#dev 2018-12-03

2018-12-03 UTC
[eddie], eli_oat, jgmac1106, [manton], tantek, [tantek], snarfed, kisik21, barpthewire, swentel, [jgmac1106], [Khurt] and cweiske joined the channel
#
petermolnar
I have a weird idea: https://cloud.google.com/vision/ can give you a list of exact matches of the image you upload. Should I assume that all exact matches for my photos are syndications, even if it wasn't done by me?
#
petermolnar
oh, that's interesting: I have a photo taken at the Barbican in London, google thinks it's Rome
[davidmead] and [jgmac1106] joined the channel
#
[jgmac1106]
Petermolnar unless you get catfished
#
[jgmac1106]
Go read Alan Levine's blod about his fruitless effort to remove his img from fake dating profiles
[tantek] and [eddie] joined the channel
#
sknebel
[eddie]: I wonder if it makes sense to try some way for a microsub server to have the client add custom buttons to a post in the reader? which e.g. could help with "how do you authenticate the magic link for the read post" etc
#
[eddie]
Hmm, I guess the question is what do the custom buttons help with?
#
[eddie]
Often times different Microsub readers will have different button options
#
[eddie]
Are you wanting to sync what button options exist to the Microsub server?
#
sknebel
with things where the current idea is "inject a link in the post"
#
[eddie]
Well, I don't think we would inject a link into a post for a Microsub reader
#
[eddie]
only external things that don't support Microsub/pub
#
[eddie]
like a Podcast app
#
[eddie]
or a Kindle article
#
[eddie]
But for Monocle, or Indigenous we never inject links into posts
#
sknebel
hm. for e.g. webmention moderation it has been proposed too
#
sknebel
I guess the podcast example you'd probably ask the reader developer to make native read post support, true
#
[eddie]
Hmmmm 🤔 I'm not sure webmention moderation should happen inside of Microsub
#
[eddie]
If anything that moderation link injection would happen on the webmention receiver side before it gets sent into a Microsub server as a post
#
[eddie]
so I think Microsub wouldn't know anything about there being a special link inside of a Webmentino that needed moderating
#
[eddie]
instead the webmention receiver would add the link, send it to Microsub (for example through Micropub) and when you clicked on the link it would take you not to a Microsub server but to your Webmention server
#
[eddie]
because it is likely people's Microsub servers will be in different locations than their Micropub/Webmention servers
#
[eddie]
I think most people's webmentions end up on their server near their Micropub post data (either in the same database or as a flat file in a folder near the post folder), which would be something interesting to research
#
sknebel
the microsub server adding display hints doesn't have to mean it actually has to "handle webmentions"
#
sknebel
but it might very well be unnecessary
barpthewire joined the channel
#
[eddie]
I'm just not sure how the Microsub server knows that there need to be webmention buttons if it's not responsible for adding the
#
[eddie]
unless it's parsing some mf2 properties from the posts it receives
#
sknebel
could e.g. be a per-feed config
#
sknebel
or that
#
[eddie]
Okay, so with the per-feed config, you would configure custom buttons essentially telling the Microsub server, for posts in this feed add a button with this title and that when clicked opens to this url
#
[eddie]
so you might say this config { name: 'Listened', redirect: 'https://micropub-server.example/?url=[URL]' } where [URL] gets replaced with the url of the feed item?
#
sknebel
maybe. ideally with some way of attaching authentication. don't have a detailed idea what one might want yet (and it's probably a bit to early to do something like this, lets try the link thing first and find its problems)
#
[eddie]
hmm interesting. Yeah, I think as you said, one we need to iron out some of the link/authentication problems first. Second, I think this wouldn't work for the podcast listen posts discussed in IndieWeb because most people will be listening to podcasts in a regular non-Microsub player
#
[eddie]
It is definitely an interesting idea to build out as we go thoug
#
[eddie]
argh! my iPad keeps messing up the last letter of my posts lol
#
[jgmac1106]
But the community would switch to a microsub player... Or in short term what if I have a podcast feed in my reader but I just use it to publish listens (all three categories) or h-review for example
#
[jgmac1106]
Maybe publish a reply post.
#
[eddie]
I don't know that *I* would switch to use a microsub player
#
[eddie]
haha hence my reservations
#
[eddie]
I do think adding podcast feeds into Microsub server should happen
#
aaronpk
I like that idea
#
[eddie]
It works pretty good I've followed some Microcast feeds
#
aaronpk
I've been doing kinda that by subscribing to the feed on microcast.club
#
[eddie]
Right now Indigenous for iOS shows a player icon to listen to audio attached to a post
#
[eddie]
but I haven't added Listen post support to it
#
[eddie]
Seems easy to say if audio exists, add a "listen button"
#
aaronpk
monocle supports audio too
#
aaronpk
but I think it'd be neat to provide a podcast feed from an aperture channel so that I could subscribe to it in a podcast app
#
[eddie]
Yeah, I think right now that is the optimal solution until another comes along
#
[eddie]
while continuing to provide better audio support within the native apps themselves
#
sknebel
aaronpk: "we could benefit from some documentation around how to handle people changing their URLs for IndieAuth apps" - http to https moves could be covered by following the redirect rules?
#
aaronpk
Yes and that should cover domain changes too, I just think it needs to be spelled out explicitly since it's often overlooked when people write apps. Even generic OAuth apps.
#
aaronpk
there was a post just the other day about this
#
[tantek]
I'm a little confused here - why would a user ever have to / want to worry about providing a specific "podcast feed" instead of just their normal "primary" feed?
#
[tantek]
or to put it another way, shouldn't a Podcast listener (AKA "app" / reader) be able to subscribe to any feed and then just download / queue up the podcasts that happen to be in that feed?
#
[jgmac1106]
Yeah but I would probe have a podcast channel in aperture
#
aaronpk
The discussion wasn't about the theoretical best podcast app, it's about how to do it right now with apps that already exist
#
[tantek]
[jgmac1106] that sounds like extra work, that's my point
#
[jgmac1106]
I might subcribe to someone's blog feed but not care for their podcast feed
#
[tantek]
right [aaronpk], that's where I was going next: another way of thinking about it, instead of trying to build podcast downloading / listening functionality into every Microsub client, how about we Microsub client support into existing podcast "apps"?
#
[tantek]
we encourage*
#
[jgmac1106]
And the feed also does the custoum button... Unless rule is if you see <audio> add button
#
aaronpk
sure but again that's not a solution that works *right now*
#
[tantek]
[jgmac1106] re: "might subcribe to someone's blog feed but not care for their podcast feed" so do you follow people on Twitter and get not care about their posts with photos / video / audio?
#
[tantek]
Note that Twitter has no way to follow someone and then mute just certain "types" of posts
#
[jgmac1106]
Following on Twitter is a h/t I never watch my main stream
#
[tantek]
so once again, I'm going to claim edge case feature request on special feeds / filtering of that sort
#
cjwillcock
tantek: re DNS short TTL - I withdraw the 600 has a special meaning of 'do-not-cache' comment. Testing showed this is not the case in the wild.
#
[tantek]
cjwillcock good to know
#
[tantek]
what is TTL
#
Loqi
TTL is Time-To-Live, a measure in seconds of how long a DNS record (such as a name-to-address mapping) is valid https://indieweb.org/TTL
#
[tantek]
cjwillcock did you find any references or examples of DNS providers with default / minimum TTLs that we could document on /TTL?
#
aaronpk
If I add a podcast feed output from Aperture then it will basically work the way Tantek is suggesting
#
aaronpk
if you follow someone's home page with a bunch of different kinds of posts, only the posts with audio would be output in the Aperture podcast feed for that channel
#
[jgmac1106]
But I do think now I could put all the podcast feeds into an Aperture channel, then Indigenous would recognize existence of <audio> add a custom button that would publish a listen post with podcast url, name,
#
[tantek]
[jgmac1106] tl;dr: "I love spending extra time micromanaging special feed subsets for every person I want to follow on the web" said no one ever.
#
aaronpk
Technically people do kinda do that, by following only someone's Instagram account or YouTube account and not their blog or twitter
#
[tantek]
Or they get annoyed at having to find/follow n different social media profiles for the same person and then get more annoyed with seeing so many dupes, but not all the time, so there's no easy way to just get everything once.
#
cjwillcock
I did not, my approach was to live-test/monitor Google's public DNS at 8.8.8.8 while I changed records at various TTL settings. I observed that any setting of seconds was respected, including 0 seconds as 'do-not-cache'. Grain of wisdom: legacy systems don't handle 0 consistently.
#
aaronpk
there are plenty of people i follow who I do that with, tho typically not people i know personally
#
[jgmac1106]
I want to publish a listen post without crashing. Not everyone just commutes by plane. Not sure it is edge case
#
[tantek]
cjwillcock that's definitely worth noting as an observaction on /TTL perhaps in a new "DNS examples" section listing "* Google 8.8.8.8: ..." explicitly
#
[jgmac1106]
[tantek] I know you are pushing the follow person model rather than person's x feed. Just not sure how it works in terms of post types... I do not want everything from everyone...
#
[tantek]
[jgmac1106] yet the default (e.g. FB, Twitter, any other "generic" social media silo) *IS* everything from everyone
#
[tantek]
seem to work ok for 100s of millions of people so maybe not a bad simple default to start with?
#
[jgmac1106]
No it isn't it is algorithmicly reduced
#
aaronpk
Also just want to point out again that people have accounts on different silos and post different kinds of content there intentionally
#
[jgmac1106]
Firehouse didn't work for Facebook and they removed it as did Twitter (though you can opt in).
#
[eddie]
tantek If you believe you should get everything from everyone try following chrisaldrich
#
[eddie]
😉 [chrisaldrich]
#
[eddie]
haha 😄
#
[eddie]
or for me that matter
#
aaronpk
or me lol
#
[eddie]
lol my watch, ate and listen posts will drive you crazy!
#
[jgmac1106]
Or me. I lose followers all the time bc I publish too much
#
[jgmac1106]
Yeah Eddie ate funions and Mountain Dew for breakfast is my fav post
#
[eddie]
lol Doritos 😉
[schmarty] joined the channel
#
[schmarty]
it's the user-centric reason to have algorithmic feeds. information overload, a need for filtering (or summarization).
#
[jgmac1106]
I think the better strategy is encouraging folks to list their feeds in their h-cards
#
[jgmac1106]
I go to someone's site or check [grantcodes] IndieWeb directory I can click on it.
#
[eddie]
I definitely think feeds linked with rel=feed (I think that's the one for h-feeds) from the identity url (like eddiehinkle.com) is the right path. Then it's up to the Microsub server/clients to present that in a friendly way about "topics" rather than technical about "feeds"
#
[eddie]
and display user friendly feed titles rather than feed urls
#
[eddie]
unless someone digs into it
#
[jgmac1106]
Yeah I had my feed links and folks told me to use regular urls with rel="feed"
#
[jgmac1106]
Would you use p-name for topic match or weight a p-category more?
#
[eddie]
by topic I am referring to a more user friendly way to refer to a feed
#
[schmarty]
jgmac1106: i don't understand how that is a dichotomy
#
[tantek]
[jgmac1106] your point about algorithmic feeds is a good one - note however that is an *automatic* way of throttling what you see from your followings, without having to manually pick and choose yourself
#
[jgmac1106]
Indieweb is about restoring control... Choosing myself not a machine
#
[eddie]
Notice though that algorithmic feeds have been DISLIKED
#
[eddie]
At one point in Facebook you could prioritize who you saw first
#
[eddie]
then with algorithmic streams they removed a lot of that control
#
[tantek]
[jgmac1106] indieweb is about independence and choice. While yes algorithmic feeds do a horrible job at this, it doesn't mean everyone wants to sign up for doing manual feed micromanagement themselves
#
[jgmac1106]
@schmarty I could had p-category=moocs and then a p-name for each course... In that instance the category higher than name
#
[tantek]
being forced to choose everything can also be a time-wasting tax of sorts
#
[schmarty]
jgmac1106: that hierarchy sounds arbitrary?
#
aaronpk
sknebel: quickly thinking through migration, if someone logs in with a new URL, how can the app know about their old URL?
#
[eddie]
If you enter eddiehinkle.com and are presented with 4 items with checkboxes (timeline, photos, videos, checkins)
#
[eddie]
You can check or uncheck any of those four
#
sknebel
aaronpk: oh right, the client knows it but not the app?
#
[eddie]
and then never have to do that for eddie again
#
[eddie]
how is that a time waste?
#
aaronpk
The app has no way to know the two URLs are linked
#
[jgmac1106]
Possibly and few communities probably get that granular... Edu world probably different we build and let feeds go dormant all the time
#
[jgmac1106]
And we write a ton.... No firehose...
#
aaronpk
I'd need some indication that this new URL is replacing the old URL, or a nudge that the old URL should be updated
#
[jgmac1106]
!tell @atronpk to change my url
#
sknebel
aaronpk: one direction it can discover, by following redirects, but the new URL needs some way to opt-in...
#
aaronpk
It can verify by following redirect, not discover
#
sknebel
user has to log in with the old url of course
#
aaronpk
which they can't as soon as they change it like in this example
#
sknebel
and into the app itself, not a client
#
sknebel
e.g. log into aperture, not a reader
#
aaronpk
Even the simple case of logging in to aperture doesn't work really
#
sknebel
it needs some verification, yes
#
aaronpk
as soon as you move domains and make the redirect, or change your Wordpress folder, nothing will report the old URL as the identity
#
[schmarty]
eddie: the person following you may have to spend time evaluating all of your feeds to understand the content in each one. also, you've carefully curated your feeds for content, but e.g. wordpress sites offer a feed per category, or by kind with post-kinds. that's a combinatorics explosion to sift through.
#
[eddie]
schmarty: that's true, wordpress could end up listing a lot of feeds
#
[schmarty]
there's an opportunity cost with time spent micromanaging which feeds you're following
#
sknebel
user logs in with old url into aperture, aperture follows the redirects, sees they are permament and applies them?
#
[jgmac1106]
Yes so me as blog author say "If you like X follow Y"
#
aaronpk
Let's say I want to change my website to aaronpk.com, so I change the domain and redirect aaronparecki.com to aaronpk.com, when I log in next there's nothing that will let me log in as aaronparecki.com because my site is moved
#
aaronpk
"Log in with old URL" isn't practically possible
#
[jgmac1106]
I am a perfect example [aaronpk]. I need to move aperture to quickthoughts.jgregorymcverry.com from jgregorymcverry.com
#
sknebel
huh? I thought discovery says to follow that and not send the old url to auth endpoint?
#
[schmarty]
also, i feel like you've put a lot of thought into how you want folks to consume your site based on the feeds you've established. more so than many! but i still might want some subset of the items you're putting into feeds, so it's not enough for a reader to let me select feeds that the author curates.
#
sknebel
If an HTTP permament redirect (HTTP 301 or 308) is encountered, the client MUST use the resulting URL as the canonical profile URL.
#
aaronpk
Oh you are saying enter the old URL in the sign in box
#
aaronpk
hm that could work but that is a very specific instruction to users
#
[jgmac1106]
@schmarty yes and you still would have ability to add a feed for me of just photos in posts tagged S #IndieWeb.... But that would be up to you
#
[eddie]
ahhh schmarty so instead you're saying we should be able to filter the feed itself?
#
sknebel
alternatively you could do that with a "claim old account" option somewhere after logging in
#
aaronpk
Yep that's more like what i was thinking because realistically that's what's going to be needed
#
[schmarty]
eddie: i think multiple approaches will be necessary!
#
aaronpk
and that's the specific guidance i was talking about needing
#
[eddie]
schmarty: makes sense to me. Aperture supports some of that in the backend right now, but I definitely think it would be useful to move some of that into the Microsub spec so that clients can update that info without someone having to go to the server
jgmac1106 joined the channel
#
sknebel
loss of a domain becomes also worse and worse the more information we access through it (especially now we are startingto talk private posts too), another interesting problem to consider
#
sknebel
cjwillcock: [tantek]: with times that low browser caches etc also could come into play in practice. e.g. I don't know if Firefox special cases TTLs that are below it's default cache time (which I think is 60s)
#
[tantek]
sknebel, aaronpk, rel=me solves this!
#
[tantek]
(plus redirects obv)
#
[tantek]
re: "quickly thinking through migration, if someone logs in with a new URL, how can the app know about their old URL?"
#
[tantek]
new URL has <link rel=me href="oldURL"/> entries. client retrieves oldURL to verify, server redirects (permanent) oldURL to newURL. presto bi-directional identity consolidation demonstrated.
#
[tantek]
no need to enter oldURL into any UI
#
[tantek]
if you need a more assertive statement that newURL is definitive, then also a <link rel=canonical href=newURL/> at the newURL
Kaja joined the channel
#
[tantek]
so when the redirect(s) terminate(s) then you can see that the final URL retrieved matches the rel=canonical, you've even more strongly confirmed that you've found the correct new URL for the user (where they migrated to)
wagle joined the channel
#
jgmac1106
[grantcodes] I meant to ask as I saw so many similar…used to bootstrap now is CSS grid classes in your directory if your started with the same Brutalism template I used for a quick copy/paste hack job\
#
aaronpk
hm what are the odds of getting people to add a rel=me to their old url
#
aaronpk
would work tho
#
aaronpk
likely multiple approaches will be necessary
#
[tantek]
was just trying to think of something that used existing tech, that people had already some experience with
#
jgmac1106
I will do it, it is a good idea for migration advice in general
#
[tantek]
the point here is that people don't have to do anything special on their oldURLs other than the redirects that sknebel mentioned above
#
[tantek]
the only place they have to "do the work" is on the newURL which they presumably have more control over anyway (since they switched to using it)
bradenslen joined the channel
#
cjwillcock
sknebel: tantek: we can't rely on layers that consume DNS to go to the DNS on each request, as sknebel says. IIRC the pancakes we were making had something to do with mobile phone as web server.
#
cjwillcock
In this case the mobile is probably behind NAT and would need port mapping on the LAN router to serve content requests coming into the public IP. Even realtime DNS updates would not be sufficient for the use-case, as port-mapping updates would need to travel between networks with the mobile device, and this isn't possible unless you are also the network admin / privileged for all those
#
cjwillcock
networks.
jgmac1106, snarfed, eli_oat, KartikPrabhu, eli_oat1, [tantek], [jgmac1106], [eddie], jgmac1106_, [Csongor], benwerd and [chrisaldrich] joined the channel
#
[eddie]
Anyone experiment with Twitter's Archives before?
#
[eddie]
It's interesting, the tweet.js file gives the old fashion RT <original tweet text> for the content, which is nice, but strangely it seems like they don't actually provide the url of the originally tweeted url
#
[jgmac1106]
I have requested a half dozen times and never not once gotten them
#
[eddie]
so there is no way to find the original tweet from the Twitter archive that I can tell unless its in another file
#
[eddie]
That's strange, jgmac1106
#
[jgmac1106]
Yeah.... Nobody ever replies to my inquiries... Will try again today... Want to import them all. My archive only goes to 2014. I lost 2009-2013 bc I used TwapperKeeper
[schmarty] joined the channel
#
GWG
aaronpk, re the migration, WordPress should return the current account URL, not the old one
[davidmead] joined the channel
#
[davidmead]
jgmac1106 eddie I asked for mine back in 2012. You can view it on my site, https://davidjohnmead.com/archive/twitter/, but I don’t think I’ve ever got another copy even though I’ve asked for one
#
[eddie]
Hmmm somehow that UI has a correct retweet you made
#
[eddie]
Is that using the current Twitter JSON export?
#
[jgmac1106]
i wonder if gdpr changed the export?
#
[eddie]
Ohhh maybe it did
#
[jgmac1106]
I don't "own" the OG tweet but I do "own" my RT
#
[eddie]
Yeah, that might be what happened
#
[eddie]
although still I feel like they could provide the URL
#
[eddie]
url != content
#
[jgmac1106]
that is position Instagram ended up taking with reactions, likes, and comments
#
[davidmead]
If I clikc on my first tweet in my archive it does go to Twitter https://twitter.com/davidmead/statuses/331332962
#
[davidmead]
though I have no URL on my domain to point to it
#
[davidmead]
This from the 2012 readme file…
#
[davidmead]
* The JSON export contains a full representation of your Tweets as returned by v1.1 of the Twitter API. See https://dev.twitter.com/docs/api/1.1 for more information.
#
[davidmead]
* The JSON export is also used to power the archive browser interface (index.html).
#
[davidmead]
* To consume the export in a generic JSON parser in any language, strip the first and last lines of each file.
#
[eddie]
hmmm interesting. I'll have to look and see if I have an index.html in mine
#
[davidmead]
As I say, I’ve never received another backup, even though I’ve requested it multiple times
#
[eddie]
oh yeah the current ones don't return an index.html anymore
#
[jgmac1106]
@davidmead.....yeah I don't get it...My requests go ignored
#
[davidmead]
I also got a `.csv` version too eddie
#
[eddie]
ohhh yeah they have definitely changed the export
snarfed, [tantek] and benwerd joined the channel
#
[jgmac1106]
I removed text decoration on ul across my site but want to add it back for article .li ...how can I do that?
#
[jgmac1106]
list-style-type worked, nvm....really need to go back and not make default be no text decoration
[eddie], benwerd, snarfed, [schmarty] and doubleloop joined the channel
#
doubleloop
Can anyone help me with what I'm doing wrong here: https://webmention.rocks/test/1
#
Loqi
[Webmention Rocks!] Discovery Test #1
#
doubleloop
I've replied to it (https://doubleloop.net/2018/12/03/4631/), but my reply is listed in the section that says "The mentions below linked to this post, but did not include this post's URL as an in-reply-to property."
#
Loqi
[Neil Mather] We’re acquainted with the wormhole phenomenon, but this… Is a remarkable piece of bio-electronic engineering by which I see much of the EM spectrum ranging from heat and infrared through radio waves, et cetera, and forgive me if I’ve said and l...
#
doubleloop
I think I have an in-reply-to property, so not too sure what I'm missing
#
KartikPrabhu
aaronpk: ^
#
KartikPrabhu
doubleloop: yeah your markup seems corret
#
aaronpk
Oh I guess it doesn't recognize the nested h-cite
#
KartikPrabhu
it should the in-reply-to>h-cite>url is correct
#
aaronpk
Yeah markup looks good I need to update the tests
#
aaronpk
file an issue?
#
KartikPrabhu
doubleloop: looks like the test is wrong. you can file an issue here https://github.com/aaronpk/webmention.rocks/issues with your example
#
doubleloop
Thanks KartikPrabhu, thanks aaronpk
jgmac1106, benwerd, [chrisaldrich], [jgmac1106], snarfed and snarfed1 joined the channel
#
aaronpk
i've got some bizarre webmention loop going on here
#
aaronpk
i think that was all triggered when I favorited chrisaldrich's reply https://aaronparecki.com/2018/12/03/10/
#
aaronpk
hmm this "mentions" page could use some cleanup... it's not particularly interesting to know that I happen to have @-mentioned chris a bunch of times https://boffosocko.com/mentions/
#
Loqi
Pingback: Doug Beal
#
aaronpk
somehow that caused a cascade effect and now a bunch of my own posts have mentions of themselves all "via boffosocko.com"
#
jacky
inception
#
sknebel
they don't have mentions of themselves, the links point to other pages
#
sknebel
other posts of yours
#
aaronpk
right but my posts don't link to those
#
sknebel
could it be that your webmention endpoint checks if the backlink is *anywhere* on the page, and then uses xray to parse the fragment?
#
aaronpk
like it's saying my post from january links to a post from august which is time travel
#
sknebel
and accepts it, even if the backlink is not in the mf2 object?
#
aaronpk
yeah that's likely
#
aaronpk
but what sent the webmention with those particular source and target URLs?
#
sknebel
superfeedr?
#
sknebel
or maybe WP does it it in some weird case?
#
sknebel
do you log webmention source IP?
benwerd joined the channel
#
aaronpk
hmm not sure
#
aaronpk
nope :(
#
jgmac1106
@aaronpk why I think webmention needs something like this: https://webmentionavatar.glitch.me/ for the generic domain level mentions
#
sknebel
it also seems like (native?) comments people left on posts on his site appear on that mentions page...
#
aaronpk
i have no idea what i'm looking at
#
Loqi
[john] Hi Greg, Not sure how you want the image to display, but I had a quick pass at this: http://johnjohnston.info/blog/flickr-faves/
#
[jgmac1106]
basically an img, with a hover effect and what would be a counter number of replies, mentions, likes
#
aaronpk
so, clustering?
#
aaronpk
that would help but isn't really the source of this problem
#
[jgmac1106]
otherwise your generic webmentions on wordpress just pile up
#
[jgmac1106]
yeah so your facepile would say, 45 likes or something....the weird part is why did the like go to his generic homepage and not to the specific post
#
[jgmac1106]
"somehow that caused a cascade effect and now a bunch of my own posts have mentions of themselves all "via boffosocko.com"" happens to me too
#
aaronpk
looking at my telegraph logs... i just sent a ton of webmentions because i support sending salmentions and that just caused a dozen of my posts to get updated so i just sent out like a hundred webmentions
#
aaronpk
yeah i don't quite understand what just happened here
#
aaronpk
here's a post from june that i just sent a bunch of webmentions for again https://aaronparecki.com/2018/06/26/6/
#
Loqi
[Aaron Parecki] Ready to kick off #indieweb summit! — with Chris, David, Marty, anomalily, gRegor, Jonathan, Manton, Tantek, microformats https://aaronparecki.com/2018/06/26/6/photo.jpg
#
aaronpk
and it just got 5 webmentions from boffosocko.com
#
aaronpk
er, via boffosocko.com with a source URL that is my own
#
[jgmac1106]
but why did your fav not go directly to the his reply? It should have never gone to his domain
#
aaronpk
like how bridgy fakes the source url for twitter.com
#
aaronpk
that's pretty bad markup because it's saying my post permalink is an h-card which is not true