2017-05-30 UTC
[cleverdevil], KartikPrabhu, KevinMarks, dougbeal|mb1, eli_oat, eli and gRegorLove joined the channel
# 04:06 gRegorLove GWG: Idea for WP webmention plugin: if pings are disabled for a URL, don't include the rel=webmention header for that URL.
# 04:07 gRegorLove Just ran into a "Pings are disabled for this post" 400 response sending a webmention to manton's post
# 04:14 GWG gRegorLove: I tried that. There was an issue and it was reverted
# 04:14 eli_oat @gRegorLove — I saw something similar when I tried to send a webmention to a post on baty.blog
# 04:19 GWG gRegorLove: Feel free to reopen the issue though.
# 04:20 gRegorLove I'll leave it up to you guys. Nothing major to me; sounds like it's already on your radar.
# 04:21 GWG I can't remember what the problem was though
# 04:36 GWG I remember there were problems, but I can't recall what they were.
# 04:37 GWG Oh well, I am knee deep in restyling my post types.
KevinMarks, cweiske, jeremycherfas, [colinwalker], loicm, dansup, deathrow1, ben_thatmustbeme, kants and miklb joined the channel
eli joined the channel
jeremycherfas, jeremycherfas_ and singpolyma joined the channel
# 14:11 ben_thatmustbeme at least getting them on the same vocabulary and bringing jsonfeed in to jf2 before is starts is better than having another seperate branch
eli_oat joined the channel
# 14:33 Loqi [cweiske] did you know that there is no way for micropub clients to tell the server that a token should be considered invalid?
# 15:13 sknebel the latter seems simple enough, although I find it interesting that they require valid auth credentials to revoke a token
# 15:16 aaronpk Looks like public clients only have to include the client id
# 15:20 aaronpk For IndieAuth I'd be inclined to reuse the token endpoint for this request to avoid defining a new rel value
# 15:34 ben_thatmustbeme wondering if it makes sense to have a seperate profile of JF2 for h-entry that sort of codifies those
# 15:34 aaronpk for each h-* object there's a list of which properties are arrays and which are single values
# 15:35 aaronpk xray supports h-entry, h-event, h-review, h-recipe, h-card, and h-product and h-item because they are used in h-review
# 15:35 aaronpk oh and it knows which properties expect URL values which used to build refs
# 15:36 aaronpk i could probably DRY up this code quite a bit now that this pattern has been established
# 15:43 aaronpk i still prefer including both html and plaintext content
tantek and leg joined the channel
# 16:22 Loqi [voxpelli] node-fetch-politely: Ensures polite outgoing HTTP requests that respects robots.txt and aren't made too close to each other
# 16:23 tantek anyone have a POSSE editing step in their publishing flow?
# 16:23 tantek e.g. when you publish a post, options in your UI to customize the automatic default text that you POSSE to Twitter, FB, etc.?
# 16:25 tantek e.g. right now I often put @ or # in my article names because I know my code will be POSSEing the article name to Twitter, but that kind of contorts / pollutes the article name
# 16:25 tantek so the use-case (for me) is: not having to use @ or # in a article name just for the sake of Twitter
# 16:29 sknebel 25 hours until sknebel: ask sgreger about POSSE UI
# 16:29 Loqi I added a countdown scheduled for 2017-05-31 7:29pm CEST (#6007)
# 16:35 schmarty tantek: keeping platform-specific @s and #s out of places where it doesn't belong (like my site) is why i only manually POSSE
tantek, [shaners] and gRegorLove joined the channel
# 17:35 aaronpk whoops, something is wrong with my webmention handling. my comment on adactio's post is causing his post to show up as a mention on my post
# 17:38 tantek aaronpk, is your comment inside the "content" on his post?
# 17:39 aaronpk no, he even properly marks up my comment in the "comment" property!
# 17:40 aaronpk oh! I bet it's because my post has a link to his post in the content, and i bet my code is only checking for that case if it's a property such as an in-reply-to
# 17:41 aaronpk yepyep that's it. I am only checking the in-reply-to property of my post. I need to expand that to check all properties as well as the post body
# 17:44 aaronpk hm, and i also just ignore webmentions that fall under that case
# 17:44 aaronpk should i be doing anything with them? I don't even store them
# 17:46 aaronpk not sure what to call this... a reciprocal webmention?
# 17:46 aaronpk when i get a webmention back from a site i just sent a webmention to because it displays my comment
# 17:46 tantek because you know your comment was "accepted" as it were
# 17:47 tantek especially when more folks start supporting / requiring Vouch
# 17:47 aaronpk yeah, it's entirely possible i get the webmention quite later than when i sent it, such as if they moderate manually
# 17:47 tantek indeed - if you do get it later, then it may be worth a notification?
# 17:48 aaronpk so that's enough of a reason to store them instead of just skipping them like i do now
# 17:48 tantek Definitely worth writing up in both "comments presentation" and "when should you send a webmention"
# 17:48 aaronpk so then i need to store some flag that says not to display this webmention
# 17:49 tantek is it possible to determine that it came from your own comment?
# 17:49 aaronpk i guess at display time i could check if my post originated it but that check is not trivial (look at all properties of the post, plus parse the HTML and look at all <a> tags)
# 17:49 aaronpk so i would want to determine that once and store the result
# 17:50 tantek it could also be a hint of salmention support right?
# 17:50 aaronpk yes! it likely indicates the site that sent it supports salmention
# 17:51 aaronpk who was it that displays the status of webmentions they sent as hover tips over links in their posts?
# 17:52 voxpelli Is it really required in Salmention to ping a page when that page pings you?
# 17:52 aaronpk voxpelli: if you follow the algorithm you will do that
# 17:54 tantek I do like the notification that the comment has been posted
# 17:54 voxpelli aaronpk: well, if adactio had been supporting Salmentions in the same way as you do right now, then you would have an inf-loop of Salmention pinging right now, no?
# 17:55 tantek that could translate into something you see in your own posting UI
# 17:55 aaronpk voxpelli: no because my second webmention would appear as an update to him
# 17:56 voxpelli aaronpk: yes, an update that your post has received a comment, and adactio will then notify you that his post has received another comment in its comments thread?
# 17:57 voxpelli It received his post as a comment of your comment – Salmention do support nested comments, no?
# 17:57 aaronpk the fact that both our sites support salmention right now and we didn't get into an infinite loop is a good sign
# 17:59 voxpelli At least it is technically possible for it to become circular so that some one has to stop confirming their webmentions
# 17:59 aaronpk i don't think so... because one or both will start seeing the other as an update instead of as a new comment
# 18:00 aaronpk so if i receive a webmention taht doesn't cause my display to change, i don't send salmentions
# 18:00 aaronpk but it didn't get a new mention, it got a notification that an existing mention was updated
# 18:01 aaronpk and that didn't cause the display of my post to change
# 18:02 voxpelli But when you first mentioned adactio, then that mention can't have been there?
creature joined the channel
# 18:03 aaronpk there are a few webmentions exchanged, but it's not an infinite loop
# 18:04 voxpelli Because one of you thankfully stopped showing mentions of mentions of mentions and so on
# 18:04 aaronpk nobody suggested ever *showing* mentions of mentions
# 18:07 voxpelli Replace "mentions" with "replies" and I guess most would like to show replies of replies
# 18:09 voxpelli If a single person in a reply thread were to accidentally or maliciously in-reply-to a post later in the thread, then the pinging would be infinite if everyone were to confirm their pings
# 18:11 voxpelli I know we have discussed this previously though, I guess we will eventually see it in practice
# 18:12 voxpelli I do know my endpoint is vulnerable to participation in an infinite pinging circle, that's why Salmentioning isn't live in it
# 18:27 aaronpk so here are some thoughts about the various states of sending webmention i plan to track
# 18:27 aaronpk the last one is when i get a webmention back like the earlier example
tantek joined the channel
# 18:38 voxpelli another note on that last one: didn't we conclude to stop sending webmentions to the comments of our posts?
eli_oat joined the channel
# 18:39 tantek that sounds familiar - possibly from IWC Brighton last year?
# 18:39 voxpelli The reason of this wording in the wiki: "resend Webmentions to everything the original post sent webmentions to"
AngeloGladding joined the channel
# 19:03 gRegorLove The Indie Author group uses "indieauth" a lot in URLs and hashtags so it shows up here.
tantek joined the channel
# 19:20 Zegnat Ah, for a second I thought you were commenting on chickens, gRegorLove ;)
cweiske, [kevinmarks], tantek and pfefferle joined the channel
eli_oat joined the channel
[colinwalker] joined the channel
# 20:40 Zegnat My PR got accepted on the Grav Webmention plugin, so any Grav users: expect an update to be rolled out soon
# 20:43 Loqi [Zegnat] #6 Correctly stop advertising Webmention endpoint on ignore_paths.
pfefferle joined the channel
[cleverdevil] joined the channel
eli_oat joined the channel
leg joined the channel
# 22:00 Loqi Ok, I'll tell them that when I see them next
# 22:23 tantek I wonder what the update semantics on such a micropub-delay-micropub service would be
# 22:23 tantek e.g. after you post to it, but before its posting time, you want to edit (or cancel!) the post
# 22:24 tantek do you just post to it again with the same permalink?
# 22:24 tantek like you can do micropub updates to it as many times as you want before it posts on your behalf, and it will only post once at that time?
# 22:24 tantek (with whatever state / update it last received?)
# 22:25 aaronpk i would imagine the service would hold on to the post until the scheduled date, and any edits would apply to the buffered post
# 22:27 tantek that would be quite handy for time-delayed POSSEing via silo-pub as well
# 22:27 tantek so you could publish, and still have 30s to catch/fix a typo, kinda like Gmail's undo send feature
# 22:28 tantek and after each update, the 30s clock resets, so you get 30s more to keep editing / tweaking
[eddie] joined the channel
dougbeal|mb1 joined the channel