#GWGmiklb: I was wondering what would be thought of a plugin that can be called by Post Kinds, Semantic Linkbacks, etc and would return parsed jf2 to be used to enhance comments and such. Sort of an x-ray for WordPress
#miklbI'm not 100% familiar with x-ray, but I love what aaronpk post for one of 100days for reposts and contexts
#miklbso I would be fully be onboard for something that moved the needle closer to something like that. And would contribute what ever I could to help
#GWGx-ray is just his service that takes a URL and returns a nicely parsed object you can use to build your response
#miklbI'm all for having as much data as possible to work with
#GWGBut, since I was using aaronpk's x-ray code to enhance my code, it occurred to me I could go in that direction
#GWGThe problem is that WordPress is horrible about dependencies.
#GWGBut I may do some more work in that area. I want to dedeplicate POSSEd copies and the the original. Which requires extending Syndication Links to comments and supporting figuring out that it is a dupe.
#miklbso you need to figure out how to have a dependent library that multiple plugins could rely on, without packaging it as a plugin or putting it in each individual plugin? Am i following?
#GWGWell, I could package it as a plugin that uses the REST API or the older admin-ajax service to allow the frontend or other plugins to query it.
#GWGSo, if it isn't there, the plugin would fall back on something else, but still...
#miklbI would be intrigued to see REST API integration there
#GWGmiklb: The webmention plugin uses the endpoint infrastructure code without being REST.
#GWGAaron Jorbin mentioned to me at IWC NYC 2 that the infrastructure could be used without using the full REST stuff. That made it more intriguing, because you can build any sort of API and there are all sorts of core functions to enhance it.
#miklbI still need to read the webmention plugin code.
#GWGI did the initial rewrite of Webmentions to use the infrastructure. I know snarfed thinks I want to rewrite the Micropub plugin in same, but it isn't at the top of my list.
#miklbI'm going to be diving deep into rest api here very soon, and authentication is one of the things I'm concerned about, so I'll be putting a lot of thought there
#GWGmiklb: What do you think about the idea of Syndication Links extending into comments?
#miklbat face value sounds good. I haven't gotten too far with comments yet, due to lost posts/lack of posts w/feedback. But working towards them, and want them to be as robust as possible
#GWGWell, Semantic Linkbacks is something that there is controversy about
#GWGThe implementation is a bit 'hacky' was what he said the last time I asked him
#GWGA lot of the stuff I worked on over the summer with the webmentions plugin and several tickets into WordPress itself was to make said implementation less hacky.
#GWGWhen pfefferle wrote the original version, he had no better way to accomplish the goal.
#GWGThere still are major things missing on the WordPress side
KevinMarks joined the channel
#GWGFor example, WordPress allows custom comment types like Webmention, but doesn't support them
#miklbyou mean like, "here, you can create this thing, but don't expect us to do anything with it"?
#GWGSome big WordPress contributors are interested. But they are focused elsewhere.
#GWGso, they put together a Github repo which got a README...and nothing else.
#GWGThe lack of registered support creates issues when you try to display webmentions if the theme doesn't support them. You basically have to take over large portions of system functionality.
#GWGmiklb: I saw you forked my old Semantic Comments plugin?
#miklbyep. I do want some semblance of facepiles, so looking over that plugin, looked like a good place to start.
#[kevinmarks]I'm thinking that jf2, rather than being a "cleaner mf2" would make more sense as a uniform encoding of an xray/post types/authorship parsing
#[kevinmarks]If the various webmention handlers could rely on a common format that implements the more content aware parts of post types/authorship etc
#[kevinmarks]I have the feeling that we are all evolving in parallel a set of heuristics for "what does this url mean in this context"
#[kevinmarks]Which is a different problem from "what does this page generally represent?"
#GWGBut the point is still well taken that if we are all looking to make it easier to go from full mf2 to something simpler to integrate, there is room for shared work
#GWGIt sounds like something to talk about at an IWC
KevinMarks joined the channel
#miklbGWG might be a good argument to keep semantic linkbacks as its own plugin then
KevinMarks, KevinMarks_, KevinMarks__, cweiske, tantek, loicm, wdfwefewvfgew and arush joined the channel
#petermolnarGWG, a sidenote: I ended extending https://github.com/vrypan/webmention-tools - essentially it's doing xpath like lookups with Beautifulsoup instead of complete mf2 parsing, so for targeted parsing it might be simpler that traditional, complete mf2 parsing
singpolyma, rMdes_, rMdes and arush joined the channel
#aaronpkpetermolnar: wow I've thought about setting up an SMS gateway ages ago and gave up and then Twilio launched and solved my more immediate problem
#petermolnaraaronpk I need a fallback when there's not internet connection at home
#petermolnarso online services won't solve this :)
#petermolnaractually, I have an 56k modem in that machine, I could initiate a backup dialup when the broadband goes does, but I don't think I want to do that :D
#sebselIn a way, this is worse than XSS or CSRF, because those two are on the client side, and here you run code on the server, so you can do anything with it.
#sebselthis also makes it less specific to webmention, I guess, and more related to user input validation
#aaronpki think the key thing here is to remind people to treat everything they get from fetching a page from webmention as user input
#aaronpkand then all the typical user input sanitization stuff applies as normal
#aaronpkif you ever handle uploaded files from users you have to take those same precautions
#aaronpkan explicit example in this case is probably good because images are such a key component of displaying useful webmentions, but i'm trying to generalize this into a recommendation so that it can be more than just listing specific attacks
#sebselyes, I don't show html at the moment, but embedded images would follow the same rules
#sebselit's good to note that they are files and need to be handled with care