikPrabhu++ Thanks. Diagram update coming up
#
Loqi kartikprabhu has 19 karma in this channel (194 overall)
[kevinmarks] joined the channel
#
[kevinmarks] Twitter puts some reply contexts below
#
kaushalmodi Rev 2 of the WM interaction diagram: https://svgshare.com/i/6Qd.svg
#
Zegnat Hmm. I feel like personifying things like webmention sender and webmention endpoint to see if it can make technical plumbing more clear to people.
#
kaushalmodi Zegnat: You mean put a stick figure for User?
#
kaushalmodi But then does that make sense for the endpoint? Because that's a script
#
Zegnat A webmention sender is someone who speaks. A webmention receiver (“endpoint”) is someone who listens. The webmention is a message saying “Page A has a link to Page B”.
#
Zegnat The listener is only interested in this information if Page B is something they own (e.g. a blogpost they wrote), otherwise they ignore the message.
#
Zegnat The listener does not know if the speaker is telling the truth, so if they are interested they go and check if Page A really links to Page B. If the speaker was telling the truth, the listener may now want to broadcast the fact that Page A has linked to Page B (e.g. by embedding a comment below their blog post).
#
Zegnat ... or something along those lines
#
kaushalmodi Zegnat: Rev 3: https://svgshare.com/i/6S7.svg
#
kaushalmodi Still not sure if/how to personify the Website or the Endpoint
#
Zegnat I was talking more of webmentions in general than your specific case.
#
Zegnat E.g. you haven't drawn the step where webmention.io checks if Bob was telling the truth.
#
kaushalmodi Zegnat: Ah OK. About "The listener is only interested in this information if Page B is something they own" though, I just do a GET for the specific URL
#
Zegnat Yeah, Hugo doesn't need to care about that in the slightest as it is handled by the listener, the endpoint.
#
Zegnat And you have outsourced the listening.
#
kaushalmodi hugo loops through all pages and does a GET for all, each specific to the page's Permalink
#
kaushalmodi The initial POST from wm.io is just to kick off the whole domino chain
#
Zegnat Yep. But that very specifically only applies to people who use webmention.io, and not webmentions in general.
#
Zegnat And my little story board explained webmentions in general.
#
kaushalmodi Correct. That diagram is specific to a blog post I am working on, on how I display wm's on my site :)
#
Zegnat I.e. everything to the right of wm.io in your diagram is no longer webmention handling. Just how the already received webmentions are being embedded into the blog.
#
Zegnat Looking forward to reading the blogpost!
#
Zegnat I should find my bed before tomorrow arobes
#
Zegnat What time is it for Zegnat?
#
Loqi In Zegnat's timezone, Europe/Amsterdam, it is currently 11:59pm on April 24
#
kaushalmodi Correct. I am not going into a low level detail (as I myself am fuzzy on that :P Kartik helped clear some stuff above)
#
kaushalmodi Good night Zegnat!
#
Loqi night
#
Loqi totally
snarfed joined the channel
#
grantcodes swentel++ for the continuing work on indigenous
#
Loqi swentel has 4 karma in this channel (8 overall)
#
GWG grantcodes: Amen
snarfed joined the channel
#
grantcodes He's making me wish I had more time to work on my reader :P
#
GWG grantcodes: If he gets a nice stable Micropub client...I know you care about Microsub...then I'll be thrilled
snarfed, tantek, tomasparks and eli_oat joined the channel