#GWGTrying to think of other ways. The scenario being one on the list I have of requests. Someone wants to have two WordPress sites, and use a single token and authentication endpoint. And I tie the endpoints into the install.
#@chrisbiscardi↩️ Nothing, comments are awful and have a mostly negative impact.
Sometimes I use Webmentions at most, which basically shows twitter discussion. I built gatsby-plugin-webmentions for this (twitter.com/_/status/1220570829622169601)
#[tantek]TBH that's enough for me to not recommend them for any webhosting, and if that's how their customer service works, it makes me question sending anyone to them for DNS registration too
#[tantek]Anyone here using Gandi for their personal domain(s)?
#jeremycherfasI used to have everything at Gandi, but over the past few years I moved away for just such reasons (except that they never lost my snapshots).
#Loqibusiness models are ways to offer indieweb products and or services that users pay for both for their own benefit, and in support an open indie web https://indieweb.org/business-model
#Loqi[Jamie Tanna] #IndieWeb folks using Indigenous for Android (https://indigenous.realize.be/) or at least allowing a Geo URI (geo:51.501476,-0.140634) on their #Micropub endpoint - do you attempt to remap that longitude / latitude ref to a place, or do you currently...
#swenteljamietanna[m], I currently don't do much it yet, I just store it. I do use the q=geo query option to get suggestions for places though
#aaronpkjamietanna[m]: i have, i would say, a lot of experience with this, and trying to do any sort of automatic venue detection is going to be mildly successful at best, and only in the most ideal circumstances
#swenteljamietanna[m], but in the future, the idea is to have a map with all checkins/geocaches/other post kinds and put markers on them
#aaronpkif you live in a low-density area where venues are spaced out 50-100 meters, then it's pretty straightforward to map a lat/lng to a place. but as soon as you get into cities, it falls apart very quickly
#aaronpkthe location reported by phones isn't always super accurate, so the dot might show you across the street. not to mention buildings with multiple floors.
#jamietanna[m]That makes sense. I seem to remember Quill allows checking into an actual place for that reason?
#aaronpkno i dont have any checkin interface in quill
#aaronpkI know that one works in some places outside the US but map data is one of those things that varies wildly due to licensing and stuff
[LewisCowles] joined the channel
#aaronpkIf by "nothing" you mean like not even a blank image of the size requested, that might be a bug, so file an issue with a sample url that causes it
#swentelaaronpk, oh, no, so, I see the marker (or the path when passing multiple markers), and an image, just no streets etc
#swentelbut I'm going to experiment with the Here SDK for android
#swentelit allows for quite some requests in the free tier
#swenteladded screenshot on the indigenous page too
#swentelI've been thinking to maybe contact Eddie too to get the iOS version going again, but I first need to get my hands on a Mac ;)
myfreeweb, [snarfed] and [jgmac1106] joined the channel
#[jgmac1106]swentel my goal would be if I have tracker active I could pin a post or photo (syndicate) to the tracker, I know trackbook wants to stay so simple and will never allow pinning
#[jgmac1106]So currently I have to do this manually after the fact by finding time and/or gps coordinates of a post and picture and then add to gpx after.....only did it once...too much
#[snarfed]has been waiting for them to announce that they're turning off old unapproved API keys, including bridgy and granary. head in the sand, mostly. no clue what to do about that yet 😐
#[snarfed]from https://developer.twitter.com/en/application/rejected : "Thank you for your interest in the Twitter developer platform. Based on our review of your use case, we are unable to approve your developer application at this time...We don’t currently allow you to appeal this decision."
#[snarfed]easy there, that was just a snippet, there's lots more detail and links to the policies
#[jgmac1106]Just poke fun at projectbluesky at that point snarfed...
#[jgmac1106]We could document successful wording that has worked for people
#[jgmac1106]I assume end is near as well..the review is automated need to learn trigger words
#[snarfed]i'm not saying twitter is bad or wrong here. i just expect granary/bridgy aren't the kind of apps they want, and don't comply with their policies, and i don't know if i'll be able to get around that
#[jgmac1106]I know it took couple tries to get Sadiks app approved for Known... We would get rejected, write different app and apply again
#[snarfed]yeah, the API use case for POSSE is very different, and more ok w/them, than backfeed
#[jgmac1106]Ohh yeah pulling out and displaying content people shared on Twitter without using their embeds... Privacy... I gave Twitter not snarfed permission...
#[jgmac1106]I think with threaderapp it was displaying @ usernames and linking to profiles
strugee joined the channel
#[KevinMarks]Right, if you built the thing Sarah Edo was asking for upstream they'd block your api key.
#[schmarty]i may have gone a liiiittle bit overboard in how i built it, but essentially i store a data file per incoming webmention. the filename is a hash of the URL that sent the webmention, so that also becomes the lookup key under site.data.webmentions.mentions[HASH].
#[schmarty]the receiver makes note of the target on my end and i have an index that maps my posts's path to a list of hashes
#[schmarty]i have some template logic that then renders those by looking up any entries in the webmention index for the current page's path and loading and rendering each one.
#[schmarty]i keep them fairly separate as i haven't had native comments on my site in a while.
#[schmarty]oh and to answer your earlier question - my webmention-catching service is separate from Hugo. the service manages the entries in data/webmention_io/
#jamietanna[m]Schmarty I'm also using JSON format for my files in `content/` in Hugo - makes it easier to publish indie post types from something automated
#[schmarty]superkuh: i only store one copy of any webmention per source URL. so i can get notified about it as many times as they want, but it won't save any new data.
#sivy[schmarty]: do you ahve the json file format documented somewhere?
#jamietanna[m]GWG interesting, looks OK for pricing - $5/device and 7500 requests a month. Much better than the 100 I get currently with PushBullef
#superkuhAh, okay. That probably mitigates most situations.
#[schmarty]superkuh: it could still definitely be abused with do-nothing URL changes.
#sivy(happy that Go has nice struct <-> json mappings :)
#[schmarty]oh another wrinkle is that i did not want to handle all the webmention verification, so Morris is actually a webhook receiver for webmention.io 😬
#[schmarty]that bit could be swapped out if you were doing your own handling.
#sivyi think I’m going to write my own receiver/verifier
#[schmarty]sivy: webmention.io sends a "jf2" format payload with some details it was able to intuit about the webmention. aaronpk has a doc'd example somewhere on the site that i'm trying to find now.
#[schmarty]if i had to do this over again, i've been thinking of ways to make it work in hosted-but-politely-cached build environments like netlify.
#[schmarty]i'd probably have the build script run some version of that refresh script, paging back through all mentions until it finds that it's overwriting cached ones. that way a fresh build will get them all, but it won't abuse wm.io by pulling them all every build (😱)