#dev 2020-02-23

2020-02-23 UTC
[dave] and [KevinMarks] joined the channel
#
[KevinMarks]
Outputting MySQL using php is fairly widely done - what's the problem?
miklb joined the channel
#
[chrisaldrich]
What is Django?
#
Loqi
Django is a python web application framework used in some IndieWeb projects https://indieweb.org/Django
#
[chrisaldrich]
[Rasul_Kireev] ^^ has a few examples of people who are using it (or variants). KartikPrabhu is frequently in the chat here.
#
KartikPrabhu
yeah I was busy at the time of the question
#
[chrisaldrich]
I'm curious what problem(s) you're having though?
#
[chrisaldrich]
Is it a webmention issue or a bridgy issue?
#
KartikPrabhu
most likely the Django-webmention plugin
#
KartikPrabhu
bridgy formats everything as HTML+mf2 so it can't be bridgy's fault
[Brian], miklb, [jgmac1106], [datashaman], tbbrown, [gRegorLove], [dave] and [chrisaldrich] joined the channel
#
[chrisaldrich]
I'm totally lost by the turbolinks session notes from earlier today. If someone who was there could flesh things out a bit more for our future selves, it would be appreciated. 🙂
#
jacky
something I've finally finished: https://i.imgur.com/yMFTy1A.png being able to 'enable' post types on my site
#
jacky
if it's not 'enabled', it'll render as a note
[Michael_Beckwit and gRegorLove joined the channel
#
KartikPrabhu
what is turbolinks?
#
Loqi
It looks like we don't have a page for "turbolinks" yet. Would you like to create it? (Or just say "turbolinks is ____", a sentence describing the term)
KartikPrabhu and [jeremycherfas] joined the channel
#
GWG
Making progress on my project, though it may be more of a Hack Month project than a Hack Day
[LewisCowles] and petermolnar joined the channel
[KevinMarks] and [datashaman] joined the channel
#
[datashaman]
I'd like to confirm my understanding of how IndieAuth works with Micropub. I"m busy implementing a micropub service. When I get the first point of entry, I get an access token, which I must authenticate against a token endpoint in order to get the packet with _me_ etc. There is no indication which token endpoint to use, so does this mean that a micropub service _must_ share a common token endpoint with any site using it?
#
[datashaman]
Up till now I've been using tokens.indieauth.com/token but this is not universal - for example aaron has his own endpoint - does this mean that he would not be able to use this service? or am I missing a piece of the puzzle somewhere...
#
GWG
All token endpoints do is issue tokens
#
GWG
They are interchangeable in that regard
#
[datashaman]
sure, but in order for me (as micropub service provider) to validate the token we must agree on a token endpoint, correct?
[LewisCowles] joined the channel
#
GWG
[datashaman]: No
#
GWG
IndieAuth mandates that the token endpoint is declared on the URL that is trying to log in
#
[datashaman]
it is, but it could be any token endpoint that's in the HTML link, when i receive a token from a micropub request all i have is the token, there's no hint at that point which token endpoint it was dispensed from
#
GWG
Wouldn't you have stored that info?
#
[datashaman]
against what tho
#
[datashaman]
there's no session in this, it's stateless
#
GWG
The token when sent to the endpoint usually returns metadata about the token
#
[datashaman]
that metadata is only accessible once you've validated it with a configured token endpoint
#
[datashaman]
the info about the token endpoint is embedded in the token, which needs to the token endpoint to decode. 🙂
#
[datashaman]
~to~
#
GWG
I may be misunderstanding the problem
#
[datashaman]
thinks how to express this
#
[datashaman]
Simplified sequence of events:
#
[datashaman]
• micropub client (MC) asks token endpoint (TE) for a token
#
[datashaman]
• TE returns an opaque token to MC
#
[datashaman]
• MC uses opaque token to query micropub service (MS)
#
[datashaman]
• MS receives request with token, but has no info at this point about which TE was used to dispense the token.
#
[datashaman]
i don't think there's any point before that where the source of the token can be stored or saved.
#
[datashaman]
i dont know which site it's for now which TE was used, that's encoded in the token and is only shown to the MS when the token is validated
#
[datashaman]
~now~ nor
#
GWG
How does your Micropub endpoint know what scopes it has if it doesn't get that info?
#
[datashaman]
when the indieauth site is registered in my service
#
[datashaman]
basic flow is this:
#
[datashaman]
• user logs in with an indieauth site, adding it to their list of managed sites
#
[datashaman]
• user logs in with github to get access to repositories
#
[datashaman]
• membership in that list of sites allows micropub clients to use the service
#
[datashaman]
i get a token and scopes back from that first login, yes
#
[datashaman]
the micropub service is using a diff token, diff client, diff requests
#
GWG
Hmm...will have to think about that. You may need someone who knows more about oauth2 flows than I do
#
[datashaman]
thanks for trying 🙂
#
[datashaman]
i've looked in headers and all the input, no hints
#
[datashaman]
it's no biggie to restrict access to sites that use a specific TE, it's just my stuff for now.
[KevinMarks] joined the channel
#
[KevinMarks]
isn't it up to your site to specify a token endpoint?
#
[KevinMarks]
with rel="token_endpoint"
#
GWG
That was my comment
#
[KevinMarks]
so, you pick one, publish that, and use it to verify tokens
#
[KevinMarks]
now maybe you use a common token algorithm that other endpoints also support, but you can't really relay on that as tokens are opaqu
#
Zegnat
From a quickread: yes, it is up to a site to decide which micropub endpoint and which token endpoint it uses. Because the site choses both, it is assumed the site can instruct the micropub endpoint behind the scenes.
#
Zegnat
People cannot bring their own token endpoints to any micropub endpoint, that would allow anyone to bring a token that gives them full access
#
Zegnat
[datashaman]: 👆
#
GWG
Zegnat: I still have to wrap my mind around how autoauth does it
#
Zegnat
The final token in autoauth is issued by your site, not by the site of the person authenticating.
#
Zegnat
Basically: you can only trust tokens issued by a trusted token endpoint. (Else scopes and me value may be spoofed.) And the only way to trust the code behind a token endpoint is for you to define the token endpoint that is used
#
[datashaman]
thanks, that clarifies the process for me, and makes total sense
#
[datashaman]
otherwise an answer could be spoofed by a dodgy token endpoint
#
[KevinMarks]
I mean it still could be if your site links to a dodgy one, but you have said that you trust it.
#
Zegnat
Sink (a test site that lets anyone login and do micropub) just supports IndieAuth alone. It lets anyone login with IndieAuth and then issues a token of itself to the Micropub client for the actual posting.
#
Zegnat
Not using the visitors token endpoint at all.
#
Zegnat
[KevinMarks]: sure, which is why I am talking about trust rather than secured. Your trust is still a personal choice, and may be misplaced ;)
#
Zegnat
[datashaman]: if you have a Micropub client and want to see a “multi-user endpoint” in action, play with https://sink.zegnat.net/
#
GWG
Zegnat: So what do I need to do to support autoauth?
[jgmac1106] joined the channel
#
Zegnat
Depends on what part of the flow (WIP flow diagram: https://www.svenknebel.de/temp/autoauth_diagram.svg) you are thinking of
#
[jgmac1106]
moving from main channel it terms of a note feed out of https://jgregorymcverry.com/poetryportpoems I may just use time as the link...in the poetry example the entire day is an h-entry the dt-published is on the day and then each poem just a h-item
#
[jgmac1106]
going to use h-entry for each note....probably not nested....not sure about the link name. Note number? Time of day?
#
GWG
Zegnat, my token endpoint does not accept remote requests, for one
#
Zegnat
GWG, the basic idea is the following. But the details are still very much WIP:
#
Zegnat
1) When the user encounters a protected resource she asks her own authorization_endpoint to get a token for her.
#
Zegnat
2) Her authorization_endpoint looks at the resource to detect a token_endpoint that the resource trusts. (This is the resource’s token_endpoint, not the user’s!)
#
Zegnat
4) The token is returned to the user.
#
Zegnat
3) Her authorization_endpoint negotiates a token with the token_endpoint.
#
Zegnat
This means the authorization_endpoint must implement the remote token request, and implement the ability to negotiate. And it means the token_endpoint must implement token requests from an unknown authorization_endpoint and the negotiation.
#
[jgmac1106]
Can I do <a class="u-url href="#"><time="dt-published">time</time></a>
#
Zegnat
[jgmac1106]: Only if you put that dt-published in a class as well. It looks like some HTML was missing there
#
Zegnat
<a class="u-url" href="#"><time class="dt-published">...</time></a> is valid HTML and microformats
#
[jgmac1106]
yeah I was short handing
#
[jgmac1106]
thanks....I think I will display notes by time of day, .oooh p-name.....crap...maybe their is a page on minimal h-entry
#
GWG
Zegnat: My token endpoint ties every token to a user. Remote sites wouldn't have a user
#
Zegnat
GWG, yeah, that would be a problem
#
GWG
It would be a redesign to address
#
Zegnat
Basically the idea was that a page could be readable only by certain people identified by URLs. And because of the remote token request flow, any IndieAuth enabled URL can get a token to use to read the page.
#
[jgmac1106]
http://microformats.org/wiki/h-entry#faq what does name can be implied mean?
#
Loqi
[Tantek Çelik] h-entry is a simple, open format for episodic or datestamped content on the web. h-entry is often used with content intended to be syndicated, e.g. blog posts. h-entry is one of several open microformat standards suitable for embedding data in HTML. ...
#
[datashaman]
i think it might be nice to include a header with the token response that indicates which token endpoint it was issued by, so that services can decide which they trust
#
Zegnat
[jgmac1106]: it means that if no p-name is given, but certain other requirements are met, the mf2 parser will automatically try and detect a name
#
[jgmac1106]
<a class="u-url" href="#"><time class="dt-published"><span class="p-name">16:30</span></time></a> that might work
#
[jgmac1106]
okay so if I use the id property, I can include start of note.....though I think I kinda like using the time as the p-name for a calender feed
#
Zegnat
[datashaman]: interesting idea. To me it feels like that might result in a lot more policing though. Do I still trust the same URL tomorrow? Or what about next month? Do I need to be wary of domain renewal dates?
#
Zegnat
You would not even need that <span> then, [jgmac1106]. Just put p-name and dt-published on the same element.
#
Zegnat
[jgmac1106]: I would focus on building the page first, with all the data you want to be presenting, and care about the mf2 second.
#
[jgmac1106]
ahh sweet, thanks, that is what i will do. time to finally build my notes feed
#
[datashaman]
all problems, agreed; i think my way gives people more flexibility; but with that comes risk, but it's self-assumed risk
#
[jgmac1106]
5 pages: october-feb
#
GWG
Zegnat: Also need to build private posts as well then
#
[datashaman]
bestest option is to be the auth provider as well i suppose
#
[jgmac1106]
..scratch that starting with February and moving forward, save backfill for later date
#
Zegnat
[datashaman]: what I am doing with sink.zegnat.net is that I am the auth and token provider, except the auth provider is not username/password but actually in itself is IndieAuth. Meaning that the Micropub endpoint only trusts stuff I generate, but I trust the user to decide what auth they want. (Which in turn means the user does not need to trust me with passwords, they only see their own login screen.)
#
[jgmac1106]
I may drop M-F, so I don't have to worry about aligning calendar with days of the week
#
[datashaman]
that's an interesting idea, but I fear I am shaving yaks here. I"ll get basic micropub working with indieauth.com as TE first. thanks for the tips.
#
Zegnat
https://tokens.indieauth.com/ ? I keep forgetting that exists, haha
#
[jgmac1106]
then realizes like a weirdo I may have estimated by dt-published to land on the fourth minute and 20 second mark...of every note..........
#
[jgmac1106]
<div><a class="event event-end h-entry u-url" href="/notes/2020-02-01-1"><time class="dt-published p-name" datetime="2020-02-01 07:44:20">7:44</time></a>
#
Zegnat
[jgmac1106]: just a warning, that HTML will not work
#
Zegnat
You cannot have the u-url and h-entry on the same element if you want the u-url to be part of the h-entry.
[LewisCowles] joined the channel
#
[jgmac1106]
yeah realized that. trying to get it as small as possible
#
[jgmac1106]
may just wrap it in a div
#
[jgmac1106]
why I kinda liked doing each day as an h-entry and each poem as an h-item but not doing that for notes
#
[jgmac1106]
working on fixing this poem first: https://jgregorymcverry.com/heavenlyremixes took a break
#
[jgmac1106]
and I haven't even looked at what today's word is for new poems...
#
[jgmac1106]
each poem could have been an h-entry, but I feel like with a 30 challenge the entry is the day.
#
[jgmac1106]
every day I wake up and have to delete 8 gigs from a folder called .git I should figure out how to stop needing to do this
#
Zegnat
why is your git folder 8GB?
#
[jgmac1106]
I dunno, it refreshes every day
#
[jgmac1106]
I think it is my entire site, maybe an extra copy?
#
[jgmac1106]
so I wake up log in and delete it
#
[jgmac1106]
rinse, wash, repeat
#
Zegnat
It shouldn’t automatically fill itself if you aren’t using git, AFAIK
#
sknebel
I assume some kind of Git backup plugin?
[jgmac1106] joined the channel
#
[jgmac1106]
I am using git
#
sknebel
well then you of course have a .git folder :)
#
[jgmac1106]
I use github desktop and an ftp action
#
[jgmac1106]
there is probably something I can add to the YAML file nto to back it all up, need to dig in
#
Zegnat
[jgmac1106]: I mean that it cannot auto-fill itself, unless you are making commits of many many hundreds MB, or are continuously `git pull`-ing from a multi GB origin
#
[jgmac1106]
Ilemme see how many commits I made yesterday
#
[jgmac1106]
253 commits this month, not sure how many yesterday
#
[jgmac1106]
14 yesterday, I will poke around what you can do with the ftp action and see if I can stop it some how
#
Zegnat
Number of commits does not matter, I think? Though I am a little hazy on how git stores its stuff
#
[jgmac1106]
yeah I will wait until tomorrow when it is there again, and poke around
[jeremycherfas] joined the channel
#
[datashaman]
yesss! micropub working
tbbrown and gRegorLove joined the channel
#
jamietanna[m]
!tell snarfed just seen your message about releases for granary and oauth-dropins. I'd be happy to help, but don't want to slow things down if you want them out there :)
#
Loqi
Ok, I'll tell them that when I see them next
[yo] joined the channel
#
[jgmac1106]
so close...just can't figure out for life of me why the note on Saturday the first showing up on Sunday
#
sknebel
there's nothing it would get a position from
#
[jgmac1106]
each week is set up as a grid, that satruday should be the 7th spot
#
sknebel
but you use display:contents on the parent div
#
sknebel
so the parent div doesn't exist for the <p>
#
sknebel
you probably don't want that display:contents
#
[jgmac1106]
sknebel++
#
Loqi
sknebel has 17 karma in this channel over the last year (52 in all channels)
#
sknebel
(generally, you have all kinds of CSS rules in there where I'd have no idea why they are there. that's not making it any easier for you)
#
[jgmac1106]
yeah there are some to span multiple days, I was just looking for a CSS Grid calendar to start from
[LewisCowles] joined the channel
#
[jgmac1106]
I might just drop the M-F and just have calendar start on one to make it easier
#
[jgmac1106]
but if I wanted to add an h-event for remoting into Austin I could us: event-start, event-end, and data-span="2"
[tantek] joined the channel
#
[jgmac1106]
that fixed it, thanks
[snarfed] joined the channel
#
[snarfed]
jamietannna great, glad to hear it! no hurry on my side, happy to help if you're interested. feel free to go ahead and try the release instructions in the readme(s) and see how far you get. o-d first
#
Loqi
[snarfed]: jamietanna[m] left you a message 42 minutes ago: just seen your message about releases for granary and oauth-dropins. I'd be happy to help, but don't want to slow things down if you want them out there :)
#
[snarfed]
also let me know your test.pypi.org and eventually pypi.org usernames, and i can add you to the packages
#
[jgmac1106]
snarfed can I use an HTML entity for spaces with Bridgy? Only tried once but I messed something up
#
jacky
[jgmac1106]: it might also depend on the place you publish to; twitter is finicky with spacing
#
[jgmac1106]
thx [yo], it works on Known and @kevinmarks taught me about option C whihc I get to work
#
Loqi
[aaronpk] #2 image alt text is lost during parsing
#
aaronpk
wow this is a mess
#
[jgmac1106]
jacky, snarfed doesn't really matter as the poem will look correct, thus mean the same thing you want it to on my website, spacing lost on POSSE copy oh well
#
aaronpk
really hairy xray issue
#
aaronpk
`<div class="e-content"><p><img src="/photo.jpg" alt="test"></p></div>` causes the entire "content" propeerty to be lost
#
sknebel
that's... interesting
loophole1 joined the channel
#
aaronpk
it's...weird
#
aaronpk
i have code that handles the case of <div class="e-content"><img src="/photo.jpg" alt="test"></div> which move the photo to the "photo" property so that it appears as a photo post
#
sknebel
(only if the image is the only content?)
#
aaronpk
tho now i am realizing it drops the alt text in that case
#
aaronpk
which is..debatable
#
sknebel
create an object like the parser would?
#
sknebel
or does x-ray not use that?
#
aaronpk
i think this would all be cleaned up with proper alt text handling, like making sure the alt text is included in the parsed result as alt text, not putting it into the text content
#
sknebel
ah, that' not in the php parser yet
#
sknebel
sigh, parser work really has stalled across the board
#
aaronpk
interesting, actually the new alt parsing rules doesn't come into play here
#
aaronpk
because there's no u-photo property
#
sknebel
sure, but your conversion pretends there was one
#
sknebel
so it should produce the same output
#
aaronpk
yeah, there's a few things going on here
#
sknebel
but x-ray probably doesn't do alt-text parsing anywhere, because the parser doesn't?
#
aaronpk
yeah i don't think it does at all right now
#
aaronpk
but in any case i need to figure out what to do with this <p><img></p> case regardless of the alt text issue
#
sknebel
presumably you had a reason for adding it
#
aaronpk
because even <div class="e-content"><p><img src="/photo.jpg"></p></div> drops the content completely right now
#
aaronpk
the problem is the plaintext version of that markup above is an empty string, which is correct
#
sknebel
if you intend to convert to a photo post that's ok (well, maybe set an empty content)
#
aaronpk
so i'm not gonna get alt text handling in here right now because that is too many steps. so right now, i need to decide whehter to try to promote that p/img/p into a photo post, or just keep the html markup as the HTML content as is
#
sknebel
that's a case where it'd be fun to have logging so you could see how often it's triggered in the wild
#
sknebel
(I generally tend to argue for less intelligence in xray because people expect all other software to behave like aaron-products, but if it causes issues regularly that's of course a good reason to keep it)
#
sknebel
(and we don't know about false positives. emoji images e.g. by wordpress would be my usual suspect here, but is that actually an issue?)
#
aaronpk
i did have to make a judgment call on my html sanitization which is that it only allows img tags in HTML content if there is no "photo" property. That allows emoji images in blog posts, but does prevent emoji images in captions of photo posts.
#
sknebel
also an interesting case
#
aaronpk
now i can't even find the code that's bumping the <img> into the photo property
#
aaronpk
oh it's not me, it's the parser
#
sknebel
right, implied photo
#
sknebel
and then your code to stop u-photos in the content from doubling up pulls it from the content?
#
aaronpk
just had a conversation here, and it sounds like the solution might be to stop handling this implied photo case completely
#
aaronpk
and instead tell authors that they should add a u-url to their h-entry which stops implied photo parsing
#
aaronpk
oh look there's an open issue for this https://github.com/aaronpk/XRay/issues/55
#
Loqi
[aaronpk] #55 incorrectly including a photo property
[Erika_W_] joined the channel
#
sknebel
fwiw, your argument in the linked parsing issue is worth discussing too (it got lost because it was just a side note in a different discussion)
#
sknebel
if implied parsing mixes with e- at all
#
aaronpk
i guess that's the other way to handle it
#
aaronpk
right now the current thinking in the room is that this case of an implied photo in h-entry should never actually exist in the wild because the h-entry should always at least have a u-url property which stops implied parsing
#
sknebel
also true
[chrisaldrich] joined the channel
#
[jgmac1106]
so if I use time as p-name for my notes am I thereby making them into articles since they now have a p-name different than e-content?
#
aaronpk
yes please don't do that lol
#
aaronpk
only include p-name if your post has a name
#
aaronpk
in general, don't add random markup or css unless you know what it's doing and have a good reason
#
[jgmac1106]
I am trying to build a feed of notes but put them on the calendar
#
[jgmac1106]
I guess I will go back to default on the id and including that
#
aaronpk
i don't see why p-name has anything to do with that
#
aaronpk
the date of the post is available already in the "published" property
#
[jgmac1106]
well I have to have some p-name for a minimal h-entry
#
[jgmac1106]
unless I go back to having it get inferred from id
#
[jgmac1106]
just a space contsraint issue
#
aaronpk
just cause you're linking to the post with the time 8:54 doesn't mean you need to also add p-name to that element
#
[jgmac1106]
but then I don't meet requirments for a minimal h-entry
#
aaronpk
what requirements?
#
Loqi
[Tantek Çelik] h-entry is a simple, open format for episodic or datestamped content on the web. h-entry is often used with content intended to be syndicated, e.g. blog posts. h-entry is one of several open microformat standards suitable for embedding data in HTML. ...
#
[jgmac1106]
the question about bare requirements
#
[jgmac1106]
only way I could see name gets implied is by using the id property
gRegorLove joined the channel
#
aaronpk
i'm not sure what you were linking to, the fragment doesn't go anywhere
#
[jgmac1106]
so in trying to build an h-feed of all my notes which each have their own canonical url...that is what I am trying to do
#
Loqi
[Tantek Çelik] h-entry is a simple, open format for episodic or datestamped content on the web. h-entry is often used with content intended to be syndicated, e.g. blog posts. h-entry is one of several open microformat standards suitable for embedding data in HTML. ...
#
aaronpk
it looks like that page hasn't been updated since the parser change that stopped implying "name"
#
aaronpk
i'm going to update that FAQ because "name" is not required or suggested for anything except articles or things with a visible name
#
[jgmac1106]
let me go check indieweb wiki for other examples of note feeds
#
aaronpk
note feeds do not need a "name" property
#
[jgmac1106]
...ohh thx, been bugging me for a long time....in terms of designing notes and titleless poems
#
[jgmac1106]
on my calendar of poetryport poems I ended up just using h-items nested in an h-entry because I couldn't figure out how to do the name and dt-published....but they actually look pretty cool in parser and readers so don't mind
#
[jgmac1106]
still may go back and fix now that I know name can be ignored
phpbooob joined the channel
#
Zegnat
h-items inside h-entry?
#
[jgmac1106]
I am gonna fix.
#
[jgmac1106]
aaronpk just let me know that what I was reading on the microformats wiki was wrong
#
aaronpk
where'd you see h-items?
#
aaronpk
i was talking about the name thing
#
[jgmac1106]
h-item allowed me to avoid having to figure out p-name and dt-published
#
[jgmac1106]
no, before you did this I chose h-item as a way to avoid having to include a dt-published and p-name
#
[jgmac1106]
I read minimal requirements....was like....that is going to be really hard to design how to fit all that in tight space...found h-item and went with it
#
[jgmac1106]
it looks great in feed previews
#
[jgmac1106]
and this is a very unique thing because each day has a theme so I used the theme for the day as the p-name and the date for the dt-published, then nested in the h-items. I will switch to h-entry
#
[jgmac1106]
it still turns to pretty rss with granary which is all I really care about in the end, I don't know many people who use social readers
[tantek] joined the channel
#
[tantek]
just expanded some deep functions in CASSIS to handle both parsing for URLs+alts at the start of a string, returning that in a structured form, *OR* returning the remainder of the string, since there are different use-cases for both!
#
[tantek]
updated it and didn't break any existing site functionality AFAIK
#
[tantek]
[jgmac1106] RSS won't ever give you the prettiness/fidelity that you get with social readers
#
[jgmac1106]
I know, but I am talking about everyone I know that uses RSS readers...which is everyone I know outside this community
#
[tantek]
also false reasoning. more people use Instagram than use RSS
#
[jgmac1106]
goal is get people to try social readers but I also need to provide the feed for my communities
#
[jgmac1106]
no our education blogging space a bit different, never stopped running on RSS
#
[tantek]
no the goal is to replace their use of Instagram
#
[tantek]
at least some of it
#
[jgmac1106]
every single person here uses RSS: https://clmoocring.jgregorymcverry.com/members/ that is my community, to not provide a way to interact with them is wrong
#
[tantek]
[jgmac1106] perhaps more for #indieweb channel, but which RSS readers is everyone you "know outside this community" using?
#
Loqi
Muses
#
[jgmac1106]
and it's not like it looks bad for social readers did you look at the feed preview
#
[jgmac1106]
inoreader
#
[tantek]
is there a bug filed on inoreader to support h-feed?
#
[tantek]
what is inoreader
#
Loqi
Inoreader is a feed reader service https://indieweb.org/Inoreader
#
[jgmac1106]
and jacket just announced a major rewrite to support titleless posts for micro.blog and it takes rss
#
[jgmac1106]
that was part of the session on microformats yesterday
#
[jgmac1106]
said he looked at adding support but the feeds were unreliable on most sites he checked out
#
[jgmac1106]
...he was looking at micro.blog sites as people kept asking him for better no title support
#
[jgmac1106]
inoreader and feedly, every educational blogger as a pro account to either
#
[jgmac1106]
if you teach through blogging and rss you tend to lean more towards inoreader for filters and rules
#
[jgmac1106]
allows you to sort student blog feeds by assignment and such
#
[jgmac1106]
and technically we are trying to replace Google+ not Instagram...pinterest...maybe but Instagram doesn't have a huge educational community
#
[jgmac1106]
CLMOOC does have a twitter hashtag and a facebook group as well though but that is much amplification as network
[KevinMarks] joined the channel
#
[jgmac1106]
in terms of inoreader..reading tea leaves of Discord comments, leaning more towards the json feed for microblogging, what was recommended to use when adding sites from micro.blog
#
[jgmac1106]
Okay moving on to see if I can fix my CSS Grid on: https://jgregorymcverry.com/mypoetry/ been driving me crazy....may just start over and build with Grid areas
#
[jgmac1106]
ohh and feedburner and other WP plugins, people build public RSS feeds with those as well
#
jacky
does anyone know a listing of 'intent' URLs for silos?
#
jacky
*know of a listing
#
jacky
my demo project today is meant to 'discover' them based on rel-mes
#
[tantek]
lol I think only Twitter tried to make 'intent' a thing as a silo
#
[tantek]
they were so poorly named they mostly confused people
#
[tantek]
Google had a dead-end effort to make it a spec, but no Google silo paid any attention
#
jacky
I could 'kinda' fake it with GitHub by sniffing out the CSRF token but I'm like 60% certain they associate those tokens to authenticated user sessions
#
sknebel
jacky: yeah, GH probably needs the API
#
jacky
_boo_
#
jacky
apis--
#
Loqi
apis has -1 karma over the last year
#
sknebel
or some really hacky stuff showing the profile page in an iframe scrolled to the right location :P
#
sknebel
(not recommended)
#
jacky
lol I feel like CORS might make that tricky
#
jacky
well not CORS
#
jacky
but like X-Frame-Options stuff
#
sknebel
they havethat
#
jacky
ugh going to give in and make a oauth client lol
oh_that_courtney joined the channel
#
oh_that_courtney
can somebody in Austin tell me ifyou can reach http://172.18.1.110:4000/ .... i'm trying to figure out if i can demo from local
macgenie joined the channel
#
jacky
what is rel=me
#
Loqi
Using rel=me on a hyperlink indicates that its destination represents the same person or entity as the current page, which is a key building-block of web-sign-in and IndieAuth https://indieweb.org/rel-me
[Brian] joined the channel
#
[Brian]
Just gonna drop this in here in case it affects anytthing y'all have set up https://www.theregister.co.uk/2020/02/20/apple_shorter_cert_lifetime/
#
[jgmac1106]
still need to figure out what to do with dtpublished on each one, I really do not want to display it
#
Zegnat
Does anything parse those nested h-entries?
#
Zegnat
[jgmac1106]: is there a reason why you need dt-published if you are not planning to display it?
#
[KevinMarks]
h-cite might make more sense given the comments precedent
#
[KevinMarks]
also you have an h-tem in there
flex14 joined the channel
#
[jgmac1106]
I didn't know it was required
#
[jgmac1106]
if* dt-published [KevinMarks] sould be write since these are technically published somewhere elese
#
[jgmac1106]
well didn't fix my grid here on time: https://jgregorymcverry.com/mypoetry....can't get the rows straight at all
#
[jgmac1106]
zegnat...I don't know but the h-items did parse well
#
[jgmac1106]
in terms of how they looked for people in inoreader
#
[jgmac1106]
Will find out what happens with granary->rss
#
[jgmac1106]
interesting the h-items worked with Granary but not with nested h-entries
#
[jgmac1106]
gonna switch to h-cite as that probably is technically correct since they were published elsewhere
#
Zegnat
microformats properties are almost never required, unless you are putting information out there for a specific consuming use-case
#
Zegnat
That’s why I always recommend to first build the page as you want it, and only care about the mf2 later :) No need to worry about it while you are designing your site
#
[jgmac1106]
Granary is my consumer use case
#
[jgmac1106]
I don't want to do my own xml for RSS
#
Zegnat
Focus on the human visitors first though. If you are focusing on granary first you may as well just use XOXO or some other outline-in-HTML :P
#
Zegnat
AFAIK granary is not going to parse those nested h-entry instances, so you may be making it more of a head scratcher for yourself than it needs to be
#
[jgmac1106]
I think it is pretty good for human visitors, all the poems under the correct day with the theme
#
[jgmac1106]
yeah I know, but it will the h-cite
courtney_ and phpbooob1 joined the channel
#
[jgmac1106]
no it doesn't with the h-cite...only the h-items worked
#
[KevinMarks]
because h-item is used for review
#
[KevinMarks]
(I think - [snarfed] can give you more)
#
[jgmac1106]
but it also makes an rss feed I like, nested h-entries and h-cites do not show up
#
[KevinMarks]
right, RSS can't nest
#
Zegnat
So if Granary is your consumer, why bother with the weird nested h-entries at all? There are no consumers for it. That is what I am saying. Feels like you are making your own HTML just tough to edit for yourself
#
[jgmac1106]
doesn't look bad actually
#
[tantek]
still struggling with debugging a new key photo extraction function
[dave], gRegorLove_, [ColinMorris] and [David_Bryant] joined the channel
#
[David_Bryant]
Looking forward to comparing photo publishing and gallery implementation notes with you!
#
@joemasilotti
Spent the weekend learning about #indieweb and meeting a bunch of awesome people at IndieWebCamp Austin! Today I added v1 of listing webmentions on my site. https://masilotti.com/webmentions/
(twitter.com/_/status/1231725298992979968)
[gRegorLove] joined the channel
#
@joemasilotti
Spent the weekend learning about #indieweb and meeting a bunch of awesome people at @indiewebcamp Austin! Today I added v1 of listing webmentions on my site. https://masilotti.com/webmentions/
(twitter.com/_/status/1231728039723184128)