2018-03-03 UTC
# 00:23 Loqi [Tantek Çelik] Happy NewCalendar Day! (first New Sunday)newcal.orgPreviously, previously, previously: * tantek.com/2010/061/t1* tantek.com/2011/061/t2* tantek.com/2012/061/t1
# 00:28 Loqi aaronpk: [eddie] left you a message 2 hours, 6 minutes ago: what’s still holding you back from going full 💯 Microsub?
# 00:29 aaronpk [eddie] up until a few hours ago, it was the ability for monocle to refresh the unread counts and automatically mark things read as they are seen/interacted with
[eddie] joined the channel
# 00:30 aaronpk oh and I spiffed up the channel list to make it always visible on desktop and an expanding side bar on mobile, like Together
# 00:30 tantek hey snarfed, do tags on issue posts get POSSE to GitHub as labels?
# 00:32 tantek if I wanted to POSSE an issue with existing label css-text-decor-4, if I have <span class="p-category">css-text-decor-4</span> in the h-entry of my issue, will Bridgy see it and translate it to the "css-text-decor-4" label (light blue background) in the GitHub issue copy?
# 00:32 tantek I guess I'll try and find out! Then if it doesn't, I can file another issue :)
# 00:39 Loqi [Aaron Parecki] Monocle is coming along nicely! It now automatically marks things as read as you scroll through the timeline. The unread counts in the sidebar update automatically when new posts come in as well! I'm getting so close to being able to completely switc... https://aaronparecki.com/2018/03/02/33/photo.png [video]
# 00:42 aaronpk [eddie]: oh the last part will be push notifications, since that still goes through IRC for me
# 00:43 Loqi [Aaron Parecki] Designing an IndieWeb Reader Based on How I Read Social Media
# 00:51 tantek ok writing issues in HTML is taking me longer than writing issues in "plaintext", and then fixing Markdown bugs
# 00:52 tantek makes me think I need to include the ability to autolink subsections (maybe on a paragraph by paragraph basis?) parts of my /article (and thus /issue) posts
# 00:53 tantek hmm maybe something like <div class="auto-link"> ... </div>
# 00:54 tantek or maybe that's what I should use p-content vs e-content in my own storage for
# 00:55 tantek <div class="p-content"> -> contents are plain text, thus should be auto-linked/escaped etc.
# 00:55 tantek <div class="e-content"> -> contents are HTML already, do not auto-link etc.
# 00:58 aaronpk I like the flexibility because sometimes I want to write html explicitly and not deal with accidental autolinking
# 00:59 tantek I have a bad feeling my home page slow loading time may be due to breaking the RECENT ARTICLES box
eli_oat joined the channel
# 00:59 tantek or rather, sometimes I want to full expressive power of HTML for a blog post
# 00:59 tantek and sometimes I just want to write a "simple" blog post with name, a few paragraphs, and maybe a few links
# 01:03 tantek goodness sakes I deployed a bug which was causing my home page to load ALL of my blog storage backwards through 2010
# 01:06 tantek ok back to less than a second before you start seeing content
# 01:06 tantek who was complaining about the loading speed of tantek.com? was that you aaronpk? or someone else using x-ray?
# 01:08 aaronpk That was me cause I stopped getting updates in my reader because it started timing out
# 01:08 tantek for the curious, the bug involved just two characters, an extra ! and a misplaced )
# 01:08 tantek aaronpk - care to retest it to see what speed you're seeing now?
# 01:11 tantek sheesh, I was loading and XML parsing 49 (X)HTML5 bim files on every home page render, instead of ~8 currently
# 01:11 tantek now to find out if Bridgy Publish an issue to GitHub supports tags-> labels
# 01:12 tantek decides to read the docs first before hitting publish
# 01:14 [eddie] aaronpk: ohhh yeah push notifications. I’m still having those go through slack until I can get that running in Indigenous as well
# 01:15 aaronpk [eddie]: you had plans to run a push notification server for the app right?
# 01:15 tantek Bridgy Publish to GitHub does not yet support tags->Labels
# 01:17 aaronpk silo.pub does and fwiw I have accidentally created a bunch of labels on github that I don't want
# 01:17 aaronpk Also you can only add labels on github to repos you have commit access to
# 01:18 tantek that's fine, I'd rather it at least add the labels that are allowed and fail on any "new" labels
# 01:18 aaronpk I keep creating the "Microsub" label on the Microsub repo because I want my website posts to be tagged that but it doesn't make any sense to tag the Microsub issue as Microsub
# 01:19 aaronpk So I'm not actually convinced that a 1:1 relationship between post tags and github issue labels makes sense when posseing
# 01:20 tantek maybe have an option (default?) to not *create* new tags
# 01:20 aaronpk I agree it is from the technical standpoint but I just gave you an example of when it doesn't make sense
# 01:21 tantek Labels *are* the UI equivalent of tags on GitHub issues
# 01:21 tantek you click them, and you see an agregation of everything with that label
# 01:21 aaronpk Yes but I want the post on my site to be tagged "Microsub" and I don't want that to be tagged Microsub on github
# 01:21 aaronpk So there's something different about tags when talking about POSSE
# 01:22 aaronpk I kee deleting the Microsub tag on GitHub to fix it
# 01:23 tantek so you only want the Microsub tag on your own site?
# 01:23 tantek Bridgy already has a mechanism for excluding content to POSSE
# 01:24 tantek no sorry, Bridgy has a mechanism for explicit subset inclusion to POSSE
# 01:25 aaronpk Also note that when it's created it has the default grey color which is also not that great
eli_oat joined the channel
# 01:26 tantek <p class="e-content"><span class="p-bridgy-github-content">Problem is ... <span class="p-category">GitHub-Label</span></span>, <span class="p-category">Site-only-tag</span></p>
# 01:27 tantek assuming p-bridgy-github-content is supported the same way that p-bridgy-facebook-content and p-bridgy-twitter-content is supported
# 01:27 tantek OR alternatively we ask for only tags -> *existing* labels
# 01:28 tantek then next we can ask for Bridgy Publish u-tag-of support to add labels (new or not) to existing issues :D
# 01:29 tantek darn it p-summary in the plain text is throwing off Loqi
# 01:41 tantek aaronpk - did you get a github mention notification?
# 01:41 tantek aaronpk - are you signed up for Bridgy backfeed from GitHub? and if so, did you get a homepage webmention?
# 01:42 tantek first q was re: a notification *inside* github
# 01:42 aaronpk oh, no I don't see anything, but I also always forget how to find github notifications because they are so confusing
# 01:48 [eddie] aaronpk: Yep, I am going to run a push notification server for Indigenous.
# 01:49 [eddie] I think it’ll need to be able to WebSub subscribe to Microsub channels in order to really make it real-time push notifications
[kevinmarks] joined the channel
# 01:49 [kevinmarks] spiderpigging tummelvision out of the wayback machine, and it hasn't got a lot of the tag pages
# 01:51 aaronpk [eddie]: interesting. would you imagine you could enable notifications for individual channels then?
# 01:52 [eddie] I still use the notifications default channel as my main one but there are definitely other banners I would want push notifications from
# 01:53 aaronpk yeah, and also maybe filtered, like only posts matching X from channel Y get a push notification
# 02:18 Loqi Ok, I'll tell them that when I see them next
# 02:21 aaronpk okay, switching my webmention notifications from IRC to Aperture, let's see how this goes
# 02:21 aaronpk at some point in the very near future i'm going to need to solve this whitespace problem
# 02:23 aaronpk my html sanitization must be removing the <br>s instead of changing them to newlines
# 02:26 tantek (I really should extract the conclusions and provide them more clearly, and move discussion to a historical section or something)
# 02:26 aaronpk this isn't even a plaintext whitespace issue, since it's actually trying to render your html
# 02:27 tantek yes it is - that section explains *why* my plain text notes have been converted to that html
# 02:27 tantek specifically because it was easier for consuming applications to consistently consume it!
# 02:27 aaronpk from my point of view, I don't care whether you authored it as plaintext or not, all I see is a <br> tag
# 02:28 tantek might be worth checking the Woodwind source for that
# 02:29 aaronpk i'm surprised I don't already have a test case in xray for thi
# 02:37 aaronpk for actual plaintext conversion, I *was* allowing <br> tags. but it was stripping them from the html version
# 02:38 tantek I wonder if my latest issue comment POSSE failed because the issue was closed? but the GitHub web UI still allows me to comment so I assumed the Bridgy POSSE would work too
# 02:39 aaronpk no error message from bridgy? the logging is usually pretty good
# 02:39 tantek pretty opaque error. see !tell above to snarfed
# 02:42 tantek except Bridgy Publish handles it just fine when POSSEing to FB or Twitter
# 02:42 aaronpk note that the html has the <br> but the plaintext conversion has no newlines
# 02:42 aaronpk so bridgy publish must be using the html version and doing its own plaintext conversion
# 02:43 tantek so that the newlines do not introduce any extra plain "space" (0x20) chars
# 02:43 aaronpk I don't follow. there is definitely a problem with the content.value in both the python and ruby examples linked above
# 02:44 aaronpk both the `name` and `content.value` values have this: "...makes sense.However there is..."
# 02:44 tantek no one handled the plain text whitespace properly, that was the problem
# 02:44 tantek and why we went with option a. instead of option b.
# 02:45 aaronpk again, if the consumer is using the html version then it's fine
# 02:46 tantek right, that's the conclusion both kylewm and I believe snarfed came to - you have to use the html version
# 02:46 tantek if you want to preserve any of that kind of "formatting" at all
# 02:46 aaronpk no, it should not be collapsing <br> to "", it should convert <br> to "\n" for the plaintext
# 02:46 aaronpk the php parser was fixed and now provides good values for both html and plaintext
# 02:47 aaronpk which is why this is an easy fix for me, since xray is doing the html sanitization, not the php parser
# 02:48 aaronpk thankfully the underlying parser is fine so I can fix this right now
# 02:48 tantek I think my microblog syndication also has this issue
# 02:49 tantek but I figured manton would fix it when he gets around to it :)
# 02:50 tantek uh that's odd - now I don't see any posts on my micro.blog
# 02:50 aaronpk hm, micro.blog's issue seems to be different. it looks like it's converting the <br>s to a space
# 02:51 aaronpk oh, no I see. it's converting it to a literal \n rather than allowing the <br> tag or using css to display the newline
# 02:51 aaronpk it's not stripping the <br> completely like the other mf2 parsers
# 02:52 aaronpk this would have been a harder issue to track down if it weren't for your autolinker not adding whitespace and literally just adding <br>
# 02:52 aaronpk cause it made posts in my reader show up with obviously missing whitespace rather than how on micro.blog it just looks like you write really long paragraphs
# 02:55 tantek I see my mb synd posts now - there was some external blocking happening
renem joined the channel
# 02:59 tantek are you POSSEing Markdown or HTML via silo.pub to GitHub?
# 03:00 aaronpk it's a bit sneaky. I have a markdown storage format for my site, and my POSSE code will send whatever the source format for my post is when POSSEing
# 03:00 aaronpk it kind of worked out great for this, even though I wasn't planning on it
# 03:01 tantek huh I can't find that issue on your home page or even linked from
# 03:01 tantek hmm - curious why you haven't enabled Bridgy Backfeed for GitHub yet then
# 03:02 aaronpk cause it launched while I was in the middle of traveling and didn't want to turn it on until I had some time to monitor how it behaved after I launched it :)
# 03:05 aaronpk okay new xray deployed to aperture... now to wait for tantek to make another post with newlines
tantek, globbot and barpthewire joined the channel
# 06:22 gRegorLove indieauth.com error "NoMethodError at /auth/twitter/callback undefined method `[]' for nil:NilClass file: auth-web.rb location: block (2 levels) in <class:Controller> line: 634"
# 06:23 gRegorLove Still have the error page open with a lot more information if you need it.
# 06:25 gRegorLove "You're seeing this error because you have enabled the show_exceptions setting."
leg1 and [tantek] joined the channel
# 06:48 [tantek] As in, grabbing (or grab your) popcorn to watch (link)
# 06:54 [tantek] As in “watching” present tense rather than “watched” / completed (which is most current uses)
tantek joined the channel
# 06:59 Loqi [gRegor Morrill] I’m experimenting with a new post footer on my homepage. I got some good feedback when someone wanted to comment from the homepage. They clicked follow my posts and were confused by the suggestions to subscribe with Feedly, Woodwind, etc. In this m...
AngeloGladding and [eddie] joined the channel
# 07:35 [eddie] Looking into live streaming video and syndication using Facebook Live. But darn it, Facebook doesn’t provide the time stamps of when comments/reactions are made. Only the total number of reactions by type
# 07:36 [eddie] I was hoping to be able to easily backfeed love reactions, but unless I figure out something clever looks like they would be turned into non-live reactions
# 07:41 tantek [eddie] do those reactions not show up in your "activities" or "notifications"?
# 07:41 Loqi Post or posts may refer to individual pieces of content published on an indieweb site such as notes, articles, & responses, or the act of creating the aforementioned content (present tense), or Posts about the IndieWeb https://indieweb.org/footer
# 07:42 [eddie] Hmmm that’s a good question. I’ll have to take a look at that endpoint real quick
# 07:47 tantek whoa just tried manually using Bridgy Publish to POSSE that issue comment to GitHub and got a 500 Internal Server Error!
tantek, mifga and raucao joined the channel
# 11:27 Zegnat Interesting discussion re: whitespace in the microformats parsets.
# 11:30 Zegnat For people reading up on logs: I responded in #microformats
[kevinmarks] joined the channel
tantek, [colinwalker], [mifga] and [xavierroy] joined the channel
# 16:17 petermolnar_ !tell cweiske "https://indiechat.search.cweiske.de/?q=https%3A%2F%2Fnews.ycombinator.com%2Fitem%3Fid%3D16505244" results in HTTP 500; I pasted "https://news.ycombinator.com/item?id=16505244" in the search field on chat.indieweb.org
# 16:17 Loqi Ok, I'll tell them that when I see them next
KartikPrabhu, leg, theReal33mhz, barpthewire, [miklb] and tantek joined the channel
# 19:29 Loqi [gRegorLove] indieauth.com error "NoMethodError at /auth/twitter/callback undefined method `[]' for nil:NilClass file: auth-web.rb location: block (2 levels) in <class:Controller> line: 634"
# 19:31 gRegorLove I think the second time hit a different server, because Twitter didn't show as an option initially. I refreshed the rel-mes and then Twitter showed up and it worked.
[tantek] joined the channel
# 19:57 dgold aaronpk: I'm going to delete that brainstorming/issue
# 19:59 dgold I was working through something,I thought the issue was with the microsub I was working on.
# 20:00 dgold I worked through the issue more fully before responding to your request for more detail
# 20:00 dgold which 'clarified' things :slightly-smiling-face:
# 20:08 aaronpk np. it's still quite possible that more of this needs to be clarified in the spec anyway.
# 20:09 aaronpk ohhh that's why the reply contexts for tantek's posts are coming up empty in monocle... there is no "in-reply-to" URL visible on tantek's home page posts. the reply context only shows up on post permalinks.
# 20:09 tantek er, changing it. it's not technically "wrong"
# 20:09 tantek but makes a series of replies on my home page less meaningful
# 20:11 aaronpk this kind of thing also leads me to wonder whether I will need to start fetching post permalinks to get better post contents
# 20:12 tantek withouth the in-reply-to, there's no way to tell it's a reply right?
# 20:12 aaronpk even visually on your home page it doesn't look like a reply other than the fact that the text doesn't make sense out of context
chrisaldrich joined the channel
# 20:13 tantek a-ha, so that makes post type discovery harder (worse result), which then means how a composite stream is displayed degrades
# 20:13 tantek right, that (visually) is the key indicator something needs to change
# 20:13 tantek do we have screenshots of how silos treat replies in a composite stream?
# 20:15 tantek perhaps we need to explicitly document (and brainstorm) in-stream reply-contexts as well as (perhaps even visually distinct from) post permalink reply-contexts
# 20:16 aaronpk interesting, twitter includes the text "Replying to _____" in grey above the post content
# 20:16 tantek that's worth screenshotting / cropping and analyzing as a minimal in-stream reply-context
# 20:48 gRegorLove Think I'm going to remove Woodwind as the suggested service from my Subtome buttons, unfortunately.
# 20:53 tantek gRegorLove: woodwind.xyz presumably? should we distinguish the project vs the .xyz deployment / service?
# 20:54 gRegorLove If another stable instance came up, I could specify it's subscribe endpoint in the button.
# 20:54 tantek if another stable instance came up with perhaps 2+ people maintaining it
# 21:18 tantek aaropnk, I'm re-evaluating my "Working On" to see about shuffling in-stream reply-contexts closer to the top, it's starting to visually bug me on my home page too
# 21:22 tantek found it - apparently already in my list with some brainstorming
snarfed and [kevinmarks] joined the channel
# 21:45 [kevinmarks] I suspect kevinmarks.com will be differently annoying as I have a p-summary on the homepage feed but not in the post pages
chrisaldrich joined the channel
# 22:01 tantek I'd really like to get On This Day working soon but I think I will prioritize in-stream reply-contexts to help with /reader iteration
# 22:04 tantek snarfed, I'm curious what I did differently to cause the 500s
# 22:05 snarfed just a silly bug on my end, half baked json serialization
# 22:05 tantek besides my usual unintentional ability to break things
# 22:10 tantek.com edited /Falcon (+620) "fixed Recent Articles box, rework remaining in section to improvements & reprioritize, reply-context iteration and reprioritize" (
view diff )
# 22:13 tantek I think the big thing I need to consider is do I try to do something minimal and not very good looking with the data I have, or do I try calling XRay synchronously at Publish time to get/store more information to use for better reply-context display later
AngeloGladding joined the channel
# 22:32 aaronpk I personally don't mind just the url as the minimal reply context
# 22:33 aaronpk And twitter's "replying to @t" isn't a bad upgrade from that too
KartikPrabhu joined the channel
# 22:44 tantek interesting, I may be able to do that with just the URL
# 22:45 aaronpk I think so... a whitelist for twitter and github to find the username from the url, then use the hostname for everything else?
# 22:45 tantek "replying to @-name" if it's a twitter.com / name URL, otherwise "replying to (domain)"
# 22:47 tantek "replying to twitter.com/aaronpk" instead of "replying to @aaronpk"
# 22:47 tantek or maybe silo-list plus exception for Twitter since I already auto-link @-names to Twitter and am not going to change that anytime soon
# 22:48 tantek that doesn't work for github tho since the replies are to a particular issue which is opened by a particular person that is not the github/name
# 22:48 tantek since that's just the name of the person/org that owns the repo
# 22:49 tantek this will take some thinking to make sense (more than just displaying the in-reply-to URL)
# 22:50 aaronpk GitHub issue URLs have enough in them to show something meaningful too
# 22:51 Loqi [aaronpk] #1 Give Microsub server access to private posts for its user?
# 22:54 tantek prefer the project name rather than the author - that could work
# 22:57 aaronpk Hm what do my reply contexts for issues do right now... I think they use the project name? Via XRay's GitHub conversion
# 23:13 Loqi [tantek] Appreciate the explanation and link to the source file; makes sense.
However there is still a fundamental usability problem of the discoverability of how to file issues and suggested improvements for CSS module test suites.
I would like to suggest ...
# 23:13 Loqi snarfed has 48 karma in this channel (362 overall)
snarfed, [kevinmarks] and [miklb] joined the channel