#eli_oatI'm interested in hosting an instance for a thing I want to try, indieDnD, or IndieRPG wherein folks can participate in a distributed game in the style of DnD
#jmac_So, my homebrew Webmention library has an object method called "content", which returns some text extracted from the wm's source document based on this algorithm: https://indieweb.org/comments#How_to_display
#jmac_Right now, it does this regardless of the "flavor" of webmention. It'll do it for in-reply-to's, and for everything else as well.
#jmac_My question is, should it apply that algorithm *only* in the case of explicit reply-mentions, and do something else (maybe just returning the p-name) for non-replies?
#ZegnatHonestly, that’s completely up to you, imo. If you are getting good content results even for not-explicitly-mf2-marked-mentions, why not use it?
strugee joined the channel
#jmac_Yeah. And on re-read, I see on that wiki page explicit instructions to display non-replies separately from replies, which I'm not doing, and which I suspect is closer to the root of my problems.
#snarfed!tell tantek fixed that bridgy github bug, feel free to try again if/when you have another issue for that repo! may still not behave ideally though since it's private though, we'll see
noah21, iasai, snarfed, [jgmac1106], jak14, Frogging10119 and [pfefferle] joined the channel
#aaronpkit seems silly to have just twitter accounts for all my other little project profiles now. i keep finding myself wanting some lightweight project i can easily set up on a new domain for accounts like @indie_auth and even @indiewebcamp
#Zegnatruns SVG through an optimizer -> SVG no longer resizes.
#Loqitantek__: snarfed left you a message 1 hour, 1 minute ago: fixed that bridgy github bug, feel free to try again if/when you have another issue for that repo! may still not behave ideally though since it's private though, we'll see
#Loqisnarfed has 63 karma in this channel over the last year (115 in all channels)
#[eddie]That said, it’s really bare bones, as I didn’t set it up to accept webmentions or anything. The websites for my cats will be more middle ground between my website and my son’s
#Loqi[Tantek Çelik] One "simple" way to do this would be to create a public repo like: https://github.com/w3c/ABpub which would thus have public issues by default.
[shaners] joined the channel
#aaronpkI've let p3k drift a bit far into the "too-specific-for-any-other-use" category so it'd take a bit to strip out some things that only I need
#aaronpkI might need to figure out how Laravel plugins work so I can move some of those things into private plugins and then open source the core
#aaronpkBut to do that I need to update to the latest Laravel which I know changed a bunch of internal things from 5.2 to 5.6
#aaronpkI just keep thinking "I should make a website for [Loqi/IndieWebCamp/BarBot/my cat/my house/etc], like okay cool I can put up an html page with h-feed easily enough.
#aaronpkBut then I want to post via Micropub. Okay that's not too hard to add.
#aaronpkNow I want to send webmentions, well that now requires background processing, especially to also receive them
#aaronpkyep and telegraph, both of which my site currently use haha
snarfed joined the channel
#aaronpkof course the other half of this conversation in my head is thinking what it would take to make p3k open source so that people could actually use it
#aaronpkand to do that, i need to make it simple to set up, which funny enough means removing external dependencies like webmention.io and telegraph, since those both require a few extra steps of pasting API keys around
#sknebelor make it easy to setup single-site versions of those, and add them to the package :D
#aaronpkanyway, I am currently leaning towards starting from new Laravel and porting all my code into it and turning it into something easy to install. Figuring out the plugin mechanism so that I can keep a few things private like my ridiculous GPS integration. That'd both make it something I could open source, as well as easy for me to spin up for small sites.
#ZegnatSVG icons by Google (or whoever owns the rights to the Apache licenced Material Design set)
#[eddie]aaronpk: That seems like a good idea. I’m working on something similar with my node.js site (which I have a goal to be able to make it serverless)
#Loqizegnat has 51 karma in this channel over the last year (140 in all channels)
#ZegnatThinking about merging it into 1 pill shape, so it looks more like a real state switch. But this was a challenge enough. Been a long time since I did anything design-y
#Loqisnarfed: tantek__ left you a message 49 minutes ago: tried to POSSE this issue comment: http://tantek.com/2018/221/t1/ and looks like got errors in the log
#[eddie]ohhh because grantcodes implemented it differently
#[eddie]😆 Not a “bad” persay. Just editing a bornstorming page that made it a bit more confusing. But I understand why you did it. You and j4y_funabashi had different approaches
#[eddie]So it would be somewhat odd for j4y_funabashi’s to be at the top of the section
#[grantcodes]It's kinda hard to notice that is all under the "Query for post list" heading. Not much of a difference with the sub headings 😛
#[eddie]Yeah, definitely. I was having a hard time telling what the headings denoted
#aaronpkif there are two implementations now, maybe time to move to the extensions page with an actual proposal?
#[grantcodes]But there was no more info on that sort of query that I could find
#aaronpkthe brainstorming page is meant to be things that are super rough ideas or not yet implemented by anyone
#[grantcodes]Well my version is definitely not a proposal. But definitely works in my specific case and good for discovering what is required of a proposal
KartikPrabhu and eli_oat joined the channel
#[eddie]Well that was super easy. I already have basic support for it added to my site. My only issue is that my internal post storage isn’t 100% mf2 json, so I just have to write a conversion function from my post storage object into complete mf2 json
#[eddie]But I’m successfully returning my internal storage objects right now, as a first step
#[grantcodes]I don't like the empty q=source version myself, as it is not very flexible
[kiai] joined the channel
#[eddie]Is it inherently not flexible or is it not flexible because we haven’t iterated on it yet?
#[grantcodes]Maybe, but also could be confusing when trying to get one post vs a list. But I think the first thing to decide on would be what parameters would be sent. Then name them from there?
#aaronpkokay i must have forgotten about some other micropub page because now i can't find the part that talks about querying for supported vocabs that both quill and micro.blog have implemented
#[grantcodes]My idea of filtering by property effectively does that
#[jgmac1106]in terms of bookmarking not making the cut [tantek] [grantcodes] and I had a long conversation at the Summit and I thought it was going to be moved into the accepted category but I think it was left at the @microformats wiki people had to move first
#[eddie]Here’s the issue I ran into with property filtering
#[eddie]name exists both on Articles and Podcasts for me
#[eddie]so Mars Edit would have to filter posts WITH a title but WITHOUT audio
#[grantcodes]I guess with post kinds as a query filter or not you are mainly choosing whether the server or client should handle the filtering, - sort of
#[eddie]Well I think the big thing about post kinds as a query is you have specialized Micropub apps
#[eddie]so for example: MarsEdit could query for only articles or articles and notes
#[grantcodes]I mean it's not super hard, just need to set up what properties match to what post type
#aaronpkif you only have notes and articles and photos then your properties map to post types 1:1
#aaronpkbut as soon as you add other things like videos, checkins, etc then you need something more complicated
#NinjaTrappeurHey indieweb, quick question: we are planning to send some webmentions from lobste.rs (a link agregator). We are wondering if sending "historical" webmentions (ie. webmentions for articles posted a long time ago) is a good idea. Do you have any input on this? Could we trigger some "anti-spam" filter by doing that?
#aaronpkalso i wouldn't send out historical ones until you're confident in how the new ones are working
#aaronpkbtw if you need anyone to review the microformats you're adding this is the place to ask as well!
#[grantcodes]Depends what you want to do. Technically on my site there are post types but in my mind they are all just the same with different properties that get rendered appropriately. There's only one template too
#aaronpkthe idea though is not to create a new URL for each comment that shows up as a comment on the original post, but instead it treats the lobsters post as "being updated" when any new comment is added there
#aaronpkfrom the receiver's POV, it looks like just a duplicate webmention if they don't also understand the list of comments
#aaronpkthat way they can basically opt in to receiving the updates of the comment thread, and if they do nothing then there isn't that "spam" effect that's described in the thread
#[grantcodes][eddie] sure, makes sense. But for sure I think there will be a need to query by specific property too. Just post type won't cover all the use cases I thought of
#NinjaTrappeurwhat kind of microformat class are we supposed to use for the comments in such configuration?
#[eddie]very cool. I’ll have to play around with that. Thanks!
#[grantcodes]The syncing is still kinda dodgy I think. But I did push all my browser bookmarks to my site. The query means it should be able to import bookmarks from your site to your browser too
#[eddie]Cool, I probably wouldn’t do much syncing, and I don’t really use Chrome or Firefox, BUT I can tweak it and open up Chrome and test the import from my website at least so I can see it working 🙂 I’ll have to mark that down as a project for this weekend
#[eddie]aaronpk: Now that we have a couple implementations, should we continue to iterate on the brainstorming page or shift to either a GitHub issue or the extensions page?
#aaronpkgo ahead and move the current proposal and example to the extensions page, then make an issue for it on github and we can continue the discussion there
#aaronpkthe extensions page is a good quick list of extensions but discussions are better on github
#aaronpkthe idea with the token method is to let implementations decide how to make it work under the hood themselves
#[eddie]Yep, exactly. Which I like because as a server person that means I can do some really simple pagination by using a string that doesn’t require me to do a lot of computation
#aaronpkyeah you could even just return page numbers there if you wanted to be lazy
[cleverdevil] joined the channel
#[cleverdevil]So, the only thing about the way this works is that Known, as far as I know, doesn't actually store the MF2 JSON once its processed.
#[cleverdevil]That means that Known would have to learn how to translate its content back into MF2 JSON.
#[eddie]aaronpk: Regarding pagination, Since I’m not currently doing any indexing with a database on my site, using an opaque string allows me to essentially use a date string and then continue loading posts that happen before or after the given date string. Which is great, almost no overhead
#[eddie]Well, I think it’s time to call it a night and go cook up dinner. I added a limit parameter and return before and after tokens to my post list query. I’m gonna have to write some more functions to actually fetch posts based on the tokens, so that’s another battle for another day
#[eddie]Not bad though. It doesn’t take almost any time/effort to implement the new extension. The hardest part is really translating posts back into mf2 json
#[eddie]I’ll have to put Client compatibility on my list of projects so I can play around with it more