[fluffy]like I was thinking, source would be http://example.com/foo and target would be /bar, and so the thing pushes a webmention to /bar’s endpoint with those params, and then the endpoint would go “oh target is relative so we’ll treat the destination as http://example.com/bar”
[fluffy]like when you do <a href=“foo”>blah</a> when you click on the “blah” you’re not doing a GET to “foo” you’re doing it to the fully resolved URL that was formed by urljoining the source page’s URL with “foo”
[fluffy]so like if you do <a href=“/foo/bar” class=“u-in-reply-to”>, the endpoint validates the target parameter by exact match to the <a>’s href attribute
[fluffy]like the two uses cases for this are 1) webmentions to pages on the same domain (without having to expand them out fully) and 2) webmentions using protocol-relative URIs like //example.com/blah
[fluffy]they are supposed to validate the URLs based on the spec but they can *handle* the URLs with relative resolution and chasing redirects and stuff (like consolidating webmention targets and stuff). I don’t believe webmention.io does, it sounds like Kevin’s thingus does some of it
[fluffy]or I guess the simple test is webmention.rocks’ case doing a relative link back to itself, if you’re just testing a standalone webmention sender (like pushl) as opposed to a publishing system (like a wordpress plugin)
aaronpkSo the flip side to this argument is maybe if you link to your own site you should link to the fully resolved url because that's more likely to actually show up right for people when viewing the post in their readers
aaronpkbecause it takes extra work for a reader to go to the trouble of resolving every relative url in the html content of an atom or rss or microformats content
[fluffy]I would think it’s up to the reader to resolve relative links in a feed like that? I know FeedOnFeeds does that (because a lot of feeds broke without it), and meanwhile Publ is friendly and lets you provide fully-resolved links for feeds.
malaclyps[m], [Rose], sfoster, jeremych_, romainbouyeDisco, [jgmac1106], [grantcodes], gRegorLove, vika_nezrimaya, dino and GelvenDiscord[m] joined the channel
jgmac1106i have an old iframe from a site that went dark and now a domain forward it is making my enitre site redriect to a squatter page…not sure how I can go and in fix
LoqiA story is a singular (one per profile) time stream collection post, that consists of ephemeral photo and video posts that are shown in sequence one at a time and disappear from the collection 24 hours after being added https://indieweb.org/story
jgmac1106wasn’t linear, just a way to layer openly licensed html elements <video><.audio><img><a><span> to make non linear multimodal remixes and stories, nto sure it needs its own page
Loqi[RimBune] websites is easy. you wanna make a checkbox? <input type="checkbox">. bam. you wanna make the text pink? css color: pink;. all it takes. you wanna put the thing in the middle of the box vertically? no you don't
[snarfed]grantcodes could you elaborate? i know that's the assumption here but i rarely hear specific examples. in-reply-to/repost-of URLs are one. got any others?
[grantcodes]Well by that I mean everything is there, but it's all in the content, not in a photo property. If that is to do with the microsub server feed parsing or not I don't know
[snarfed]but yeah i'd like to keep enumerating these to see where the actual problems are. so far it's not that data is missing, but that indie readers handle mf2 different (better?) than atom
aaronpkthe "content" in atom needs to be only the text content of the instagram post, and the photos/videos would need to be "attachments" or whatever atom does
aaronpkthere is 0 way you can convince me to extract these photos/videos out of the HTML content as is right now lol, i've been down that road before and it was not fun
[fluffy]In Publ I differentiate between things where order matters and where it doesn’t. If it doesn’t, it’s just, each instance gets its own line. So like `Tag: foo` `Tag: bar` `Tag: baz` for tags.
[fluffy]Like if I’m replying to someone else’s entry I’ll have `In-Reply-To: http://example.com/bad-take` as a header, which will generate `<a class="u-in-reply-to" href="http://example.com/bad-take"></a>` in the entry’s hcard section.
[fluffy]basically I don’t see the point to doing everything as semantic data structures with specific HTML-generating stuff when the end result is, to users, the same as having it formatted in markdown 🙂
[fluffy]I guess I kind of capitulated to the whole “lowest common denominator” thing because I have my feeds set up to support what exists now, rather than what I’d like to see.
[fluffy]like now FoF will follow a rel=“self” link and, if it’s different than the feed URL, will try loading it as a feed, and only if that’s valid will it update the subscription URI.
[jgmac1106]fluffy pardon neophyteness but a serial view in a social reader or site be neat. Like drop a url in and be able to jump to 6 out of 7, nav back and forth
[jgmac1106]But if you used a p-category inside a data class, parsed a collection this way, couldn't fluffy build views that allow you to jump issues, see them descending and ascending?
[fluffy]the “show unread only, oldest-to-newest” view happens to work well for me and my reading habits but being able to specify some things as being serial and some things being in-the-now would be good
[fluffy]like I don’t want to have separate readers for reading all my subscriptions, but I also don’t want to have to worry about which things are oldest-first and which things are newest-first (or at least “it’d be nice to see newer thins first”)
[fluffy]I mean for now I’m pretty happy with FoF as a reader. And I could theoretically add micropub to it. Microsub would be a bit harder but I also don’t see as much of a use case when it’s, like, already an integrated reader and subscriber.
[fluffy]I also want to add RFC5005 to FoF although that means also trying to square the circle of how FoF is pretty aggressive at trying to purge older content too. Which is configurable at least and RFC5005 backfilling could just be another toggle.
[Jamey_Sharp]If you're reasonably confident you can get old posts again from the origin server, because it provides metadata per section 2 or 4 of RFC5005, then you can keep just a minimum amount of information for all old posts, with enough to remember which feed document you probably have to re-fetch if the user wants to see it again.
[fluffy]FoF’s data model is fairly simplistic. It also doesn’t lend itself well to things like feeds with conditionally-visible entries (like multiple subscribers to feeds with private content that differs between the subscribers)
[fluffy]feeds with “magic links” are supported and maybe I should add a thing to express those in Publ. I’m just especially concerned about privacy leakage in that form though.
[fluffy]it’s very easy to imagine a situation where someone tries sharing a public entry from a “magic link” feed in a way that also exposes said magic link.
[Jamey_Sharp]Ah, okay, if you have the tombstones and you put them in the newest feed when you add them, then you're set, as far as standards conformance goes.
[grantcodes]Fixed that author issue fluffy. It was automagically caching stuff based on the author url, so it was working in feeds where authors have their own urls.
Loqi[Aaron Parecki] I've reached the point that a day trip to the Bay Area from Portland is effectively the same as a "long commute to work". I'm also pretty sure this is still shorter than some peoples' commutes within the Bay Area. https://aaronparecki.com/2019/08/28/59/photo.jpg
@mrejfoxThere is no reason that your personal website couldn't ping my personal website when you post an awesome new photo, and yet the entirety of some artists' work is held in an Instagram profile- these are choices we have made as developers, creators, and as a society (twitter.com/_/status/1167181611340849155)
jgmac1106always hesitant now to @ mention a blue check that doesn’t follow me…think it takes many (2) unsolicted @ mentions to trigger the harrassment bots
KartikPrabhu why would one send a relative URL in a POST parameter anyway?
[fluffy] like I was thinking, source would be http://example.com/foo and target would be /bar, and so the thing pushes a webmention to /bar’s endpoint with those params, and then the endpoint would go “oh target is relative so we’ll treat the destination as http://example.com/bar”
[KevinMarks] The naive check of "grep the source page for the target string" will work for a relative url
[jgmac1106] Both look quite useful
@janboddez On IndieAuth, and using one’s own site as their online identity. “As with owning your content by owning your URLs, even if it’s hosted elsewhere, we use our own site as our identity, not our profile on a 3rd-party site, even if we… https://mastodon.social/@janboddez/102700270261166620 (twitter.com/_/status/1167052109918986240)
Loqi It looks like we don't have a page for "zeega" yet. Would you like to create it? (Or just say "zeega is ____", a sentence describing the term)
[schmarty] jacky++ posting videos on peertube.social
jacky I might point people to follow the tag https://v2.jacky.wtf/tags/video on my site
[snarfed] interestingly if you scrape IG with a user session cookie, that overrides IP blocks. eg https://instagram-atom.appspot.com/ is still working fine
[grantcodes] Ahh, well I would move my feeds to the atom suite but the mf2 feed is so much richer in indie readers
aaronpk [grantcodes]: see the thread here https://github.com/indieweb/microsub/issues/21
@jgmac1106 ↩️ I get all the webmentions but yes I got the @ mention on Twitter too. Glad to hear that @withknown meets most accessibility (beyond the minor photo title as both syndicated content, as alt text).
Never thought about all the other tools… https://quickthoughts.jgregorymcverry.com/2019/08/29/cswordpress-i-get-all-the-webmentions-but (twitter.com/_/status/1167159159126732801)
@RonautDreams ↩️ @mrejfox i haven't used it but i think http://brid.gy tries to solve for this. (twitter.com/_/status/1167197963887812609)
@mrejfox There is no reason that your personal website couldn't ping my personal website when you post an awesome new photo, and yet the entirety of some artists' work is held in an Instagram profile- these are choices we have made as developers, creators, and as a society (twitter.com/_/status/1167181611340849155)
[tantek] Was going to say, someone could respond to this tweet with a POSSE reply invitation. https://twitter.com/mrejfox/status/1167181611340849155
@brendt_gd ↩️ @MetalArend @krakjoe Have you heard of webmentions though? (twitter.com/_/status/1167211031925596160)