#dev 2018-09-03

2018-09-03 UTC
#
upon2020.com moved /Indie_Box to /UBOSbox "Renamed the project"
#
upon2020.com edited /UBOSbox (-7) "Updated" (view diff)
KevinMarks, [eddie], AngeloGladding, renem, [keithjgrant] and iasai joined the channel
#
gregorlove.com edited /Leaders (+703) "/* Evaluate barrier of entry to wiki / wiki use-case */ +my comment" (view diff)
#
gregorlove.com edited /Leaders (+560) "/* Issues */ Evaluate if/how to apply Code of Conduct for outside events" (view diff)
oodani, barpthewire, ketralnis15, Zegnat, cweiske and TripFandango joined the channel
#
petermolnar huh. an idea: auto-check in. Once you've been at a place for X amount of time, and you already have a check-in for that location, it checks you in. is anyone doing something similar?
iDanoo16, renem, decode1 and [kevinmarks] joined the channel
#
[kevinmarks] Google location history does that
#
[kevinmarks] Foursquare/Swarm kind of does - it knows where you've been, and offers you 'checkins you have missed"
cweiske and walle30320 joined the channel
#
@jgmac1106 ↩️ Do not know  screen reader experience but for me  two extra clicks. Get url, new>reply. @microdotblog can syndicate to Twitter. I do not know how well  webmentions thread. I never tried, but if  like everything else the answer is probably well (http://jgregorymcverry.com/6416-2/) (twitter.com/_/status/1036589975096242176)
#
vanderven.se martijn edited /IndieWebCamps (+298) "/* During */ Specifically call out the presentation of [[code of conduct]] and having volunteer points of contacts available." (view diff)
#
vanderven.se martijn edited /Leaders () "(-1784) Remove previous code of conduct issue: all specifics contained within have been resolved and mentioned wiki pages have been updated." (view diff)
[jgmac1106] joined the channel
#
aaronpk petermolnar: a long long time ago I rigged that up with the foursquare api and geoloqi
#
petermolnar so the rule of the wiki kicks in again; either tantek or aaronpk thought about it and had it implemented 2+ years ago :D
#
aaronpk That was more like 7 years ago :D
#
petermolnar 2+ :P
#
petermolnar but yeah
#
petermolnar I never liked the idea of public check ins
#
aaronpk I eventually disabled it because it meant I couldn't add a comment to checkins anymore
#
petermolnar they imply I'm not a home
#
petermolnar so others could feel like they want to be in my home
#
aaronpk Now that swarm lets you edit them I wonder if it would work again
#
aaronpk eh there's easier ways to know if you're not home
#
petermolnar fair point
#
sknebel I still have retroactive checking on my to do list... And more as a reference of "I was there, and it was good/bad for X"
#
aaronpk Like mini reviews?
#
petermolnar h-mini-review
#
KartikPrabhu -nooooooo! don't make up *-mini* now
#
Zegnat Haha
#
Zegnat h-sknebel-review-of-checkin
decode1 joined the channel
#
petermolnar KartikPrabhu: I was joking, sorry if I took it too far
#
KartikPrabhu petermolnar: I know :P
#
petermolnar btw, nesting microformats (and schema) is not always trivial
#
sknebel aaronpk: yep. Kinda imagine a checkin UI with a star rating and maybe some tags for details
#
petermolnar would it be a checkin within a review or a review within a checkin
#
petermolnar ?
#
aaronpk Nice
#
sknebel Really need to figure out background logging on my Android, now that I have one where the battery might survive that
#
petermolnar sknebel: https://play.google.com/store/apps/details?id=com.llamalab.automate
#
Zegnat My backitude logs are still running and they do not kill my battery. IRC is a much more dangerous battery killer.
#
petermolnar for location, backtitude
#
petermolnar that's without question
#
Zegnat Though I do wish I had continuous tracking instead of polling, as the polling just can’t get the same accuracy, really
#
petermolnar for other sensor logging, automate can do basically anything, but it needs patience to put it together in it
#
petermolnar Zegnat: I tried, that is certainly a batter killer
#
petermolnar essentially I tried to do if-location-changes-25m in automate
#
petermolnar that devoured battery
#
sknebel I somehow thought backtitude wasn't around anymore
#
sknebel Was ist Just the website going away?
#
Zegnat I think it may be gone from the store ... but I still have it.
[grantcodes] joined the channel
#
[grantcodes] Oooh I have microsub notifications working!
#
[grantcodes] Will need to secure it then I'll push it online 🙂
#
[grantcodes] aaronpk: What's a reasonable polling time do you think? I have every minute for now.
#
aaronpk wow
#
aaronpk Do every minute. That'll encourage me to make a streaming API for it :-)
#
petermolnar sknebel: https://drive.google.com/open?id=1uOZqQHG7KDnwzAFFITYWP2iB8hoGktQ_
#
[grantcodes] Ha ok will do!
#
sknebel petermolnar: thx, saved
#
jeremycherfas For locations, I ought to work out a way to bring the data from Wanderings back home to my site and display.
#
sknebel What is location tracking?
#
Loqi It looks like we don't have a page for "location tracking" yet. Would you like to create it? (Or just say "location tracking is ____", a sentence describing the term)
#
sknebel Hm, thought we had an article, will search it later
#
Zegnat How do those GPS animal trackers work? Are they just super expensive or also not continiously tracking?
#
petermolnar most of them are not gps but cell tower triangulation
#
petermolnar real gps ones are rare
#
petermolnar but they log to a microsd card in most cases
#
petermolnar and polling
#
Zegnat cell tower triangulation? The trackers for fish, bird, and herd migrations?
#
Zegnat To be fair, logging to flash memory is all I want. It is all I do right now. I just want accurate logs, don’t need live internet connections. That just introduces attack vectors
#
sknebel The long running ones do just track one a day or so
#
sknebel There's pretty small loggers available for hiking and stuff though
#
sknebel Ping me tomorrow evening and I can go dig for info, we should own one I check details on
#
Zegnat I’ll remember :)
#
Zegnat I would prefer carrying an extra (small) device along over draining my phone battery
#
sknebel Yeah
#
petermolnar oh, not those, the dog/cat collar tag things
#
petermolnar I'd but a proper gps device to be honest
#
aaronpk has done a lot of research into this problem over the years
#
Zegnat aaronpk, ever run into any devices you could recommend?
#
aaronpk nope haha
#
Zegnat Hahaha
#
aaronpk it depends on your goals tho
#
Loqi hahahaha
brackets joined the channel
#
petermolnar not even some kind of luggage sized old military equipment?
#
aaronpk the longest lasting standalone device that works well is the spot trace https://www.findmespot.com/en/index.php?cid=128
#
aaronpk it has a GPS receiver and talks to a satellite network, so it works literally everywhere
#
aaronpk downside is it only tracks at predefined intervals, 5 minutes is the most you can do without paying for an upgraded plan which gives you 2.5 minutes.
#
aaronpk the battery lasts a few months though which is nice
#
Zegnat Goal: create a log of where I (the device) was at any time. Actual battery life isn’t super important as long as it is over 24 hours. I will most likely have a chance to recharge every day.
#
aaronpk what resolution of data do you want? because that's a huge factor
#
aaronpk e.g. here is a day at ~1-5 seconds logged from my phone https://media.aaronpk.com/Screen-Shot-2018-09-03-07-08-50-2IkmjzmHcN.jpg
#
aaronpk vs the same logged from the spot: https://media.aaronpk.com/Screen-Shot-2018-09-03-07-09-12-lXx2qD2l6C.jpg
guest3546 joined the channel
#
Zegnat Oh I am definitely looking for something closer to what your phone does. My curreny backitude polling is a higher resolution than that spot log.
#
aaronpk you won't find a device that does a better job than a phone really
#
Zegnat Guess I’ll have to get a second phone.
#
aaronpk the separate devices usually are either only GPS or only cellular tracking, both of which fail in different circumstances
#
aaronpk the phone gets to use both to do smart things, plus is always getting software updates from google/apple to make it work better
#
Zegnat Living on the Swedish country side I can not risk running the battery on my main phone dry, ever. Especially as someone who doesn’t have the option of getting in a car and driving to the closest town. So I am very hesitant to move to anything that does more regular polling or continuous tracking than my very conservative backitude settings.
cooled joined the channel
#
jeremycherfas Zegnat why don't you get a battery booster instead?
#
petermolnar get a waterproof dumbphone with a huge battery that lasts weeks if not months
#
aaronpk o.O cweiske you're a maintainer of GeSHi?
#
aaronpk just stumbled across https://github.com/GeSHi/geshi-1.0/pull/100
#
Loqi [suve] #100 Remove usage of create_function()
#
petermolnar create_function, aaargh, no! /me runs away screaming in terror
#
aaronpk hacks the source since apparently that code isn't released yet
[eddie] joined the channel
#
www.svenknebel.de edited /location (+54) "/* Location Tracking Apps */ add link t location category" (view diff)
iasai joined the channel
#
cweiske aaronpk, kind of
#
cweiske now and then I do a release
#
aaronpk can you do one now? :D
#
aaronpk that bugfix isn't in the current release
#
cweiske now, no. maybe this week
martijnvdven and KevinMarks_ joined the channel
#
aaronpk just piped his aperture "notifications" channel into Pushover to turn them into mobile notifications
#
[eddie] Nice
#
aaronpk we'll see how this goes
#
[eddie] Yeah, definitely report back how you feel about it
#
martijnvdven Also, lets see how long you can go before you have to pay for more messages
#
aaronpk pay??
[grantcodes] joined the channel
#
[grantcodes] aaronpk: I'm getting a 500 error from aperture depending on my access token
#
aaronpk [grantcodes]: uhoh! which endpoint?
#
[grantcodes] channels
Kaja joined the channel
#
[grantcodes] It's weird, works with one token I already have, if I put a fake one in I get the message I expect, but if I use another valid one it properly breaks.
#
[grantcodes] As in html error, not json error
#
aaronpk let me see if I can find that in the logs
iasai joined the channel
#
aaronpk [grantcodes]: oh yeah, there was an xray bug I fixed the other day and forgot to deploy here
#
aaronpk [grantcodes]: try again
[kevinmarks] joined the channel
#
[grantcodes] Oh, hang on, might be an issue with scopes
#
aaronpk are you still getting the error?
#
aaronpk ah new error this time
KartikPrabhu joined the channel
#
aaronpk it is a scope error, but I should fix that 500 error
#
[grantcodes] Yeah. My bad, but should still return something different
#
aaronpk try now, you should get a real scope error
#
[grantcodes] Yep better error now
#
aaronpk great, thanks for catching that :)
#
[grantcodes] No problem, double debugging
#
[grantcodes] Also while I'm at this whenever I save a users url that they enter for indieauth I should follow any redirections and save the final url right?
#
aaronpk Almost
#
aaronpk all the combinations are listed out here https://indieauth.spec.indieweb.org/#redirect-examples
#
aaronpk You should store the url that the server returns at the end, but you do need to keep track of the initial redirects at first too
#
Loqi I agree
#
aaronpk And read that 4.1 section too https://indieauth.spec.indieweb.org/#discovery-by-clients
#
[grantcodes] Ok perfect thanks.
eli_oat1 joined the channel
#
[grantcodes] I have been bad and just stored whatever the user entered in the form. But means if you write a different url it will make you a new user. Which obviously isn't quite right
#
[grantcodes] !tell schmarty Been using your indieauth lib, but I can't open issues on https://git.schmarty.net/schmarty/indieauth-authentication So 1. It should support scopes, 2. Should be able to get a token 3. Resolve and return redirected user urls. Happy to write pull requests, but don't want to create another separate npm module
#
Loqi Ok, I'll tell them that when I see them next
sknebel, Zegnat and [iambismark] joined the channel
#
sknebel [grantcodes]: it's even in the name of schmarty's library that it doesn't do those things though
#
aaronpk it should be able to do #3 tho
#
sknebel true
KevinMarks, eli_oat, [tantek] and xset joined the channel
#
aaronpk Zegnat: you're a good test of my nicknames cache because you have so many different usernames/urls everywhere
[grantcodes] joined the channel
#
[grantcodes] Hmm yeah I know it's just minimal auth but if it doesn't support scopes you need another library / to do them manually, which seems like a waste
AngeloGladding joined the channel
#
[tantek] That makes sense
#
[tantek] (Oops was replying re: redirects)
#
[grantcodes] schmarty's library does handle redirects internally, just needs to expose the resolved url 🙂
eli_oat and AngeloGladding joined the channel
#
[tantek] Thinking about UI / community friction/barriers/filters and Derek Powazek’s Argument Machine post which makes a good case for *some* UX efficiencies amplifying more negativity than positivity
#
[tantek] Implications for IndieWeb posting UX design
#
[tantek] (Thus some friction/barriers may actually create a more net positive UX and/or community)
#
vilhalmer it seems to me that forcing yourself to think before posting, even briefly, is a good idea
#
AngeloGladding https://www.wired.com/2008/10/googles-mail-go/
#
AngeloGladding google's mail goggles prevents drunk emailing (2008)
#
AngeloGladding wow i never really associated inebriation with ranting on twitter
#
vilhalmer I like to think my slightly tipsy twitter rants were actually MORE thought through
#
vilhalmer one time I designed a window manager
#
AngeloGladding obligatory https://www.xkcd.com/323/
barpthewire joined the channel
#
tantek.com edited /MediaWiki:Sidebar (+0) "next hwc 09-05" (view diff)
#
aaronpk ooo I might be able to launch auto-transforming @-mentions to service-specific names very soon!
#
aaronpk the risk of course is someone like Zegnat whose twitter username is longer, and might cause me to go over the character limit without realizing
tantek___ joined the channel
#
Zegnat That's an interesting edgecase
#
aaronpk of course it works the other way with tantek! I usually type @tantek and then it'll change it to @t
#
vilhalmer wasn't there one point that twitter was going to make names not count as part of the character limit? that would be nice except for all the spam implications
#
vilhalmer aaronpk: could you use the longest username in your cache for any given person being mentioned to calculate the character limit?
#
tantek___ vilhalmer: yes Twitter switched a while ago to not counting username length as part of char limit (pretty sure)
#
tantek___ and even more so (not counting additional usernames at all) as part of their canoe chaining update
#
tantek___ Aside re: income being eliminated, at Technorati we used to get very nasty hate emails (personal, and through support) from spam blog farm maintainers saying Technorati blocking their spam blogs were hurting their income and were to blame for their kids not getting healthcare etc. It was some really nasty stuff.
#
tantek___ sorry that should have been -meta
dougbeal|mb1, eli_oat and TripFandango joined the channel
#
pstuifzand grantcodes[m], I just saw that you posted the relScraper function as a library
#
Loqi pstuifzand: [grantcodes] left you a message 19 hours, 30 minutes ago: I've made an independent rel scraper library. https://www.npmjs.com/package/rel-scraper Feel free to give it a go
[eddie] and tantek__ joined the channel
#
tantek__ grantcodes[m]: minor naming nit - looks like your library does actual HTTP header and HTML parsing, didn't see any regex use in there. So you have a relparser technically, not a relscraper.
#
tantek__ making the distinction because there are *actual* scraping libraries out there that use regex to get rel values from headers and/or say the first 1k of an HTML document
#
tantek__ (that is, calling your library "scraper" is unnecessarily and inaccurately undervaluing/underselling your work :)
[grantcodes] joined the channel
#
[grantcodes] tantek: I did not know the difference, thanks. I will see if I can rename it
#
pstuifzand it works great, I just used it in my reader
#
sknebel I'd say "scraping" is extracting information that's not specifically encoded to be machine readable, but doesn't say anything about what tools are used. But since rel= is clearly dfinied and intended to be parsed, parser is the better label, yes
#
tantek__ sknebel: that is a reasonably precise definition yes. In coarse terms, usually when code is using a regex to extract apparently (rather than explicitly) machine readable information, then it's "scraping"
#
tantek__ if code is following (even roughly) a specification for a format in order to extract explicitly machine readable information, it's parsing
#
tantek__ what is scraping
#
Loqi Scraping is extracting data from an unstructured web page, opposed to from an API or embedded structured data like microformats https://indieweb.org/Scraping
#
tantek__ what is parsing
#
Loqi It looks like we don't have a page for "parsing" yet. Would you like to create it? (Or just say "parsing is ____", a sentence describing the term)
#
tantek__ heh
#
tantek__ grantcodes++
#
Loqi grantcodes has 20 karma in this channel over the last year (28 in all channels)
#
tantek__ sknebel,grantcodes there is another nastier subtler misconception lurking here from the old XML days, and that's the myth that if you're machine parsing information from HTML, you're inevitably "scraping" (as opposed to XML, which of course is clean, well defined, always valid, and thus always "parsing")
#
tantek__ another thought on UX (in)efficiencies by design, has anyone implemented "Undo" in their "Publish
#
tantek__ " UI?
#
tantek__ what is undo
#
Loqi It looks like we don't have a page for "undo" yet. Would you like to create it? (Or just say "undo is ____", a sentence describing the term)
#
[grantcodes] Not sure it count's as ui, but my chatbot micropub client has a delete button as soon as you make a post 😛
#
tantek__ delete is definitely very useful (especially if it's quick & responsive immediately after a publish). undo implies very different semantics. like Gmail's undo send.
eli_oat1 and [manton] joined the channel
#
tantek__ [manton], is indie /event post (or at least RSVP) support somewhere on the list for micro.blog?
#
tantek__ if it is, that would be a motivation to figure out how to syndicate an indie /event via RSS so that it showed up in my micro.blog, and allow folks to RSVP to it from micro.blog
#
tantek__ s/RSS/Atom
#
tantek__ even if it is low on the list for mb I'm willing to help work on it, to answer questions, make it easier etc. until enough folks ask for it that it's worth building/shipping/maintaining
eli_oat joined the channel
#
[manton] [tantek] Thanks! I'm not sure where on the list it is, but I'd like to think about it. I've been hand-coding my RSVPs so far.
[tantek] joined the channel
#
[tantek] Right. My use-case is to allow / encourage RSVPs *from* mb backfed to my indie events
#
[tantek] Easiest way I could think of enabling that is to make it so when I posted an indie event, my existing Atom feed syndication to mb cause the creation of an indie event post on my mb
#
[tantek] Then anyone on mb could reply to that with an rsvp
#
[tantek] Cross-site rsvps from mb would be even better of course
#
[tantek] Not sure of a good UX for cross-site RSVPs though
KevinMarks joined the channel
#
[manton] Interesting. I'm looking over your feed now... Is there currently enough information in it to be able to recognize events?
KevinMarks, tantek__, willnorris, voxpelli and ancarda joined the channel
#
tantek__ Manton, not currently no because I there was no obvious way to do "Atom event posts" or "RSS event posts" for that matter so I just named them "Event" (which I realize is not very helpful)
#
tantek__ However if you know of anything resembling a decent/simple way to post an "event" in an Atom entry or RSS item, I'm willing to do the experimental dev work to publish it as such and post one to try experimenting with
#
tantek__ for reference, my most recent indie event: http://tantek.com/2018/234/e1/homebrew-website-club-sf
#
tantek__ and minimal tweet equivalent https://twitter.com/t/status/1032112905301516288
#
@t Event: Homebrew Website Club SF! At: 2018-08-22 17:30 @MozSF RSVP: http://tantek.com/e/4w51 (twitter.com/_/status/1032112905301516288)
#
tantek__ (yes that that tweet summary of the event is 100% auto-generated from my (longer) indie event post)
#
vilhalmer nice
#
[manton] I'm not aware of a convention for events in feeds yet, but I haven't looked closely either.
#
tantek__ one approach would be to post a fully h-event as the only child element of an <atom:entry>
#
GWG Ah, RSS discussion
#
tantek__ e.g. <content type="xhtml"><div xmlns="http://www.w3.org/1999/xhtml" class="h-event">...</div></content>
#
tantek__ atom entry *content* element that is. (also "full h-event")
leg joined the channel
#
tantek__ I know that's not the "XML way" (no <event> element etc.) but I'd rather not extend Atom with additional XML at this point anyway.
#
Zegnat You could also add an (auto-generated) iCal file as an enclosure to the item. So if the item (whether in Atom or RSS) has a text/calendar enclosure it probably denotes an event.
#
tantek__ Manton, would it be reasonable for an Atom feed consumer to look for a sole <div> child like that with a class name of h-* and then treat that entry as that h-* ?
#
tantek__ Zegnat: ical parsing is a MUCH worse thing to try to get right than XML or HTML
#
tantek__ as well as ics generation
#
Zegnat True. It was just an alternative that wouldn’t introduce magic marks within the entry that need parsing
eli_oat1 joined the channel
#
tantek__ (read: no open test suite for ics, no actual testable interop, you have to test one calendar user agent at a time, worse than browsers in the early 1990s)
#
tantek__ Zegnat: nothing magic about it. you can even define some XPath for detecting what I suggested :)
[jon] joined the channel
#
Zegnat Assuming everyone puts valid xhtml in the feeds, XPath could work, yeah
#
Zegnat As soon as someone puts HTML in a CDATA block as the contents, that goes out the window and the consumer will need both an XML and HTML parser.
#
tantek__ Zegnat, HTML as CDATA is a strawman, no one is advocating for that.
#
tantek__ also, escaped HTML is an RSSism, not really a thing in Atom (AFAIK)
#
[grantcodes] Ok my microsub notification service is online now: https://microsub-notifier.tpxl.io/
#
[grantcodes] Only does webhooks for now but it's a solid proof of concept
#
Zegnat Isn’t WP by default all HTML-as-CDATA in feeds? I am just thinking of practicality here in case your proposal becomes the defacto standard for recognising events in feeds.
#
Zegnat Wow. We’re up to 5 separate update RFCs to the original RFC 5545 for iCal. Yikes.
#
GWG I need to read up
#
tantek__ Zegnat, do you mean the original RFC 2445 for iCal? ;)
#
grant.codes created /microsub_notifier (+368) "Create microsub notifier page" (view diff)
Jovan0 joined the channel
#
Zegnat I guess then you need to do +1 :P
#
Zegnat I think RFC 5545 obsoletes 2445, I think, and starts with a clean slate ... or as clean as they could ever make iCal.
#
Zegnat The other 5 all just introduce updates, rather than obsoleting
#
tantek__ Zegnat, I'm ok with restricting this to Atom-only (no RSS), well-formed XHTML div inside Atom <content> only as a way to get this started. Then if the only consumers of such "feed events" also require that, then publishers will comply (otherwise their stuff won't get picked up).
[kevinmarks] joined the channel
#
[kevinmarks] Atom often has cdata around the html, yes
#
[kevinmarks] Because xhtml is fragile
#
tantek__ as a fallback, publishers can publish "simple" events (like my tweet example above) if they can't handle publishing "well formed XHTML" in general
#
[kevinmarks] Feedparser will extract I Al from hCalendar in feeds
#
[kevinmarks] *ical
#
GWG I am having trouble converting non mf2 to jf2 for Microsub type purposes
#
tantek__ kevinmarks, can you point to the docs that explain how that works and what they're expecting for "hCalendar in feeds"?
#
GWG I really need to look at what feed readers do
#
tantek__ (I'd be willing to put in some backcompat for feedparser if that helps, e.g. class="h-event vevent")
#
grant.codes edited /Microsub (+229) "/* Clients */ Add microsub notifier to the list of clients" (view diff)
#
tantek__ otherwise I'll figure out the aforementioned XPath for "h-event in Atom content" detection, and then feed consuming code can hand-off the h-event div to a mf2 parser to get mf2json
#
tantek__ (otherwise as followed-up to request to kevinmarks docs request)
#
[kevinmarks] Ah, I was wrong, just hCard https://github.com/kevinmarks/feedparser/tree/develop/feedparser/tests/microformats
#
[kevinmarks] I think h-event in the content is reasonable
#
tantek__ cool.
#
tantek__ if I get the thumbs-up from manton on that proposal then I'll try to prototype something for my next HWC SF indie event post
#
tantek__ so at least we have a published Atom example to then analyze / dissect / iterate
#
Zegnat FWIW: /feed/entry/content[@type="xhtml"]/div[contains(concat(" ", @class), " h-") and count(../*)=1]
#
Zegnat That’s probably what you need. XPath engine may need to have the default namespace set to Atom’s though. That can be a little wiry.
#
tantek__ Zegnat++
#
Loqi Zegnat has 57 karma in this channel over the last year (151 in all channels)
#
Zegnat Should match all divs that are an only child of xhtml type content element where the class="" contains atleast 1 classname starting with h-
#
tantek__ this is starting to get defined enough to probably write-up as a Brainstorm on /event
#
tantek__ Zegnat yes that looks good to me
#
@ton_zylstra ↩️ dat snap ik, maar wat ik niet snap is dat die posting in de RSS wel een titel krijgt (‘microsub post’) terwijl op je site niet. (twitter.com/_/status/1036706652370690050)
leg and [manton] joined the channel
#
[manton] [tantek] An extra div with h-event to wrap the post content seems fine to me too. It wouldn't have to be Atom-only, would it? Seems like a similar approach would work with RSS or JSON Feed.
#
[kevinmarks] Sure, anywhere you can put html works
#
loqi.me edited /NewBase60 (+426) "[kevinmarks] added "base58 by kellan and Satoshi Nakamoto https://twitter.com/kellan/status/1036725567574433793" to "See Also"" (view diff)
#
miklb__ aaronpk any suggestions on how to phrase bug report for bridgy regarding my reply posts?
#
aaronpk ah yeah
#
aaronpk i'd comment on that previous issue with a link to your post saying that it looks like it's not working anymore
#
aaronpk clearly it looks like that previous fix should have made this wrok
#
miklb__ OK
#
miklb__ I tagged you in the issue too
#
aaronpk 👍
#
sknebel what's weird is that the bridgy log doesn't show it even having the property in the parser result
#
sknebel and if you feed it in the preview function, it works
#
aaronpk huh
#
miklb__ ok, I just tested from my admin and it worked, so maybe something in posting from Indigenious/micropub firing in wrong order?
#
miklb__ dont know if this needs to move back to #wordpress or not
#
aaronpk yeah now this sounds like something about the order the wordpress plugins fire
#
miklb__ I'll test Quill desktop now
#
miklb__ confirmed it's micropub issue
#
aaronpk great
[tantek] joined the channel
#
[tantek] Manton, if you think there’s value in specifying how it works across all three then yes.
#
[tantek] I assume JSONfeed already specifics exactly how to include HTML post content so that should be a direct mapping
#
[tantek] IIRC, RSS had a few different ways of putting HTML into <item>, in particular I recall HTML escaping. Or maybe CDATA? Is there only one way that micro.blog extracts HTML content from RSS item elements, or multiple?
snarfed, eli_oat and [manton] joined the channel
#
nikhiljha.com edited /Grav (+142) "/* IndieWeb Examples */ add myself" (view diff)
#
[manton] Yep, Micro.blog supports either escaping the HTML or using CDATA.
#
[manton] I think most feeds are just escaped HTML. It is ugly but works.
#
aaronpk wow dealing with html is hard
#
aaronpk i'm trying to get my autolinker to work with @mentions in HTML
#
aaronpk can't seem to get it to work without breaking other things
#
aaronpk damn, I thought I could get it to work with markdown too, but since markdown can contain html I need to solve it for html first!
#
AngeloGladding just got v3 .onion addresses working and stumbled upon another interesting consequence -- even if I have HTTPS enabled on my (non-shared-hosting) site a user's ISP will know that they are interacting with me (reverse DNS lookup)
#
AngeloGladding owning your own domain is bad for privacy
[kevinmarks] joined the channel
#
[kevinmarks] I think Tantek's autolinker is idempotent
#
aaronpk crap yeah, I need to run my autolinker *after* the markdown->html conversion, using the same rules I use for html autolinking (traverse all the text nodes except the contents of <a> <pre> <code> and autolink the nicknames inside)
strugee joined the channel
#
[kevinmarks] Hmm. If the mentions came from Instagram, are they linked already? Or do you need a different mapping?
#
aaronpk i'm trying to run everything through my nicknames cache
#
aaronpk to start I just need to turn my own @-mentions into the appropriate nickname based on the destination and autolink it in my own post
#
aaronpk secondly, I want to take a @-mention from a PESOS'd post like from instagram and transform that back to a link to the person's website when my site displays it, but that's an even harder one
snarfed joined the channel
#
aaronpk anyway, the reason I can't autolink markdown *before* the html conversion is that [this mention of @tantek](http://example.com) should not be autolinked
#
aaronpk in order to do that pre-html conversion I would need to do that crazy trick of substituting all the non-replaceable bits with random strings, then do the replacing, then swap them back
#
aaronpk other non-replaceable bits include code blocks in `` as well as triple multi-line ```, plus there's the question of what to do for embedded HTML